Political Reasons and the Limits of Political Authority

Abstract Authority is a normative power to create duties in others. The most plausible accounts of this general power relate it to existing reasons the subjects of authority have with which authoritative directives can help them comply. Such accounts lead some theorists to ascribe a morally ambitious function to political institutions. This article argues against such theories. It defends political authority as a modest normative power, constrained by the type of reasons with which it can help its subjects comply. This modest account differs from other liberal views in the limits it imposes on the exercise of political authority. It casts doubt on familiar limits that protect an individual private sphere. Instead, it imposes a condition of moderation. It suggests that legitimate exercises of political authority should leave space for individuals to be motivated by reasons that political institutions do not and should not mediate for them.


I. Introduction
There is an old and venerable view of political authority that attributes an ambitious moral function to political institutions. 1 The view holds, roughly, that political institutions can help their subjects discharge their existing moral duties and that, when they do, this provides the basis for their political authority. The view concedes that some matters-private or personal matters-might be beyond the scope of authority, but insists that authority's function is morally salutary where it legitimately applies. Since authority mediates for its subjects reasons for action they already have, it helps individuals do what they ought to do.
Contemporary variants of this old view can be found in unlikely places, among both Benthamites and Kantians. Joseph Raz locates this ambitious moral function at the heart of the phenomenon of authority. For Raz, legitimate authority helps its subjects comply with reasons that already apply to them. Legitimate political authority is merely an instance of this general phenomenon. This is why, Raz explains, when law claims to hold legitimate authority, it claims to "occupy, as all authority does, a mediating role between the precepts of morality and their application by people in their behaviour." 2 Other theories-with different visions of political morality-sometimes employ this Razian structure. When considering the compatibility of Kant's political theory with Raz's framework, Arthur Ripstein explains that the exercise of political authority "can be described as enabling people to 'do better' at conforming to the application of their underlying duties of right." 3 Although in this Kantian vision, political institutions can only help individuals comply with some of their moral duties (i.e., duties of right), the account still shares with Raz's the attribution of a morally salutary function to political institutions. It is still the case that the mediation of unconditional duties to the authority's subjects enables them to do what they are morally required to do. 4 Such views of political authority have both appealing and objectionable sides. Their main virtue is that they seem to offer the right kind of answer to a difficult question. They ask: "how can the mere say-so of some create a duty for others?" and they answer: "there were already moral reasons there, before the exercise of authority, and it is from these existing reasons that the authoritative directive draws its normative force." This sort of explanation seems necessary if we are to explain political authority as a genuine normative power, that is, a power to create a pro tanto duty in others. The explanation allows us to persuasively deny that this normative power emerges out of thin air by locating its roots in nonmysterious reasons individuals already have. And yet, it is easy to see why people often balk at such visions of political authority. Even when it is offered only as an account of legitimate authority, which may only be claimed by existing institutions without being possessed by most, it seems to associate too much power, or the wrong kind of power, with political authority. It is not only that people are skeptical about the moral service that their institutions actually provide; most people would reject this view as an ideal of the service that political institutions should provide.
Or at least so I will argue. Much of what follows is concerned with defending political authority as a genuine normative power while denying a morally ambitious function to political institutions. This requires two arguments. The first argument is positive, identifying a particular service that political institutions should provide for their subjects in order to have legitimate authority. This service is narrower than the one morally ambitious accounts argue is the basis for legitimate authority. Political institutions do not and should not help their subjects better comply with reasons for action generally, but they can and should mediate for them a particular type of reasons they have: their political reasons.
Political reasons are reasons triggered for people by the fact of their living under legitimate political institutions, such as reasons to respect the rights of others and reasons to do one's part in a scheme of cooperation. These are discussed in Section II. It is easy to defend the claim that the subjects of political authority are more likely to comply with these reasons if they follow authoritative directives. However, the same does not go for other types of reasons for action. Section III illustrates how many criticisms of Raz's conception of authority lose their force when the normative function of authority is confined only to the mediation of political reasons.
The following sections offer a negative argument relating to types of reasons that political institutions can never authoritatively mediate for their subjects. Section IV discusses reasons from personal value, which emanate from a person's individual will, and Section V addresses small-scale moral reasons to which individuals respond in their interactions and exchanges. Together, these two sections make two related points. The first is that political authority cannot defensibly mediate for its subjects these other morally relevant reasons, which are not political reasons. Political authority mediates for its subjects reasons they have as people living under political institutions. It cannot mediate for them reasons for action they have as distinct individuals, relating to nonpolitical aspects of their practical identity. The second point is that the full mediation of political reasons is not morally desirable. Political reasons often conflict with other morally relevant reasons, and their mediation restricts people's responsiveness to other reasons for action they have.
The resulting account understands legitimate political authority as a modest normative power, incapable of mediating for the subjects of authority any approximation of the balance of their reasons for action. It is a power that helps one comply with one's political reasons, which may be different from helping them do what they ought to do, even prima facie, given reasons emanating from their own attachments and other moral duties they have. Section VI explains why this introduces a practical limit on the exercise of political authority: a condition of moderation. The condition differs from other familiar liberal limits on political authority, which track public subject matters or protect an individual private sphere. Rather than dividing the realm of action into private and public domains, this condition acknowledges that different types of reasons-political, moral, personal-apply to almost every choice or decision. It demands that, regardless of subject matter, political institutions refrain from being overly demanding of their subjects. If they try to achieve too much-too much justice, too much fair distribution, too much equal autonomy-they cannot be defended as generating genuine duties for their subjects.

II. Political Institutions and Political Reasons
The phrase "political authority" can have different technical meanings. I will not go through the list of possible distinctions, 5 except to stress that I am not talking here about authority in the sense of a right to rule. 6 The political authority that I am interested in is a type of normative power. This particular normative power makes the say-so of some, intended to create a duty for its addressees, succeed in creating a pro tanto duty for them. 7 Whether such a normative power-or normative powers in general-exists and what exactly its nature would be has been the subject of a longstanding debate. 8 In the next section, I will argue for a nonmysterious notion of political authority as a bona fide normative power. The present section sets the stage for the argument by explaining the normative circumstances characteristic of life under political institutions. To do so, it uses the vocabulary of reasons and draws distinctions between types of reasons.
Political institutions are institutions that, among many other things, effectively coordinate and enforce: they issue directives that purport to be authoritative and use coercive means to uphold them effectively. The one feature common to all those who live under political institutions is that they are subject to these institutions' decisions and coercive power. This feature is different from and more common than the exclusive statuses of citizenship or membership in a political community. Many of those who live under political institutions-in some perverse instances, most of them-will be unnaturalized residents, migrant workers, or foreign investors. Still, as people who live and interact under political institutions, they have reasons that they share with (and only with) others living under such institutions.
The theory of political authority is plagued by the conflation of two sets of reasons that individuals have, both related to political institutions. The first set of reasons includes those reasons people have to establish, accept, or support political institutions. These reasons bear on the legitimacy of these institutions and determine whether their existence can be justified to those under their power. The second set of reasons includes reasons that only become relevant to people in the circumstances of coordination and enforcement that characterize life under legitimate political institutions. I will refer to this second group of reasons as "political reasons" and will argue that these are the reasons from which political authority derives its normative force.
There is a long list of contenders for being good reasons for establishing, accepting, or supporting political institutions. 1. Reasons to establish, accept, or support political institutions that would allow people to pursue collective ends and promote the common good of their community (and also: reasons to withhold support if they fail to do so); 2. Reasons to establish, accept, or support political institutions that would make it possible for people to pursue their disparate, individual ends simultaneously (and also: reasons to withhold support if such individual pursuit is excessively impeded); 3. Reasons to establish, accept, or support political institutions that would constitute just relations between people (and also: reasons to withhold support if the constituted relations are excessively unjust); 4. Reasons to establish, accept, or support political institutions whose decisions are in some way "ours" (and also: reasons to withhold support if institutional decisions are not "ours" in the relevant sense). 9 It might seem that showing that people have good reasons to have and support political institutions, which claim to exercise authority over their subjects, settles the question of political authority. 10 This, however, would be too quick. For example, the fact that institutions bring about material prosperity, either individual or collective, might mean that they deserve respect, acceptance, and support, but it does not in itself justify a normative power to create duties for their subjects. 11 This is so because not doing as one is told does not necessarily undermine the achievement that collective enterprises of cooperation bring about. It might be that the beneficial function of political institutions gives people reasons not to undermine them. It is also possible that, in many circumstances, this might mean that people have reasons to act as if the directives of these institutions create duties for them. However, this is not the same as having such duties. Consider circumstances in which one can defect without anybody else knowing. In such circumstances, reasons to accept or support political institutions do not entail treating their directives as authoritative. 12 Showing that institutions are legitimate and that people have a duty to support them falls short of showing that these institutions have the power to create duties in their subjects. 13 To understand the grounds of political authority, we should focus on another type of reasons related to political institutions: political reasons. Political reasons are reasons triggered for individuals by the fact of their living under legitimate political institutions. They are triggered in this way as life under political institutions allows for 9 These categories of reasons are not mutually exclusive. Under certain interpretations, some of them can be seen as overlapping or even identical. Even if they are, and even if some of these contenders are spurious, there is still likely more than one (or one primary) achievement associated with having political institutions or only one overarching reason to support them. The same applies, mutatis mutandis, to procedural arguments from democracy. See infra note 31. 12 For an analogous criticism of arguments that defend the normative power of promising by reference to a beneficial social practice of promise-keeping, see Murphy, supra note 8. 13 Edmundson, supra note 6, at 256. Cf. JOSEPH RAZ, THE MORALITY OF FREEDOM (1986), at 101-104. new ways for individuals to relate to others. I use the word "triggered" here in a technical sense. 14 Sometimes, people have reasons to act in certain ways, which only come into effect if there is a change in their nonnormative circumstances. These reasons have an "if . . . then . . . " structure, and their trigger is the "if . . . " clause. In the case of political reasons, it is the factual participation in life under legitimate political institutions that triggers certain reasons people have to behave in certain ways.
One instance of reasons that are triggered in such circumstances are reasons to respect other people's rights. Kant's delineation of law and right is premised on the insight that we require political institutions to constitute most of the rights we assert against others in our personal interactions. 15 In the absence of political institutions, a principle of equal freedom between people may demand that they respect each other's property, refrain from certain types of harmful conduct, and do what they undertake to do. And yet, in the absence of enforced coordination, it is impossible to say that people have a right to be treated in a way that is consistent with these reasons. This is so because having such rights depends on enforced coordination.
People need institutions to set out a common rule for them to follow, which will specify-among other things-what it means to respect another person's property, what it means to injure another, and under what conditions we should do what we have undertaken to do. What respecting others' rights means in practice would depend on the way actual institutions delineate rights. In one political community, for example, there may be a rule that requires people to make and carry out their contracts in good faith, and in another community there will be no such rule; in one community there may be a rule requiring people to compensate for bodily harm they negligently cause, while in another there will be no such rule, as these losses are covered by a no-fault state insurance scheme. While people generally have reasons to perform contracts and compensate for tortious behavior, they can only act on these reasons if they live under legitimate political institutions; and although in all political communities people will have reasons to respect the rights of others, respecting these rights may entail very different courses of action in different places.
People also need institutions in order to ensure that their own right-observing behavior is reciprocated. If the common rule under which we both live is not enforced, I have no assurance that you will respect my property, perform your contracts with me, or avoid injuring me. In such circumstances, a duty for me to respect your claimed rights would not be an instance of us sharing in equal, reciprocal limitations on our freedom. If reciprocity is not ensured, the observance of another's so-called right would not be an instance of participating in a system of equal freedom under a common rule but a unilateral submission to a rule that others might not observe in turn.
In the absence of institutions that promulgate rules and effectively secure conformity, there would be no single standard applied to all and no assurance that others 14 I am following here Enoch, supra note 6, at 299. will reciprocate one's right-observing behavior. People may still make claims against each other that look like claims they would make in a rightful condition (e.g., "stay off my land," "perform your part of the bargain"), but these claims would fall short of being right-claims since they would lack the necessary validity on which rights depend. 16 If the people making these claims are powerful enough, the addressees may have prudential reasons to comply with them. However, it is only life under legitimate political institutions that triggers reasons to respect this sort of rights.
The scope of political reasons extends beyond a general duty to respect the rights of others in personal interactions. People also have reasons for action that derive from legitimate demands others have against the political institutions under which they live. In circumstances of enforced coordination, distributive justice demands that people get their fair share of socioeconomic resources and opportunities. Claims based on such demands are not, strictly speaking, addressed to other individuals or correlate with individual duties. 17 They are addressed to political institutions, which are necessary for a just allocation of burdens and benefits on a societal scale. Individuals are not the addressees of such claims because no individual action can establish a just scheme of cooperation. The determination of each person's fair share must be systemic-encompassing the determination of private law rights and duties as well as a range of benefits and public duties. Although people living under political institutions are not the direct addressees of distributive justice claims, they do have related duties derived from the duties of political institutions. The availability of legitimate political institutions triggers reasons people have to do their part within a fair scheme of cooperation established by these institutions, secure in the knowledge that others will do their part as well. 18 Such reasons are triggered by life under legitimate political institutions. They are political reasons.
Political reasons are attached to the potential of justice, equality, and fairness that legitimate political institutions can bring. 19 This is why political reasons are deontic: they require action and do not merely recommend it. If people living under political institutions neglect their political reasons, they fall short of acting on duties they have within the particular scheme of cooperation they inhabit, and others may complain. This does not mean that political reasons constitute the only duties people have, or 16 Cf. Joel Feinberg, The Nature and Value of Rights, 4 J. VALUE INQUIRY 243 (1970). even the most important ones. Among the many duties each of us has, emanating from different normative sources, political reasons are those we distinctly incur as people living under political institutions. This distinguishes political reasons from the aforementioned duties to support legitimate political institutions (which can be shared by people who do not live under them) 20 and from reasons based on one's love for their institutions or community (which will not be necessarily shared by all those who do live under them). 21

III. Authority and the Force of Political Reasons
There is a special connection between political reasons and political institutions. Not only are political reasons triggered by life under political institutions, they are also the reasons from which the authority of political institutions derives its normative force. This is so because following the directives of legitimate political institutions is normally the only way for people to act on their political reasons. The relation between political reasons and directives of political institutions is thus reminiscent of (although not dependent on) Raz's general account of authority. Raz argues that legitimate authority is characterized by its ability to help its subjects comply with reasons that already apply to them. According to his Normal Justification Thesis (NJT)the normal way to establish that a person has authority over another person involves showing that the alleged subject is likely better to comply with reasons which apply to him (other than alleged authoritative directives) if he accepts the directives of the alleged authority as authoritatively binding and tries to follow them, rather than by trying to follow the reasons which apply to him directly. 22 Applying this conceptual thesis about authority to the special case of political authority, Raz explains: [State] authority should be recognized to the extent necessary to enable it to secure goals, which individuals have reason to secure, for which social co-ordination is necessary or helpful, and where this is the most promising way of achieving them. . . . Beyond that it will have authority regarding other issues which is based on other considerations, such as superior expertise, economy of effort, immunity from temptations and blackmail. 23 20 Reasons to support institutions apply to anyone in a position to support (or undermine) them.  , 1996), at 101. Note that it is also possible to be attached to political institutions without living under them, as some members of ethnic or national diasporas are. Moreover, as will be explained below in the discussion of reasons from personal value, reasons that emanate from such attachments are not deontic. These two qualities set them apart from political reasons. 22 RAZ, supra note 13, at 53. 23 Id. at 100. Later I will offer an argument against Raz's conception of political authority. I will argue that authoritative directives cannot derive their normative force from the range of reasons that Raz envisions and that are not political reasons. But for now, I want to stress some commonalities between the Razian framework and the account of political authority I offer here. If confined to political reasons, something like the NJT might be a good way to understand how political authority can plausibly be justified and how its normative force can be related to the force of certain reasons its subjects share. Consider, then, a more limited half-sister to the NJT, which can serve as a plausible account of legitimate political authority:

The Political Reasons Account
The exercise of political authority makes it the case that the subject of authority is likely better to comply with political reasons which apply to her if she accepts the directives of the authority as authoritatively binding and tries to follow them, rather than by trying to follow directly the political reasons which apply to her.
The Political Reasons Account explains political authority as a genuine normative power much along the lines of how Raz's account explains the influence of legitimate authority on the practical reasoning of its subjects. Raz argues that authoritative directives give their subjects a new reason for action, which also acts as an exclusionary reason, that is, a reason not to act on other reasons. The exercise of legitimate authority thus has two implications for its subjects. First, it demands that they follow the directive (in order to comply with reasons that already apply to them); second, it demands that they will not be directly motivated in their actions by the sort of reasons on which the authoritative directive is supposed to be based (the authoritative directive replaces these underlying reasons and is not added to them). This is indeed the case when it comes to political authority. Legitimately enforced coordination triggers reasons for action, with which people can normally comply only by following the directives of political authority. 24 The authoritative directive is simply the best path, and often the only path, for following these reasons.
Consider, for example, a legal rule that determines the obligations owed by the owner of one piece of land to the owner of an adjacent lot. The enforced determination of such obligations is necessary for rightful relations to exist between them. Since it allocates burdens and benefits, the determination also has broader distributive implications beyond the two owners' bilateral relations. It is part of the determination of what each owner owes others in the overall scheme of cooperation. The existence of such an enforced determination does two things. First, it triggers conditional reasons to respect the rights of others and do one's part within a scheme of cooperation. 24 Although the account invokes the idea of coordination, it does not contend, as Raz does, that "the case for having any political authority rests to a large extent on its ability to solve co-ordination problems and extricate the population from Prisoner's Dilemma type situations." Id. at 56. As we have seen, the ability to promote individual interests can be a good reason to establish or join life under political institutions, but, by itself, cannot give rise to political authority. The directives of legitimate political institutions draw their normative force from duties one incurs toward others as someone who lives under legitimate coordinating institutions.
Second, it mediates those reasons so that the only way each owner can respect the rights of others and do their part is by following the directive.
To say that political authority not only triggers reasons but also mediates them implies a standard of success. From the perspective of political reasons, perfect success would be creating a fully just scheme of cooperation, which correctly specifies the rights and duties of participants, enabling them to live in rightful relations with others while discharging their collective distributive obligations. This means that there would be a degree of failure beyond which directives can no longer be said to assist people in complying with their political reasons. 25 But it also means that some degree of failure would be tolerable. In reasonably just societies, people living under legitimate political institutions would ordinarily still be more likely to comply with their political reasons if they were to follow imperfect institutional directives than if they were to ignore them. Without following these directives, they will not be able to respect the rights of others or do their fair share in discharging their collective responsibilities at all. They will not be able to do these things since doing them requires deferring to an authoritatively determined point of coordination, legitimately enforced by political institutions.
The exercise of political authority emerges in this account as an instance of what David Enoch calls "robust reason-giving." 26 Political authority genuinely gives its subjects new reasons for action. It does not perform a purely epistemic function, merely showing how one might best comply with the reasons one has. Nor does it merely change the nonnormative circumstances of the subjects of authority in a way that only triggers conditional reasons the subjects of authority already had. It does more than this. It involves an intentional communicative activity in which it is intended that the directives of political institutions be taken as a new reason for action by their addressees because they are the directives of political institutions. To go back to the example of the two neighbors, the exercise of political authority intentionally gives them reasons to behave in a certain way toward each other. Had the directives of political authority been different, the neighbors would have had a reason to behave differently. In both cases, following the directives of their political institutions would be the best way for them to comply with their political reasons. They will be more likely to observe each others' rights, and do their part in the overall scheme of cooperation, if they follow the directives of their institutions than if they act on their own judgments on what these mean. Each can continue to believe that her views on what is right and just are better than the views underlying the directives. However, if they are to comply with their political reasons, their best course of action would be to follow the directives of political institutions.
So much for the affinity between the Political Reasons Account and Raz's conception of authority. The account is meaningfully different from Raz's in at least one way, which goes to the heart of liberal political morality: it is much more discriminate in the moral claim it attributes to political institutions. This discrimination has some 25 For a canonical expression of a similar threshold, see Gustav Radbruch, Statutory Lawlessness and Supra-Statutory Law (1946), 26 OXFORD J. LEGAL STUD. 1 (2006). 26 Enoch, supra note 6, at 301-302. Significantly, in the account offered here, robust reason-giving does not depend on mysteriously deriving a deontic normative power from nondeontic reasons. See infra note 44. immediate advantages, making the Political Reasons Account immune to some cogent criticisms, to which Raz's theory is sensitive. I want to mention a few of those (four in the text and two in the footnotes).
One long-standing criticism of Raz notes that there are many circumstances where individuals can better comply with reasons that already apply to them by heeding the say-so of others. However, many of these circumstances do not seem to give rise to relations of practical authority. 27 Such counterexamples seem to miss the mark when it comes to the Political Reasons Account. Some counterexamples involve situations in which better compliance is promised only regarding nondeontic reasons. This is the case, for example, when an uninvited expert on Chinese cooking barges into the kitchen of a Chinese-cooking enthusiast and orders him around. Since there is no underlying duty to cook a better Chinese meal, better compliance with one's reasons to cook a good meal cannot justify authority. The Political Reasons Account, however, is a "duty-in, duty-out" model of political authority. 28 It finds the root of the normative force of authoritative directives in underlying political reasons, which are deontic reasons.
Other counterexamples involve moral duties accompanied by sound advice regarding the best way to discharge them. One such example is the person who owes a duty to his family to take care of their financial affairs and receives the advice of an expert financial adviser, who does not seem to be exercising authority. 29 Marmor suggested that the force of such counterexamples is that they invoke relations outside institutions and therefore underscore the uninstitutional nature of Raz's theory. 30 But the Political Reasons Account is fully institutional. It pertains only to directives of legitimate political institutions based on reasons that arise from life under such institutions. Therefore, it can distinguish political authority from noninstitutional instances of advice and also answer two related criticisms of Raz, from jurisdiction and from democracy, which I address in this footnote. 31 27 RAZ, supra note 13, at 25-26; Stephen Darwall, Authority and Second-Personal Reasons for Action, in REASONS FOR ACTION 151 (David Sobel & Steven Wall eds., 2009); Stephen Darwall, Authority and Reasons: Exclusionary and Second-Personal, 120 ETHICS 257, 259 (2010). 28 Cf. Hershovitz, supra note 10, at 10; Marmor, supra note 10, at 126-127. But see Enoch, supra note 6. 29 For a discussion, see Hershovitz, supra note 10, at 10. 30 Marmor, supra note 10, at 129-130. 31 Raz's theory has been accused of an inability to explain jurisdictional limits on political authority. See id. at 254. The Political Reasons Account explains why it is normally the case that political authority does not create a duty for people in other jurisdictions. Let us assume that the rules of the English common law of contracts are superior to those of the German BGB and would have helped bring about a better and more just scheme of cooperation than German law. This would still not be the case for their having political authority over Germans. Germans cannot act on their political reasons by following English law. They will not be able to do their part and respect the rights of others in Germany by following the directives of political institutions under which they do not live. Cf. Waldron, supra note 18, at 18. Relatedly, Raz's conception of political authority is sometimes criticized for its failure to account for so-called "democratic authority. " See Scott Hershovitz, Legitimacy, Democracy, and Razian Authority, 9 LEGAL THEORY 201 (2003). The Political Reasons Account maintains, like Raz, that democratic processes do not by themselves give rise to a genuine political authority, understood as a normative power. However, it acknowledges that if democracy is a condition for the legitimacy of political institutions, its absence may negate their authority. On the effect of illegitimacy on authority, see supra note 19.
There is another feature of the Political Reasons Account that sets it apart from the Razian theory of political authority. Raz's theory differentiates between purported subjects of political authority. 32 Since legitimate authority depends on its ability to help people better comply with reasons that already apply to them, and since people differ in their ability to act on these reasons unassisted by authority, the same directive might be authoritative for some and not for others. Stephen Perry argues that the notion that political authority emerges as an aggregation of multiple "one-on-one" relationships is hopeless: [T]he largest failure of the NJT, considered as an exercise in conceptual analysis, concerns its aggregative aspect. If the NJT has a conceptual core, this is it. . . . [T] he NJT applies only to the normative relationships between an authority and the individuals subject to it, considered one by one. There is no room for the idea of the enterprise as a whole possessing legitimate authority in any sense other than the conjunction of statements about all such relationships. 33 The Political Theory Account shares some of the "conceptual core" to which Perry objects. It maintains that political authority applies to people one by one, allowing each to act on reasons they have. However, it does not suffer from the flaw that Perry identifies. The account allows us to speak meaningfully about reasons for action, which always operate at an individuated level, while at the same time recognizing a political phenomenon. It does so by identifying a political dimension to who we are, a political dimension to our practical identity, which requires us to attend to our shared political reasons. This means that political authority does not apply differently to different people living under the same political institutions. All such people depend equally on the exercise of political authority for acting on their political reasons, which apply to them as people living under political institutions-an aspect of their practical identity they equally share.
The Political Reasons Account differs from Raz's conception of authority in a final, related respect. It fits better with practice and prevalent beliefs. Raz, contemplating the consequences of his more expansive and demanding theory of legitimate political authority, concludes that "political authorities are likely to have a more limited authority than the authority many, perhaps all of them, claim to have, and that people generally believe they have." 34 Indeed, if to claim legitimate political authority meant to make a claim to mediate between people and an unspecified set of reasons that generally apply to them, one could see why Raz's skepticism would be warranted. But this skepticism raises doubts about the soundness of Raz's conception of authority and its attribution of either deception or mistake to almost all those who are party to political practices. The Political Reasons Account does not give rise to a similar concern. If limited to political reasons, the claim attributed to political institutions is much more 32 RAZ, supra note 13, at 73. plausible. It is a claim that can be defended in relation to many political institutions and, as I shall argue below, aligns with the ordinary understanding of their normative function.

IV. Reasons from Personal Value
The argument so far has been a positive one, establishing a connection between the normative force of political reasons and the normative power involved in the exercise of political authority. This section and the next offer a negative argument along two lines. They argue that political authority does not carry the normative force of other morally relevant reasons for action that apply to individuals. They also reject the notion that political reasons embody unconditional duties, whose mediation orients the subjects of authority to proper action. In this section, I make this argument about reasons from personal value. First, I explain why the normative force of authoritative political directives cannot be derived from the normative force of such reasons. I then reject the idea that there is a clear hierarchy between reasons from personal value and political reasons and that the role of political authority is to enforce this hierarchy.
Reasons from personal value are reasons for action emanating from one's attachments and valuing of her projects, chosen aims, and relationships. These reasons also bear on what individuals should do. 35 As Stephen Darwall explains: the fact that someone values, wants, cares about, or is devoted to something can be a source of reasons that add weight to that of whatever reasons there might be for him to value it in these ways, that is, to the object's value independently of his valuing it, whether this be in terms of well-being or values of other kinds. 36 The idea here is not (only) that, in our everyday lives, individual desires and wants affect our practical reasoning. It is that, upon reflection, the fact that a person values something, is attached to something, or is devoted to something can be an independent source of reasons for action. The value of having people acting on reasons from personal value is distinct from the value of acting in one's own best interest. 37 It is also different from the (doubtful) value of simply acting on one's desires or promoting one's preferences. 38 It is the value associated with asserting one's will, attachments, and valuing as a distinct source of reasons for action. associated with deliberate responsiveness to one's reasons, self-expression, and individuality. 39 The exercise of political authority may help people act on their reasons from personal value. 40 One possible use of political authority is to create circumstances in which more people can act on reasons for action emanating from their own valuing and attachments. For example, as an employee, my powerful employer might coerce me to make decisions that do not best serve my family and me in living our lives as we see fit. 41 The sophisticated behavior of credit card companies, to give another example, may have a similar negative effect on my responsiveness to reasons when taking out a card or when using it. 42 The exercise of political authority can help in such situations, as it sometimes does through employment law and consumer credit law. When it does, it expands the range of choices available to individuals in society and improves these individuals' ability to act on their reasons from personal value.
Notwithstanding this possible use of political authority, it is still wrong to think that authoritative directives can derive their normative force from reasons from personal value. I offer two independent arguments for thinking they cannot. The first goes to the deontic nature of authoritative reasons. We can imagine a directive that would help a person better comply with her reasons from personal value. For example, valuing your career gives you a reason to invest in it disproportionally to its impersonal worth. It might be that a directive, based on expertise, can help you get promoted, become better at what you do, or be better recognized in your industry. Some believe that such a directive can be authoritative. David Enoch suggests that if a directive gives you a reason for action together with a reason not to act on some conflicting reasons, this constitutes an instance of authority. 43 As the examples above suggest, this view seems overinclusive. 44 In this example, the reasons with which authority helps you comply are not deontic. How can a directive, whose only claim is to help you comply with such reasons, give rise to a duty? In order for the directive 39 Cf. Shiffrin, Egalitarianism, supra note 35, at 291; Smith, supra note 35. 40 Some examples of such uses of political authority are elaborated in Shiffrin, Paternalism, supra note 35. On a possible affinity between reasons from personal value and Raz's notion of autonomy, see id. at note 51; Shiffrin, Egalitarianism, supra note 35, at 291-295. On a possible connection between reasons from personal value and Raz's independence condition for legitimate authority, see Section VI.  Enoch, supra note 6, at 317-322. Enoch is careful not to commit to saying, as Raz does, that this can give rise to a duty. He chooses to remain agnostic about the use of the term. The reason for this is clear: Raz's identification of duty with protected categorical reasons is idiosyncratic (to say the least). However, it is hard to see how Enoch can be silent on this matter, given that his definition of authority involves the notion of duty. This is also related to the following footnote. 44 Enoch defends this view as necessary. He suggests that unless one is willing to admit that duties can arise from other "normative materials" that are not duties, authority cannot be understood in terms of robust reason-giving. Id. at 322. Enoch offers no other defense of the plausibility of this normative transubstantiation. I believe that my argument in this paper succeeds in offering a counterexample for practical authority that is based on preexisting duties (i.e., political reasons) and still involves robust reason-giving. See note 26, supra. The feature in my account that allows for this is the recognition that compliance with some duties conceptually depends on the exercise of authority.
to be authoritative and create a duty, the reasons on which it is based must have the appropriate normative force. If my claim that you have a duty to do as I tell you is based on your better complying with reasons that already apply to you, these reasons must themselves have deontic force.
The second consideration against thinking that political authority can base itself on the force of reasons from personal value is a concern from pluralism. 45 Since people have different attachments and wants, their reasons from personal value can diverge considerably. It makes little sense that the exercise of unitary authority over a large group of people with diverging reasons can find its normative roots in such plurality. A directive or a rule that would help one person conform to her reasons from personal value might not help another. Moreover, it is hard to imagine either a general rule or a system of tailored directives that would get the right mélange of priorities and attachments for any single person and be able to guide her on how to comply with the proper balance of her personal reasons. Our responsiveness to reasons from personal value is, at least in part, what makes us distinct individuals. 46 This is also why they cannot be the sort of reasons from which political authority derives its normative force.
So what should we make of the examples above-and many others-that suggest that the exercise of political authority can be justified in reference to its contribution to individual responsiveness to reasons from personal value? The answer to this involves an important nuance. There is a difference between (a) helping people comply with a duty to respect and promote the ability of others to act on their reasons from personal value and (b) assisting the subject of authority in acting on their own reasons from personal value. Through our private property and contracts, for example, we might be able to pursue our reasons from personal value independently from others. 47 But this does not mean that the reason I have to perform my contract or respect your property is that it enhances my compliance with my reasons from personal value. What underlies the authoritative force of the rules of property and contract is that they help me comply with deontic reasons I have to respect the rights of others. These are political reasons, emanating from the potential I have to live with others in relations of equal freedom under a common rule-not reasons from personal value.
The exercise of political authority may be aimed at creating a more just distribution of people's ability to act on their reasons from personal value, or it can be aimed at creating a society in which people's "abstract opportunity" to act on such reasons is maximized. 48 But this does not mean that the reasons political authority mediates for its subjects are reasons from personal value. The following example illustrates the point. It is possible to think of the regulation of employment relations as ensuring that an imbalance in the power relations between employers and employees does not undermine the latter's ability to act on reasons from personal value. But one should not conflate this with the claim that the normative power associated with 45 Cf. Perry's objection from plurality, discussed in the text accompanying note 33. 46 Smith, supra note 35. political authority can emanate from the service it provides to an employee's own ability to act on such reasons. It is not that the employee should act on authoritative directives because it would help her better comply with her nondeontic reasons from personal value. Rather, the law helps the employee better comply with her duties toward others in her community, demanding that she refrain from being employed below the fixed standard. 49 It helps her comply with her political reasons, which are deontic reasons, not her reasons from personal value. Now, one might think that the argument so far is missing an important point. It might seem that there is a hierarchy between political reasons and reasons from personal value. If there were such a hierarchy, then political reasons would set the proper limits to what reasons from personal value should be and how people may act on them. One may hope that, with the proper exercise of authority, people's valuing would be shaped so that it does not conflict with the demands of their political reasons. After all, most people's attachments, projects, and goals are shaped in light of impersonal reasons they have. 50 However, when there is conflict, it may seem that the role of political authority is to mediate the sort of duties that trump reasons from personal value.
In contemplating such suggestions, it is important to distinguish between the effect of political authority and the normative force of political reasons. The distinction is important and will be the focus of Section VI. For now, it suffices to note that the exercise of political authority would sometimes mean that the subjects of authority must not act on some of their reasons from personal value. This is often what it means to be subject to political authority. However, this does not necessarily imply a hierarchy between political reasons and reasons from personal value. Suggesting such a hierarchy must be premised on one of two illicit moves. It either discounts the fact that personal valuing and attachment can be an independent source of reasons (and value)-a point I addressed above-or it underestimates the potential demandingness of political reasons.
Political reasons can be highly demanding. In our everyday lives, almost every action bears on the position of others and the share they get from the scheme of cooperation. Most of our decisions-in the workplace, in our community, and in our family-affect how resources, burdens, and benefits are distributed. Many of our behaviors have direct externalities on the well-being of others, on their autonomy, and on relationships in society: driving a car, having kids, not getting a checkup, eating unhealthy food, taking out a mortgage, and so on. If we were to observe completely our collective duty to create a fully just society, and if we all were to act on the derivative duties that we would have in such a society, there would be little space for us to act on reasons from personal value at all. Seana Shiffrin contemplates a society that fails to accommodate individual responsiveness to reasons from personal value in this way. In such a society-49 Consider the classic example of an employee who is not allowed to be employed below a minimum wage. It is not clear that her reasons from personal value are served by this prohibition, say, if this means that she remains unemployed. If this prohibition is attached to reasons she already has, these would be duties owed to others in her community not to take up employment below a minimum wage. agents may feel constrained by the sense that everything they do impacts on others and is subject to accounting. This may constrain or dominate the experience of choice. . . . The goods of purely free choice may be overly compromised. Some of the more important goods of self-expression may be sacrificed, particularly in arenas in which agents feel especially susceptible to social pressure. 51 The insight here is that it is wrong to think about the relations between political reasons and reasons from personal value in terms of hierarchy or subjugation. Since responsiveness to reasons from personal value is morally worthwhile, when it conflicts with respecting impersonal reasons for action, this presents individuals with an internal conflict. 52 It is not that political reasons and reasons from personal value always pull in different directions. Often, as noted, our personal attachments are shaped in light of what is valuable and worthwhile from the political perspective. But sometimes there is a conflict, and when there is, there is no simple hierarchy between the two sets of reasons. Although political reasons seek to introduce limits to how or how much I can act on my reasons from personal value, it is wrong to think that they should always prevail.
One thinker who had a deep appreciation of the possible conflict between our attachments and our impersonal duties was Rousseau. 53 In one passage, Rousseau suggests that by ensuring congruence between our reasons from personal value and our impersonal reasons, it may be possible to resolve this internal schism: while men cannot be taught not to love anything, it is not impossible to teach them to love one object rather than another, and to love what is genuinely fine rather than what is malformed. If, for example, they are taught from sufficiently early on never to look upon their individual [self] except in its relations with the body of the state, and to perceive their own existence as, so to speak, only a part of its existence, they will at last succeed in somehow identifying with this larger whole, to feel themselves members of the fatherland, to love it with that exquisite sentiment which any isolated man has only for himself, to raise their soul perpetually to this great object, and thus to transform into a sublime virtue the dangerous disposition that gives rise to all of our vices. 54 What is discounted in this passage from Rousseau is not the value of attachments as such, but the value of love and attachment that stems from one's own individuality, even when inconsistent with one's character as a citizen. If extinguished, either by coercion, by indoctrination, or by some other means, this fountain of value rooted in individuality will be lost. Thomas Nagel forcefully makes a similar point when 51 Shiffrin, Egalitarianism, supra note 35, at 287-288. 52 Cf. Williams, supra note 50, at 17-18. reflecting on utopian visions of politics. 55 Like Rousseau, Nagel believes that there may be a schism within individuals between their personal and impersonal reasons for action. He also holds that if an individual were to act on all of her impersonal reasons and discharge all of her duties toward others (including those from political reasons one would have in a fully just society), there would be little room for her to follow her attachments. However, Nagel is much more careful in insisting that personal value is given its due within any system. "To deal with the problem," says Nagel, "by attempting to shrink the domain of the private to a tiny compass by an assault on individualism is foolish, and to the extent it succeeds it will destroy most of what is valuable in human life." 56 All of this is important for underscoring a crucial fact about political authority and the sort of moral service it can plausibly provide to its subjects. Political authority does not draw its normative force from the mediation of reasons from personal value, nor does it mediate reasons whose normative force always trumps individual attachment. According to the Political Reasons Account, political authority is a modest normative power. Political authority cannot plausibly claim to mediate for its subjects more than a sliver of their morally relevant reasons for actionwhich may well conflict with other morally relevant reasons they have. The mediation of political reasons does not guide the subjects of authority toward proper action or impose proper side constraints on them. In short, political authority, even at its very best, is sectional rather than comprehensive in the normative service it offers to its subjects.

V. Other Deontic Reasons
The rejection of a clear hierarchy between reasons from personal value and impersonal reasons leaves unresolved the problem of the possible contradiction between them, in a way that continues to trouble political philosophy. Like Rousseau, Nagel also finds in this schism the germ of politics. Nagel's preferred solution, however, is an institutional division of labor rather than the taming of one's attachments. The great challenge of political theory, says Nagel, is to "design institutions which serve an ideal of egalitarian impartiality without demanding a too extensive impartiality of the individuals who occupy instrumental roles in those institutions." 57 So, instead of a hierarchical relationship between reasons from personal value and impersonal reasons, we will have institutional roles-including those established by political authority-that will harness individual motivations (including those from individual attachments) into a workable system that serves impersonal ends.
There is some commonality between the Political Reasons Account of political authority and Nagel's division-of-labor model. They share the idea that the service political authority provides to people, in helping them act on their reasons, is sectional rather than comprehensive. However, Nagel's model remains an instance of assigning an ambitious moral function to political institutions. According to his 55 THOMAS NAGEL, EQUALITY AND PARTIALITY (1991), ch. 3. 56 Id. at 32. 57 Id. at 61. model, people would be largely free to act on their reasons from "self-interest, . . . personal attachments and commitments" in their personal lives. 58 What allows for this freedom is the fact that the institutional framework would ensure compliance with the rest of their reasons-their impersonal reasons for action. Compliance with the broad range of one's impersonal reasons in this institutional reality would turn on each playing their role(s), as determined, in part, by the coordinating and enforcing function of the state. The duties they incur by dint of their living under political institutions would thus be their channel for complying with a broad range of their duties toward others (beyond the more limited subset of political reasons). 59 The allocation of this morally ambitious role to institutions is the result of Nagel's alignment of motivational questions and questions of institutional function. 60 One may think that if people living under political institutions are ordinarily motivated by their reasons from personal value (or even their selfish desires), the role of social and political institutions must be to help them discharge the broad range of their impersonal reasons for action. But this view ends up ascribing to political institutions a function that is too broad. Apart from political reasons, which it is the role of political institutions to mediate, people also have other impersonal reasons for action that need not be mediated by political institutions and do not require the use of collective action for us to comply with them.
The idea that there are moral reasons that are not political is fundamental to liberalism. As Samuel Scheffler explains: Egalitarian liberalism insists on the autonomy of the values and norms governing human interpersonal relations . . . Those values are part of a conceptual and normative repertoire that we are given by virtue of our participation in forms of human interaction and aspiration that are so basic that we cannot really envision what human social life would be like without them. . . . Yet, [egalitarian liberalism] denies that our normative repertoire is exhausted by the values that govern our personal lives and small-scale interpersonal relations. . . . The familiar commonsense principles of individual conduct are not an adequate guide to institutional design. 61 There are several reasons for thinking that political institutions should not mediate for their subjects the kind of small-scale moral reasons that attach to the values that Scheffler identifies here. One reason is that such mediation degenerates and denigrates people's moral agency. Practically, if individuals stop trying to ascertain and act on their impersonal reasons, they will lose the capacity to develop the powers of judgment that their moral agency requires. Conceptually, relying on institutions to take care of large parts of the impersonal dimension of our moral psychology is no less offensive to our individuality than relying on political authority to determine our objects of attachment. Unmediated responsiveness to impersonal reasons in our personal lives and our small-scale interpersonal interactions is part of what gives value to individual lives. Figuring out for ourselves our moral reasons in these interactions is part of what constitutes each of us as a distinct individual. 62 This is why there are only limited grounds on which we would be willing to accept moral authority. It is also a reason to avoid consigning the responsibility for figuring out these moral reasons to political institutions, which will only make everybody more like everybody else on matters in which we value difference and plurality. 63 The idea that political authority claims to speak in the name of the broad range of our impersonal reasons is not only normatively indefensible but also radically implausible. Consider Mark Greenberg's example of statutory rape, where, he believes, political authority helps specify a moral duty. For Greenberg, "before action by legal institutions, the content of the moral prohibition will be relatively vague, perhaps something along the lines of: sex with children is prohibited." 64 This changes, however, after the exercise of political authority. "Once the legal institutions have acted," says Greenberg, "the content of the prohibition will typically be much more precise." I take Greenberg's example to prove the opposite point. The criminalization of sex with people younger than sixteen does not help clarify an ordinary moral duty not to have sex with children. From the perspective of this small-scale moral duty, sex with someone two days before their sixteenth birthday and sex with someone two days after their sixteenth birthday do not become distinct after the exercise of political authority. 65 Nor is it part of the implicit claim of institutions exercising political authority that they do.
What exactly went wrong in Greenberg's reasoning? Greenberg's argument confuses reasons from small-scale morality with political reasons. For an individual considering having sex with a young person, there would be strong, decisive reasons from small-scale morality not to. Importantly, however, there would also be relevant political reasons. The criminalization of sex with people below a certain age helps people comply with the latter type of reasons, not the former. Criminalization delineates rights and duties, distributing burdens and benefits between members of society in a particular way, and it can do so also when it comes to sex. It is in this sense (and in this sense alone) that legislation changes its addressees' normative circumstances and helps them better comply with reasons that already apply to them. The point is not that the statute draws the line in the wrong place. Had the age limit been twenty-one, that might stave off much of the moral concern associated with having sex with children but would create the opposite worry, of creating a moral prohibition where none exists (as if criminalization made sex with a mature person of twenty years and 364 days immoral in the same way that sex with a fourteen-year-old invariably is).
Here is another example. Shiffrin, contemplating the relations between the (smallscale moral) category of promise and the (legal-political) category of contract, suggests that we should understand contract law as "the public complement to the private promissory relationship." 66 This is a good model for thinking about the relations between law and small-scale moral duties generally. It is often the case that smallscale moral reasons and political reasons apply to the same circumstances (e.g., making a promise/contract; sexually exploiting a young person/statutory rape). The directives of political authority, mediating for subjects political reasons, can be seen as capturing the public aspect of this plurality of morally relevant reasons. The exercise of political authority does not include an implicit claim to mediate for people their reasons from small-scale morality. Nor does it establish a minimal moral threshold above which people can be supererogatorily virtuous. There is no reason to think, as some Kantians do, that a person's political reasons take precedence over her ordinary moral reasons. 67 Political reasons can be more or less demanding than reasons of small-scale morality, depending on the circumstances, and the same goes for law and other directives of political authority.
Still, the examples from statutory rape and contract law can only take us so far in understanding the limits of political authority's moral function. Both examples are hospitable to some notion of congruence between small-scale moral reasons and political reasons. They might support an implicit assumption that there can always be a single right answer to what morality-both small-scale and political-requires in any particular set of circumstances. Something akin to this belief in harmony leads Shiffrin to argue that contract law should, by and large, accommodate moral reasons that individuals have and that are related to the relations that this law regulates. 68 Political authorities should make sure their rules do not undermine the conditions "necessary for moral agency to flourish," 69 and, in the case of contract law, protect the moral culture surrounding the small-scale moral institution of promising.
I will say more about accommodation in the following section when I evaluate different strategies of moderation. The point I want to make here is that Shiffrin's framework downplays the possible tension between reasons of small-scale morality and political reasons. The tension is harder to spot in the context of the abstract questions of contract law theory but becomes clearer in matters with more pronounced distributive implications. Take, for example, types of contracts whose societal distributive implications are unmistakable, such as employment contracts or tenancy contracts. In such contracts, the small-scale moral reasons that arise in the context of promisemaking between two individuals can be in tension with political reasons these same individuals have, emanating from their relationship's societal and distributive significance. In such contexts, legal rules often depart considerably from the interpersonal norms that govern promising. They do so for good (political) reasons. 66 Seana Valentine Shiffrin, The Divergence of Contract and Promise, 120 HARV. L. REV. 708, 750 (2007). 67 Timmermann, supra note 4. The possible conflict between political reasons and other deontic reasons and the lack of hierarchy between them are crucial elements missing from the standard Kantian understanding of political authority. Cf. Arthur Ripstein, Authority and Coercion, 32 PHIL. & PUB. AFFS. 2 (2004). 68 Shiffrin, supra note 66. 69 Id. at 710.
Political reasons can diverge from and conflict with reasons of small-scale morality that apply to the same circumstances. The question of the right balance between the two sets of reasons has no easy answer. Strategies that establish a hierarchy between political reasons and other reasons we have, seek to replace our small-scale moral reasons with institutional political reasons, or prescribe harmony between the demands emanating from different normative sources are either dangerous or naïve. Since there is no guaranteed convergence and no hierarchy between the different types of reasons, their diverging demands are difficult to compare wholesale, and no hard choice between them could be made on a general basis.

VI. Moderation and the Limits of Political Authority
The discussion of reasons from personal value and small-scale morality aimed at showing two things. First, political authority does not mediate such reasons for its subjects and therefore does not derive its normative power from them. Second, the relations between these reasons and political reasons are not hierarchical; one's political reasons do not always take precedence over these other reasons for action individuals have. Notwithstanding the fact that people live together under legitimate political institutions, they are also individual choosers and individual moral agents in their small-scale interpersonal relations. As such, they also have nonpolitical reasons for action, which stand in nonhierarchical relations with their political reasons.
The conjunction of these conclusions points to a potential gap between the normative grounds of political authority and its practical demands. On the one hand, since different reasons individuals have can conflict, and since political authority only mediates a limited subset of these reasons, its directives' authority is inherently sectional. It speaks to its subjects in the name of only one aspect of their practical identity-the aspect relating to their life under political institutions. On the other hand, the demands of political authority are peremptory and imperative. Political authorities speak with a bottom-line ought, not a weak conditional prescription. If there is a claim associated with the exercise of political authority, it is not "this is what your political reasons require that you do and nothing more"; it is "this is what you ought to do because your political reasons require it." 70 It calls upon its subjects not to act on conflicting reasons from personal value and small-scale morality and heed their political reasons.
Note that this gap is only a potential one. The exercise of political authority can create pro tanto duties for people based on some of their political reasons for action, which would be congruent with their overall balance of reasons and can be expected to withstand competition from some conflicting reasons. However, if the scheme of cooperation put in place by political institutions is so demanding that it does not leave enough space for people to reason as individuals and as moral agents, this undercuts the practical service provided by political institutions for their subjects. 70 Cf. Edmundson, supra note 6, at 215-216. It will not do to say that political authority only claims to create a pro tanto duty for its subjects and is silent about the potential conflict between its directives and other, nonpolitical reasons its subjects have. Even a pro tanto duty must withstand some pushback from conflicting reasons.
In such an overly demanding system, political institutions may be telling people what "doing one's part" means around here, but the system they create would be unsuitable for the moral psychology of their subjects. The topic of this section is the practical condition this poses to the exercise of political authority.
One possible limit to the exercise of authority, motivated by a similar concern of authoritative overreach, is to restrict political authority to some activities or subject matters, allowing for a "protected sphere of citizen autonomy." 71 This familiar liberal limitation distinguishes between a realm of action that must be left to reasons from small-scale morality and personal value and a realm of action in which political authority reigns supreme. Raz defends this sort of limit to authority, spelled out in his "independence condition." 72 According to this condition, the exercise of authority cannot include "the sort of personal decisions regarding which I should decide for myself rather than follow authority." 73 In some matters, says Raz, it makes no sense to defer to authority because they are the sort of matters on which individuals must decide for themselves. Raz gives the example of marriage in a culture in which decision on the choice of one's spouse are expected to be made solely based on the person's judgment. In such matters, he explains, "there are reasons that can only be satisfied by independent action." 74 The main flaw of the independence condition is its focus on a distinction between types of matters and decisions, rather than between types of reasons. It falsely assumes alignment between a protected set of "personal decisions" and the set of "reasons that can only be satisfied by independent action." In reality, various types of reasons usually apply to the same subject matter (either "personal" or "public"). Raz's own example of the decision involved in choosing one's partner in marriage is not "a sort of decision" on which one cannot heed political authority (for example, in following rules against bigamy). Political authorities do instruct people on political reasons that apply to them and are choice-of-spouse related. They delineate rights and distribute burdens and benefits in such a way that gives individuals good deontic reasons in this context. More generally, almost every "personal matter" has implications or externalities that affect distribution and shape relationships in society. 75 This means that alongside the personal reasons that apply to such decisions, there may also be political reasons that apply to them.
It is not only that political reasons can be relevant to a host of "personal" decisions. Reasons from personal value or small-scale morality are also relevant to "public" decisions and subject matters. Sometimes, so-called "public decisions" are the primary locations where our reasons from personal value or small-scale morality demand we act. Our personal valuing of relationships and projects is relevant to almost every decision we make about work, spending, education, consumption (and pollution), political activism, and so on. And the same goes for small-scale moral reasons. If we were to accept Raz's independence condition, we might have asked whether the decision to perform a promise in an employment contract is a private moral decision or a public political one. But stating the question in this way is unhelpful. The question is not, as the independence condition suggests, what type of decision or subject matter is concerned.
The question is what type of reasons apply to the decision, and over what kind of reasons legitimate political authority can be claimed.
Let us return to the problem at hand, which concerns the necessary condition authority must meet if there is to be no problematic gap between the normative force of the reasons underlying its exercise and its ultimate demands from its subjects. Any condition that is premised on the delineation of a set of personal decisions or a sphere of private matters misses the ubiquitous relevance of different types of reasons and, therefore, misfires. A better solution would identify a condition that deals directly with the core problem of political authority: the potential conflict between political reasons and other, potentially incommensurable reasons for action individuals have. The solution I defend here-a condition of moderation-does precisely that.
There are two possible strategies for moderation. The first is to make a judgment about reasons from small-scale morality and predictions about people's attachments in the circumstances that law aims to regulate, and then make the demands of political authority conform, or conform in part, to these reasons. Shiffrin's model of accommodation, which I have mentioned above, follows this path. In the context of contract law, for example, Shiffrin demands that political institutions should not treat promises in ways that are inconsistent with their moral value, thus allowing people to respond to their small-scale moral reasons. 76 When it comes to reasons from personal value, the law can predict individual attachments-to a holy day of rest, to the care of one's children, or to unhealthy personal choices such as smoking and drinking-and protect people's ability to act on such attachments. 77 The second strategy of moderation is not accommodation but abstinence. Following this path means having political authority prescribe less, and be less stringent in mediating the demands of political reasons for its subjects. Any exercise of political authority should be aware of its limitations, acknowledging that it can only mediate a subset of the relevant reasons for action in the circumstances it aims to regulate. It should put in place a system that delineates rights, distributes burdens and benefits, and creates roles and positions in a way that leaves enough room for people to orient themselves toward their reasons from personal value and smallscale deontic reasons, not only their political reasons. Modern law has various mechanisms that help ensure this sort of moderation in the exercise of political authority. They include procedural rules that curb demanding lawmaking, substantive limitations on the content of legal rules and their enforcement, a traditional preference for the imposition of negative rather than positive duties, and other formal constraints that impede an authoritative override of the range of reasons individuals have, apart from their political reasons. 76 Shiffrin, supra note 66, at 749. 77 Shiffrin, Egalitarianism, supra note 35, at 275-278. Shiffrin's work on accommodation does not focus on accommodation as a moderating mechanism for political authority. On the contrary, it emphasizes the need to use political power to insulate individual decisions from the pressure of other nongovernmental actors such as employers, landlords, or medical care providers. As we have seen, securing individual capacity to act on reasons from personal value can be a legitimate end of political authority. See text accompanying note 49. However, it is important to remember that the exercise of authority itself constrains the very same capacity. This is the sense in which I use the notion of accommodation here.
There is no reason to think that the choice between these two strategies of moderation should be made wholesale within any modern legal system. The choice depends on several parameters. The first strategy asks more of political institutions, assuming, for example, that these institutions would be able to ascertain the proper demands of small-scale moral reasons that apply to individuals in the circumstances regulated by law. It also presumes a certain homogeneity in individual attachments and the interpersonal moral practices of the subjects of authority. The more pluralist the society and the more controversial the subject matter, the more hesitant we should be in employing the accommodation strategy, even with competent institutions at our disposal.
Of course, the moderation condition is not so much a solution as it is a form of a solution to the challenge of political authority. Moderation is not a panacea, nor is it satisfactory from the perspective of political reasons. Particularly for those who believe that duties from justice can be very demanding, it means that individuals subject to political authority will be unable to fully comply with their political reasons, that political institutions will not be fully just, and that they will not delineate rights and will not distribute burdens and benefits equitably and fairly. How stringent the demands of political authority should be or how much "space" should be left for motivation by other morally relevant reasons are not questions that can be answered wholesale, and perhaps not even questions that have a single right answer in any particular set of circumstances. 78 And yet, sufficient moderation-either by way of accommodation or by abstinence-constitutes a practical condition for the exercise of political authority. In the absence of sufficient moderation, political institutions exceed the authoritative power they can derive from political reasons. Since political authority is a modest, sectional normative power, it must be exercised accordingly if it is to be of practical use to its subjects. Political institutions that are too demanding of the people who live under them, whose directives leave too little room for people to be responsive to their reasons from personal value and small-scale morality, simply fail to exercise political authority in a way that would bind those who live under them. 79

VII. Conclusion
Part of the argument I have offered here builds on a familiar liberal aversion to the conflation between law, understood as the product of political institutions, and morality. Similar sentiments and convictions, close to those I have tried to express here, have led others to believe that the normativity of law is not moral at all but is of a 78 Note that the condition of moderation complements the Political Reasons Account of authority, which already introduces a limit on how authority ought to be exercised and what sort of reasons it can mediate for its subjects. Therefore, despite its flexibility, it differs from Patrick Devlin's "elastic principles" that pay lip service to liberal concerns. Cf. PATRICK DEVLIN, THE ENFORCEMENT OF MORALS (1965), at 16-20. The flexibility does not emerge here, as it does in Devlin's work, from a perceived need to balance private interests against those of the amorphous community. It stems from the need for egalitarian liberal political morality to respect the tension between its prescriptions and certain dimensions of individuality. 79 This does not necessarily mean that such institutions or their directives are illegitimate or unjustified. On the distinction, see Section II. different-perhaps, sui generis-kind. 80 It is important to note that the emphasis on political reasons points away from views attributing an alternative normativity to law. Political reasons are, ultimately, a limited subset of our deontic reasons and have the same normativity as other deontic reasons we have. The challenge of having political institutions in liberal, pluralist societies is to correctly delineate this limited set and ensure that it does not dominate our normative universe.
Raz sometimes referred to his theory of authority-the theory that captures the conceptual heart of the view I criticize here-as "the service conception of authority." I think it is right to insist that the normative power associated with political authority depends on a certain valuable moral service political institutions render to their subjects. The problem with the service conception is that it is not sensitive enough to the unruly nature of our normative landscape, and I think that analogous accusations can be made against others who exaggerate the relations between law and morality. The problem with such ambitious approaches is not only that they pose too high a bar for political authority, assigning it a task it cannot plausibly perform. It is also that they exaggerate the moral role political institutions ought to play, or claim to play, in their subjects' lives.
A modest view of political authority does not resort to solutions of alternative normativity, nor does it ascribe authority to political institutions over everyday, smallscale moral questions or individual attachments. 81 This entails a practical prescription for those exercising political authority. First, if the normative power of political authority is premised on its ability to mediate political reasons to its subjects, then this should be the primary business of those exercising it. Although they are only a subset of our reasons for action, the promise of political authority is that it can help us comply with these reasons, not others. This does not mean that considerations of small-scale morality, or plausible generalizations about people's reasons from personal value, have no role to play in lawmaking. But this understanding should dampen the enthusiasm of those who think that the role of law is to reflect small-scale moral duties and that the ascertainment of these duties is the first part of legal reasoning or the prelude to any law reform. Second, if political authority can plausibly claim to mediate only political reasons for its subjects, then this is the extent of the authority of its directives. This means that law at its best-or even as an ideal-is not only a rough and incomplete guide on what ought to be done but that it is rough and incomplete in a skewed way. Following it helps us comply with some duties but neglect others, act on some reasons and discount the rest. If the directives of political authority are to serve as pro tanto guides on what their subjects ought to do, they must acknowledge this structural incompleteness and be moderate in their demands. 80