Justice Holmes dissent in Lochner bears the mark of an authoritative relationship.

The Nature of Law: A Philosophical Inquiry: Proto-Chapter 3 A General Account of Authority and Its Justifications Stefan Sciaraffa McMaster University April 4, 2012 Draft This case is decided upon an economic theory which a large part of the country does not entertain. If it were a question whether I agreed with that theory, I should desire to study it further and long before making up my mind. But I do not conceive that to be my duty, because I strongly believe that my agreement or disagreement has nothing to do with the right of a majority to embody their opinions in law. Lochner v. New York, 198 U.S. 45 (1905), Justice Oliver Wendell Holmes (dissenting). Justice Holmes dissent in Lochner bears the mark of an authoritative relationship. Holmes holds that certain decisions are not his to make, but rather, they belong to the legislature, and he and his fellow judges must comply with them irrespective of their merit. In the language of theorists who write on authority, Holmes acknowledges a duty to surrender his judgment to the legislature. Putative authoritative relationships of this sort are commonplace. Employees everywhere face and oftentimes accept claims that they should abide by their superiors decisions concerning workplace strategies and tasks irrespective of what they think of such decisions merits. And, the same can be said for citizens vis-à-vis the state and members of various institutions within civil society vis-à-vis such institutions officials. Not only are claims to authority commonplace, they also profoundly impact those subject to them. For example, the legislation that Holmes felt duty-bound to enforce prohibited bakers from working more than 60 hours per week. Had the majority joined Holmes in surrendering judgment to this legislative decision, the effects of this law would have been deeply felt by New York state bakers and their employers alike. Because claims to authority are so pervasive and significant, it is imperative that we be able to reason clearly about whether such claims are justified. Do I have reason to surrender my judgment to the 1

legislature, my employer, and the officials of the various organizations that I have joined, and if so, what are the normative limits of such claims? Answering this question requires understanding what, exactly, such relationships entail and the grounds, if any, that can justify such claims. Here, I develop and defend an account of the content of a putative authority s claims and three conditions under which such claims to authority should be accepted. I do so in the hopes of structuring and guiding our reasoning about the normative reach of the various claims to authority we face. Joseph Raz s account of the content of a putative authority s claims and such claims conditions of justification is highly influential, 1 and it is the most sophisticated treatment of the subject to date. Raz characterizes such claims in terms of the notion of surrender of judgment described above. On his view, for A to claim authority over S entails that A claims that S must surrender his judgment to A. As we shall see in more detail below, Raz further specifies his account of authority by more fully describing what surrendering judgment 1 Prominent theorists who accept his account include Leslie Green (1988) The Authority of the State, Oxford: Oxford University Press, 38ff; Andrei Marmor (2001) Positive Law and Objective Value New York: Oxford University Press, 34-37; John Finnis (1980), Natural Law and Natural Rights, Oxford: Oxford University Press, 233-38, and; perhaps, Brian Leiter (2007) Naturalizing Jurisprudence, New York: Oxford University Press, 128-30. Those who do not accept it, often feel compelled to respond to it at length. See, e.g. Thomas Christiano (2004) The Authority of Democracy, Journal of Political Philosophy 12: 266-90, 278-80 and Wil Waluchow (1994) Inclusive Legal Positivism, New York: Oxford University Press, 136-40. 2

entails. 2 Raz also develops an account of the condition under which claims to authority are normally justified. He enumerates this condition in his well-known normal justification thesis. Together, Raz s account of the content of a claim to authority and his account of the condition under which it is normally justified constitute Raz s service-conception of authority. I join Raz in focusing on the idea of a surrender of judgment. However, I join in this focus with no commitment as to whether we must understand authority solely in terms of this idea. Rather, my thought is that there are commonplace and important relationships that we tend to describe as authoritative and in which one party is expected to surrender judgment to the other. This thought is not inconsistent with the claims that there are yet other important social relationships that we also tend to describe as authoritative that do not feature the claim that one party must surrender judgment to the other or that there are further important features of authoritative relationships in addition to the surrender of judgment claim. 3 2 See (1985) Authority and Justification Philosophy and Public Affairs 14: 3-29, reprinted in (1990) Authority (Joseph Raz, ed.) Oxford: Oxford University Press, 115-41, 119-22 and (1975) Practical Reasons and Norms, New York, Oxford University Press, 62-65. 3 For discussions that characterize authoritative relationships in terms of features other than or in addition to a duty of obedience (e.g. a moral right to enforce directives or the duty to support the authority s directives), see Robert Ladenson (1980) In Defense of a Hobbesian Conception of Law, Philosophy and Public Affairs 9 and William Edmunson (1998), Three Anarchical Fallacies, Cambridge: Cambridge University Press. See Waluchow, Inclusive Legal Positivisim, 136-40 for an account of authority that does not feature one party surrendering judgment to the other. 3

The core argument of this paper is threefold. First, I argue that Raz specifies the notion of a surrender of judgment too narrowly, and I defend a broader specification of this idea, and hence, a more general account of authority than Raz advances. Second, I argue that Raz s unduly narrow account of authority leads him to join Leslie Green in the mistaken conclusion that an authority s ability to coordinate its subjects realization of important goods is not a distinct ground of legitimate authority. Third, I argue that Raz overlooks a third distinct condition under which authority is justified: the condition in which the authority embodies a democratic process that gives each an equal say with respect to the community s rules. In sum, I explicate and defend a general account of authority based in the notion of a surrender of judgment. This general account of authority encompasses at least three forms of authority distinct in the sense that their respective conditions of justification are distinct: Raz s service-conception, coordination, and democratic authority. As we shall see, this argument has an important implication: When assessing whether a claim to authority is justified, we must look to three rather than one possible grounds of such a claim. More pointedly, we must take care to determine which ground or combination of grounds apply to a particular claim, for, as we shall see, the limiting condition of each ground, and hence, the normative reach of each form of authority, differs significantly. 1. Raz s Account of Authority s Structure and the Conditions of Its Justification Raz argues for his service-conception of authority in discussions strewn throughout a number of works. 4 Here, I distill from these works what I take to be his core argument. As I 4 See Practical Reasons and Norms, 62-65; (1979) Legitimate Authority, in The Authority of Law, Oxford: Oxford University Press, 3-27; (1985) Authority and Justification ; (1986) The 4

see it, Raz has developed a powerful two-prong approach to theorizing about authority. First, he enumerates a number of widely accepted platitudes about authority, and he draws from these platitudes further non-platitudinous implications about authority s structure. Second, in light of his account of authority s structure, he adduces what he takes to be the normal or typical condition under which authority so understood could be justified. As a necessary preface to my critique of Raz s account of authority and the defense of the general account I propose here, I describe Raz s account of the structure of an authoritative relationship and its legitimating condition. 1.1 The Platitudes and the Specification of Authority s Structure As Raz puts it: A person has (practical) authority, we have concluded, only if his authoritative utterances are themselves reasons for action. 5 Here, Raz describes an aspect of legitimate authority rather than merely claimed authority. A has legitimate authority over S only if A s directives are themselves reasons for action. A claims authority, but may not have it, insofar as A claims its directives are themselves reasons for S s actions. By contrast, an agent does not have legitimate authority if her subjects reasons for conforming to her utterances are based only in the merits of the directed action irrespective of the putative authority s directive to so act. For instance, if a sailor s only Morality of Freedom, Oxford: Oxford University Press, chs. 2-3; (1989) Facing Up: A Reply, Southern California Law Review 62: 1153-1235, and; The Problem of Authority: Revisiting the Service Conception, in Authority and Justification, Oxford: Oxford University Press, 126-66. 5 See The Morality of Freedom, 35-37. See also The Authority of Law, 11-13. 5

reason to conform to his captain s command to swab the deck is that swabbing the deck would be fair and conducive to the good functioning of the ship, then the captain does not have legitimate authority over the sailor. Rather, the captain has legitimate authority over the sailor only insofar as his directives are themselves reasons for the sailor to perform the directed actions. To put this same thought in the language of theorists of authority, if an authority is legitimate, its directives provide its subjects with content-independent reasons to act as directed, and if A claims authority over S, A claims, among other things, that its directives are content-independent reasons for S. Raz s second platitude holds that authoritative directives are qualitatively distinct from pieces of advice and requests. 6 For example, if I ask you to open the door, my request may very well be a content-independent reason for you to open the door. One reason you have to open the door is the fact that I requested that you open the door. However, we would not say that my request is authoritative in this case. The idea, then, is that authoritative directives must be more than content-independent reasons to act as directed, for otherwise they would not qualitatively differ from pieces of advice and requests. Raz credits Richard Flatham with a third platitude about authority. Flatham states: There has been a remarkable coalescence of opinion around the proposition that authority and authority relations involve some species of surrender of judgment on the part of those who accept, submit or subscribe to the authority of persons of rules and offices. 7 6 See Practical Reasons and Norms, 63-64, The Authority of Law, 11-15, and The Morality of Freedom, 35-37. 7 Authority and Justification, 118. 6

Let us call this platitude the surrender-of-judgment thesis. As Raz characterizes it, this thesis holds that those who surrender their judgment, and hence, accept the claims of an authority accept that they should obey even if their personal belief is that the balance of reasons on the merits is against performing the required act. 8 The surrender-of-judgment thesis incorporates the first platitude described above, for it holds that the subjects of an authority are expected to do what the authority directs for the reason that it so directs. It also incorporates the second platitude, for the surrender-ofjudgment platitude describes how an authority s directives are distinct from requests and pieces of advice. Whereas the recipient of a request or advice is expected to take the request or advice into consideration as one among many considerations relevant to how she should act, a subject of an authoritative directive is expected not only to take the directive to be a content-independent reason to act as directed but is also expected to follow the authority s judgment about how she should act rather than her own. For example, if someone asks George to play Taps at a funeral, George may thereby have a reason to play Taps at the funeral that he should consider along with various other considerations for and against doing so. By contrast, if George is a soldier given an authoritative directive by his commander to play Taps at the funeral, then he not only is expected to play as directed, he is expected not to act on his independent judgment about the merits of playing at the funeral. From George s perspective (insofar as he accepts his commander s claim to authority), whether he should play Taps at a particular funeral is a matter for his superior officer to decide; it is the superior s decision to make and not his. 8 Id. 120. For further discussion of this feature of authority, see Practical Reasons and Norms 63-64, 193. 7

Raz holds that to surrender one s judgment to an authority is to take the authority s directives to be content-independent reasons to act as directed and exclusionary reasons not to act for some of the reasons that weigh for or against the directive. 9 In turn, Raz specifies authority in terms of surrendering judgment. 10 On his view, A has legitimate authority over S only if S has reason to surrender his judgment to A. Thus, Raz holds that A has legitimate authority over S only if A s directives give S a content-independent reason to act as the authority directs and exclusionary reasons not to act for some of the reasons that weigh for or against the directive. 11 1.2 The Putatively Normal Condition of Authority s Justification Why should we ever surrender our judgment to another? Robert Paul Wolf argues that this question admits of no good answer. 12 He holds that [t]he moral condition demands that we acknowledge responsibility and achieve autonomy wherever and whenever 9 Supra n. 3. 10 Id. 11 The fact that an authority requires performance of action is a reason for its performance which is not to be added to all other relevant reasons when assessing what to do, but should exclude and take the place of some of them. (Italics removed). Id. 124. See also, Green, The Authority of the State, 40 glossing Raz s account of authority as follows: A has authority over B only if the fact that A requires B to phi gives B a reason to phi and a reason not to act on some of the reasons for not phi-ing. 12 Robert Paul Wolff, In Defense of Anarchism, Berkeley: University of California Press. 8

possible. 13 Wolff adds that [t]he responsible man is not capricious or anarchic, for he does acknowledge himself bound by moral constraints. But he insists that he alone is the judge of those constraints. 14 Thus, for Wolff, to act autonomously is to act in accordance with one s judgment about the requirements of morality and, presumably, all other reasons that apply to one. From these premises, Wolff argues that legitimate authority is incoherent, for it incoherently posits the possibility of an agent justified in violating the moral duty of autonomy, i.e. an agent justified in surrendering judgment about reasons requirements to an authority. One of Raz s signal contributions to political philosophy is a powerful argument that accepts Wolff s basic premise, yet denies his conclusion. Raz accepts that moral agents are bound to act in accordance with their understanding of the reasons that apply to them. However, Raz adds that Wolff overlooks a crucial distinction between two levels of reasons. 15 Some reasons are reasons for acting and some reasons are reasons not to act on other reasons. On Raz s view, a moral agent might have an all-things-considered reason not to act for some of the reasons that apply to her. In such a case, a moral agent would act in accordance with her reasons by not acting on her judgment with respect to the requirements of some of the reasons that apply to her. Moreover, in doing so she may very well act in accordance with her understanding of the overall requirements of her reasons, for she might rightly judge that she has all things considered reason not to act in accordance with her understanding of some of her reasons. Raz argues that justifying authoritative relationships 13 Id. 14. 14 See id. 12. 15 Raz directly replies to Wolff in Legitimate Authority, 26-27. 9

requires showing that the subjects of the authority are subject to such second order exclusionary reasons. Prefatory to such a showing, Raz distinguishes between conforming to and complying with one s reasons. For Raz, to conform to one s reasons is to act as one s reasons require. To comply with them is to act out of one s understanding of their requirements but not necessarily to act as they require, for one may be wrong about one s reasons. 16 Raz s idea is that (1) when it is more important to conform to one s reason than to comply with them, authority is justified insofar as (2) the authority s subject is likely better to comply 17 with reasons which apply to him (other than the 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. 18 Famously, Raz refers to condition (2) as his normal justification thesis. In sum, Raz argues that when conditions (1) and (2) are met, a subject does not violate a moral duty nor does she act irrationally by treating an authority s directives as content-independent and exclusionary reasons (i.e. by surrendering judgment to the authority), for she likely better conforms to the reasons that apply to her by doing so. Thus, Raz argues that under certain conditions, conditions (1) and (2), an agent has reason to 16 See Raz, Practical Reason and Norms, 190ff. See also Raz, Authority and Justification, 133 and discussion of condition (1) under the name of the independence condition in Revisiting the Service-Conception, 136-40. 17 Raz confusingly uses the term comply here. To track his compliance/conformity distinction, he should have used the term conform. 18 Raz, Morality of Freedom, 53. See also, Raz, The Problem of Authority, 136-37. 10

surrender judgment to another. Note that such a surrender of judgment is not a complete surrender of judgment, for the agent who acts on his reasons to surrender his judgment need not surrender his judgment with respect to his judgment whether he should surrender judgment. In other words, Raz purports to show that surrendering judgment can be rational, and hence, accepting an authority s claims that one must surrender one s judgment to the authority may also be rational. 2. Coordination Authority John Finnis states: Authority (and thus the responsibility of governing) in a community is to be exercised by those who can in fact effectively settle co-ordination problems for that community. ***The fact that the say-so of a particular person or body or configuration of persons will in fact be, by and large, complied with and acted upon, has normative consequences for practical reasonableness; it affects the responsibilities of both ruler and ruled, by creating certain exclusionary reasons for action. 19 Below, I explicate Finnis thesis in more detail and explain why Raz rejects it. Moreover, I argue that though Raz correctly rejects part of Finnis thesis i.e. coordination agents directives are not sources of exclusionary reasons by virtue of their capacity to coordinate Raz mistakenly concludes from this observation that coordination agents are in no case authoritative by virtue of their capacity to coordinate. Note that the key plank in my defense of the distinctiveness and legitimacy of coordination authority is the claim that Raz 19 Finnis, Natural Law and Natural Rights, 246. 11

mistakenly holds that an authority s directives necessarily provide its subjects with exclusionary reasons. I argue that though the subjects of a legitimate authority necessarily have a reason to surrender judgment to the authority, such reason does not entail that the authority s directives are exclusionary reasons. In short, I argue that Raz specifies the structure of an authoritative relationship too narrowly and, as a result, fails to see that an authority s ability to coordinate is a condition of the authority s justification that is distinct from and no less likely to obtain than the condition set out in Raz s normal justification thesis. 2.1 Modified Game-Theoretic Coordination Problems and Razian Authority To help understand Finnis basic point above and why Raz and Green reject it, consider the following very simple modified coordination game. 20 (Alfred, Bert) Pattern A (left side) Pattern B (right side) Pattern A (left side 1 st, 1 st 2 nd, 2 nd ) Pattern (right 2 nd, 2 nd 1 st, 1 st B side) 20 As originally formulated, a pure coordination game models the strategic relationship among actors with a particular configuration of preferences. For present purposes, I have moralized the coordination game, defining each cell of the game-matrix in terms of value. 12

In the foregoing matrix, each cell represents pairs of actions or courses of actions that Alfred and Bert might take. The action-pairs in the top-left and bottom-right corners are coordination points in which Alfred and Bert each conform to a coordinated pattern of behavior that realizes a good that can only be realized through coordination. The standard illustrative example of a coordination game concerns the good of safety on the public roads. Alfred and Bert can safely navigate the roadways together if they both drive on the left (topleft cell) or both drive on the right (bottom-right cell). The top-right and bottom-left action pairs are non-coordination points in which Alfred and Bert court disaster by driving on different sides of the road. In the diagram above, I rank the value that Alfred and Bert respectively realize in the coordination cells more highly (1 st ) than the value they realize in the non-coordination cells (2 nd ). The main impediment to coordination in a coordination game is mutual uncertainty within the group about the coordination point others will follow. Alfred will realize the greater value by driving on the same side of the road as Bert, but he does not know what side Bert will choose. The same can be said of Bert vis-à-vis Alfred. In some cases, uncertainty of this sort can be resolved by custom. For example, it may become established over time that drivers drive on the right side of the road. A second possibility is that a coordinating agent establishes the coordination point for the group. For example, Sweden might pass a law requiring that all drivers on Swedish roads drive on the right. Finnis idea is fourfold. First, certain very important goods, such as safety on public roads, can only be realized if the members of the relevant group coordinate their action. Let us call these coordination goods. Second, in some cases, the coordination necessary to secure the relevant coordination good can be best or, perhaps, can only be achieved and maintained by a coordinating agent through its directives to the group. Third, in such cases, the 13

members of the relevant group have a reason to follow the directives of the coordinating agent. Thus, the coordinating agent has a kind of authority grounded in the value of the relevant coordination good. Though each participant in a coordination game has reason to do as the coordination agent directs, this is not yet to say that each participant has reason to treat the coordination agent s directives as authoritative. On Raz s view, for the coordination agent to have legitimate authority, its directives must also provide the participants with exclusionary reasons; however, a coordination-agent does not necessarily provide such reasons. 21 As Leslie Green puts it: Conventionalism presupposes that conformity with convention is always conditional: each should conform if and only if he expects the others to do the same. When he does expect this, he has no reason to do anything else. It is part of the concept of authority, however that it is binding. 22 To see Green s idea, consider again the coordination problem of picking which side of the road to traverse. If a coordinating agent chooses the right side of the road for the group, then each member of the group (in this case, Alfred and Bert) can be confident that the group s other members will drive on the right. Hence, by virtue of the coordinator s directive, each individual member of the group has reason to drive on the right, namely, she will achieve the good of public safety on the roads by doing so. Thus, the coordinator s directive alters the reasons that apply to the players in the coordination game. Before the 21 See Green, The Authority of the State, ch. 4. Raz concedes this point at Facing Up, 1189. 22 Green, The Authority of the State, 121. Note that in Green s terminology, a binding norm provides content-independent and exclusionary reason to its subjects. 14

directive, considerations of public safety did not require that each member of the group drive on the left or the right; rather, either choice would have been equally perilous. After the directive, and because of the directive, such considerations weigh in favor of driving on the right. However, and this is Green s key point, the coordinating agent s directives are not exclusionary reasons in this case. Such directives alter the reasons that apply to the coordinating agent s subjects, but they do not exclude any reasons. 23 Because such directives are not exclusionary reasons, Raz and Green conclude that they are not authoritative, for on their view, a necessary feature of an authoritative directive is that it be an exclusionary reason for its subjects. 2.2 The Authority of Coordination Though coordination agents do not provide their subjects with exclusionary reasons by virtue of their capacity to coordinate their subjects and, hence, do not thereby possess Razian service-conception authority, they nonetheless do have a kind of authority by virtue of this capacity. To see this, it helps to add some complexity to the simple coordination game we are considering. Many actual coordination games comprise many actors, perhaps millions, and such games range over a virtually infinite number of coordination points that 23 Though Raz agrees with Green that coordination agents are not legitimate authorities by virtue of their capacity to coordinate their subjects, he would add that coordination agents oftentimes are authoritative because their subjects are likely to do better with respect to certain reasons i.e. their reasons to participate in justified coordinated social behavior by following the authority s directives rather than acting directly on those reasons. Joseph Raz, Facing Up: A Reply, 1991-92. 15

vary in the degree to which they realize the relevant coordination good. To take a moderately complex example, consider an orchestra. Here, we have about 100 players in the game who seek to play music together as an orchestra. The coordination problem that such players face is to decide what piece of music to play, and, once this is determined, how to interpret the piece. Thus, the players face a virtual infinity of coordination points, for there are countless numbers of pieces any given orchestra might play multiplied by countless visions of how any particular piece should be played. Failure to coordinate with respect to the foregoing would either completely undermine the possibility of playing any piece at all as an orchestra, or it would result in a cacophonous, bedraggled joint effort. Conductors solve such coordination problems; they choose the music and provide the orchestra s members with a unifying vision of how the chosen piece should be played. As we have seen, Green and Raz would hold that the conductor s directives are not authoritative, for though the conductor s directives alter the player s reasons i.e. once the conductor has selected a piece and a vision for playing it, considerations of the value of playing a harmonious orchestra piece favor playing the selected piece in accordance with the conductor s vision they do not provide the members with exclusionary reasons. However, as we shall see below, the conclusion that the conductor s directives are not authoritative is mistaken. To see this, we must revisit the basis of Raz s account of authority, his formulation of the surrender of judgment thesis. As he precicifies the thesis, it holds: S surrenders her judgment to A if and only if she takes A s directives to be reasons to act as directed, and she takes such directives to be reasons not to act for some of the reasons that weigh for and against her so acting. The foregoing formulation not only clarifies the folk platitude concerning the surrender of judgment, it also adds content to it. Moreover, this addition is not innocuous. 16

To see this, consider the following alternative account of surrender of judgment that I propose. S surrenders her judgment to A if and only if S would acknowledge that considerations C weigh for and against the merits of A s directives D to S, S holds that S has reason not to act (directly) for considerations 24 C, and S takes herself to have a content-independent reason to conform to A s directives D. Whereas Raz holds that S surrenders his judgment to A only if S takes A s directives to be exclusionary reasons, the general account I propose does not require that S take A s directives to be exclusionary reasons; rather it more generally holds that that S must take himself to have a reason not to act for certain considerations C, considerations that S would acknowledge determine the merits of A s directives to S, yet it is indifferent as to whether 24 I use the term considerations rather than reasons here to mark a peculiar feature of coordination games. From the perspective of the coordinator, certain considerations are reasons that weigh for and against its subjects acting in a particular way, whereas from the subjects perspectives, they are not reasons, for the subjects are unable to act as they require. For example, from the individual player s perspective it might not be possible for him to play Mahler in concert with his fellow players because his fellow players will not coordinate with him on this. Hence, from his perspective, the considerations that weigh for and against his playing Mahler in concert with his fellows are not reasons to do so. By contrast, from the conductor s perspective, the individual player could play in concert with his fellow players Mahler, Bruchner, or The Scorpions, among many other possibilities, for the conductor can secure the necessary coordination at her pleasure; hence, from the conductor s perspective these considerations are reasons that weigh for and against her players playing Mahler. 17

A s directives are the source of such exclusionary reasons. It would seem, then, that Raz has introduced a detail into his specification of surrendering judgment. Moreover, Raz s introduced detail has a significant theoretical consequence. To see this, consider again the conductor and her orchestra. A player might disagree with the conductor s judgment about which piece would be best for the orchestra to play or the best style to play it. She might hold that for various considerations, some alternative pieces or interpretive visions would have been better than the conductor s selection. Nonetheless, such a player has reason not to act on her own judgment with respect to such considerations, but rather has reason to follow the conductor s judgment instead. Her reason not to act on her own judgment with respect to such considerations is that she cannot secure the coordination with others that is requisite to the possibility of so acting. For example, a player cannot act for the considerations that favor the orchestra and her playing Bruchner together unless she can secure that her fellows will join her in playing Bruchner. The player has reason to follow the conductor s judgment with respect to these considerations, for the conductor is capable of securing the coordination of the group requisite for the realization of the relevant coordination good (in this case, the performance of a harmonious orchestral piece). 25 25 Some have argued that a coordination agent is not authoritative because its directives to Φ are not content-independent reasons to Φ, for the subject s reason for Φ-ing in such cases is that doing so contributes to the realization of the relevant coordination good and not that the agent directed the subject to Φ. See, e.g. Donald Regan (1989) Authority and Value: Reflections On Raz s Morality Of Freedom, Southern California Law Review 62: 995-1095, 1024-25. Note that a similar worry afflicts the normal justification thesis. One might say that 18

Whereas the individual player in the foregoing example has reason to surrender her judgment to the conductor according to the general account of this notion, according to the Razian account, she does not. The reason for this difference is that according to the Razian account, S has reason to surrender judgment to A only if A s directives provide S with exclusionary reasons, yet in the present example, A s directives are not exclusionary reasons. The player s reason not to act for certain considerations is not the fact that A gave her a directive based on such considerations; rather, it is that she cannot secure the coordination requisite to acting on them. By contrast, the general account is indifferent as to the ground of the exclusionary reason, and hence holds that the player in the present example has reason to surrender her judgment despite the fact that the source of her exclusionary reason is not the authority s directive but rather is the fact that she is not capable of acting on the excluded considerations. a putative service-conception authority is not authoritative, for its subject s reason to Φ is not that the authority directed her to Φ but rather that she likely better conforms to the reasons that apply to her by following the authority s directive to Φ rather than acting directly for such reasons. The response to each of these worries is the same. The fact that an agent s directives secure the group s coordination of its subjects with respect to the relevant coordination good enables it directives as reasons for its subjects. Similarly, the fact that the subject will better conform to her reasons by following the authority s directives rather than acting directly for such reasons enables the authority s directives as reasons for its subjects. See Jonathan Dancy (2004) Ethics Without Principles, New York: Oxford University Press, 38ff for a discussion of enabling conditions. 19

The point in the foregoing example generalizes. According to the general account of surrendering judgment, and pace the Razian account, the subjects of coordination agents have reason to surrender their judgment to the coordination agent insofar as the agent is able to secure the coordination necessary for the realization of important coordination goods and the subjects are not similarly able. Hence, according to the correspondingly general account of authority and pace the Razian account, coordination agents are authoritative with respect to their subjects by virtue of their capacity to secure group coordination. We should accept the general account of authority over the Razian account because it is more general. I here rely on a methodological premise that general, less restrictive characterizations of platitudes that inform the analysis of concepts, such as the concepts of authority and the surrender of judgment, should be favored over more specific characterizations unless reasons can be given for favoring the narrower, more restrictive specification over the more general one. 26 More specifically, the proponent of the more restrictive characterization must point to implicit or explicit commitments of the users of the concept in question to justify the attribution of the more restrictive features to the concepts in question. To bolster this claim, it helps to recapitulate the structure of the relationship between subject and coordination agent in the coordination cases at issue. In the coordination cases described above, the subject has reason to acknowledge that there are considerations that determine how a group to which he belongs must jointly act, and hence, determine how he as a participant in the group, must act. Moreover, the 26 Cf. Terence Horgan and George Graham (1991) In Defense of Southern Fundamentalism, Philosophical Studies 62:107-34, 121-22 for a defense of a similar methodological principle. 20

subject has reason not to act upon such considerations, for he cannot act on them. Instead, the subject should act in accordance with the judgment of some other agent, the coordination agent, concerning the requirements of such considerations, for this agent can secure the coordination of the group in accordance with such judgment. It seems capricious to hold that the subjects in such cases do not surrender judgment on the ground that the putative authority s directives (the coordination agent s directives) are not themselves sources of exclusionary reasons. Why must these relationships meet this further condition to count as cases of surrender of judgment and authority? Absent a good answer to this question, and I can think of no such answer, we should count such cases as authoritative in accordance with the general account. Allow me to conclude this section with two final points about coordination authority. First, the disagreement about whether coordination agents are authoritative by virtue of their capacity to coordinate is significant, for a recurrent feature of the human condition are situations in which a group can realize some important coordination good via a number of approaches that are each mutually incompatible, and a coordination agent is uniquely positioned to select and set the approach the group takes. To name just a few, instances of this sort of coordination game include playing well as a sports team, executing a business strategy, maintaining public safety on the roads, providing fair and efficient public education, creating systems of social insurance, maintaining public defense and order, and creating a fair and efficient market for goods and services. Second, a limiting condition of a coordination agent s authority is that her judgment must not be so bad that it would be better for its subjects not to coordinate at all than to coordinate as directed. So long as the coordinator s judgment clears this bar, its subjects have reason to follow its judgments about how they should act, irrespective of their views of such judgment s relative merit. For 21

example, the player should play Mahler even if he thinks the conductor should have chosen Bruchner, unless Mahler is such an atrocious choice that it would be better not to coordinate at all. 2.3 A General Conception and Two Forms of Authority We can now see two ways in which Raz s conception of authority is too narrow. First, though the subjects of an authority are expected to surrender judgment to the authority, surrendering judgment need not be specified in terms of taking the authority s directives as exclusionary reasons. Rather, we should replace Raz s characterization of authority s structure with a broader conception one based on the general account of a surrender of judgment: A has legitimate authority over S only if S has reason to surrender judgment to the authority, which entails that (i) considerations C weigh for or against the merits of A s directives D to S (ii) S has reason not to act (directly) for considerations C, and (iii) S has a content-independent reason to conform to directives D. Second, Raz mistakenly holds that the condition he enumerates in the normal justification thesis is the typical or most commonplace condition of authority s justification. On his view, the normal way to justify authority is to show that the subject is likely to conform better to reasons that apply to her by following the authority s directive rather than following those reasons directly. However, we have seen above a condition of authority s justification that seems no less likely to obtain than the condition enumerated in the normal justification thesis, for it is a recurrent feature of the human condition that certain important goods can only (or best) be realized by coordinated group action secured by a centralized person or institutional body, i.e. a coordination agent. Let us call this feature the coordination 22

condition. When this condition obtains, the coordinating agent s directives (those that institute the relevant important good) are reasons for its subjects to surrender judgment to the coordination agent with respect to the precise form the group s coordinated actions, including the subjects actions, shall take. One might object that Raz s normal justification condition and the coordination condition are not distinct, for one might think that the coordination condition implies the condition described in the normal justification thesis. According to this objection, when the coordination condition obtains, members of the group in need of coordination will likely conform better to certain reasons that apply to them specifically, achieving the relevant important goods that only be realized via coordinated activity by following the authority s directives rather than trying to follow such reasons directly. To see why this objection fails, it helps to contrast the kinds of service that Razian and coordination authorities provide their subjects. Raz labels his conception of authority the service conception because on this view, a legitimate authority serves its subjects by providing directives that enable them to better conform to the reasons that apply to them. However, the authority can only provide the service of improving the subjects conformity with their reasons insofar as its subjects refrain from acting directly for such reasons. By contrast, the directives of a coordinating agent serve their subjects by enabling them to act for certain considerations that they could not have acted upon were it not for the authority s coordinating directives. To see this, recall that an orchestra s players are not able to act directly for the considerations that weigh for and against the various pieces of music and interpretive visions that the orchestra might execute because they cannot secure the coordination requisite to so acting. However, once the coordination authority issues a directive, the individual player is able to act for a subset of 23

these considerations, the considerations that weigh for or against performing in concert with others the piece the conductor has chosen in the style the conductor has selected. In other words, she is able to act directly for considerations for which she could not have acted were it not for the coordinating agent s directives. In sum, when a coordination agent coordinates its subjects realization of coordination goods, it serves them by enabling them to act for considerations that they otherwise could not have acted upon. This service is distinct from and does not entail the service characteristic of Razian authority that is enumerated in the normal justification thesis. Thus, the condition under which coordination authority is justified does not entail the condition enumerated in Raz s normal justification thesis, and, hence, coordination authority and Razian service-conception authority are distinct forms of authority. 3. Democratic Authority Democracy refers to a decision-making process in which each member of the relevant group has an equal say with respect to the group s decisions. This equal say takes the form of an equal vote with respect to the decisions before the group. 27 Positions that win a majority of the group s votes are the democracy s decisions. My question here queries whether and under what conditions a democracy s decisions are authoritative. The key to answering this question lies with a central concern of democratic theorists: What is the nature of the underlying value of democracy? 27 The arguments in this section are intended to apply to representative democracy as well. 24

Many theorists of democracy hold that its value is, at least in part, instrumental. That is, democracy is valuable because it tends to produce very good rules 28 or because it has felicitous energizing and affirming effects 29 on the population so governed. Here, I focus on the intrinsic value of democracy. Democracy is intrinsically valuable insofar as it is valuable for its own sake and not merely for the good decisions it produces or the good effects it has on the population so governed. A number of theorists hold that democracy enjoys intrinsic egalitarian value. On this view, democracy is valuable because it is a form of governance that provides every competent adult in the relevant community with an equal say with respect to the rules that govern the community, and, as such, this form of governance embodies appropriate equal respect for the competent adults of the democratically governed society. 30 Let us call this thesis the democratic egalitarian thesis. The proponents of this thesis differ with respect to their views of the considerations that support the democratic egalitarian thesis. For example, Thomas Christiano argues at length that providing each with an equal say with respect to the 28 See, e.g., Richard Arneson, "Democratic Rights at National and Workplace Levels," in The Idea of Democracy, ed. by David Copp, Jean Hampton, and John Roemer (Cambridge: Cambridge University Press, 1993), pp. 118-148. 29 See Stephen Holmes (1993) Tocqueville and Democracy in The Idea of Democracy. 30 See e.g., Thomas Christiano, The Constitution of Equality; Jeremy Waldron (2002) Law and Disagreement, New York: Oxford University Press, particularly ch. 11, section 3; Harry Brighouse (1996) Egalitarianism and the Equal Availability of Poltical Influence, Journal of Political Philosophy 4: 118-41, and; Christopher G. Griffin (2003) Democracy as NonInstrumental Just Procedure, Journal of Political Philosophy 11: 111-21. 25

rules that govern the community is a way of manifesting equal appropriate respect for each because so doing equally advances the interests of each in a public way, 31 whereas Jeremy Waldron seems to take the democratic egalitarian thesis as a foundational normative thesis. 32 Here, I set to the side debates about the further grounds of the democratic egalitarian thesis, and I argue that those who accept this thesis should also accept the authority of democracy. 33 Moreover, I argue that this ground of authority is distinct from the ground of Raz s service-conception authority, and, hence, is a distinct form of authority. 3.1 The Argument for Democratic Authority The democratic egalitarian defense of the authority of democracy holds that it is intrinsically valuable to surrender one s judgment to the democratic institutions that determine one s community s rules, for by doing so one acknowledges in one s deliberative practices and actions that certain matters are for the community to decide in a process that gives each an equal say, and, hence, one manifests equal respect for one s fellow citizens. Conversely, failing to surrender one s judgment to the democratically organized group with 31 See Christiano, The Constitution of Equality, 85 for this argument. 32 See Thomas Christiano (2000) Waldron on Law and Disagreement, Law and Philosophy 19:513-43 for criticisms of Waldron for not providing further justification of this democratic egalitarian thesis. 33 Christiano and Scott Hershovitz defend a kind of authority of democracy based on its intrinsic value. However, these authors do not defend the claim that such authority entails a surrender of judgment, as I do here. See Christiano, The Constitution of Equality, 241 and Scott Hershovitz (2003) Legitimacy, Democracy, and Razian Authority, Legal Theory 9: 201-20. 26

respect to certain issues is to deny the equality of one s fellows. It is to indefensibly hold that one s judgment and will with respect to such matters should count for more than the judgment and will of one s fellows. To develop and strengthen this democratic egalitarian defense of democratic authority, it is helpful to consider three powerful challenges to it. Wolff supplies the first challenge to the democratic egalitarian defense of democracy s authority. Recall Wolff s claim that all agents have a moral duty to act autonomously, meaning that they have a duty to act in accordance with their own judgment of reason s requirements rather than to defer to a putative authority s judgment. Thus, the Wolffian anarchist would reject the egalitarian democratic defense of democratic authority, for the individual is morally obligated to act on his own judgment rather than to defer to the collective judgment of democratic institutions. The egalitarian democrat s strongest response to this challenge relies on the notion of interdependent interests. 34 Some decisions that an individual makes have significant effects primarily on her own interests and only lesser if any effects on the interests of others. Perhaps, an individual s choice of religion or associates are such decisions. Other decisions effects are felt more widely; they significantly affect the fundamental interests of many. It is with respect to decisions such as these that the democratic egalitarian case for the authority of democracy is strongest. To illustrate this point, consider the case of a landowner, Chris, who would like to dam a waterway in order to irrigate the farmland that he cultivates. If Chris dams the waterway, he would subject everyone downstream to flooding risks. He calculates the risk that the others will bear and weighs that against his personal gain 34 See Christiano, The Constitution of Equality, 78-83 for a similar notion of fundamental interdependent interests. 27