We do not choose our rulers by the toss of a coin—tossing coins is associated with the low-status activity of gambling—but who would dare to claim that the world would be in a worse way than it is if rulers had from the beginning of time been chosen by the method of the coin?
— J. M. Coetzee,Diary of a Bad Year
It is easy to think that democracy rests on shaky foundations.Footnote 1 Anyone who has been given pause by the argument from expertise at the centre of Plato’s Republic will have a vivid sense of democracy’s natural implausibility. If it is uncontroversial that only those people with the requisite expertise should be permitted to operate airplanes, say, or perform root-canal surgeries, or engineer skyscrapers, why should not the analogous political experts (whoever they are) be justified in having the authority to make crucial decisions about state power and policy?Footnote 2 A common answer is that there are no such experts, but this answer does not vindicate democracy as such. For one could still accept the anti-democratic premise that if there were experts of the appropriate political kind, they would be justified in having the right to rule. The answer makes democracy look like a default position, a politics of retreat, the only feasible system of government left in a world without Philosopher Kings.
Joseph Raz’s view is that we ought to accept democratically made decisions as binding because democracy has (or may sometimes have) legitimate authority, by which he means that it has the moral power to require action.Footnote 3 Democratic governments are vindicated on this view because they are deemed, like all legitimate authorities, “more likely to lead to action supported by sound reason than any alternative method of deciding what to do.” (Raz Reference Raz1994, 116) This is an instrumentalist justification of democracy. It is energized by the premise that democracy leads, by and large, to good government.Footnote 4
Democracy is best understood as a political system allowing individuals opportunities for informed participation in the political process whose purpose is the promotion of sound decisions. Democracy is justified inasmuch as it is necessary to serve the well-being of people. It shares the general structure of authority and relies, for its legitimacy, on its ability to deliver sound decisions. (Raz Reference Raz1994, 117)
I want to argue here that Raz’s “service conception” of authority, while brilliant and laudable in a number of other respects, suggests an incomplete account of the nature and source of democratic authority. While an ability to reliably deliver sound decisions (if it has it) may well provide a partial explanation of the ways in which democratic governments enjoy legitimate authority, it will be my argument that there are other non-instrumental reasons for regarding democratic decisions as authoritatively binding that Raz’s service conception simply cannot accommodate. If the fundamental questions for a theory of legitimate (political) authority are, as Raz frames them: “What can justify holding some people duty-bound to obey others?” and “Under what conditions can some have a right to rule others?” (Raz Reference Raz1994, 356), it seems doubtful that acceptable answers to either question can be wholly divorced from non-instrumental procedural considerations. If democracy has legitimate authority, if it does in fact yield moral reasons for action and compliance, then that is to some extent because of the morality of its procedures. Readers will not find a full argument for that conclusion here, but it will suffice to mention in passing one serious obstacle that any instrumentalist account of political authority must overcome. It is widely believed that the burdens of political justification are uniquely stringent since the exercise of political power involves coercion, and coercion requires for its legitimacy, justification “at least in principle, to every last individual”. (Waldron Reference Waldron1993, 37) The main idea is that no citizen shall be legitimately coerced in the name of a comprehensive ideal he or she could reasonably reject. This raises a profound challenge for instrumentalist accounts of political authority. Instrumentalist theories, those that say that a political authority is legitimate in virtue of its tendency toward optimal outcomes, will not have a justification that that is reasonably acceptable to all qualified points of view. Given what Jeremy Waldron has called “the circumstances of politics,” it is assumed that reasonable people will not only disagree substantively about optimal political outcomes, but also (a fortiori) about who can rightly claim the relevant political expertise, and about the kinds of laws, institutions, and policies that genuine experts would recommend on various occasions. “Correctness theories [of democratic authority],” writes David Estlund, “cannot meet the qualified acceptability requirement.”(Estlund Reference Estlund2009, 99) This is by no means some daring new idea. It is merely a reaffirmation of the quintessential liberal requirement that political legitimacy lies (in whole or in part) with the consent of the governed. Beyond any expertise claimed on its behalf, the objection goes, the argument for democracy’s legitimate authority involves showing that it is a system of government that is generally acceptable in the right way.
Whatever is to be made of this kind of argument, the more important point is that Raz cannot even entertain the possibility that non-instrumental considerations play any role in establishing the legitimate authority of democratic governments.Footnote 5 The very shape of the service conception, I shall be arguing, precludes that possibility. This is significant not only because the Razian theory of authority has become so influential among scholars and theorists,Footnote 6 nor simply because, as I believe, its applicability to democracy has been insufficiently explored.Footnote 7 Most importantly, it is because the service conception is advertised as a doctrine that purports to explain the nature of authority in general—conceptually, tout court, everywhere, in all contexts. It would surely be much more than an innocuous shortcoming, then, if, as I shall be arguing, it cannot properly or fully account for the authority of democratic government. Indeed, the service conception’s inability to satisfactorily account for the authority of democracy would raise serious questions about Raz’s seminal work on legal authority and the “exclusive” legal positivism he has defended on that basis.Footnote 8 I will not explore those wider implications of the argument here, however. Rather, I shall be arguing that if we want to get clear about the theoretical underpinnings of democratic authority—about the bases upon which democratic decisions constitute moral reasons for action and compliance—Raz’s service conception is not fully up to the task.
1. The Service Conception of Authority: A Brief Overview
It is commonly assumed that “having authority” is interchangeable in meaning with “being in charge”. On this view, if Jones can make Smith eat potato chips, Jones has the authority to make Smith eat potato chips. Raz repudiates this (“Hobbesian”) conception on the grounds that it mistakenly runs authority together with power and influence.Footnote 9 If Jones gets Smith to eat potato chips by bribing or threatening him, it would be eccentric to claim that Jones thus exercises authority over Smith.Footnote 10 On Raz’s view, merely being in charge is not even sufficient for de facto authority, let alone its morally superior legitimate counterpart.Footnote 11De facto authority “either claims to be legitimate or is believed to be so and is effective in imposing its will…perhaps because its claim to authority is recognized by many of its subjects.”(Raz Reference Raz1994, 215) Legitimate authority by contrast is a kind of moral authority, soundly mediating between people and the right reasons for action that apply to them. We learn nothing about legitimate authority, in other words, by observing what passes for or what is merely claimed to be legitimate authority. On the contrary, the legitimacy of an authority does not depend on anyone’s recognition or acceptance; an authority can be legitimate even when its claim to legitimacy is widely rejected. As Raz notes, “[Authority] is not about what people think it is like to have authority or to be subject to it, but of what it is to have it or be subject to it.”(Raz Reference Raz2006, 1006) More particularly, to say that Smith has authority over Jones on certain matters is to say that on those matters if Smith tells Jones to do something, then Jones is, for that very reason, required to do it. The force of “required” here is moral: a person ought to obey an authoritative command even when they have the power not to—much as one ought to follow sound advice (all things being equal), even when they possess the freedom to ignore it. The service conception casts legitimate authority in the same light: “It is the truth or soundness of the decisions which counts ultimately. Truth and soundness provide the argument for the legitimacy of the authority. Honest belief is merely a necessary means to the goal. An authority is legitimate only if its honest belief is, at least in the long run, a reliable indicator of the correct course of action to take.” (Raz Reference Raz1994, 116)
As its name suggests, the service conception draws upon the idea that the role and primary normal function of legitimate authorities is to provide a service. It is grounded, says Raz, in two mutually reinforcing theses: the dependence thesis and the normal justification thesis. The dependence thesis requires that, “all authoritative directives should be based on reasons which already independently apply to the subjects of the directives and are relevant to their action in the circumstances covered by the directive.” (Raz Reference Raz1986, 47) That is, authoritative directives should be based on what Raz calls dependent reasons—reasons which antecedently and independently apply to their subjects. Consider a case in which two people refer a dispute to an arbitrator. According to the dependence thesis, the arbitrator’s decision should be based on all the reasons pertinent to the dispute, reasons which already apply to the disputants. The arbitrator’s decision is not one more reason to be counted alongside the others when pondering which course of action right reason would support. On the contrary, the arbitrator’s decision is meant to be based on the other reasons, to depend on them, to “sum them up and to reflect their outcome.”Footnote 12 (Raz Reference Raz1986, 42) This is a normative thesis about how authorities should use their powers. Whether and to what extent existing authorities actually meet this requirement is another question entirely.
Democracy may very likely do well on this score—though I won’t argue for that here—given that its decisions are at bottom nothing more than the aggregative value of a citizenry’s votes. Yet as Raz himself points out, while this makes a democracy an excellent tracker of expressed preferences, it may render it a poor tracker of actual reasons.
The reason a vote counts is different from the reason for which it is cast. My reason for voting to abolish capital punishment is that it is morally wrong. But the “system’s” reason for counting my vote is that I cast it. The system is alleged to be indifferent to my reason, indifferent to the true moral merit of capital punishment and concerned only with counting heads. (Raz Reference Raz1994, 115)
Whatever is to be made of such concerns, it is enough to point out that, even if it does not do so perfectly, and even if it fails to differentiate satisfactorily between preferences and reasons, democracy appears to satisfy the requirements of the dependence thesis at least as well as any other feasible alternative.
The service conception’s second feature is more germane to our theme. It is expressed by the normal justification thesis, which holds that:
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… 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. (Raz Reference Raz1986, 53)
Raz has us consider the case of a person who accepts his friend’s advice because he fears his friend will be hurt if he does not. Raz admits that this may well be a fine reason for accepting the advice—there is much to be said for not hurting one’s friends after all—yet it is not the normal reason. “The normal reason for accepting a piece of advice is that it is likely to be sound advice. The normal reason to offer advice is the very same.” (Raz Reference Raz1986, 54) The legitimacy of an authority is established, then, when it can be shown that the purported subject is more likely to act in accordance with the right reasons that apply to her, if she accepts the directives of the authority as binding and tries to follow them. Roughly put, an authority is legitimate if an agent will do better by following its directives than by working out what to do on her own. As Waldron explains, “I recognize the State of New Jersey as an authority on traffic speed if I reckon I will drive more safely and efficiently on the Princeton Turnpike by following the State’s instructions and speed limits than by trying to figure out the optimal mix of safety and efficiency in every locality for myself.” (Waldron Reference Waldron1999, 84-5)
So-called coordination problems notwithstanding, there is something eerily Platonist about the normal justification thesis.Footnote 13 This is because the justification of legitimate authority is inseparable—for Raz as for Plato—from its purported epistemic reliability. Just like Plato’s argument from expertise, the service conception is animated by the premise that the likelihood of its knowing better is a condition for the proper identification of an authority. It is important, however, not to press the comparisons between Raz and Plato too far here. For one thing, Raz is a democrat and Plato is not.Footnote 14 Second, Raz’s value pluralism stands in stark opposition to Plato’s monistic view according to which all worthy ends fit harmoniously together—born as all good things are from the same eternal source. And diametrically unlike Plato, for whom the good exists independently of anyone’s attempt to apprehend it, Raz defends a quintessentially liberal conception of the good according to which it is dependant upon each person’s exercise of autonomy, within the bounds of justice that makes this possible.Footnote 15 My argument is simply that both Plato and Raz offer a “service conception” of government in the sense that the justification of political power is entirely instrumental or teleological. Both can unite around the basic (paternalistic) idea that the normative knowledge of those who know better is a warrant for their having authority over others. The Platonic parallel is unmistakable in the following passage:
I regularly confront a decision…whether or not to sell certain shares of stock in varying circumstances. Suppose that it is known that a financial expert reaches the “right” decision (whatever that may be) in twenty percent more cases than I do when I do not rely on his advice… In cases about which I know only that his performance is better than mine, letting his advice tilt the balance in favour of his solution will sometimes, depending on my rate of mistakes and the formula used, improve my performance.Footnote 16 (Raz Reference Raz1986, 67-68)
The parallels between Raz and Plato here add substance to the conclusion that the idea of expertise and the idea of democracy are conceptually at odds, that they tilt and strain in different, sometimes antagonistic, directions. Whatever else might be said about democracy, it is clear that it embodies at least two fundamental principles—universality and equality—that a conception of government based on expertise will have difficulty accommodating. Democracy is universalistic, first, because it requires that all persons in the relevant political community be included: everyone (from among the relevant group) is supposed to count. Democratic procedures apply, universally, to each person. Democracy is egalitarian, second, because it requires that each person’s voice count equally; no one’s vote shall count more or less than anyone else’s. It is essential to democracy that, at least at some fundamental level, political power be shared equally among citizens. It is not difficult to see that the logic of expertise upon which the service conception is based is neither universalistic nor egalitarian.Footnote 17 First, the very idea of expertise renders it incompatible with universality. Expertise entails exclusion: to be an expert on some matter is precisely to enjoy a status that it would be unintelligible to extend to everyone. “Universal expertise” is incoherent, a contradiction in terms. Second, the logic of expertise contradicts democracy’s egalitarian dimension because expertise itself suggests that some voices shall count more or be heard louder than others. In claiming expertise on some topic one is implicitly claiming a unique right to be heard on that topic, a right to be consulted. Expertise thus entails authority, both in the sense in which “to be an authority on X” just means that someone has the relevant expertise on or about X—viz., “Jones is an authority on the poetry of Yeats”—and the sense in which “to be an authority on X” implies that one ought to be listened to on matters concerning X. What is the good of experts, after all, if not to provide expert guidance? Raz’s service conception ingeniously brings these two common notions of “authority” closer together: to be an authority is bound up with expertise, and expertise of the right sort is bound up with a right to rule.
2. The Foundations of Democratic Authority
What does all of this suggest about the legitimate authority of democratic government? According to the requirements of the normal justification thesis, it must be shown that democratic decisions tend, more reliably than any other method of deciding what to do, toward the best political outcomes. As Raz says, in a passage from which I have already quoted, democracy “relies for its legitimacy on its ability to deliver sound decisions”. Yet this seems not only a proposition for which there is no test of verification, but one which is patently false. As John Dunn opines, rightly in my view,
[Democracy] wholly fails to ensure any regular and reassuring relation to just outcomes over any issue at all. As a structure of rule, within any actual society at any time, it makes overwhelmingly probable that many particular outcomes will turn out flagrantly unjust. The idea of justice and the idea of democracy fit very precariously together. (Dunn Reference Dunn2005, 149)
To be sure, that democracies tend toward the best decisions is by no means some audacious, new idea. Readers will recall that it was Aristotle who argued that, “the many are better judges than a single man”—that “each individual among the many has a share of virtue and prudence, and when they meet together…[they] may very likely be better than the few good.” (Aristotle Reference Aristotle2000, 119-20) I am less optimistic about the so-called “wisdom of the multitude”.Footnote 18 For one thing, Aristotle’s argument is only plausible on the assumption that each additional voice in the democratic process has a better than random chance of being correct. Only on that assumption could the positive correlation between the numerical quantity of political participants and the epistemic quality of their decisions be affirmed. It is doubtful, however, whether we are entitled to any such assumption. Indeed, as David Estlund has recently pointed out, there are powerful reasons to think the very opposite.
[I]f you were to ask, ‘How could a person be dumber than a coin flip?’ the answer would be ‘easily’. People have more or less systematic views about many issues. If their system is bad, so to speak, then they could easily be wrong all the time. If, for example, people in some time and place were systematically racist, or sexist, or both, it would not be surprising if their political decisions were worse than the performance of a coin flip would be on political matters involving race or sex. Who knows what other important biases or errors people might have in their systematic thinking on issues?Footnote 19 (Estlund Reference Estlund2009, 16)
In short and to sum up, even if we grant that there is some kind of epistemic power in numbers—something I do not wish to deny here—the claim that majority rule tends toward the best answer more frequently and reliably than any other kind of decision procedure is by no means obvious. It might be thought even less obvious given the empirical facts, endemic in so many contemporary democracies, of low or declining voter turnout, the average voter’s poor knowledge of public affairs, and her susceptibility to propaganda and manipulation.
I want to suggest that democracy’s legitimate authority cannot be accounted for in wholly instrumentalist terms, i.e., in terms of the quality or soundness of outcomes it tends to lead toward. For if democracy has legitimate authority simply in virtue of the sound results it reliably yields, then there is no basis upon which to rule out as less authoritative, political decision-making based on highly accurate anonymous polling, or indeed, any other method whatsoever that could be shown to reach the right result as frequently and reliably as democratic procedures did. The problem for Raz is that the service conception cannot differentiate between (a) sound political decisions reached democratically, and (b) perfectly identical decisions reached by any other non-democratic procedure. This is because, on Raz’s view, an authority is legitimate never because there is anything inherent in the authority that confers this status, but merely to the extent that obeying it brings about better compliance with reasons that are independent of the authority. (Christiano Reference Christiano2004, 278) Recall that for Raz, “It is the truth or soundness of the decisions which counts ultimately. Truth and soundness provide the argument for the legitimacy of the authority.” If my argument is on the mark, then, Raz must be committed to one of the following three claims:
(1) Democratic regimes rarely (if ever) enjoy legitimate authority because legitimate authority is accounted for instrumentally, and democratic regimes are rarely (if ever) instrumentally optimal in the relevant sense.
(2) Legitimate political authority is not accounted for solely in instrumentalist terms. Those democratic regimes that enjoy legitimate authority enjoy it at least in part because of the fairness of its methods and procedures, quite independent of its instrumental performance.
(3) Some democratic regimes have legitimate authority and some do not. Those that have it really are instrumentally optimal (at least in comparison to any other feasible alternative).
The Republic puts forward a version of (1). I am arguing here for a version of (2). Raz himself appears committed to something like (3), which is unsurprising, given that it is the only option which leaves the service conception’s applicability to democracy un-injured. Yet this instrumentalist conception of authority again puts democracy on shaky footing. It makes a democratic regime with legitimate authority look like a merely temporary consolation prize, a system of government we would be justified in abolishing the moment a more competent team of “epistocrats” became available.Footnote 20
3. Why Democracy?
In the Euthyphro, Socrates famously asked whether something is pious because it is loved by the gods, or whether the gods love it because it is pious. A parallel question can be usefully posed here about democracy: Are democratic outcomes sound merely because they were democratically chosen, or are they democratically chosen because of their soundness? The two horns of the dilemma are also familiarly Socratic. Embracing the latter horn, viz., They are democratically chosen because they are sound, is tantamount to conceding that true or sound political outcomes (whatever they might be) enjoy their truth or soundness quite apart from the procedure through which they are reached. The implied picture is that true or sound political outcomes are out there all along, waiting patiently for some system of government to come along and pick them out accurately. More problematically, it leaves room for the chance that a democratic vote may sometimes get the wrong answer on a difficult political issue, while consulting a ouija board gets the right one. The former horn of the dilemma, viz., Democratic outcomes are sound because they are democratically chosen, is even more problematic for Raz, since, on this view, the truth or soundness of a political outcome cannot be measured independently of the democratic procedure through which it was reached.
This puzzle about democracy can be further brought into focus by recalling John Rawls’s important distinction between “pure” and “perfect/imperfect” proceduralism. In “Kantian Constructivism in Moral Theory” he expressed it this way:
The outcome of the original position defines…the appropriate principles of justice. This contrasts with perfect procedural justice, where there is an independent and already given criterion of what is just (or fair) and where a procedure exists to ensure a result that satisfies this standard…The essential feature of pure procedural justice, as opposed to perfect procedural justice, is that there exists no independent criterion of justice; what is just is defined by the outcome of the procedure itself. (Rawls Reference Rawls and Freeman1999, 311)
Rawls’s idea is that there are some procedures whose outcomes can be evaluated according to standards that are independent of the procedures themselves—that is, defined separately from, and prior to, the specification of the procedures. If two people want to divide roughly equal shares of a piece of cake, for example, then a fair procedure would be to have the person who cuts the cake choose her piece last. This is an example of “perfect” proceduralism: the quality of the procedure is evaluated in accordance with its ability to lead—perfectly, or nearly perfectly—to the independently desired result. Perfect proceduralism stands in contrast to “pure” proceduralism. So-called pure procedures—Rawls thought of his “Original Position” abstraction as one such example; a fair bet is another—do not rely on any independently desired outcome for their acceptability. The justification of the outcomes is simply in terms of the fairness of the procedure that produced them. Judgments about the acceptability of the procedure depend, in other words, on the characteristics of the procedure itself, rather than on the quality of its results. If the procedure itself is acceptable, whatever outcomes it happens to generate are also acceptable.Footnote 21 As we have already seen, Raz wants to depict democracy as a species of imperfect proceduralism. Democracies, like all legitimate authorities, will sometimes get things wrong. Still, they are justified by their comparatively superior tendency to get the (procedurally independent) right answer. Indeed, we should treat democratic decisions as binding and try to follow them, Raz has it, precisely and solely on the basis of this tendency.
Is democratic authority better articulated in “pure” or “perfect/imperfect” terms? That is a complicated question to be sure, the right answer to which probably involves a nuanced combination of the two. Yet even if a fully “pure” or “procedural” justification of democratic authority is untenable—a conclusion around which a growing consensus among democratic theorists has started to emerge—it does not follow, as Raz would seem to have it, that non-instrumental procedural values have no part to play in the justification of democracy’s legitimate authority. By way of illustration, consider the large number of citizens in any normal liberal democracy who have never voted for a winning party or candidate in an election. This generates a puzzle: if those citizens vote for the party or candidate that they think most likely to make right decisions and promote sound policy—leaving cases of “strategic voting” aside—then they must believe that each time they cast their vote the democratic system produced the “wrong” outcome. That much follows as a matter of course. Now, it is true that on the service conception, it is the soundness of the outcomes (over the long run) that justifies an authority’s legitimacy, not a citizen’s (or anyone else’s) belief in the soundness of the outcomes.Footnote 22 Yet how could anyone earnestly regard as legitimately authoritative a system of government that consistently produced (what were in their eyes) unsound outcomes? I would suggest that democratic decision-making can be accepted as authoritatively binding, not on the faith that the voice of the demos has a better chance of reaching the right result—though that may count as a reason too—but because democratic decision-making represents, in Rawls’s phrase, “fair terms of social cooperation” (cf. Rawls Reference Rawls1993, 15-22). Scott Hershovitz has eloquently put the general point this way.
If a government’s electoral system favors some interests in society, or appears corruptly financed, or causes portions of the population to be marginalized and voiceless, we are quick to judge it illegitimate, or at least less legitimate than it might be otherwise. Where these deficiencies are present, it counts for little that a government may produce substantively good decisions, decisions that the normal justification thesis would hold authoritative. . . . This shows us one way in which the normal justification thesis is incomplete as a theory of legitimacy for political authorities: Governments that fulfill it may fail to be legitimate on procedural grounds. (Hershovitz Reference Hershovitz2003, 216)
To put the same point in Razian terminology, the “normal reason” for accepting a democratic decision as binding is not (or not entirely) because it is more likely to be a sound decision, or because democracy gets things right more often than any other system, or because democracy tends to bring about this or that desired outcome, but rather because it was a decision reached through a fair and free participatory procedure—a decision reached on fair terms, under conditions of political equality. Here we return to the issue concerning the justification of political authorities that I flagged at the outset: the so-called “qualified acceptance requirement”. The main idea is that a necessary but insufficient condition for its claim on legitimate authority lies in democracy’s representing a mode of decision-making that can be accepted by all qualified points of view. This does not suggest that a political authority is rendered legitimate in virtue of its satisfaction of the acceptability requirement alone; only that a political authority cannot be legitimate without satisfying that requirement. In short and to sum up, one can at the same time regard democracy as a system of government that quite regularly produces abysmal results (as incidentally I do), while also regarding its decisions as legitimately authoritative—i.e., as constituting moral reasons for action and compliance—partially on the grounds that they were arrived at democratically. I am not suggesting that considerations of fairness and “qualified acceptance” tell the whole story about the justification of a democratic regime. A procedure that holds a vote and chooses one person’s vote randomly, or a coin flip, say, are surely no less fair (and hence, perhaps, no less acceptable by all qualified points of view) than regular democratic procedures. There is nothing to prevent a perfectly fair procedure from being perfectly ludicrous! Raz of course recognizes clearly that the case for democracy cannot be grounded entirely in procedural fairness. “[I]f my subjective belief is all that matters [to a democratic procedure],” he asks rhetorically, “why should I, in forming it, agonize about the rights and wrongs of the issues concerned? Why shouldn’t I just express any belief or preference that comes first to mind? Why should I, as we all believe I should, try to form an informed judgment, i.e., one responsive to the truth?” (Raz Reference Raz1994, 115) Raz is certainly right that democracy, if it is to be plausible, cannot be substantively innocent. Yet one can accept all of this and still maintain that democratic authority cannot be disconnected, as the service conception would have it, from the fair and participatory procedures constitutive of democratic decision-making. If it could, we democrats concede the crucial terrain in the dispute between the rule of the epistocrats and the rule of the demos.
Raz is by no means blind to the sort of argument I have been advancing. Focusing specifically on the relationship between democracy and the service conception, he writes, “Some people believe that one has a duty to obey anyone who is elected by a majority…that is no problem for the service conception. If that is so it simply shows that the conditions of the service conception are met regarding anyone who is so elected.” (Raz Reference Raz2006, 1031) Raz goes on to suggest in a corresponding footnote that democratic governments “in some countries” do possess “unique claims to enjoy some qualified or limited authority, either through their ability to produce beneficial results or because of their ability to give expression to people’s standing as free, autonomous agents, or whatever other values they serve.” (Raz Reference Raz2006, 1031, n.20) This certainly gives the impression that, contrary to what I have been arguing, the service conception has no trouble accommodating democracy’s non-instrumental virtues.
An authority is legitimate for Raz, let us recall, when “the subject would better conform to reasons that apply to him anyway (that is, to reasons other than the directives of the authority) if he intends to be guided by the authority’s directives than if he does not.” (Raz Reference Raz2006, 1014) Legitimate authorities, as we have already seen, meditate between moral reasons and the subjects to whom those reasons apply. So, when Raz asserts that it is no problem for the service conception to grant that we have a duty to obey anyone who is elected by a majority (assuming the proper conditions are met regarding anyone who is so elected), he seems to be suggesting that that will be one of the dependent reasons relevant to the correct application of the normal justification thesis. If we have good dependent reasons to obey anyone elected by a majority, the service conception will have no problem acknowledging and promoting those reasons.
While this admittedly makes the normal justification thesis look less instrumentalist than most readers of Raz have usually supposed,Footnote 23 it is still tempting to agree with Christiano when he writes that, “Raz seems entirely instrumentalist regarding political authority.” (Christiano Reference Christiano2004, 278n12) While Raz’s instrumentalism is capacious enough to allow the instrumental realization of non-instrumental goods—giving “expression to people’s standing as free, autonomous agents,” for instance, or treating citizens as equals, or any number of other things—it remains an instrumentalism nevertheless. This interpretation can be buttressed, I think, by recalling that Raz employs a rather cautious tone about democracy. “I do not believe that democracy is the only regime that can be legitimate,” he writes, “nor that all democratic governments are legitimate.” He goes on to warn that “we should preserve our ability to recognize the limitations of democratic regimes as well as acknowledge the possibility that what pass for democratic regimes could completely lack legitimacy.” (Raz Reference Raz2006, 1031, n.20) But surely Raz’s vigilant tone about the legitimacy of democratic regimes—his admonition that “we should not fall prey to the current, and much abused, democratic rhetoric, and maintain a clear-sighted and critical perspective on the nature of democratic institutions…” (Raz Reference Raz2006, 1031, n.20)—is evidence of the extent to which the legitimate authority of those democratic regimes that have it is instrumentally determined. If the service conception really could give credence to the view that democratic regimes enjoy a special authoritative claim on account of their non-instrumental virtues—as Jürgen Habermas and scores of so-called “deliberative democrats” contend—surely Raz would be less cool and cautious about democracy than he actually is.Footnote 24
In short, it remains unclear how the service conception can do justice to the thought that there is something legitimating and validating about decisions reached through fair and free political procedures, procedures that treat all participants as free and equal—even when or if those procedures fail to serve its subjects by assisting them in conforming with right reason. It remains unclear, that is, how the service conception can make sense of the widely held belief that decisions reached democratically enjoy some claim to legitimate authority (even if not an irrefutable or decisive claim) quite independently of how well, or even if, they assist subjects in better conforming to the reasons that apply to them. People rightly care not only about the quality of political decisions—about how well an authority’s directives lead them on the sound moral path—but also about how decisions are made, about whether they are made fairly, in a manner that respects all citizens as equals. The point is that sometimes we demand and expect more from political authorities than conformity to right reason. This is something the service conception cannot allow.
In the course of a discussion on the intrinsic value of democracy, Raz notes, “[B]ecause participation is valuable in itself, providing opportunities for it through democratic constitutional arrangements is a value which may justify putting up with some shortfall in other dimensions of performance.” (Raz Reference Raz1994, 117) This is not the place for a full inventory of democracy’s virtues, but the intrinsic value of participation is merely one of many. Raz might have also mentioned any number of items on Robert Dahl’s list of democracy’s benefits (cf. Dahl Reference Dahl1998, 44-61), none of which bear directly on democracy’s ability to produce sound decisions (at least, not in the sense that is of concern to the normal justification thesis).
(1) Democracy helps to prevent government by cruel and vicious autocrats.
(2) Democracy guarantees its citizens a number of fundamental rights that nondemocratic systems do not, and cannot, grant.
(3) Democracy insures its citizens a broader range of personal freedom than any feasible alternative to it.
(4) Democracy helps people to protect their own fundamental interests.
(5) Only a democratic government can provide a maximum opportunity for persons to exercise the freedom of self-determination—that is, to live under laws of their own choosing.
(6) Only a democratic government can provide a maximum opportunity for exercising moral responsibility.
(7) Only a democratic government can foster a relatively high degree of political equality. The inherent egalitarianism of democratic political arrangements has at least two dimensions. One is the more formal idea that democracies allocate political power equally among all citizens. The second idea, often stressed by adherents of “deliberative democracy”, is that it allows citizens to participate as equals in political decision-making. Indeed, as Amy Gutmann and Dennis Thompson point out, democracy helps bolster their status as equals. “When citizens deliberate in democratic politics, they express and respect their status as political equals.” (Gutmann and Thompson Reference Gutmann and Thompson1996, 18)
(8) Democracy tends to produce a more engaged and reflective citizenry. On Rawls’s view, it “enhances the self-esteem and the sense of political competence of the average citizen.”(Rawls Reference Rawls1971, 234) The service conception might welcome this argument on the grounds that a more engaged, enlightened, and reflective citizenry is more likely to make sound political decisions. Even if it is more likely to do so, however, it is not farfetched to claim that such a citizenry is valuable for its own sake too.
(9) Countries with democratic governments tend to be more prosperous than countries with nondemocratic governments.
While there is nothing to prevent Raz from recognizing all of democracy’s non-instrumental virtues—and its other instrumental virtues that do not bear directly on its ability to deliver sound decisions—the trouble from the standpoint of the service conception is that such virtues turn on a separate normative axle from democracy’s claim on legitimate authority, neither adding to nor subtracting from that claim. To repeat what is claimed above, democracy’s non-instrumental virtues only contribute to the legitimate authority of a democratic government if and to the extent that they enhance its instrumental performance—”only if… [they] lead, by and large, to good government.” (Raz Reference Raz1994, 117) Qua non-instrumental virtues, they are irrelevant to the service conception of authority.
It may well be that having legitimate authority is not the only encouraging thing one can say about democracy; that the question of whether democracy has legitimate authority is distinct from the question whether it is, all things considered, a desirable form of government. I am not convinced, however, that this line of argument is available to Raz. For it is not possible on Raz’s view to have legitimacy without authority, to drive a wedge between the questions “Does democracy have legitimate authority?” and “Is democracy a justified form of government?” For Raz, an affirmative answer to either question will entail an affirmative answer to the other. In part, this is because the conceptual coupling of “legitimate authority” with “a duty to obey” is a definitive hallmark of the service conception.Footnote 25 The problem of the legitimacy of an authority, Raz says, “is but one aspect of the wider issue of the justification of government.” (Raz Reference Raz1994, 356) There is simply no space within which to entertain the possibility that a justified form of government actually turned out not to have legitimate authority. Where there is justified government for Raz, there is ipso facto legitimate authority too.
4. Conclusion
It pays to remember that one of the original aims of the service conception was to respond to a set of problems raised by the so-called anarchist challenge. The best defence of philosophical anarchism, according to Raz, is the version set forth by Robert Paul Wolff in his In Defense of Anarchism. The central problem addressed in Wolff’s book is “how the moral autonomy of the individual can be made compatible with the legitimate authority of the state.”(Wolff Reference Wolff1988, xxvii) As is well known, Wolff argued powerfully that no such compatibility exists, that neither the state nor the law has genuine authority, and with that, the conclusion that there is no general obligation to obey the directives of either. The service conception explicitly responds to Wolff’s argument by making a necessary feature of legitimate authority the thesis that an individual will do better, will conform more faithfully to the reasons that apply independently to her, by obeying its commands and trying to follow them than by working out what to do for herself (viz., the normal justification thesis), thereby generating a moral obligation to obey.
Viewed strictly as a rejoinder to Wolff’s anarchism, as a solution to the problem of how to combine autonomy and authority, I would argue that the service conception is spectacularly successful. Nonetheless, a tool designed for one purpose may not always be ideal for another purpose. I have argued that the service conception fails to properly account for democracy’s legitimate authority largely because of its refusal to endow democracy’s non-instrumental procedural values with any authoritative weight, apart from the extent to which such procedural values—like fairness, participatory parity, political equality, equal respect, and so on—contribute to democracy’s instrumental performance. While I have not explicitly argued for the conclusion that a political authority depends in part for its legitimacy on certain non-instrumental considerations, my main ambition has been to show that, by its very nature, the service conception is unable to even consider the possibility that the legitimacy of political authorities depends on anything more than their ability to provide an instrumental service.
