Law, Legitimacy, and Consent

LAW, LEGITIMACY, AND CONSENTt LESLIE GREEN* ,The democratic tradition has long maintained that the legitimacy of political authority depends on the consent of the governed. It is easy to see why the tradition has not always kept faith with that thesis, however, for if taken seriously it has radical potential. So consent has been tamed: first, by inflationary theories which make it almost automatic; and second, by deflationary theories which deny it any important role. Inflation set in as early as Locke's claim that consent is given "tacitly" by any peaceful behavior that takes place within the territory of a just government. Kant went even further, maintaining that actual consent is unnecessary so long as it would be rational to consent to the authority in question. The worst damage, however, was caused by Hume. Both Locke and Kant knew that consent is not sufficient for legitimacy because consent does not bind except when it is free and informed and when it does not exceed the agent's powers to consent. Hume doubted that anyone actually consents at all, but the truly subversive aspect of his thought was the deflationary claim that consent is not even necessary. He held that governments may make law, coercively enforce it, and have a right to obedience provided only that they are generally beneficial. In his hands, as in those of the utilitarians, the problem of legitimate authority dissolved into the question of the good society. Modern legal philosophy has inherited those traditions of argument. It is not piety to great books that so implicates political theory in the theory of law. Rather, here, as at other points,' there is a mutual dependence between the two subjects. Many contemporary theorists are, in different ways, exploring this relationship; but no one has done more than Joseph Raz to illuminate its consequences for the legitimacy of law. In his analytic writings on the nature of authority and in his critical discussions of the obligation to obey, Raz has deepened our understanding of the problem while defending a distinctive and controversial position. He denies that everyone has an obligation to obey the law but maintains @ 1989 Leslie Green. All Rights Reserved. Associate Professor of Philosophy and Law, York University, Toronto. 1. See Green, The PoliticalContent of Legal Theory, 17 PHIL. Soc. SCL. 1 (1987). t * SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 that law may be valuable and thus merit our respect. Those who do respect the law have an obligation to obey even if they do not consent to be ruled by it. For all its vaunted importance, Raz denies that consent is central to legitimacy: "[lit can have no more than a marginal ceremonial, as well as an auxiliary and derivative, role. So Hume understood the matter better than Locke." 2 This Article explores some of these claims. I shall defend the view that there is no general obligation to obey but respectfully criticize Raz's argument and offer in its place a modified version of consent theory. I. LAW'S CLAIM Theories of the nature of law and theories of its moral force are different, though related, parts of legal philosophy. Contrary to common supposition, however, there is no positivist or natural law "line" on legitimacy. First, positivists need not hold that there is a general obligation to obey, for valid law may be so wicked as to merit disobedience, even counting the risks of disorder. It is true that a complete theory of law must offer some explanation of its "normativity," some interpretation of the language of rights and duties used in making and mentioning legal statements. But there are many possible accounts of that apart from the thesis that legal obligations are also moral ones. Second, the most stringent forms of the "social thesis" 3 are perfectly compatible with the view that there is an obligation to obey. Even if the existence and content of law can always be determined solely by reference to social facts, there nonetheless may be moral considerations requiring general obedience to law so defined. Compare the partly analogous case of promises. One might consistently hold that the questions of whether a promise was made and what it required can be fully answered by considerations of social fact, while also holding that the values of autonomy, fairness, or reciprocity mandate that promises be kept. Likewise, thinking that there are morally neutral tests for law does not in the least commit one to holding that obedience to law is morally neutral. Hobbes and Bentham, for example, were positivists who advocated very stringent duties of obedience. Finally, if natural lawyers are committed only to the thesis that law necessarily has moral value, then they too can deny any obligation to obey. There may be necessary connections between law and morality 2. Raz, Government by Consent, in AUTHORITY REVISITED: NoMos XXIX 76, 93 (J. Pennock & J. Chapman eds. 1987). 3. See J. RAZ, THE AUTHORITY OF LAW 37-52 (1979). 19891 LEGITIMACY AND CONSENT which stop short of imposing such obligations.4 Only on one particularly crude account does the nature of law settle the problem of legitimacy. If only those requirements which there is a moral obligation to obey count as law, then the problem of legitimacy is solved by sheer definition. That implausible thesis is now rarely defended. While the nature of law and the problem of legitimagy are thus partly autonomous, we must remember that they are only partly autonomous, for legitimacy involves the question of whether we should accept law's claims as valid and it therefore imports some view about what these claims in fact are. Most abstractly, we could say that law claims authority and that its subjects owe a correlative duty of obedience.' But what does that amount to? Raz's answer is constrained by his general perspective on the interpretation of legal statements: To understand legal statements we should interpret them as meant by those who take them and accept them at face value, those who acknowledge the law in the way it claims a right to be acknowledged. The decisive argument concerning the meaning of statements of legal duties is that the law claims for itself moral force. No system is a system of law unless it includes a claim of legitimacy, or moral authority. That means that it claims that legal requirements are morally binding, that is that6 legal obligations are real (moral) obligations arising out of the law. Here, "legitimate" is used as a full-blooded normative term to be distinguished both from its purely legalistic use to mean the lawful7 government of some territory and also from its sociological use to mean an approved or accepted government. Law is legitimate only if it justifiably claims to impose moral obligations or duties8 on its subjects. An adequate theory of legitimacy must therefore show how law can have this kind of force. The problem is more complex than is sometimes realized, 4. Id. at 240. Cf. J. RAZ, PRACTICAL REASON AND NORMS 167-68 (1975). Raz writes of David Lyons that: "[flie advances, albeit tentatively, the view that the facts which determine the existence and content of law do not guarantee it any moral value. This, as I said, seems implausible. What does appear true is that the necessary connection between law and morality which is likely to be established by arguments of the kind canvassed by the above mentioned authors [Hart, Fuller, Finnis, Soper, and Lyons] is a weak one. It is insufficient, e.g., to establish a prima facie obligation to obey the law." Raz, Authority, Law and Morality, 68 MONIST 295, 323 n.13 (1985) (footnote omitted). 5. For a discussion of the correlativity of obligation and authority, see J. RAz, THE MORALrry OF FREEDOM 23-28 (1986). 6. Raz, Hart on Moral Rights and Legal Duties, 4 OXFORD J. LEGAL STUD. 123, 131 (1984). 7. Cf Raz, On Lawful Governments, 80 ETHICS 296 (1970) ("the lawful government is that authorized by the positive law of the land"). 8. 1 draw no distinction between them. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 for law can guide behavior without obligating at all. Not all reasons for acting are obligations. Some are purely prudential. Others, though moral, indicate what is merely desirable or meritorious. Legal requirements purport to do more than this. They tell us not what it would be in our interest to do, or what it would be nice to do, but what we must do, what we are bound to do, what we are obligated to do, what others have a right that we do, and so forth. The most influential accounts of this stringency have been either sanction-based or rule-based. Austin and Kelsen, for instance, held that one has a duty if and only if one may be subject to sanctions for nonperformance. In contrast, H.L.A. Hart maintained that one has a duty if and only if the behavior in question is required by a social rule which is enforced by serious pressure to conform, thought important to social life, and which may conflict with immediate self-interest.9 There are wellknown objections to both theories.'0 Raz attempts to avoid them by basing his account not on the sort of social structures which sustain or enforce claims of duty, but on the reason-giving character of the claims themselves. He identifies obligations by a formal and a material feature. First, they purport to be exclusionary reasons for acting: they defeat some considerations against acting as one is obligated, not by outweighing them, but by excluding them from consideration altogether. As applied to obligations imposed by authority, Raz calls this the "pre-emptive" thesis: "[T]he fact that an authority requires performance of an 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."" Second, obligations are inherently goal-independent;they "derive from considerations of values independent of the person's own goals .... ,, 2 Raz's theory of exclusionary reasons has received critical attention 9. H.L.A. HART, THE CONCEPT OF LAW 84-85 (1961). 10. For some objections to sanction theories, see id., at 79-88 and Hacker, Sanction Theories of Duty, in OXFORD ESSAYS IN JURISPRUDENCE 131-70 (A.W.B. Simpson ed. 2d 1973). For some objections to rule-based theories, see R. DWORKIN, TAKING RIGHTS SERIOUSLY 48-58 (1977) and N. MACCORMICK, H.L.A. HART 55-70 (1981). 11. J. RAZ, supra note 5, at 46 (emphasis omitted). One who takes something as an exclusionary reason does not make obedience conditional on an assessment of the merits of the case, but rather acts for the reason that it is required. Acceptance of authority is therefore different from mere compliance with it. One may regularly comply with the law because one judges that what it demands is, on independent grounds of prudence or morality, for the best. But only those who regard the very fact that law requires something as a reason for compliance obey it. 12. Raz, Promises and Obligations,in LAW, MORALITY AND SOCIETY 224 (P.M.S. Hacker & J. Raz eds. 1977). 1989] LEGITIMACY AND CONSENT elsewhere.13 Though it can no doubt be refined, I think that it is broadly correct, and shall therefore focus only on the material condition. In the first place, as Raz concedes, voluntary obligations are an exception to it. The justification for individuals having the power to create, extinguish, and vary duties is something that can be plausibly explained by reference to the way having such power serves their goals. This is, however, no minor exception, for it includes promises, consent, contracts, vows, and oaths. Whether it is the only exception depends on controversial moral views. If there are duties owed to oneself, such as a duty to develop one's talents, then they too are most likely to be explained by reference to the agent's goals. Or consider the duty not to cause animals unnecessary harm. Some people do think that the ground of this duty lies in the interests or rights of animals or in the need not to develop bad habits which may infect inter-human relations. But others hold that it is wrong just because it is inhumane, because it corrupts the person who does it. (Some moralists believe that this also explains certain deontological restrictions on how we may treat people.) If not corrupting oneself is a possible goal then these too would be an exception to Raz's analysis. I wish to express no view on these substantive controversies, but a general account should not foreclose their outcomes. Even though we should abandon Hart's practice theory of rules in favor of Raz's idea of exclusionary reasons, I think we would do well to retain Hart's two material conditions for obligations: they regulate matters thought important to society and they leave room for backsliding in the sense that they may conflict with one's immediate self-interest. With respect to the first point, it is true that some particular obligations are trivial, so obligations cannot simply be reasons of great weight. However, Hart's claim can accommodate this. The fact that the obligations to keep a lunch date or not to walk on a flower bed are trivial does not show that obligations of promising or property are trivial. Importance applies to general types of obligations and not necessarily to their particular tokens. The difference between the second point and Raz's notion of goal-independence is this: Something can be independent of immediate self-interest though not of one's own goals because pursuit of such goals may itself require some restrictions on self-interest. If we say that obligations are grounded in considerations which do not appeal directly to the agent's immediate selfinterest, then we leave room for the conflict between duty and desire even 13. See Moore, 62 S. CAL. L. REv. 827 (1989); Perry, Second OrderReasons, Uncertaintyand Legal Theory, 62 S. CAL. L. REv. 913 (1989). Other valuable discussions include Perry, Judicial Obligation, Precedent and the Common Law, 7 OXFORD J. LEGAL STUD. 213, 239-43 (1987) and Gans, Mandatory Rules and Exclusionary Reasons, 15 PHILOSOPHIA 373 (1986). SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 in the cases of obligations to oneself and the deontological views described above while allowing a unified account of voluntary and nonvoluntary obligations. Accepting with this slight modification Raz's analysis of obligation, let us turn to the content of the obligation to obey the law. Raz's view here is a striking one: law does not just claim to obligate, it claims unlimited power to do so. Legal systems, he argues, are comprehensive, supreme, and open normative systems. 14 Of these, the first is of special importance: We should be careful to see precisely the nature of this feature of comprehensiveness. It does not entail that legal systems have and other systems do not have authority to regulate every kind of behaviour. All it says is that legal systems claim such an authority whereas other systems do not claim it. Furthermore, legal systems do not necessarily regulate all forms of behaviour. All that this test means is that they claim authority to regulate all forms of behavior, that is, that they either contain norms which regulate it or norms conferring powers to enact norms which if enacted would regulate it.15 This is said to be true of all legal systems and not just those, like that of Britain, which have doctrines of legislative supremacy. Many governments have constitutionally limited legislative powers. But Raz maintains that this fact does not refute his thesis because even such constitutions may be changed by law, and in any case limited governments claim authority to regulate action even if only by explicitly permitting it. Every legal system by its nature thus claims the right to legislate on anything whatsoever and to define all possible exceptions to its own jurisdiction. To put it more bluntly, "[w]hatever activity or pursuit you may think of, whether or not it is now forbidden by your law, your government claims the right to forbid it. In just about all states there are legal means to change any law, and to pass any conceivable or inconceivable law." 16 We should scrutinize this account with some care because it foreshadows grave difficulties for some common views about the obligation to obey the law. Most liberal theorists suppose that a theory of legitimacy need only explain our duty to obey reasonably just governments or those which do not violate human rights. If Raz is correct, this amounts to wishful thinking. Because every government demands more, "[a]ny 14. 15. 16. at 76-77. J. RAZ, supra note 3, at 116-20; J. RAz, PRACTICAL REASON AND NORMS 149-54 (1975). J. RAz, supra note 3, at 117; J. RAz, supra note 14, at 151. Raz, supra note 2, at 85-86. CompareJ. RAz, supra note 3, at 31 and J. RAz, supra note 5, 1989] LEGITIMACY AND CONSENT conditional or qualified recognition of legitimacy will deny the law the authority it claims for itself."17 Such authority seems more likely to inspire resistance than obedience. We must satisfy ourselves, then, that the analytic claims are sound and do not repeat Austin's mistake of getting trapped within a tyranny of ill-grounded definitions. As Raz rightly says, "One can derive a concept from a theory but not the other way round." 18 Raz claims that the law is more exigent than even the most demanding of our other social institutions. The duties of marriage, for example, are bounded by both law and convention. Unless we retreat to crude Austinianism, however, we must admit that this is also true of law. Even if every government necessarily claims the power to regulate anything that does not show that every government has such power. Law governs its own creation, but it can do so by limiting its own authority. (There is also the somewhat esoteric point that the identity conditions for legal systems over time limit what a given system can claim: massive changes in the actual authority of some legal system may amount to changes of a system.) And custom is equally important. It is often said, particularly in systems with legislative supremacy, that in political culture lies the best and, in any case, the only ultimate guarantee of civil liberties. Does the mere claim to comprehensive authority flatten out such empirical facts about legal systems so that they become morally irrelevant? It does not, but other factors magnify its importance. Consider the weakest claim that a comprehensive authority might make: it may claim power to regulate some area only by explicit permission. (Raz does not count implicit permission, i.e., silences in the law, as a mode of regulation.) This is a good deal milder than the assertion that every government claims power to forbid anything. But before we breathe easily again, it is useful to examine what the actual effects of explicit permission may be. Once some activity, say freedom of expression, becomes constitutionally guaranteed in this way, it is embodied in certain words or principles rather than in others. A constitution which protects expression only under the banner of "speech", for example, leaves some other modes of expression open to restriction. Explicit permission is thus regulation in the serious, and not just logical, sense. Next, consider the fact that legal systems of modern states govern the behavior of large societies. This being so, there is an important asymmetry between citizens and officials. Individual citizens, even in a 17. 18. J. RAz, supra note 5, at 76. Id. at 16. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 democracy, are law-takers just as individual consumers in a competitive market are price-takers. This is not to deny the possibility of genuine popular control of governments; it is to remind us that such control is essentially collective in form. To exercise it therefore requires organization and the ability to bear substantial transaction costs. The modem state is thus quite unlike the family, tribe, neighborhood, or other face-toface communities. The state claims obedience from individuals who can control the state only in groups. Thus, the notion of self-rule conceals an important equivocation: the people (jointly) rule the people (severally). The actual limits on the authority of law therefore reflect the collective rather than the individual will. Finally, the actual powers of the most liberal of governments are both large in absolute terms and larger than those of any other normative system in their societies. All of them regulate-and not just permissively-personal relationships, property, freedom of action, citizenship, as well as the process of legislation and the administration of justice. This applies also to proposals for minimal states. A Nozickian government will protect some kinds of property, prohibit others, enforce certain agreements, ignore others, define offenses, punish offenders, and so forth. Even if one favors that regime there is no point in having any illusions about what it actually does. Thus, while it is not strictly correct to say that all governments claim power to forbid anything, it is true that even the weaker forms of regulation may have profound social consequences and that even limited governments are institutions claiming vast power. Raz's general view thus seems well-grounded. Why then have political theorists (anarchists aside) failed to notice what legitimate political authority involves? In part, their blindness may be due to a division of labor which has crept into the subject. Many writers distinguish between the theory of political obligation which is meant only to apply to reasonably just societies and the theory of revolution, which specifies conditions under which a change of the entire system is warranted. Such distinctions may be heuristically useful. However, they can also conceal the fact that a tyrannical state and a just one each claims comprehensive authority and does so with a straight face. Because the problem of legitimacy involves the question of whether we should concede the state's claims, this familiar distinction also has its risks. 1989] LEGITIMACY AND CONSENT II. HOW AUTHORITY IS JUSTIFIED Raz advances two claims about how any authority should ideally operate. The first is a condition of rationality which he calls the "Dependence Thesis": "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."19 That is, law should base its requirements on those things we should do anyway, such as respect rights, strive for the common good, promote efficiency, and so forth. But if these are indeed already the right things to do, then what do we need law for? The answer is that we are not always able to do them, or to do them well, without authoritative regulation. That is the basic intuition behind Raz's second claim, the "Normal Justification Thesis": [T]he normal way to establish that a person has authority over another person involves showing that the alleged subject is likely better to comply with reasons which apply to him (other than 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.20 Together, these theses direct us to take authoritative requirements as exclusionary reasons. If authority is already based on reasons which apply independently, and if following authority will indeed produce better compliance with those reasons, then we should not take those reasons into account again upon receiving a directive. The directive itself is grounded in them and, if followed, will serve them better than would fresh judgment. In structure, this is like the familiar indirect argument for following rules. Many common justifications for authority take this form, but the following are probably the most influential in legal philosophy. A. THE ARGUMENT FROM EXPERTISE In certain areas, for example in the sciences, some people are more knowledgeable than others. We often do better to follow their advice in the relevant matters than we would do either by acting on our own views or by trying to second-guess and correct their views. It has occasionally been argued that this also applies to political authority. 19. Id. at 47 (emphasis omitted). 20. Id. at 53 (emphasis omitted). SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 Scientific judgments are often important to public policy, and there may be expertise in moral matters as well. Although some actions are inherently wrong, or mala in se, it does not follow that individuals acting on their own will always be able to comply with the dictates of morality. Even if law cannot make things right or wrong, legislators, judges, and bureaucrats may have a keener insight into morality or into what best secures compliance with it. If so, we would do better by taking their directives as binding than we would by following our own lights. That argument, though capable of greater subtlety and elaboration, is in all forms an unattractive justification for the authority of law. First, it is unlikely that the relevant sort of expertise is available in political affairs. Scientific experts may be authorities on what we should believe but not on what we should do. There are experts on whales but not on whether we should save the whales. Second, even if policy-relevant expertise is available, there is little reason to suppose that officials have uniquely rich reserves of it. Universities, research organizations, and private individuals often have greater or equal knowledge in certain matters and thus deserve a hearing as well. However, political authorities do not just put their voice in, leaving it up to us what weight to give it and whether to follow it. They purport to settle the question. Finally, and this is in any case conclusive, there are profound moral objections to a society run by experts, objections which are rooted in the values of selfgovernment and political equality. In view of this, and because the argument from expertise plays no important part in Raz's theory, I shall say no more about it. B. THE ARGUMENT FROM COOPERATION In some matters expertise is irrelevant. Consider the traditional notion of mala prohibita. One must drive on the right side of the road, and one must pay a certain amount of income tax. In each case, the law seems not only to identify but also partly to create the correct standards of behavior. This creative effect is consistent with the dependence thesis provided that there is reason for everyone to drive on the same side of the road, and for a certain total amount of income tax to be paid. However, each of these requirements can be satisfied in many different ways. In such circumstances the law does not act as an expert-there is nothing 1989] LEGITIMACY AND CONSENT for it to be an expert on-but instead creates a framework for cooperation where no particular one is independently mandated. It makes concrete and supports desirable ends that might otherwise be frustrated.2 1 The argument from cooperation has a long history, including the Hobbes's contractarianism and Hume's conventionalism, and it has recently been revived by a number of legal positivists and natural lawyers. At one point, Raz advanced a very strong version of it claiming that "[a]ll political authority rests on this foundation (though not only on it)." 2 2 His most recent view is only a little weaker. He says, "The case for having any political authority rests to a large extent on its ability to solve co-ordination problems and extricate the population from Prisoner's Dilemma type situations."2 3 It is uncertain why Raz holds that all or most political authority depends on the argument from cooperation. It is true that a major justification for having governments at all is to help promote certain forms of social cooperation (and to prohibit others, such as conspiracies and pricefixing). It is also true that an institution is a government only if it claims to act with authority. But that does not entail that the reason for governments having authority is in order to promote cooperation, for they may be able to do that in ways which do not rely on the exercise of authority. (In Britain, for example, the government is now the second largest purchaser of advertising which it deploys to promote certain behavior and attitudes.) Raz may be attracted to the above view partly because he sees law through the lens of the "communication model":24 law guides human behavior primarily by communicating to its subjects certain standards of appropriate behavior. In order to exercise that technique successfully, law must be able to resolve certain coordination problems involving, for example, the interpretation of legal statements. Although this reasoning is sound, it fails to support the argument from cooperation, for these higher order coordination problems are themselves created by law. Shared interpretations are needed if law is to guide behavior at all. (Even a coercive threat must be understood if it is to succeed.) The 21. Cf. J. FINNIS, NATURAL LAV AND NATURAL RIGHTS 284-89 (1980) (positive law as a "determination" of natural law); id. at 351-52 (law's function in solving co-ordination problems); Boardman, Law and Co-ordination,97 ETHICS 546, 552-53 (1987). The most sophisticated account in the positivist version is Postema, Coordinationand Convention at the Foundationsof Law, 11 J. LEGAL STUD. 165 (1982). 22. J. RAz, supra note 14, 64. 23. J. RAz, supra note 5, at 56. 24. Raz, Dworkin: a New Link in the Chain, 74 CALIF. L. REv. 1103 (1986). SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 communication model is neutral with respect to the nature and justification of the guidance which law purports to give. Even if cooperation is not a universal justification for authority, it may be a possible and important one. This contention is also open to doubt. I have attempted to refute the argument from cooperation in its various forms, 5 and I will not elaborate the case here, save for one polemical point which I relegate to the footnotes.26 The general thrust of those arguments, however, is as follows: Authoritative directives are not necessary for cooperation because whenever people have reason to cooperate, other ways of guiding behavior-including exhortation, persuasion and coercion--can more flexibly achieve their ends without the risks of generalized authority. Moreover, when directives are issued, their success in promoting cooperation does not depend on them being taken as authoritative. Rather, success depends on people generally conforming to what they require but not on conforming as the directives require or because they require it. Authoritative directives are only sufficient for securing cooperation in certain rare circumstances: when there is no temptation to defect from the desirable pattern of behavior and when there are no competing frameworks for cooperation. (Suppose two people were directing traffic, one with authority and the other with a gun. Who would you obey?) Be that as it may, I think the following is common ground between Raz and myself, and shows why the cooperation argument does not in any case warrant taking the law at its word. First, not all legal requirements serve cooperation in any sense, so the argument does not apply to them. Second, if the existence of a legal system as a whole is taken to be necessary to securing cooperation, then the argument applies only to those actions which threaten its existence as a whole, which not all disobedience does. Third, it has never been shown that everyone must accept 25. L. GREEN, THE AUTHORITY OF THE STATE 89-157 (1988). See also Green, Law, Coordination and the Common Good, 3 OXFORD J. LEGAL STUD. 299 (1983). 26. As Raz's mention of coordination problems and prisoner's dilemmas suggests, some elementary concepts from the theory of games are useful in presenting and analyzing the argument from cooperation. However, neither that argument nor its refutation depends on that apparatus. This seems to have been misunderstood by, for example, Finnis, The Authority ofLaw in the Predicament of ContemporarySocial Theory, I NOTRE DAME J. OF LAW, ETHICS & PUB. POL'Y 115 (1984), and also apparently by Raz, The Obligation to Obey: Revision and Tradition, 1 NOTRE DAME J. LAW, ETHICS & PUB. POL'Y 139, 152 (1984). Both Raz and Finnis now qualify their use of the term "coordination problem" with the rider "in the ordinary sense." I am not familiar with this sense. Perhaps the more general notion of "cooperation" will capture what we all have in mind. The substantive point for which I contend is that authority is neither a necessary nor generally desirable way of securing cooperation among people who take account of each other's expectations. 1989] LEGITIMACY AND CONSENT an obligation to obey the law in order for it to be able to promote cooperation. Finally, even if it were true that only the government can secure general cooperation, it does not follow that the government can only do so by issuing authoritative directives. In many cases it should or must use other techniques. Whatever the success of the arguments from expertise or cooperation in justifying other forms of practical authority, I conclude that neither can ground the extensive claims of law, for these justifications must be context-sensitive in a way that the law is not. Not everyone is inferior to the officials in wisdom, and not every pressing social problem is one of cooperation, so there are bound to be people and circumstances to which the justifications do not apply. It is often said that a policy of obedience would be better than a policy of deciding each case on its merits. However, those are merely two extreme ideal-typical strategies which individuals might follow. A policy of general watchfulness would probably not even be best for criminals and dedicated revolutionaries. In some matters they would do better simply to obey the law and leave more free time for deliberation when it really counts. When one begins to consider the full range of available policies, the arguments quickly unravel. One might, for instance, adopt the policy of general obedience when the party for which one voted is in power while judging each case on the merits when it is not. One might adhere to a policy of obedience only to governments which act in good faith. Or again, one might divide laws into core and peripheral concerns and obey only the former while complying with the latter only when one judges it to be best. The possibilities are endless. How could it be that, among all of them, the policy of general obedience to law commends itself to everyone? That is the general reason indirect forms of consequentialism (such as rule-utilitarianism) fail to justify the authority of law. They may show that acting on some set of rules would always be morally superior to acting on individual judgment, but they cannot plausibly show that acting on the actual set of legal rules would always be superior. Thus, even when laws are on the whole good and the government is actiig within its proper bounds, its claims to authority will be larger than some individuals should grant. Raz asks, "does following the authority's instructions improve conformity with reason? For every person the question has to be asked afresh, and for every one it has to be asked in a manner which admits of various qualifications." 7 This is an important point and goes far to establishing his thesis that there is no general obligation to obey the law. 27. J. RAZ, supra note 5, at 74. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 We can, I think, go farther still. On grounds of expertise and cooperation, probably no one has an obligation to obey the law just as it claims to be obeyed. III. THE ARGUMENT FROM CONSENT Consent theorists, however, hold that some citizens do indeed have an obligation to obey, namely those who agree to do so. Some consent theorists also think that all or most citizens have in fact made such an agreement. Raz, like Hume, denies this. He writes: "Promises and other voluntary commitments are created by an expressed intention to be bound. It is clear that the ordinary life of normal citizens includes nothing amounting to a promise or a voluntary undertaking."2 8 I fully agree that some things often taken to be signs of consent- voting in free elections, inheriting property, not emigrating, etc.-do not, in our societies, bear such meaning. On the other hand, freely pledging allegiance, taking oaths of public office, or becoming a naturalized citizen normally do. The intermediate cases are the most difficult and also the ones on which political theory casts the least light. Which acts give consent is a matter which varies among cultures. Political theory can only suggest certain constraints on interpretation. In any case, the absence of consent need not trouble a consent theorist. If consent theory correctly states a necessary condition for individuals to owe an obligation to obey the law, and if they do not consent, then the conclusion is clear enough-they owe no such obligation. Some, including Hume, regard this as the reductio ad absurdum of consent theory. This, however, depends on taking universal political obligation as a basic presumption of any theory of legitimacy. I have argued against this presumption elsewhere, 9 and Raz's defense of the flexible justification of authority points in the same direction. Universal obligation is not a pervasive and unshakable feature of moral thought which must be explained by any competent theory of legitimacy. It is a controversial political position. Consent theory is not therefore a failed explanatory hypothesis. It is a critical theory of the conditions under which we would have such an obligation. In ordinary speech, "consent" often means little more than approval or consensus. Raz adopts a stricter definition which is as follows: Consent is given by any behavior (action or omission) undertaken in the belief that 1. it will change the normative situation of another; 28. 29. J. RAz, supra note 3, at 239. L. GREEN, supra note 25, at 24047. 1989] LEGITIMACY AND CONSENT 2. it will do so because it is undertaken with such a belief; 30 3. it will be understood by its observers to be of this character. The first criterion refers to things like altering fights, duties, powers or liabilities, the second to the reflexive role of belief, while the third makes consent a public act. Only the second needs further comment here. Consider the following example. If Ann gratuitously punches Bob in the nose, she changes their normative situation by creating a duty to compensate Bob and correlatively by giving him a right to compensation. But even if she believes that as she swings, it will still not be a consensual obligation because the fact that she believes it is not part of the justification for holding her to be under the duty. The duty is binding because she is at fault in harming him. A consensual act, in contrast, creates duties because it is done with the belief that it will do so.3 1 Promissory obligations fall under this conception of consent as a special case. Unlike consent in general, promises are always intended to undertake obligations and confer rights. In contrast, "acts undertaken for another purpose and not in order to consent can constitute consent if undertaken in the belief that they will confer a right or impose a duty and if the fact that they are undertaken with such a belief is the reason for them having this result." 32 Raz exemplifies the principle by suggesting that those who enter a park and are given notice that in doing so they must abide by the posted rules thereby consent to the authority of those rules. One does not enter the park in order to get obligations, but entering the park creates them and makes one liable to the authority of others because it is undertaken with that belief. Of course, in most real cases there may be other beliefindependent grounds for holding the parties to be under the duty in question. A rule prohibiting noise in a park may be justified by the potential nuisance imposed by noise-makers, irrespective of any beliefs they hold about the consequences of entering the park. The question therefore arises why duties should ever be made responsive to individual beliefs. The answer is simply that, under certain conditions, it is desirable that agents have the power to control their obligations in this way. These conditions govern the validity of consent-giving acts. Consent must, for example, be free and informed; it must also be an agreement which citizens are normatively competent to make. If I promise to obey the law in 30. J. RAz, supra note 5, at 81. 31. Legal relationships such as contracts may create duties of both consensual and nonconsensual sorts. The distinction does not, therefore, coincide with any particular legal institutions. 32. J. RAZ, supra note 5, at 83. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 order to avoid being tortured, then I am plainly not bound by my promise. Holding this way of creating duties to be valid would leave people prey to all kinds of extortion. But neither can I validly promise to murder someone. Such promises are void ab initio. Locke's famous argument in the Second Treatise exploited this truth to show that we cannot owe a duty of obedience to an absolute government, since we lack the power to put our lives in jeopardy. He defended this claim on the grounds that we are God's property and thus it is for Him alone to decide when we should die. That step no longer persuades many people, but the structure of his argument is correct: the validity of consent is bounded by our normative powers. Raz approaches this problem by asking when the power to consent to obey a government would be justifiable. He holds that such consent must satisfy what I shall call conditions of rationality and autonomy 3 3 : The Condition of Rationality Acting on the government's directives will promote one's success in following right reason. The Condition of Autonomy The government's directives only regulate matters in which acting on right reason is more important than deciding for oneself how to act. I will assume that these are both necessary conditions for the legitimacy of political authority and for the validity of consent to obey it. Consent to be ruled must therefore be qualified at least to this degree. Are these conditions also sufficient for legitimacy? Raz suggests: [C]onsent to be governed can be binding only if limited so as to be consistent with autonomy. That entails that it must be consistent with the two conditions of legitimacy. But it seems that the two conditions are in themselves sufficient both to show that the governments that meet them are consistent with autonomy and that they are legitimate. There seems to be nothing that consenting to be governed can do. It imposes no duty and confers no right except those that exist independently anyway.34 As we shall see, Raz refutes the hypothesis that consent is idle by establishing a minor role for it. Despite the somewhat tentative tone, however, he says no more about the hypothesis then that the two conditions are sufficient for legitimacy. Do they in fact show that a citizen subject to a government which fully meets them has a duty to obey it? Consider 33. Raz, supra note 2, at 89. The first condition is just the Normal Justification Thesis as applied to governments. I do not follow Raz in calling these conditions of "legitimacy" because I argue below that they are not sufficient for it. 34. Id. at 89-90. 1989] LEGITIMACY AND CONSENT an analogy. Suppose Carol is an excellent investment counselor and that conformity to her advice is certain to be optimal. Suppose further that there is no intrinsic value in David managing his own financial affairs. Does that show that Carol has legitimate authority to act for David? No; although it does show if she were given such authority she would be justified in having it, and it also shows that David would be justified in giving it to her. Why should we think that a government is in a different position? It is true that the second condition insures that accepting its authority does not compromise autonomy, but the same applies to David-if he grants Carol authority to handle his portfolio then his autonomy is still secure. However, his autonomy is not secure if Carol unilaterally takes charge of his affairs and then administers them in a way that assures both rationality and autonomy. The fact that there is no intrinsic value in managing one's finances does not show that there is no intrinsic value in deciding whether to do so. For one thing, one may have a choice between different potential authorities. Similarly, it seems correct to say that a government satisfying the two conditions deserves authority but not that it already has it. There may be other actual and possible regimes which equally deserve one's allegiance, or one may decide that in spite of the merits of all of them one would still prefer to retain one's moral independence. One might, perhaps, object that for any political authority to serve reason it must also be effective, that is, able to secure a substantial degree of compliance.3 5 A legal system cannot satisfy the first condition unless it can generally coordinate behavior, so every legal system which satisfies the two conditions therefore not only deserves authority but has it. This invokes the argument from cooperation which I rejected above. Moreover, if efficacy is necessary for legitimacy, that establishes only that a legal system which deserves authority also deserves to be efficacious. It would be excessively conservative to maintain that our choice among authorities must be drawn from the set of those which are already in power (normally a set of one). All that is warranted is the weaker conclusion that a potential authority must be such that it is feasible for it to become efficacious, given our present circumstances. Suppose now that someone does consent to the authority of a legal system which fully meets the above conditions. In what way does this change things? It clearly does not make it a better legal system any more than David's grant of authority to Carol made her a better counselor. Any improvement would not be a direct consequence of consent, though 35. J. RAz, supra note 5, at 56, 75-76. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 it could be a by-product of the level of information and critical awareness needed if consent is to bind citizens. The difference it purports to make is a normative one, but it actually makes that difference only if consent is valid. Consent may sometimes be validated by instrumental considerations, for instance, when it serves to protect the consenter (as in consent to surgery) or to reinforce independent duties (as in promising to tell the truth). In view of the demanding character of the law's authority and the need for flexibility, however, consent to obey it can only be justified by such considerations in the case of some people. But, as Raz says, consent has a noninstrumental function as well: it may express or create intrinsically valuable relationships. Consent to be married, for example, is not merely a device for protecting people against below optimal allocations of partners or reminding them of their preexisting moral duties. It is a constitutive feature of the relationship itself. Raz concedes that consent to obey the law may also have expressive value but even then only when it is limited by the two conditions. Those who consent to the rule of the Chilean or South African governments perform an action which is morally inert. Thus, the most consent can contribute, Raz says, is the minor and marginal role of closing the gap when the two conditions are nearly but not quite fulfilled and of reinforcing and expressing one's commitment to a system which already satisfies them. Consent is valuable only as the exception rather than the rule. It is useful for officials and the like to declare publicly their duties but unnecessary and uncommon in the lives of most citizens. They may be bound to a legitimate government just in consequence of its value and of their having a certain attitude towards it, even if they do not express that attitude through consent-giving acts. I have rejected the claim that the two conditions are sufficient for legitimacy, but it remains to be seen whether non-consensual attitudes can have the same legitimating function as consent. IV. THE IDENTIFICATION THESIS What would we think of people who say that they are obligated to obey the law "because they are citizens of the state or members of the society, because the government is their government or the law their law"? 3 6 We might well be inclined to suppose that they must have no coherent view about legitimacy at all. To hold that one must obey the law because it is one's law sounds more like evasion than explanation. Raz's most interesting thesis is that this reaction might be wrong. He 36. Id. at 97. 19891 LEGITIMACY AND CONSENT says that although there is no general obligation to obey, one is entitled to respect the law in such a way that a personal obligation to obey flows from the very fact that one respects it. This is surprising because it seems more natural to hold that one is entitled to respect law only if there is an obligation to obey it. Let me offer an analogy. The old "expressive" theory of punishment said that we are justified in punishing criminals not as retribution or deterrence, but just because that is an appropriate expression of social indignation at crime. That argument is unsound because it fails to show why we are entitled to feel such indignation or why punishment is an appropriate expression of it. The normal justifications for punishment warrant indignation only as the upshot of a theory of criminal justice. They establish that one is entitled to feel indignant at those who deserve to be punished. Thus, the expressive element has no independent force. What preserves Raz's claim from the same objection? He says, "if the attitude of respect is itself the source of the legitimacy of the authority then it is a self-referential attitude. It respects the authority because it is legitimate, and it is legitimate because it is respected." 37 The way out of this hall of mirrors is marked by three attitudes, which Raz calls "respect" for law, "identification" with the community, and "trust" in government. These terms arise in slightly different contexts in his writings, and the relationships among them are not fully explained. The following, however, seems consistent with most of what he does say. Identification with one's community, provided that it is reasonably just, is of intrinsic value as "an attitude of belonging and of sharing in its collective life" which may be more or less intense but which necessarily "excludes indifference to the group as well as alienation from it."3 8 It is good (though not obligatory) to have such an attitude, and it is also good to express it because that affirms and strengthens the attitude itself.39 When the legal system is an aspect of the community's life, one way of expressing that identification is by behaving as one who trusts the law, that is, by acknowledging its moral authority and the correlative obligation to obey it. One who trusts the law exhibits a necessary condition of respect for the law; one recognizes that there are moral reasons for obeying it.' Thus, the duty of obedience is partly constitutive of practical 37. Id. at 99. 38. Id. at 91. 39. Raz, supra note 2, at 92. 40. J. RAz, supra note 3, at 261, 253. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 respect, an attitude which it is good to have as an expression of identification with the community. There are other ways of expressing one's identification, but at least in our societies, this is one conventional and important way of doing so.4 Notice that the thesis bristles with qualifications which include or entail all of the following: 1. Law may not be an aspect of the community. This is true to some extent in all societies. While law supports some customary practices, it diverges from or attempts to replace others. The gap between law and custom is greatest in times of rapid social change and in those countries in which the legal system is a foreign imposition. Moreover, this divergence is not always a bad thing. Customary norms may be inefficient or unjust or in other ways undesirable. 2. A community may be worthy of respect while its legal system may not, and vice versa. This follows from the previous point. The law may be the worst aspect of some society-its conservative, conflictual, alienated, litigious side. More rarely, it may be more progressive, just, or integrated than social outlook in general. Typically it will be better in some respects and worse in others. 3. It is permissible not to respect or identify with a community or legal system even when that community or system is worthy of respect. Respect for law is thus like friendship inasmuch as it is valuable, but never obligatory, to have any friends or to be friends with any particular person. 4. Expressions of identification are valuable only to the extent that they reflect, support, and foster identification. Some attitudes are harmed by their public expression. Whether loyalty to governments is nurtured by being expressed partly depends on how others understand and react to it and its value partly depends on whether it has undesirable side effects, such as promoting xenophobia and the like. 5. Whether obedience to law is in fact an expression of respect is a matter of social convention. A community may deserve respect, its law may reflect its collective views, its legal system may deserve respect, and yet obedience to law may not be recognized in that community as an appropriate expression of commitment to the community. Taking all of these qualifications into account, we are left with what I shall call the identification thesis: If one respects a legal system, and if that legal system is worthy of respect, and if it is worthy of respect as an aspect of a community 41. J. R.Az, supra note 5, at 91. 1989] LEGITIMACY AND CONSENT which is itself worthy of respect, then one has an obligation to obey it because one respects it, provided that this is conventionally recognized as an appropriate and desirable expression of identification with the community. In order for this thesis to help establish legitimacy there must be both a certain connection between communal relationships and their formal expression in law as well as a practical or normative component to the attitude of respect. Let us take these in turn. "Respect for law" is not respect for the law qua law; it is respect for law qua aspect of the community or, to put it another way, respect for law to the extent that it is an aspect of the community. The identification thesis is thus only meant to apply to non-alienated communities whose legal systems are a faithful expression of what Raz calls "social conventions and outlook." Note, however, that it is not sufficient for the legal system to be mostly in step with society, for the identification thesis is offered as an account of why some people rightly think they are bound by all their laws, provided they meet the conditions of rationality and autonomy. A patchwork of concordance and divergence from social outlook would not warrant this, for even in a reasonably just state law does not claim to bind only to the extent that it is an aspect of the community. Let us now turn to the normative character of respect for law. It may, as Raz says, be possible to respect a legal system only in the cognitive sense of admiring it without thereby feeling bound to it at all. (This is a common attitude to take towards attractive foreign legal systems.) And respect may comprise practical attitudes amounting to something less than recognition of law's binding force. More problematic, however, is the relationship between obligation and identification. Raz first introduces the identification thesis in an analogy with friendship, which is also a complex cluster of cognitive and practical attitudes. There, he confronts the possible objection that while friendship implicates a whole gamut of human relations, respect for law seems much narrower; it is "first and foremost a recognition of a reason to obey the law. Therefore while it is possible to conceive of certain actions as appropriate to express the friendship (i.e., the other aspects of the relationship) there is nothing which obedience for law can express." 4 2 He replies that practical respect is not really an independent attitude but "one aspect of a complex attitude and style of life, relating not only to the law but to the community whose law it is." 4 3 This seems to interpret the objector as holding that 42. 43. J. RAZ, supra note 3, at 259. Id. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 respect for law is narrow because one's attitude to law can express nothing further. That is clearly wrong. The objection may be refined, however. It may mean instead that respect for law is narrow because the normative attitudes comprising it, principally the obligation to obey and support the law, are themselves narrow ones. The obligations of friendship-loyalty, reciprocity, caring, etc.-have a content and character which makes them appropriate expressions of friendly relations. It is, I think, significant that we do not think of obedience as expressive of friendship. Respect for law, however, is identified by a duty of obedience. Is that an appropriate expression of the underlying communal relations of belonging and identification? On this point, compare Raz's account with the more familiar argument that one must obey the state as an expression of one's gratitude for all that it makes possible. Raz of course denies that it is obligatory to express identification through obedience, or even at all. In this respect the arguments are not analogous. But there is also a point of similarity: both arguments assume that obedience is an appropriate expression of another, non-obediential, social relation. Why is an obligation of obedience ever an appropriate expression of gratitude? Can we answer: "It is just a matter of social convention that gratitude may be expressed in this way"? We normally take evaluative attitudes towards these conventions, however, and interpret "appropriate" to mean a "desirable" or "fitting" expression and not just a "usual" one. What Locke says in the Second Treatise about the biblical injunction to honor one's parents is suggestive here: A Man may owe honour and respect to an ancient, or wise Man; defence to his Child or Friend; relief and support to the Distressed; and gratitude to a Benefactor, to such a degree, that all he has, all he can do, cannot sufficiently pay it: But all these give no Authority, no right to any one of making Laws over him from whom they are owing." Locke thus distinguishes obligations imposed by the legal system from the other kinds of duties we typically owe. Social relationships may bring very stringent obligations without amounting to an obligation to obey their commands. According to Locke, the latter must be grounded in either necessity or consent. Below the age of reason children need to obey their parents; after that they are only bound by and to the extent of 44. J. LOCKE, Two TREATISES OF GOVERNMENT § 70 (P. Laslett ed. 2d ed. 1970). 1989] LEGITIMACY AND CONSENT their own agreement." His point cannot be that all the enumerated obligations are less demanding than the obligation of obedience. Indeed, Locke supposes that they are not. But they are all appropriate to the relationships in question. One thanks a benefactor, respects the wise, defends a friend. One does not obey them as part of the relationship. For Locke, obedience is out of place in all these contexts whereas it would be fitting in the relationship of Creator and created or of parent and child. However, this is because there are other, primary, grounds sufficient to validate authority in these contexts. Until we find independent sufficient conditions for legal authority, the fittingness of obedience as an expression of the communal relationships of membership, solidarity, identification, and so forth is liable to similar doubts. Before leaving the identification thesis, we should notice one last, mainly polemical, point. The alleged marginality of consent is due to two things. First, consent is said to be a rare event in the lives of ordinary citizens. Second, it only binds where the two necessary conditions for legitimacy are already met or at least nearly so. This seems to give the identification thesis a certain rhetorical edge since, unlike the oddity of consent, it is advertised as capturing a more normal social relationship. But that is illusory. Because all real communities are to some degree alienated from their legal systems, identification of the sort Raz discusses is an equally, if not more, rare and marginal occurrence. And the necessary conditions for legitimacy continue to play their role. One who trusts the law believes there is reason to do as it requires, but the law must deserve that trust. Above, I argued that the conditions of rationality and autonomy, though necessary, are not sufficient for legitimacy. They are, however, sufficient for the government to be trustworthy. Whereas one who consents to the rule of a trustworthy government is bound by that consent, one who respects a trustworthy government is bound by that respect only if the law is also an expression of social convention and outlook. Certain qualities of the government are necessary to validate consent-based obligations, but these same qualities of government as well as other qualities of the society are necessary to validate respect-based obligations. Since it is therefore more difficult to validate 45. Cf. J.J. Rousseau: The oldest of all societies, and the only natural one, is that of the family; yet children remain tied to their father by nature only so long as they need him for their preservation. As soon as this need ends, the natural bond is dissolved. Once the children are freed from the obedience they owe their father, and the father is freed from his responsibilities towards them, both parties equally regain their independence. If they continue to remain united, it is no longer nature, but their own choice, which unites them; and the family as such is kept in being only by agreement. THE SOCIAL CONTRACT 50 (M. Cranston trans. 1968). SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 respect-based obligations, it is hard to see why they should be thought central to legitimacy while consent is relegated to the ceremonial world of national anthems, barristers' wigs, and royal families. V. CONSENT AND THE ROLE OF CITIZENSHIP The defense of any complex philosophical position, such as consent theory, consists partly in making it plausible and partly in detaching it from the implausible claims with which it has become associated. Consent theory emerged with doctrines of limited government, moral individualism, and political instrumentalism. In extreme versions, it celebrated the notion of the atomic will as the source of all values and duties. That picture of political life is a profoundly misleading one. Thinkers as diverse as Hegel, Marx, Burke, Durkheim, and T.H. Green were all agreed in rejecting it, and Raz is right to join their numbers. But it is possible to deny that political life is a mere tool or that all values are purely individualistic and still maintain that it is only by agreement that good laws become authoritative ones. Any adequate version of consent theory must be, in a sense, social. My aim is not, therefore, to defend Locke against Hume. Indeed, on one crucial point, both of them are wrong about legitimacy. They are political instrumentalists who regard legal and political institutions as means to essentially private ends-for Locke, the protection of property and for Hume, the promotion of our mutual interests. Fortunately, consent theory has more potential than that. Many special obligations are attached to social roles, and the obligation to obey may be among them. Parents have a duty to care for their children, doctors to respect the confidentiality of their patients, judges to decide impartially, and so forth. Here, I use the term "role" in the sociological sense to mean a fairly stable cluster of deontic expectations which receive social recognition, roughly what the English moralist F.H. Bradley meant by his more stuffy talk of "stations" and their attached duties. Not all sets of duties cohere into roles. Those who negligently injure others owe them duties of reparation, but there is no social role of "tortfeasor." Which do cohere varies among cultures. Within each culture, however, the relation between a role and its duties is a constitutive one. It is a matter of fact that we have undertakers and they have certain duties; but it is a matter of logic that an undertaker has got to do what an undertaker has got to do. 1989] LEGITIMACY AND CONSENT Role-duties thus have a relative objectivity, by which I mean, not that they are self-justifying or uncontroversial, but that they are recognized as being partly beyond the control of their subjects. Roles are social and not individual creations. One can choose not to be an undertaker, but one cannot choose to be an undertaker with no duty to be decorous. Not all roles impose valid duties, however. The traditional gender roles in our societies are among those which do not. When are they valid? Here, we need a two-part explanation. First, one needs to justify the clustering of duties into a role at all. Why not instead have complete flexibility in these matters and insist that people construct their own roles as they go and according to their circumstances? Radical individualists would no doubt favor this as giving fullest expression to the sovereignty of will. They might of course concede some minor instrumental benefits to having standard patterns of social interaction, and they might permit roles as default options. However, they would see ideal human relations as polymorphous and contractual. Against such a view, considerations like those which Raz elaborates with great power in his discussion of the social nature of value4 6 seem to me persuasive. Common roles may be valued because they are common and embody social relations which are valuable for their own sake. Beginning with the ancient world, part of the value of citizenship was recognized to lie in its character as a shared status purporting to transcend other divisions among people. The individualist ideal of a multiplicity of different social contracts does not adequately capture this. Nonetheless, to defend common social roles does not show why they bind their incumbents. Thus we come to the second problem. Why should socially constituted roles create valid duties? There is no single answer here but rather a variety of different considerations. Some roles-duties bind nonvoluntarily, most plausibly when others are nonvoluntarily dependent4 7 on their performance (such as infants on their parents) or when the duties benefit others but are modest in their demands (for example, the duty of teachers to be punctual). This supposes, of course, that there are valid obligations which we never agreed to bear. Could a radical consent theorist deny even this? The difficulty with this view is that we need some explanation as to why agreements bind, and that explanation cannot itself appeal to an agreement. One can be as hard-nosed as one pleases about obligation, resisting any commitments to which one has not specifically agreed, until one reaches the obligation to keep agreements. Here, 46. 47. J. RAz, supra note 5, at 289-366. For a vigorous defense of dependence as a ground of obligations, see R. GOODIN, PRO- TECTING THE VULNERABLE: A REANALYSIS OF OUR SOCIAL RESPONSIBILITIES (1986). SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 one will need to appeal to reasons of a different sort which, on their own, will typically justify some nonvoluntary obligations as well. That, in essence, is how Hume tries to subvert consent theory in his essay, Of the OriginalContract.4" He says, "The obligation to allegiance being of like force and authority with the obligation to fidelity, we gain nothing by resolving the one into the other. The general interests or necessities of society are sufficient to establish both."4 9 Hume's argument fails because his view of what these general interests amount to is incompatible with the kind of categorical force that voluntary obligations are normally felt to have. But that need not be argued here. Consider, however, the general argument given above. Does it establish that citizenship binds nonvoluntarily? I think not. Other citizens are not nonvoluntarily dependent on our accepting the state's claim to authority. At most, they are dependent on our complying with the law in those cases when it would harm them not to do so. Nor can it be said that what citizenship requires of us is only modest. As we have seen, every modem state makes extensive claims to our allegiance. The social role of citizenship is not, then, liable to the same justifications as the familiar sorts of nonvoluntary roles. Roles can, however, also come to bind with our consent, even when that consent stops short of creating the roles themselves. It may seem paradoxical that consent could stop just here, but in fact this is just a generalization of the point made earlier about promises. Promising creates valid obligations because it is desirable for people to have the power to do so; but it is not desirable for people to be able to determine whether or not the promising principle applies to themselves. The obligation to keep voluntary obligations is, therefore, not a voluntary obligation. Similarly, it is desirable that people have the power to choose for themselves which social roles will structure their lives, but it is not always desirable that the existence and content of roles should depend entirely on choice. The validation of consent to role-duties thus proceeds in the same way as the validation of consent in general and is subject to the same limitations. However, consent to role-duties proceeds via the assumption of a role known to bring obligations with it. Out of an abundance of caution, I emphasize that this is not meant to establish that everyone has consented or should consent to the role of citizenship. Nor does it show that every modem society has such a role. 48. D. HUME, Of the Original Contract, Miller ed. 1985). 49. Id. at 481. in ESSAYS: MORAL, POLITICAL AND LITERARY (E.F. 1989] LEGITIMACY AND CONSENT I believe that many of them do and recognize it as differing from the role of a tourist or even a resident, occupants of which the law also claims to bind. This however, is a matter which cannot be established abstractly and in advance. The claim is a narrower one: we can reject moral individualism at the structural level by conceding the inherent value of shared roles yet without marginalizing consent. This proposal can only succeed if it makes sense to think of people consenting to roles. In some cases it plainly does. For example, judges and doctors take oaths which bind them to do what their stations require. However, parents and undertakers do not. In discussing the relational obligations inherent in friendship, Raz makes some observations which seem to suggest that social roles and the like cannot sensibly be seen as the objects of consent. Raz maintains that, unlike consensual obligations, relational obligations have the following features: 1. They cannot be created or terminated by a "single act" but only through "the normal habit-forming processes of education and habituation";" ° 2. They are partly constitutive of the social relation in which they occur and are not formed by the act of assuming such a relation;5" 3. They are "by-products" of the relationship rather than its "point or 52 purpose"; 4. They are not fully voluntary; at best they are "semi-voluntary '5' 3 submission[s]. performative These observations are suggestive and might easily be applied to social roles. They do not, however, distinguish all consensual obligations from relational ones. Only the final point does that, and, when correctly understood, it casts doubt on the validity of nonconsensual obligations to obey the law. Raz's first point combines two different ideas: there is the contrast between a single act and a process, and the contrast between deliberation and habit. Is a consent theorist committed to holding that consensual obligations are always assumed by one single act? This may ultimately depend on one's view about the individuation of actions: what, after all, is a single act? At the very least, the necessity of including omissions among ways of giving consent complicates matters. At what moment does an omission take place? How many omissions were there? These 50. J. RAZ, supra note 3, at 257; Raz, supra note 2, at 93. But cf.J. RAZ, supra note 5, at 94 ("it would be wrong to regard consent as a one-off act of identification"). 51. J. RAz, supra note 3, at 257. 52. Id. at 257-58. 53. J. RAz, supra note 5,.at 98. SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 questions are pointless. Moreover, can a process not be a consensual one, emerging through stages of agreement? The contrast between habitual and deliberative processes is thus the salient one. We shall return to it below. The second point, contrasting constitutive and assumed obligations, also cuts across the distinction between relations and consent. The traditional marriage vows are consensually assumed by the parties, but at the same time they partly constitute the marriage. Moreover, they are internally related to a whole web of other valuable human relations. There are, indeed, other ways of getting many of the obligations in question, but there are no other ways of getting married. Consent may occasionally be justified by the very fact that it has such a constitutive role. The third point draws too much power from the highly instrumental flavor of the term "by-product." Raz writes, "People may create a friendship in order to have someone to care for, but not, normally, in order to have an obligation to care for someone." 54 That is true. However, those who consent to being searched before boarding a plane do not normally do so in order to give someone else a right to search them but in order to board the plane. Even in the special case of promising, where the assumption of an obligation is an essential aspect of the transaction, it is ill-described as being its purpose. The purpose of promising to meet a friend for lunch is normally to meet a friend for lunch, not to have an obligation to keep a promise. Relational and consensual obligations do not, therefore, necessarily differ in their ends, as the third point seems to suggest, but only in their means. A promise essentially invokes a certain means in order to achieve purposes which may be identical with those of nonvoluntary obligations. There is to be sure, a somewhat weaker connection between ends and means. Some normative relationships are poorly served by being created through agreements. There is, for example, something nightmarish in imagining social life being taken over by extensive use of explicit contracts. However, this is not because it signals the abandonment of a particular set of aims but because other side effects of formally regulated relationships can undermine some of the aims we seek to achieve. Raz's final point is, I think, the nub of the matter and accounts for what is plausible in the others. Relational obligations are not fully voluntary. By this Raz means that while one is free to assume or terminate 54. J. RAZ, supra note 3, at 258. 1989] LEGITIMACY AND CONSENT them, one normally does so without believing that one's acts attract obligations. Consent, in contrast, always incorporates as a reflexive but nonredundant element the fact that certain actions are believed to bring about certain duties because they are done in that belief. One cannot consent by accident. But one's awareness of the normative consequences of one's actions may be more or less complete, so voluntariness is a matter of degree.5 5 In purely relational contexts, that awareness is either wholly lacking or is redundant in justifying the obligations in question. This is why we naturally describe the emergence of such relations in terms of habituation rather than choice and why the theory of socialization offers a plausible explanation of how such relations take hold. When we thus distinguish the fact that consent depends on such awareness from the mistaken claim that it must be given by a single act whose purpose is to assume obligations and whose connection to social roles is purely instrumental, consent begins to look more attractive. The theory which emerges, however, differs significantly from many of the classical versions. Its centerpiece is not the social contract but rather the notion of people assuming roles in awareness of the duties that they bring. It is true that this will rarely be a single act of commitment. The choice sometimes cannot be made as a whole; opportunities for commitment often present themselves only one by one as the role gradually emerges. Consider becoming a lawyer. Some stages on the way involve direct choices-applying to law school, writing the bar examinations, taking an oath. Others are simply omissions, such as bypassing other opportunities. No doubt other aspects, such as adopting distinctive modes of dress and speech are matters of socialization. In such a case, occupying the role involves a whole cluster of commitments and choices such that it makes no sense to ask which was the commitment or the moment at which one became a lawyer. Nonetheless, it makes perfectly good sense to say of one person that he or she chose to become a lawyer and of another that he or she blundered or grew into it. A choice-laden process can be consensual provided only that it meets the usual conditions for consent. Perhaps it is too crude to put socialization all on the side of habit and choice all on the side of deliberation. Socialization may include rational awareness of its consequences, and choice often takes place within a framework of options and preferences which are not themselves the subject of deliberation. However, the general contrast remains sound. Note also that the above remarks do not establish that lawyers are bound by their roles only if they choose to become lawyers. On the 55. See Raz, Promises in Morality and Law, 95 HARV. L. REv. 916, 929 (1982). SOUTHERN CALIFORNIA LAW REVIEW [Vol. 62:795 contrary, their case fits easily within the usual justifications for nonvoluntary duties. The claim was merely that roles commonly thought of as emerging through processes can also be the results of consent. The social role of citizenship can be acquired in various ways ranging from habituation to conscious choice. Role-bound consent restricts its validity to those cases falling well towards the deliberate end of the spectrum. For many people their feelings of loyalty to the law are results of convention, custom, inculcation, or habit. Raz holds that "where the attitude they give rise to morally valuable it is so even if it is not acquired as a product of deliberate choice." 56 But to show that it is valuable does not show that it is obligation-imposing, as our discussion of the identification thesis revealed. Moreover, where the stakes are so very high there is a connection between the presence of deliberation and the likelihood that the attitude will in fact track right reason. The willing acceptance of the role of citizen thus remains central to transforming a valuable legal system into a legitimate one. Nonetheless, as Raz shows, on any account its legitimacy must be qualified and on no account will it conform to the classical doctrine of political obligation. Did Hume, after all, understand legitimacy better than Locke? I doubt it. Locke knew that consent had its limits, and he knew that it had to be validated. It is true that he did not do much to explain why consent binds, and he was almost certainly wrong about which actions, even in his society, were signs of consent. Here, Hume's skepticism is welldeserved. But Hume's own account was deeply flawed. And, interestingly enough, even he could not resist the pull of the democratic tradition. In revising his famous essay he felt compelled to add the following remark: "My intention here is not to exclude the consent of the people from being one just foundation of government where it has place. It is surely the best and most sacred of any.""7 Hume never saw that his own theory could offer no account of why this should be so. Yet the truth is not far to seek. Our attitudes to law, like our feelings of identification with our own communities, are encrusted with habit and custom. In this area, good habits are fostered by the high degree of awareness that deliberation requires. By limiting the individual's obligation to obey to those circumstances in which the role of citizenship is willingly assumed, consent ensures that obligations track right reason. By acknowledging the value of social roles, it avoids extreme individualism. The normal 56. 57. Raz, supra note 2, at 93. D. HUME, supra note 48, at 474 (emphasis added). 1989] LEGITIMACY AND CONSENT 825 processes of socialization may provide an explanation of common feelings of loyalty, but they do not justify them, at least not when they have as their object the authority of law.