In his subtle discussion of the continuity of law after revolutions, John Finnis advises that it is “usually reasonable to accept the new rules. . . proposed by successful revolutionaries who have made themselves masters of society and thus responsible for meeting the contingencies of the future.”Footnote ¹ In Natural Law and Natural Rights he situates that idea in the context of a general justification for political authority: “Authority (and thus the responsibility of governing) in a community is to be exercised by those who can in fact effectively settle co-ordination problems for that community.”Footnote ² In his sympathetic treatment of Aquinas's legal philosophy, he comments that “public authority is not merely a moral liberty but essentially a responsibility (a liberty coupled with, and ancillary to, a duty).”Footnote ³ These thoughts mark a persistent and distinctive theme in Finnis's legal philosophy: the idea that there is a primary duty to govern and that its ultimate justification lies in the rulers' effectiveness at a certain morally urgent task.

I. A PROBLEM RECOVERED

The problem of governance is the most neglected aspect of a theory of legal and political authority. Contemporary legal philosophers have mainly directed their energies to two other issues: the problem of legitimacy, or the right to rule, and the problem of obligation, or the duty to obey.Footnote ⁴ This cannot be explained by considerations of practical urgency. Based on the volume of writing on political obligation, you would be forgiven for supposing that compliance in Western states is actually at risk and that people need to be reminded of the virtues of obedience. The truth is pretty much the opposite. Even in the mature democracies there is no shortage of sheeplike subjects, too many of whom have already ended up in the slaughterhouse.Footnote ⁵ Nor are we experiencing anything that could properly be called a crisis in legitimacy. While there are always particular complaints that this government or that is meddling where it has no business, serious general doubts about the state's right to rule are confined to philosophers, political theorists, and (other) fringe groups on the margins of society.

Were we to direct our energies in proportion to the urgency of the question, the problem of governance would be a better candidate. Fashionable political theories display an instinctive bias against anything that could be properly be called “governing,” that is, setting and supervising authoritative rules for the guidance of people in a society. Indeed, some do not even grasp that this is what governing involves; they blandly refer to anarchic ordering (bargaining, lotteries, self-help, and so on) as so many forms of “regulation” to be considered in our “choice of governing instruments.” That is a bit like regarding atheism a form of heresy.

Then there is the much-discussed disengagement from public life. You may say that this is not surprising, for most people do not aspire to rule and they have more important priorities in their lives. No doubt. But the worry is not that ordinary subjects do not seek political engagement above all or that they are unwilling to sacrifice central personal interests to it. It is that even very modest burdens of citizenship are shirked. People vote less, seek more exemptions from jury duty, and massively neglect the elementary knowledge and capacities they need to acquire if they are to do their part in governance. Nor are things much better among officials, including politicians who aggressively seek public office at spectacular expense. Absenteeism in legislatures is routine, and those who cannot go AWOL eagerly subcontract their work to private entities, from “outsourcing” policy development to publicity firms and think tanks to contracting for enforcement with private prisons and armies.Footnote ⁶ The idea that our political leaders might actually have a fundamental duty to govern and to bear the responsibilities of doing so seems quaint.

Legal philosophy is unlikely to influence these trends, which reflect cultural shifts in the relationship between the public and private realms. But our inherited theories of governance are not even much help in understanding the issues. They tend to lurch between high-minded views that make the duty to govern seem impractical and narrow-minded views that make it seem impossible. Many of the former descend from one interpretation of Aristotle's idea that the free person is one who rules and is ruled in turn. Sharing in governance is not, on this account, merely one of the liberties of citizenship; it is a fundamental duty constitutive of it.Footnote ⁷ In the modern era it was Rousseau who carried this conception furthest, concluding that the general will is therefore incapable of representation and that the duty to govern is a duty to govern directly. The attractions of civic republicanism have never faded, and there is much that remains admirable in that tradition. But Rousseau also identified its difficulties: it depends on social foundations (tiny states, social equality, and common experiences) that are very remote from modern political life. Large, inegalitarian, multicultural, and anonymous societies are not environments in which direct and continuing engagement in politics is likely to thrive, and the attempt to achieve it in those circumstances can, as Rousseau himself warned, easily backfire.

An opposing current denies the very coherence of a duty to govern. Blackstone had held that not only does the supreme power have the right to make any law whatever, “but farther, it is its duty likewise. For since the respective members are bound to conform to themselves to the will of the state, it is expedient that they receive directions from the state declaratory of that its will.” To fulfill this duty to legislate, the sovereign has at least to promulgate general rules. Bentham balked. If the sovereign has a right to govern, “Its duty then is to do—what? To do the same thing that it was before asserted to be its right to do, to make laws in all cases whatsoever. . .” Bentham was certainly not unfriendly to the principles of legality that Blackstone derived from the duty to rule—in fact, he tended to make a fetish of them. But Bentham could not see how a sovereign could both have a right to legislate as it pleases and also be subject to an obligation to legislate. Sovereigns are not subject to legal obligations (or else they are not sovereign) and they are constituted by the fact that they govern, so, concluded Bentham, they cannot have either a legal or a moral duty to do that without which they do not exist in the first place.

There are replies, qualifications, and modifications available to both civic republican and imperatival theorists of governance. John Finnis's account refreshingly bypasses all such fiddling. He argues instead that some have a nonvoluntary duty to govern, grounded in their effectiveness at a morally necessary task. In doing so, Finnis not only recovers a central but neglected problem of legal philosophy, he proposes an answer unencumbered any nostalgia for the town meeting or the unitary sovereign.

II. THE PRIMARY DUTY TO GOVERN

The problem at issue is one of justifying a primary duty to govern. This is a duty not derived, directly or by delegation, from any positive right to govern. In functioning legal systems, many people have derivative duties to govern (or, more commonly, to do their part in a system of governance). American presidents have a duty to execute their office and to preserve, protect, and defend their Constitution. Subjects of common-law systems have a duty to serve on juries if asked. Australian citizens have a duty to vote. Were there a general moral duty to obey the law, there would be a duty to obey these laws and thus a derivative duty to govern in these ways. But this does not reach our problem. It may reasonably be doubted whether there is a prima facie duty to obey, or at any rate, one that applies to all of law's subjects and on all occasions on which their obedience is required. More important, the sort of duties just enumerated depend on the contingent existence of laws that attempt to secure them. But we are seeking a justification that covers also the primary duty to make and apply laws, possibly including laws such as these (or to provide a setup through which laws can be made if necessary). I shall also assume that the duty at issue is a positive one: not merely a duty not to interfere with morally justified governance that is already up and running but a duty to get it going in the first place. Put another way, this duty is supposed to bind as much in “the state of nature” as in any functioning legal system. So it comes to this: Who, if anyone, has a primary, positive duty to govern, and on what grounds?

This problem is clearly related to the two other main problems of political authority—legitimacy and obligation—but it remains conceptually distinct from them. A theory of legitimacy may provide warrant for someone or other ruling—it may show they are at liberty to do so and even that they have good reason to do so—without going so far as to claim that they violate any duty by failing to do so. And a theory of political obligation presumes that someone is already ruling, and whatever justification it offers for obeying their directives need have no bearing on the primary duty to govern. Suppose, for example, that fairness requires complying with the law whenever you are assured that others will too, or even that it requires doing one's share in ruling subject to similar assurance. None of this proves it to be somehow unfair if nobody rules at all or if social order is maintained without resorting to governance. It does not follow that the problem of governance must have a solution independent of our arguments for legitimacy or obligation, only that there is no conceptual necessity for them to fall together or all to stand on the same principles. Whether they do or not is a matter for substantive argument.

III. ON EFFECTIVENESS

Finnis says that “Authority (and thus the responsibility of governing) in a community is to be exercised by those who can in fact effectively settle co-ordination problems for that community.”Footnote ⁸ This is a normative thesis. Political authority, including legal authority, is in fact exercised on many other bases, including self-interest, charisma, superstition, and so on. Those who make themselves “masters of society” may do so in order enrich themselves and their friends and not even try to settle any sort of problems. But in the realm of political authority and obligation there is a connection between the actual and the ideal: self-interest, charisma, and so on may stand in a causal relation to the capacity to settle problems effectively. The fact that someone is a charismatic leader, for example, Churchill during the Battle of Britain, may contribute to his effectiveness in settling problems of strategy by sustaining morale and inspiring obedience during wartime. An authority is not deficient qua authority if its effectiveness rests on such bases, and that applies even to superstitions about the divine rights of kings, the virtues of aristocrats, and so on. (There is at least that much truth in Hume's attack on the Whig contractarians, though there may be other objections to authority exercised on such grounds.) The claim is that authority should be exercised when there is a certain kind of problem to be solved, and that it should be exercised by someone who has the effective capacity to solve it. At this point, normative power depends on actual social power.

It is important to grasp how radical Finnis's version of this idea is. Others have suggested that effectiveness is a necessary condition for justified political authority.Footnote ⁹ If authority's role is to secure some valued end, be it justice or finality in social ordering, then it is bound to count against a putative authority that it lacks any capacity to do so. It may be that in 1745 the Young Pretender still had the best right to the British crown; but it is certain that after the disaster at Culloden the political claims of the Jacobites became a fantasy shrouded in tartan. Few legal philosophers (and fewer courts) would now doubt that. What is striking is not only that Finnis regards effectiveness as in such ways necessary for justified authority but that he also regards it as defeasibly sufficient: “the sheer fact of effectiveness is presumptively (not indefeasibly) decisive.”Footnote ¹⁰ Indeed, a casual reader may be shocked by Finnis's repeated insistence that raw power plays a pivotal role in both the right to rule and the duty to govern. It sounds uncomfortably close to the claim that might makes right—and that is not the sort of thing we expect to hear from a natural lawyer.

We should learn to tolerate this discomfort. Effectiveness (and, more generally, social power) is a feature of law of which analytic jurisprudence seems embarrassed. It is as if, having exfoliated those reductivist theories that identify law with a certain structure of power or with predictions about the deployment of power, we have become reluctant to explore other relations that may hold between social power and law as a normative system. We acknowledge that law is not a matter of what we are obliged to do but of what we are obligated to do; but it is also true that what we are obliged to do can affect what we are obligated to do. We say that law consists not only of duty-imposing rules but also of power-conferring rules; but normative power is not the only sort of power that those rules confer. We hold that a society of angels could have still have need for lawFootnote ¹¹ and thus that coercion is not of law's essence; but there is also a pervasive association of the ultimate capacity for coercion with the concept of law.

Even Hans Kelsen, who asserted that legal systems are distinguished from all other forms of social organization by their deployment of coercive force, dithered on the contribution that effectiveness makes to law, ultimately offering the somewhat opaque suggestion that “Effectiveness is a condition for validity—but it is not validity.. . . right cannot exist without might and yet is not identical with might.”Footnote ¹² Whatever a “condition” for validity is supposed to be, it cannot here amount to a presumptive justification for it. Kelsen regarded moral justification as an ideological matter to be excluded from a “pure” theory of law. But neither can it contribute to a legal authorization, for, as Kelsen also insisted, the reason for the validity of a norm can only be another valid norm. After hesitation, he eventually came to see effectiveness as a requirement of the basic norm of international law that confers authority on de facto municipal authorities. This norm supposedly directs states to treat as binding the norms they customarily treat as binding, including the norm that those exercising persistent, effective control over a territory have lawful authority.

Of course, the normativity that customary rules contribute is social normativity, and if you think that there is a sharp divide between “is” and “ought,” you are likely to put them in the former box. But then they seem not to be norms at all, and that was the deeper source of Kelsen's ambivalence and his reason for postulating the reduplicative basic norm, “Do what your customary norms tell you to do!” H.L.A. Hart skirted that problem by sticking to social normativity and by arguing that effectiveness is neither a condition nor a precondition for the intelligibility of normative discourse in law; it is simply a presupposition of the normal contexts in which we use it.Footnote ¹³ We do not normally help ourselves to terms such as “validity,” “authority,” or “obligation” unless the legal system in question is a going concern, any more than we talk about champions of matches not actually played. This is not meant to solve the justificatory problem. But it is unclear how far it even solves Hart's own problem of explaining the use of normative language in law, for the “normal” context is not the only context. We also talk about valid rules under extinct legal systems and about what rights and powers a proposed constitution would confer, in each case free from any presuppositions about effectiveness. It is unclear why these are not among the “normal” contexts of legal discourse, unless we fix the norm according to the interests of lawyers and their clients, whose typical focus is indeed on what is rather than what was or what could be. But that seems as arbitrary as fixing it according to the interests of Holmes's “bad man,” whose only concern is when to duck.

Finnis's argument is very different from these. He considers that the nature of law is best seen in its “focal” case, and that is the case in which legal obligations are truly morally binding. His argument for treating them as such gives to effectiveness a key role in the justification for authority, not in its logical preconditions or its background presuppositions. And the effectiveness he invokes is not Kelsen's or Hart's, namely, the fact that the norms of the legal system are generally obeyed and applied.Footnote ¹⁴ Finnis has in mind effectiveness at a certain nonlegal task: that of settling coordination problems for a whole community. When that sort of effectiveness settles on certain people, they acquire not only a right to rule and a correlative duty to be obeyed but, more important, a prior duty to rule. The task is for them not optional. Contrary to Kelsen and Hart, Finnis thus holds that under certain conditions, might (of a certain kind) does indeed make right.

Before panic sets in, we had better say something about these conditions. First, they are limited to the broad task of “coordinating” human activity for the common good. According to Finnis, this consists in maintaining certain framework relationships in our pursuit of a list of objective, mutually irreducible, self-evident values that include life, knowledge, play, aesthetic experience, friendship, practical reasonableness, and religion. These values are themselves “common goods” in that they are good for anyone and everyone but are not subject to scarcity or rivalness.Footnote ¹⁵ Their reasonable pursuit by individuals gives rise to a problem of securing the common good, by establishing a set of framework conditions within which individuals can pursue life plans oriented to the common goods. And in securing this effectively, rulers must conform to a general requirement of practical reason that no violence is to be done to any one of the common goods. So while might is indeed a presumptively sufficient condition for legal right, it is so only in a context in which it is actually oriented toward the common good and in which all the fundamental goods, such as life and religion, are properly respected.Footnote ¹⁶

It is central to this argument that while no unique set of framework conditions is required by reason, failure to secure some such framework is unreasonable because it will frustrate the pursuit of the basic goods, leading not only to gridlock but also to injustice. We need to settle with finality such issues as how children are to be educated, how natural resources are to be managed, and how claims of right are to be reconciled or adjudicated. Most of these can be fairly settled in more than one way, so to eliminate indeterminacy—to ensure that there is actually a settlement—we need to fix on one of the (satisfactory even if suboptimal) solutions to the exclusion of all others.Footnote ¹⁷ That is the role of political authority. It is on this structure of practical reason that the positivity of law rests, and it is a structure closely related to Aquinas's notion of determinatio. The dual contribution of the common good and the capacity to select and secure a satisfactory-if-not-necessary means to it explains the justified normativity of positive law, at least in its focal case.

I have previously raised doubts about some elements of this theory,Footnote ¹⁸ including its claim that resolution of coordination problems depends on treating one alternative as authoritatively binding, as providing “a reason for judging or acting in the absence of understood reasons, or for disregarding at least some reasons which are understood and relevant and would in the absence of the exclusionary reason have sufficed to justify proceeding in some other way.”Footnote ¹⁹ It is doubtful whether that is typically needed in the situations Finnis has in mind, and doubtful too whether a “co-ordination problem” is a helpful model (as opposed to a permissible name) for these situations. There is also some uncertainty about the idea of overall social coordination. (Even if every coordination problem needs a solution, it does not follow that there is a [common] sort of solution that every problem needs.) And the law notoriously sustains a fair amount non-meshing, pockets of anti-coordination, and so on, so the overall picture usually looks more modular than homogeneous. But it may be possible to reformulate the core thesis to deflect some of these worries, so I simply bracket them here.Footnote ²⁰

Finnis rejects the idea that political authority is solely a matter of what I have been calling legitimacy: the liberty-right to set general rules and to enforce them against their subjects.Footnote ²¹ He accepts instead the common view that it involves treating directives as claiming an obligation of obedience. In a sympathetic exposition of Aquinas's position, he writes:

Note that in this formulation the common good calls for lawmakers. This is the key move: in ordinary (modern) circumstances there must be law, and the task of making it, or at least providing the means by which it might be made, if necessary, inherently falls to those who are able to do so. It is true that if law is to serve the common good, lawmakers also need a right to rule and a correlative right to obedience commensurate with their task. Hence, as a heuristic or expository matter, we might begin to explore these ideas at any point: at the grounds of legitimacy, or the right to be obeyed, or the duty to govern. But as a matter of the logic of justification, there is a priority relation among them. Finnis is committed to the view that there is justificatory primacy to the duty to govern. The liberty-right to do so and the other normative powers necessary to accomplishing it must be ultimately explained by reference to grounds and limits of that duty.

IV. TASK-EFFICACY

Finnis's theory of governance falls into a class that I call arguments from necessity.Footnote ²³ Elizabeth Anscombe sets out the archetype: “If something is necessary, if it is, for example, a necessary task in human life, then a right arises in those whose task it is, to have what belongs to the performance of the task.”Footnote ²⁴ Anscombe's preoccupations being with the right to rule and the duty to obey, she does not say much about how we are to determine “whose task it is.” Finnis offers an answer: the task falls to those who can as a matter of fact effectively settle problems of coordination for the common good. I am going to call this a task-efficacy justification for the duty to govern.

There are cases in which task-efficacy has this force. Suppose A arrives at a crossroads and finds a serious automobile accident dangerously obstructing traffic. A could drive on but is in no rush and can without danger to herself pull off the road and safely wave oncoming traffic around the collision. She phones for an ambulance and directs traffic until it arrives. Here, A's right to direct traffic depends on her capacity to do so, which in the circumstances gives others a reason to comply with her directives.Footnote ²⁵ Moreover, in view of the nature of the benefit to others and the small inconvenience to herself, she also has a moral duty to so as a specification of the duty of beneficence (whether in general or toward those with whom she shares a territory or even just a road). The hard-hearted might deny this, but to assert that A has in these circumstances no moral duty to direct traffic would probably also force the conclusion that she does not even have a duty to phone for an ambulance, which seems more than a bit severe. (Remember that we are not worrying about whether anyone ought to have an enforceable remedy against A in case of her breach of her duty; we are just worrying about what she is bound to do.) We could therefore say that owing to the universality of these duties and the general importance to everyone of being able to count at least on an easy rescue, the duties have adequate foundation in the requisites for the common good in Finnis's sense (a good that is not a common destination but is a common framework for any destination). There is a task that needs to be done, and here A finds herself the only one in an easy position to perform it.

Does this suggest that task-efficacy generates duties only in cases of very “easy rescue”? The case of belligerent occupation suggests otherwise. Suppose that in the course of war B invades C and occupies part of its territory. The citizens of C are entitled to go on as far as possible with their ordinary lives, but to do so they need someone to organize the supply of food, water, and medicine and to suppress looting and violence. Their own government has been completely disabled by the occupying forces. In these circumstances B is plausibly bound by morality, and plainly bound by law,Footnote ²⁶ to rule C, and in particular to maintain law and order. The occupant's duty to govern derives from its duty to preserve the rule of law (if any remnants remain) or to establish it—even if it destroyed it in the first place. This derives from the natural duty to support institutions necessary to prevent the situational harms of anarchy or the harms of governance radically deficient in the virtues of legality. Common moral thought (in unison with international law) holds that this duty applies whether the invasion itself was just or unjust, and it binds even in the face of serious costs to B. Indeed, to occupy a country and then to permit mayhem because keeping order is now hazardous or expensive to the occupant may be morally worse than waging an illegal war in the first place. B may thus be expected to bear very substantial costs not as some kind of punishment for occupation but because now B and only B is in any position to reestablish the order on which ordinary life in the occupied territory depends and because securing this outcome this is of the greatest importance.

Note how task-efficacy figures in the traffic and occupation examples. If A cannot direct traffic or if it is unlikely that anyone will obey her, she had better just get out of the way. If B cannot actually keep order (for example, because B's forces and agents are incompetent or corrupt or because B's rule provokes massive resistance), then B should withdraw from the territory and let someone else try to rule. Whether task-efficacy is here sufficient depends on the force of the underlying duties of beneficence and the duty to support just institutions, together with the absence of any defeating conditions. It does not depend on any voluntary undertaking by A or B (though in these cases it falls to them as a consequence of their voluntary acts—the act of driving to that very crossroads, the act of occupying that very territory).

Waiving any doubts about the stringency or scope of the duty of beneficence, there are three other problems to face if we are to make this argument work. The first is obvious. It will often be difficult to judge whether anyone has the capacity for the task, and that judgment may have to be made before the task is even attempted. In many cases it may involve a leap in the dark, the only safety net being the fact that one may sometimes get the capacity to perform the task ex post actu. The mere act of getting out of the car and moving one's hands as if directing traffic may make A someone capable of providing a solution to this problem or at least alerting others to the fact that a problem lies ahead. Here, nothing will succeed like success.

The second worry has to do with an artificial feature of these examples. In my sparse descriptions, things were fixed so that there was but one person or group in a ready position to govern (the first driver to arrive, the only army in occupation). When there are multiples ones, we have not only the problem of working out what (set of) people have the requisite capacity but also the problem of selecting among them. At this stage we will need to think about allocation procedures (such as elections or lotteries) and therefore about the existence of higher-order coordination problems. (How do we select among possible rulers or among possible decision-rules to select rulers?)

The final worry runs deeper. As Anscombe herself notes, the concept of a “task” already imports some degree of necessity—a task is something that must be done, needs to be done, or for some reason ought to be done. So in describing the issue in this way, perhaps we have stolen the conclusion that needs to be purchased by argument. What turns a possible governing function—to use a neutral term—into a necessary task, one capable of defeating the familiar reasons we have for being let alone to tend our own gardens? Most answers turn on the putative importance of the function. Anscombe has in mind tasks that serve “general human needs,”Footnote ²⁷ Finnis, those that serve “the common good,” and George Klosko, ones that provide “presumptively beneficial public goods”—goods that anyone would want but which require cooperation and mutual restraint to produce.Footnote ²⁸ Without going so far as to suggest that theories amount to much the same thing, there are obvious commonalties among them, and even a family resemblance to Hart's idea that anything we would be prepared to count as a full-blooded legal system must aim to secure a “minimum content” of protection for the most urgent interests of at least some of its subjects.

To choose among these variants would require an argument that cannot be pursued here. But there is an interesting feature of all of them: however we understand “necessary tasks,” they are likely to carve out a duty to govern that is much narrower than the claims of modern states or the scope of legitimate governance. Yet the fact that a certain activity is not a necessary does not entail that it is not permissible. While one needs a right to do whatever one has a duty to do, one may have no duty to do things that one has every right to do. A legitimate government may establish national holidays and official languages, protect historic buildings and support the arts, and declare the trillium to be the provincial flower of Ontario. If we say that the right to rule must be ancillary to and strictly commensurate with the necessity of the task, we will be driven to conclude either that the task of governance is a lot narrower than most people think (and certainly narrower than what most states claim) or that these seemingly nonnecessary, permissible activities actually do fall under some suitably abstract conception of a necessary good. (Play?)

This marks an interesting asymmetry between the adequacy criteria for a solution to the problem of governance and the adequacy criteria for a solution to the problem of obligation, and this in turn gives us an additional justification for distinguishing them. It always counts against a theory of political obligation that it cannot justify the duty to obey actual and permissible directives of a reasonably just state. The directives set the standard of success: the duty to obey the law is a duty to do what the law requires. But it may well be that the duty to govern is narrower than the right to govern and that a satisfactory theory of governance is therefore not hostage to the spectacular failures of scope that plague most justifications for the duty to obey. This is one of the more interesting conceptual payoffs in isolating the problem of governance. A theory that cannot hope to justify a general obligation to obey the law as it claims to be obeyed may nonetheless succeed in justifying the existence of a primary duty to govern. We may therefore discover that some failed theories of political obligation turn out to be adequate theories of governance.

V. WHAT IS THE TASK?

Let us take a closer look at the putative task. According to Finnis, the authorities' typical job is to “adopt” one of the set of reasonable solutions to the framework problem of serving the common good. If the thought is that they need to select, mark, or make salient one of the options, then we need to ask whether this requires authoritative regulation. Many familiar coordination problems (which language to speak, what currency to accept, which keyboard layout to use) are solved without authority of any kind, and so it must be at the foundations of law, since the ultimate sources of legal authority cannot themselves rest on any authoritative legal act.

But perhaps this misconstrues what I have been calling the “task.” Finnis, along with many others, writes as if the problem is how to select one solution from an acknowledged list of satisfactory ones, that is, how to effect a determination. But this may have things backwards. Raz argues that the subjective conditions that characterize coordination problems—the fact that people acknowledge at least two alternatives, either of which could further the fundamental goal but neither of which can do so without general conformity—are often part of the solution, not part of the problem. “[T]he problem is to get people to realize that they are confronting a coordination problem.”Footnote ²⁹ That accomplished, the solutions may take care of themselves (in the way that many coordination problems do). On this view, duty to govern (and the derivative right to rule) rests not on the capacity authoritatively to solve problems, as in the case of directing traffic or belligerent occupation, but on the capacity authoritatively to identify problems of a soluble sort. Practical authority will then be only one of the tools to which wise rulers need resort. Their primary duty to govern rests on their wisdom (and perhaps on their theoretical authority) in spotting coordination problems that need a solution. This will apply, however, only to certain types of problems. In the case of prisoner's-dilemma situations, full information about the nature and structure of the problem may even make cooperation less likely (for people will see very plainly the advantages in defection). But in some circumstances, it will help meet the objection that many acknowledged coordination problems, even socially important ones, need no authoritative intervention whatever.

Common knowledge of our circumstances cannot simply be assumed. One of the hardest tasks in law and politics is to get people to understand the need for cooperation, especially when it is very complex or involves people unlike or remote from themselves. Two sorts of error are common. First, there may be a need for cooperation that is not adequately felt. Managing climate change poses a coordination problem in Finnis's sense if anything does, and maintaining the planet as a viable habitat for Homo sapiens is as clear an example of a humanly necessary task as we have. Any nation or group that could come close to providing an effective solution, or even steps toward an effective procedure leading to a solution, would have a powerful obligation to do so. But some people, owing to ignorance, self-deception, or willful blindness, do not see that this is a task calling for cooperation of an unprecedented kind. The second type of error involves deeply felt coordinative “needs” that are in fact illusory. In some societies there is a felt need to organize the ritual cutting or amputation of parts of children's genitals without anything that could possibly count as their consent or that would be a morally adequate substitute for it. It would be much be better if that “need” were not felt (which is not to say that others are thereby authorized to eliminate it).

Does the possible role of theoretical authority in reducing such errors sound the alarms about Platonic guardians? Jeremy Waldron seems to think so. He says that a theory of authority ought to orient itself to “questions of common concern” wherever there is a need for a settlement, but also that “Which questions actually are questions of common concern in this sense is not something which a theory of authority ought to settle.”Footnote ³⁰ This depends, I think, on what one takes a theory of authority to comprise. If it includes, as Finnis and I both accept, a theory of legitimacy, then it is hard to see how it can avoid taking a view about what sort of things governments can and should undertake and thus a view of what the questions of common concern properly are. Is the worry that we do not want authoritative settlements of those questions? Do we fear that as the supremely salient coordinator, the ruler is always in a favored position to discover “tasks” that only it is well positioned to fulfill—a sort of moral Keynesianism? The short answer is that one who does so makes moral errors and, if they are frequent enough, loses theoretical as well as practical authority.

In these cases, duty to govern rests not on the capacity authoritatively to decide what needs settlement but on the reliable capacity to know what needs settlement, how to assist in achieving it, and when to apply authoritative guidance to do so. Admittedly, people may disagree about all of this, and even when there is a right answer to the question, there may be no reliable method of reaching it. Once we get beyond simple examples like traffic direction or belligerent occupation, things can get murky. (Some might even query Finnis's assertion that “someone must decide” how children are to be educated:Footnote ³¹ if that means that someone must settle the substance of a common educational policy, then J.S. Mill doubted it.) Should we therefore say that in the face of such disagreements, we need to retreat to some favored procedure—for example, majority rule or bargaining—and allow the question of a duty to govern to arise only when that procedure delivers a decision about the appropriate scope of governance? You can see why this is tempting. If only it could work. Part and parcel of the disagreement about the scope of governance is disagreement about what sort of procedures are a reasonable response to this disagreement. And then we need to know when to invoke any such procedure, and that is in turn liable to further disagreement, which cannot itself be settled by bargaining or voting. As Montesquieu says, not everything can be decided by a vote, including the question of when we should decide things by a vote.

VI. THE QUESTION OF PRIORITY

Such are the foundations of Finnis's account of the ultimate duty to govern (or, to be safe, of an account inspired by Finnis's theory). What, then, are the relations among the duty to govern, the right to rule, and the duty to obey? As I understand Finnis, governance comes first. (He is not alone in thinking this.)Footnote ³² I do not mean that the issue has any methodological priority; as I said, we can begin thinking about political authority at whatever point is convenient. Nor do I mean that we can fully understand the concept of governance without reference to what it implies for related notions. But the question is not about logical correlates, it is about justificatory reasons. On a task-efficacy account, the justification for a duty to rule also influences its proper scope, and that in turn affects the permissible responses of its subjects.

In his various formulations of the theory, the first normative relation Finnis derives is always the responsibility of governing, whether that falls to the revolutionary who is now among the “masters of society” and therefore needs to think about the ground rules that will henceforth regulate the legal order or to the practically reasonable person who has the capacity to select or notice solutions to society's coordination problems. This is reflected in Finnis's insistence that “public authority is not merely a moral liberty but essentially a responsibility (a liberty coupled with, and ancillary to, a duty).”Footnote ³³ Now, if A is ancillary to B, then A is not merely a logical correlate or conceptual reflex of B; they are not just two sides of a coin. A is ancillary to B only if A is in some way auxiliary or subordinate to B. This is no minor point. To say that the right to rule is ancillary to the duty to govern is to take sides against that strand of the liberal tradition that puts legitimacy first. The emergence of this distinction (and of the related idea that there are moral rights that are not simply the logical the reflex of duties) was pivotal in the intellectual revolution of seventeenth-century political theory.Footnote ³⁴ It set in motion a debate that was protracted and possibly inconclusive: Are human rights something that fall out of an independently intelligible theology of divine command or teleology of the common good, or do they have some kind of moral primacy that is susceptible to independent explanation?

The passage just cited represents, I think, Finnis's official position. But at other points he seems influenced by the idea that there are efficacy-independent foundations of political legitimacy. For example, he says that ‘the goods which define the range of lawmakers’ and other rulers' responsibility—say, the goods of peace and justice—can be called the common good of, or specific to, the political community or state.”Footnote ³⁵ That suggests the duty to govern, to which the right to rule is ancillary, is not an open-ended duty responsive only to the mandates of the common good. There are things that are Caesar's, and there are things that are not. So while governance must serve the common good, it may not serve just any and every aspect of the common good but only those roughly demarcated as the provision of “peace and justice.” Why should that be?

One familiar explanation holds that to transgress such limits is to trench on the legitimate interests of individuals, interests that should remain under their control even if they are interests in common goods (such as religion or friendship) and even if retaining such control means that they will achieve these goods less well than they otherwise would. Simply put, the explanation flows not from limits on effective governance but from the independent force of a doctrine of legitimacy that prizes something in the neighborhood of individual (or group) autonomy. That suggests a different understanding of the priority relation: we need to explain the kind of goods that fall within in the legitimate sphere of governmental authority, and only then and subject to that constraint can we work out who has any duties to provide them and what sort of attitudes their subjects should take.

How much tension this idea creates within a task-efficacy argument depends on how we explain the boundaries of the rulers' special responsibility. I am not confident of Finnis's position on this point. He clearly asserts the ancillary character of the right to rule. He also endorses the principle of “subsidiarity”: the state should not try to rule where smaller, intermediate institutions (or individuals) can adequately govern. At the same time, the fact that “particular individuals and groups have as their prior concern (as they should) their respective interests” is part of his reason for thinking that overall coordination requires authority.Footnote ³⁶ How, then, should we interpret the idea that individuals have their own particular interests as their proper prior concern? Might this hint at some efficacy-independent constraints on legitimacy?

There is in our political tradition one familiar line of argument that makes legitimacy firmly subordinate to the duty to govern. Its paradigm is parental authority, and for some time it cast a long shadow over theories of governance and legitimacy. The most plausible source of parents' rights to instruct and discipline their children flows from their prior duty to attend to the well-being of their children, together with certain facts about the intimacy of the parent-child relationship that make the duty especially theirs. As long as rulers felt safe that there was likely to be no earthly enforcement of such duties, they were quick to exploit the analogy. James VI of Scotland wrote that:

The duty to govern may be part of a “mutual and reciprocal bond,” but James considers that his end of the deal derives solely from his divine duty as pater patriae, not from any contract with his earthly subjects. In Locke's view, such an analysis of parental authority is fair; the question is only how far it carries over into the case of political authority, which has to apply to adults who confront each other as moral equals:

The limited parental right to rule over children is wholly ancillary to the duty to govern. In fact, it does not amount to much more than that duty as seen from the parents' point of view. Though it is for the benefit of their children, parents are accountable to God for their performance of this duty. That coincides with the classical view of how natural law can generate subjective rights. But for Locke this distinguishes parental from political authority. He maintains that the latter can be established only by actual (though perhaps tacit) consent of the governed and is therefore subject to all the validity conditions for such consent (including, famously, the condition than no one may consent to any sort of rule that puts his life at the whim of his ruler). It is this rather than any functional appeal to the things that are somehow Caesar's that explains the differences between parental and political authority.

In emphasizing the difference between these two traditions of argument, I am not trying to hint that Finnis's theory is latently absolutist or even especially paternalist. It is a theory of limited government in which the limits are traceable upstream to the common good and have downstream effects on both the right to rule and the duty to obey. The only issue is whether there are further limits that flow directly from constraints imposed by the interests of individuals independent of the inefficacy of authority at certain tasks, including the task of securing the interests of individuals as participants in the common good. It is a fine point, but on it turns the question of which of the relations constituting authority is the primary one.

VII. THE ALIENABILITY OF THE DUTY TO GOVERN

This paper has been mostly exploratory, examining the many ways in which Finnis's account of governance is both illuminating and distinctive, and flagging some issues for further inquiry. I want to conclude with a few remarks about its capacity to handle the worry I raised at the outset: the widening delegation of powers of governance to private entities. The issue is complex, and I can make only a few remarks about it here.

As we have seen, a task-efficacy theory of authority has to explain how the need for governance becomes the duty of a particular person—your duty, for example. Sometimes, there may be only one feasible candidate; more often, choosing among them will be resolved by chance or custom or convention; sometimes it will fall to an explicit decision procedure. Suppose that in some way or other, we determine that the duty falls to X (who may, of course, be a person corporate). Does this mean that X must govern, or is it enough if X sees to it that someone governs?

I assume that any plausible theory will allow for a division of labor in governance and, contra Rousseau, there is no principled objection to limited and controlled delegation of some legislative authority. There are also cases in which private arrangements are more efficient or transparent than public ones and where wisdom therefore dictates that we should not be governing at all. But there also comes a point where the core responsibilities of necessary governance are delegated to a degree or in a direction that gives rise to the worries I mentioned earlier. We should, I think, have grave reservations about any broad delegation of the primary duty to govern. Even if Dicey was right in thinking that the U.K. Parliament could lawfully commit suicide by transferring all its powers to the Manchester Corporation, it can scarcely be denied that this would be regarded as an immoral alienation of its duty to govern.Footnote ³⁹ Or if a municipal government seems a tolerable recipient of sovereign powers, imagine instead that Parliament transferred them to the Disney Corporation.

Actually, we begin to have serious worries well short of fantasy cases of total and permanent alienation. A significant delegation of the U.S. government's primary duty to legislate to the Brookings Institution or its primary duty to adjudicate to the American Arbitration Association—to say nothing of its primary duty of enforcement to Blackwater USA—would be properly deplored quite apart from any doubts about the constitutionality of such delegation. This suggests that there may after all be a nondelegable core to the duty to govern and that Rousseau may have been right in part.

If so, this presents a difficult hurdle for a task-efficacy theory. In general there is nothing wrong with delegating a task, even a necessary one, to someone who can do it at least as effectively as you can. Finnis's case for the presumptive sufficiency of the effectiveness criterion suggests as much. Here, the best a task-efficacy account can do is to say that one may not delegate governance in a way that would substantially interfere with effectiveness, for instance by reducing oversight, distorting incentives, or making people uncertain as to where authority really rests. Does this adequately explain our sense of what would be wrong with radical delegations of primary authority? I doubt it.

Locke maintained that parental governance may properly be delegated, as it often was, to nursemaids, governesses, and teachers.Footnote ⁴⁰ But that is because he thought the parents' duty to govern their children is really a duty to see to it that this task is accomplished. Our own moral attitudes toward the family are somewhat different: we expect a certain amount of hands-on, first-personal care of children, even from parents who could easily afford to farm their children out.Footnote ⁴¹ In any case, Locke also recognizes non delegable familial obligations: children are perpetually and indispensably obliged to honor their parents, and in the nature of the attitude required, that cannot be delegated at all. An adult child may hire someone to care for his aged parents, but he cannot hire someone to honor or love them. Perhaps the fundamental duty to govern shares some characteristics with the duty to honor a parent, to the extent that it, too, is anchored in a special relationship. Locke writes:

This is the idea that takes a more radical form in Rousseau's hands.

Along with most contemporary legal philosophers, Finnis refuses any foundational role for consent in a theory of authority. That fact that someone enjoys the consent of the people may contribute to his effectiveness as a ruler; it may make it more likely that his rule will be accepted; it may mark him as a source of salience in solving coordination problems. But these are just about the only roles consent has, and they flow, significantly, from its attitudinal rather than performative nature: they flow from consent as contentment rather than commitment.

The common philosophical view that consent theory is bankrupt contrasts with its popularity among subjects of modern states. This may be irrelevant, although those who resist consent theory on the ground that they think its consequences out of line with popular opinion about the scope of the duty to obey might want to be cautious before embracing that view. Three main arguments are offered to undermine a consent theory of obligation: there has been no social contract; more people believe they have a duty to obey than have actually promised to do so; and consent is valueless unless it is valid, and when we give full weight to its validity conditions, we see that they do not require any transmittal of authority. If consent would have been valid if given, then that is enough; if it would not, then actual consent cannot help.

I tend to think that these objections can be met. (Why should consent be a contract? Why assume that everyone does have an obligation to obey? Why do we not think that people are bound to do what it would have been reasonable for them to promise, even if they did not actually promise to do it?) But put all that to one side. Once we distinguish the duty to govern from the duty to obey, the route is open to a consent-based theory of governance, even if we reject a consent-based theory of obligation.

Even if it is false that every subject has promised to obey, it may yet be true that many become rulers because they have promised (or otherwise committed themselves) to undertake the task of governance. Surely one reason we think fundamental delegations of the duty to govern are wrong is that we have nominated, elected, and appointed people with the intention that they should govern and that we have done so because they have deliberately put themselves in the way of being nominated, elected, and appointed to govern. Whatever we decide about native-born subjects and naturalized immigrants, our rulers are generally volunteers, not conscripts. Through a variety of procedures, we have placed our trust in particular people or particular political parties, not just in anyone who turns out to have the capacity to solve coordination problems for the common good. Of course, the fact that we have done so contributes to their effectiveness in solving such problems, but now the order of explanation is running the other way round. Some such special relationship between ruler and ruled—whether we think of it as flowing from consent, trust, or perhaps honor—is commonly thought central to governance, and it resists both delegation and explanation in terms of task-efficacy.

This is not to deny that the duty to govern may be supported by more than one type of consideration. Even if these worries about delegation make us wonder whether task-efficacy is the whole story, it may well be part of the story. And in any case, by calling our attention to the existence and importance of the duty to govern, Finnis has reopened important questions about power, authority, and law—questions that have too long been neglected in our obsession with legitimacy and obligation.