Rules and Rulers:  Demanding, Commanding and Assuming Responsibility

I. COMMANDS AND DEMANDS

Let’s start with a set of binding social norms which have yet to be institutionalized and so are devoid of officials entitled to issue instructions. Call it a system of customary rules.⁵ Imagine a village where neighbours feel obliged to relate to each other in certain ways e.g. to greet each other and lend each other tools. Over time the villagers have evolved a practice of following these rules and evaluating each other’s behaviour in terms of them. This process involved no lawmaker and no collective agreement; without any design these rules came to be generally accepted. The rules are not codified, there is no canonical statement and they form no integrated system. Furthermore, there is no procedure for adjudicating breaches of the rules, neither for settling disputes of fact nor for applying the rules where the facts are agreed, nor for collectively enforcing compliance where a rule might be violated. Since the rules are vaguely sensible, they genuinely bind the villagers but these norms set up no directive authority; we have rules without rulers.⁶

Customs can be onerous: one villager may be obliged to lend a valuable tool that another villager has requested. Where the former shows reluctance, the latter can demand the tool and where the demand is refused, blame, indignation etc. both from the neighbour and from others is to be expected. In an undeveloped society people greatly depend on those around them and their norms of neighbourliness may authorize many such demands. Not all villages are egalitarian; perhaps age gives you high status and the village elders enjoy special privileges. Even so one important form of hierarchy is absent: the customs of the village have yet to require one neighbour to obey the commands of another.⁷

Command and obedience enter the picture once authorities are established either to apply customary norms or to create new norms. A de facto directive authority exists in a community where a sufficient proportion of its members regard themselves as being bound by the commands of the authority. There are other ways of ensuring that people reliably do as you tell them. The protection racketeer does not try to convince their ‘customers’ that there is any obligation to pay up; without claiming any authority over them, they simply make it clear that it would be wise of them to pay.⁸ Directive authorities operate differently; they issue commands in the expectation that their subjects will feel obliged to comply.⁹ Directive authorities may also claim the right to enforce their commands – we’ll return to that – but any such claim is based on the prior idea that we are bound to obey them. No doubt habit, fear or even positive enthusiasm often play a role in ensuring that the orders of an employer, a teacher or a state official are followed but, to function effectively, hierarchical organisations of any complexity must also inculcate the idea that subordinates are obliged to obey their superiors.

A directive authority is legitimate where its subjects are genuinely obliged to obey at least some of its commands. In this paper I shall not be offering a full account of legitimacy, of exactly when de facto authority constitutes de jure authority, of which (if any) of the commands of a socially recognised authority bind. In the first two sections, I’ll be assuming that some commands issued by some authorities do bind, that it makes sense to obey them simply because they have been issued rather than for some extraneous reason (e.g. because disobedience will ferment social disorder, perpetrate an injustice or entail punishment). Though falsifying a tax return would cause no social harm, would likely remain undetected and would be to their advantage, many people still complete the return correctly because they feel so obliged and without asking themselves how wise or fair the government’s tax policy is.¹⁰ In the third section, I’ll begin to sketch an account of the function of such directive authority but I’m not out to settle the proper origin (consent?), scope or limits of that authority.

What does it mean to say that we are bound to obey? Obligation has two aspects: motivational and reactive. On the one hand, obligation motivates action in a distinctive way; on the other hand, we react in a distinctive way to those who breach their obligations. Neither aspect can be explored in detail but I should briefly outline the approach I favour.¹¹

As to the motivational aspect, it makes sense to fulfil an obligation – i.e. you can intentionally and sanely do what you are obliged to do simply because you are obliged to do it (e.g. keep your promise) – but obligation does more than render compliance intelligible: it requires it. Views differ on what that means. Some think that a genuine obligation must provide you with a decisive reason for action, showing that you ought to fulfil it all things considered. This would rule out conflicts of obligation, cases in which you are genuinely obliged to perform incompatible actions, conflicts which are in my view a regular occurrence.¹² In the third section I shall go further, suggesting that the proposition that you are obliged to do something need not support the proposition that you ought to do it, need not help to justify your doing it.

Obligation involves two other features. When we are obliged to do something, it is not just that it makes sense to do it; it is also the case that one should not act on certain other considerations which count against doing it, that one should not treat such considerations as a factor in one’s practical deliberations.¹³ For example, in deciding whether to comply with their officer’s order to confiscate a private car, a soldier should not be influenced by the inconvenience to the motorist, the adverse effect of confiscation on relations between the army and citizenry etc. Respect for the officer’s authority should prevent the soldier from acting on such considerations – and so from including them in their deliberations about what to do – thereby clearing the way for the soldier to comply with the command.¹⁴ The effect of this exclusion is to leave it up to the authority to weigh the excluded considerations on their subordinate’s behalf. An officer must take account of all the relevant factors before issuing their order but ‘mine not to reason why’ is, within certain limits, the attitude expected of a soldier.

I just spoke of what our soldier ‘should’ do rather than what they ‘ought’ to do and this introduces obligation’s reactive aspect. In a case of obligation, to say that I should do something is to say that blame would be apt for my not doing it.¹⁵ Feelings of blame, resentment, indignation from others and guilt or compunction from the agent are an apt response to their breach of an obligation.¹⁶ They are also an apt response to any manifest willingness to act on considerations excluded by the obligation. These reactions differ from mere criticism of the agent. For example, when you fail to do something that you have good reason to do, you might be regarded (or regard yourself) as a bit of a fool but you won’t be blamed (or blame yourself) unless you were obliged to be more sensible.

Having explained what ‘obligation’ means, we can distinguish commanding from demanding:

To command someone to do something is to communicate the intention of hereby putting them under a fresh obligation to do it, of obliging them to do it by the communication of this very intention (i.e. by declaration).

Commands, instructions and directives, such as the orders issued by military officers to their subordinates, often employ imperatives such as ‘do this’, ‘stop that’ but, as I use it, ‘command’ also covers the rulings of an adjudicator such as a judge’s verdict or an umpire’s call.¹⁷ What such speech acts have in common is that they are exercises of deontic power, imposing new obligations by declaration even when they purport to be applying pre-existing obligations.

I shall use ‘demand’ to cover any request that is somehow associated with an obligation. Though a command implicitly asks you to do something and so could be understood as a species of demand, demands as I am understanding them are a much more widespread phenomenon. Sometimes a demand simply highlights a pre-existing obligation. Perhaps my lawnmower is not working, you have the only tool that can fix it and must hand it over. Should I demand the tool, that does nothing to alter the normative situation but simply draws your attention to it; such a demand could be made by a stranger on my behalf. Alternatively, suppose several such tools are available but I decide to ask you for yours and so place you under an obligation to provide one. I could have approached someone else but I asked you and thereby refined the normative situation. Still, I did not command you to help me; a request and an obligation to do the thing requested are connected other than via the exercise of deontic power. What makes my request for a tool bind is not that I communicated an intention of hereby obliging you to provide one. Even if I did communicate that intention, this is not why my request binds. You are obliged to help because I asked for a tool and, according to the rules of neighbourliness, requests of this sort are binding. The source of the bindingness of the request is the rule and the situation, not some declared intention to bind you.¹⁸

This difference between commands and binding requests may seem largely presentational, after all the result is much the same: my speech act obliges you to comply. Furthermore, don’t we speak of your obeying the rules of neighbourliness when you comply? Indeed we do but that is not the same as obeying your neighbour, as complying with their orders. Instructions to help are out of place between friends and neighbours, even though demands for help are not. If your neighbour tried to order you to hand over the tool (i.e. communicated the intention of hereby obliging you to do so) their presumption might well provoke defiance.¹⁹ But why should the distinction between commands and demands matter? What hangs on the character of the obligation generating mechanism here?

A tempting answer is that commands can be enforced whilst demands cannot. In our village there is no system for enforcing compliance with requests for help, however binding. A villager in dire straits may be entitled to seize another villager’s water hose when the owner is either obstinate or unavailable but neighbourly requests are not enforceable simply because the request binds. That depends on other factors – the gravity of the situation, whether the owner has been made aware of it (perhaps by being asked), whether the hose is needed elsewhere – and in extremis it may not matter who owns it; you’d be entitled to seize any available hose. Things seem different once an authority has been established, an authority whose function it is to resolve disputes about tools between neighbours. Such authorities generally do claim the right to enforce their rulings simply because they bind. True enough but tracing the objectionable character of commands to their enforcement just pushes our question back a step. Why should the enforcement of commands seem appropriate in a way that the enforcement of mere demands does not? We’ll address this issue in the final section. In the meantime, let’s first clarify the relationship between the institution of a ruler and the recognition of a social rule. I shall argue that the power of command requires not merely the recognition of a social role defined by rules (like ‘friend’ and ‘neighbour’) but the creation of an office.

II. COMMANDS, OFFICES AND CULPABILITY

Though binding social rules need establish no directive authority, the power of command presupposes recognition of a rule that institutes an office.²⁰ This seems true enough of the social authorities who come immediately to mind – police, military officers, employers, teachers, judges etc. – but can binding commands sometimes be issued by one who occupies no office?

An injured person is lying in the street and a small crowd has gathered; though anxious to help, they mill around aimlessly because the phone network is down and they are stumped without it. Soon after I join the crowd another stranger shows up and begins telling people what to do: some of us are to tend to the injured person whilst others are to disperse and return once they discover where the nearest hospital is. The stranger claims no official position, nor do they manifest any special expertise; they are simply more confident and charismatic than the rest of us and so able to take charge of the situation. Perhaps dispersal is not the best strategy for discovering where the hospital is. Still, reasoning that anything is better than chaos and anticipating that others will think the same, I conclude that we are all obliged to implement the stranger’s plan.

Several authors have argued that such speakers exercise a form of social authority without occupying any office.²¹ I agree for our stranger means to get us to act in a certain way and their utterances oblige us to comply. Furthermore, many of us may feel so bound even though we suspect that it would have been better to do something else and this is so precisely because we all know that enough of us will comply anyway. Given this, doesn’t our stranger rule over us, however briefly, in virtue of their possession of such authority? I think not. Directive authority is exercised by issuing commands and not demands. Our stranger is not issuing commands, at least not as we defined them in the last section.

When the charismatic stranger ‘tells people what to do’ they are best understood as asking people to do something, a request with which their audience will feel obliged to comply. Our interpretation of the stranger’s words here should not turn on whether they happen to use the imperatival form. Demands for help can be as peremptory as you please. The real issue is how the speaker obliges us to comply. We are losing valuable time, someone needs to ‘take charge’ and so much is clear to us all; what makes the charismatic stranger’s words bind is this common knowledge. The stranger need not intend to impose an obligation on anyone. If they do intend to bind us, their communication of this intention does no normative work. What does the work is the need for action coordinated around their speech. Regarded as commands their utterances are normatively inert; the obligation-generating mechanism here is that of a binding request.²²

Once the authority is in uniform, a different obligation-generating mechanism comes into play, one which does not depend on their personal charisma.²³ Take a police officer directing traffic. They occupy an office defined by a battery of social rules – one of which requires drivers to do as the office holder tells us once they invoke the authority of their office – and they use the powers of their office by communicating the intention to oblige us to act in their capacity as a police officer.²⁴ To assess their claims, we must ask: are they really a police officer? Is this a situation which falls within their jurisdiction? Are they requiring us to do something illegal? And in so doing we treat the police officer’s words as utterances communicating the intention to bind us qua commands, regardless of whether there is any further reason to act as they suggest. Of course we usually expect them to be an effective coordinator of the traffic but that fact is not the source of our obligation to obey their instructions, rather it is the other way around: we expect them to be an effective coordinator qua police officer because their orders bind us. Confronted by a charismatic stranger and a police officer with little presence each trying to direct the traffic, we might find ourselves with conflicting obligations but the horns of this deontic dilemma would arise by different mechanisms.

Our example indicates that Oakeshott is right: only the police officer can bind us by declaration because only they can appeal to a social rule identifying them as the occupant of an office with a power they are currently invoking. ²⁵ So why does command require an office i.e. an artificial personality assumed by the speaker in issuing their command? Our answer will employ two ideas: (a) that a personality is a locus of responsibility and (b) that when a subordinate obeys the command of a superior, the subordinate acts on the authority of the superior and so responsibility for what they do is assumed by the superior.²⁶

An office may be held by a natural person or a group of such people equipped with some procedure for making collective decisions like a court or a legislature.²⁷ Whatever the nature of its occupant, the essential feature of an office is that we can distinguish the acts which its occupant performs qua officer from their other actions. Being addressed by someone in their capacity as a police officer is quite different from being addressed by that same person in the guise of a fellow citizen. My interlocutor chooses whether to assume the role of police officer and their choice will determine the significance of their words. ‘Friend’ and ‘neighbour’ are not in that sense offices. Like offices they are social roles defined by a battery of rights and obligations but whether my request obliges you to help does not depend on whether I choose to address you as a friend (or as a neighbour).²⁸ ‘I’m asking you this as your neighbour’ calls attention to your request’s significance, it does not imbue it with that significance.

Why should issuing a command (as opposed to a mere demand) require that one adopt some conventional persona? Social rules impose obligations, differentiating right from wrong. I maintain that social rules can also determine who is responsible for a given action. When it sets up a directive authority with the power of command, the rule does two things: firstly it institutes a power both to create an obligation by declaration and to assume responsibility for its fulfilment and secondly it creates an artificial personality – an office – whose occupant serves both as the possessor of the power and as the locus of responsibility for obedience once the office is invoked.

What does it mean to say that a police officer ‘assumes responsibility’ for my compliance with their orders? For one thing I can pass the buck, referring complaints about what I have done to the officer. The form of responsibility here is culpability rather than liability. The agent responsible for a given act is the person who is blameworthy (or guilt-worthy) for that act should the act turn out to be wrong rather than the person who must suffer some penalty or pay compensation (though that may also be true).²⁹ Note responsibility so understood is not merely shared between the parties, rather the authority assumes the entire responsibility for the action they command.³⁰ Thus, when a soldier is ordered to confiscate a car by their officer and obeys, they act on the officer’s authority and can refer on any complaints about the wisdom or propriety of the confiscation.

Our soldier may retain responsibility for various aspects of their agency here. For one thing, they are responsible for the implementation of the order in so far as that is left to their discretion. If our soldier can choose which car to confiscate, they are to blame if the wrong choice is made. Furthermore, they may be praised for obeying an order that is especially difficult to carry out. Someone can fail to carry out an order without being disobedient, without deciding not to obey e.g. where they encounter serious obstacles or lack the skill required or where their will is weakened by fear and other temptations. When lack of compliance would be excused by such factors, compliance might be praised. The soldier gains no credit simply for deciding to do the thing they are ordered to do – credit for that decision goes to the officer – but they might gain some credit for following through in the face of difficulties.

There is no assumption of culpability for compliance with a binding social rule. Suppose I lend a valuable tool to my neighbour and my brother asks why I risked damage to something that we might need. Here familial and neighbourly obligations conflict and though it may be that I ought, all things considered, to loan the tool as requested, guilt is apt whatever I do. My neighbour may share responsibility with me for whatever happens but I can’t evade responsibility by passing the buck to them. Now suppose a manorial court is established to adjudicate issues of this sort and they instruct me to loan the tool to my neighbour. Like the rule of neighbourliness, the court’s instruction obliges me to comply but something further is true: perhaps to my relief, the matter has now been taken out of my hands. Not that the previous conflict of obligations is thereby resolved, rather it is no longer my problem.³¹ If my brother has a valid complaint about the risk involved in the loan of the tool, that complaint can now be referred to the court. Having determined that the court is acting within their jurisdiction I simply pass the buck on the matter of where the tool should go. Note the point is not that the court’s verdict gives me an extra reason to hand over the tool which settles the matter in favour of my neighbour – I might have already decided that, all things considered, my neighbour ought to have the tool – rather being ordered by the court to hand it over means that I can now do so, perhaps regretfully but without compunction.³²

In the next section, I’ll argue that the assumption of culpability involved in obedience to a binding order reflects a deeper transfer of responsibility, namely an assumption of imputability.

III. COMMANDS, DISAGREEMENT AND IMPUTABILITY

A community governed solely by customary rules is a credible state of nature and yet, as a way of regulating human social life, customs have some obvious inconveniences memorably noted by Hart. There is (a) the uncertainty as to the identity and precise content of the rules and (b) the static character of the rules i.e. their immunity to deliberate change.³³ Authoritative rules of recognition (e.g. the promulgation of a code) reduce uncertainty whilst powers of legislation enable the deliberate creation, amendment and repeal of social rules. Both codification and legislation involve the creation of obligations by declaration and so are exercises of deontic power but for the moment we’ll focus on what is needed to deal with a third inconvenience identified by Hart, what Hart calls the inefficacy of rules. This is addressed by recognising a power of adjudication exercised by norm-applying officials who issue rulings i.e. authoritative determinations as to when the rules, whether customary or created, have been violated.³⁴

Once the authority of a set of rules has been recognised there remain disputes about questions of both fact and interpretation, all of which becomes evident when we seek to apply the code to individual cases.³⁵ Judges and umpires purport to settle the matter correctly whilst also settling the matter authoritatively i.e. by using a deontic power whose efficacy does not depend on their being right. To end the controversy, they speak qua judge or qua umpire, deliberately invoking the authority of their office. Could one deal with such differences with something less than this? Suppose there are elders well acquainted with the customs of our village and with what goes on in it, elders we all are obliged to consult once disputes about neighbourly obligations arise. Where the parties to the dispute are ignorant or at least uncertain of how the rules apply to them, they can frequently resolve matters by accepting the mediation of the elders and trusting their understanding of the normative situation.³⁶ What though of cases, common enough, in which each party is convinced that they are in the right? To achieve a resolution here the villagers must regard themselves as bound by the views of the elders even if they know them to be wrong. Once this happens the elders are no longer trustworthy people offering expert advice but judges issuing verdicts.

Courts make rulings on points of fact and norm and once they pronounce you guilty, all of us must accept the verdict. Though, unlike legislators, judges represent themselves as interpreting or applying rules which they do not create, they issue their verdicts by communicating the intention of hereby binding the relevant parties to comply. I am obliged to stay away from a certain tree when the manorial court’s interpretation of custom and practice leads them to declare that it is owned by my neighbour and not by me. If I object that their verdict is misguided, they can reply that I must follow their understanding of the rule and not mine.³⁷ So is the batter bound to leave the field once the umpire declares them to be out and regardless of whether the umpire’s eyesight is at fault.

Our villager should not regard the verdict of the manorial court about either the tree or the tool as providing them with an extra reason to comply, a reason which makes this the right thing to do. The effect of a binding order is rather to relieve them of the task of justifying compliance: it is no longer their decision. The loan of the tool is, as I shall say, no longer imputable to them and so even where loaning the tool was indeed the wrong thing to do, we can’t infer that they act wrongly in handing it over. The court may have misunderstood or misapplied the norms of neighbourliness but since the loan is not imputable to the villager, they do not act wrongly in complying.³⁸ Imputability is assumed by the manorial court and this explains the assumption of culpability noted earlier.³⁹ That is why (unlike in the case where they have only an excuse) our villager is relieved of the burden of guilt. The appropriate object of blame shifts along with the proper venue of decision-making.⁴⁰ It is also why our solider is not to blame for confiscating the car under orders.⁴¹ Action in obedience to a command is imputable to the authority and it is no longer up to the subordinate to determine whether the commanded act is the right thing to do.

Our soldier does have a story to tell themselves (and others) about why they do what they do – after all obedience is a form of intentional agency, not an automatic reflex – a story which takes the following form: it is someone else’s decision and they have so decided. Having determined that they are bound to obey a given order, our soldier need no longer consider whether the car ought to be confiscated. An objector may reply that it must remain up to the soldier to justify obeying the order whilst it is up to the officer to justify issuing the order – there is no assumption of imputability here – and the officer assumes culpability only because the soldier is better placed to avoid blame by justifying their obedience to a misguided order than the officer is to avoid blame by justifying their issuing it. This response elides the distinction between thinking that you are blameless because you did the right thing and thinking that you are blameless because though you may have done the wrong thing, doing it is not imputable to you. To use the terminology introduced in the first section, the obedient soldier is not deciding that confiscating the car is what they ought to be doing; rather they confiscate because this should be someone else’s call (hence they are not culpable for obeying).

A partial analogy between the way in which commands shape practical deliberation and how future directed intentions do so may clarify what I have in mind.⁴² The point of forming a future directed intention is to block reconsideration of whether one ought to do what one has just decided to do. The ‘blockage’ here is normative. It changes the normative situation by making room for a case in which (i) one ought (as one might say) not to be reconsidering a decision one has previously made and yet (ii) if one were to reconsider that decision, one would realise that one ought to be doing something other than what one earlier decided to do. In such a situation we seem forced to say both that one ought and that one ought not to do what one presently intends to do, that this would be both the right and the wrong thing to do. I suggest we resolve this linguistic muddle as follows. First say that one ‘ought’ to do the thing that one would be justified in doing if one deliberated correctly about what it is one ought to do. Second, say that one ‘should’ do the thing the doing of which would be immune to criticism. In the situation just described one would be immune to criticism for sticking to one’s decision whilst one would be subject to criticism for abandoning it by reopening the question of what one ought to do. Hence one should do the thing one ought not to do.

Having decided last month where to go on holiday next month by assessing the merits of various destinations and having settled on France, if I feel tempted to do otherwise, often I should simply remind myself of my earlier decision to visit France. This is so even though the brute fact that I earlier decided on France provides no justification, no recommendation at all for a trip to France. I would be criticised as incontinent or fickle were I persuaded to reconsider the plan by some attractive advertising of other destinations, whether France was (or remains) the best destination or not. Perhaps the fact that I settled on France recommends France once combined with other information about the reliability of such intentions etc. but the whole point of making up your mind in advance is to block later consideration of the quality of your earlier decisions. Very often you should stick with what you decided to do without asking either whether the decision is likely to have been a good one or whether you ought to reconsider; you should ask only whether you have remembered the decision correctly, whether you really did decide to go to France. In such cases you might (or might not) be criticised for the earlier decision – depending on whether that decision was sensible when taken – but not for sticking to it once it is taken. True when war breaks out in France you should then re-open the issue of whether you ought to go – like obligations, intentions come with a limited zone of exclusion – but absent such extremities you should simply defer to your earlier self.⁴³

The propensity to act on our intentions without assessing them is not merely useful but essential to intra-personal coordination over time given the limitations on cognitive resources like our memory and attention (restricting our capacity to re-open deliberation effectively) and the build-up of various pressures and temptations as the time for action approaches.⁴⁴ Asking what one ought to do – or even whether one ought to reconsider – in the heat of the moment is usually a bad idea, even when the result of such reconsideration ought to be a change of mind. A command affects the normative situation in a similar way. Return to our soldier ordered to confiscate a car. It might be true both (i) that that the soldier should not seriously consider whether they ought to confiscate the car and (ii) that if they did consider this question, they would realise that all things considered, they ought not to confiscate the car.⁴⁵ Very often you should obey a binding command (i.e. would be blameworthy for disobedience) regardless of whether the thing commanded is what you ought to be doing.

The analogy with intention casts light on command but so does the disanalogy between them. Once I realise (perhaps without having to reconsider the matter at all) that fulfilling my intention is not what I ought to be doing, I cannot reasonably fulfil it and would be criticisable if I did. By contrast, my knowledge that obeying this command is not what I ought to be doing need not release me from my obligation to obey it. The deference involved in fulfilling your own intention entails no transfer of the decision to another person and so must be constrained by what you presently know. The deference involved in obedience to a binding command entails precisely such a transfer. What you owe to a legitimate ruler cannot be tightly constrained by your own views about what you ought to be doing if directive authority is to enable you to live alongside those who ‘know’ something quite different.⁴⁶

Is such a radical form of deference really required to deal with the disagreements that trouble the state of nature? Suppose a binding command instead altered what an agent is justified in doing by adding an extra consideration to the mix whilst leaving it to them to take the decision. That might not enable the parties to converge on a common view of what ought to be done all things concerned. It might even make things worse by turning easy decisions into hard calls. On my understanding of how adjudication settles disputes, there need be no consensus on what ought to be done all things considered; we need only agree on whose call it is. Of course, it is perfectly possible to disagree about the latter issue also, to dispute the identity of the authority, their powers and jurisdiction, the content of their instructions, just as it is possible for me to doubt that I earlier settled on a holiday in France. But, at least when the system of adjudication is well designed, these matters will generally be easier for people acting in good faith to settle than the question as to which of two conflicting obligations ought to be discharged or of what ought to happen all things considered.⁴⁷

The account of directive authority offered in this section raises some familiar worries. First, does it serve to vindicate the ‘just obeying orders’ defence? Does it absolve soldiers from responsibility for committing atrocities? That depends on whether the relevant commands can bind them and that in turn depends on our theory of legitimacy. I have been assuming that some seriously unjust commands do bind. Having one’s car confiscated is no small matter and our soldier may think it most unfair that this person is to be deprived of their car, yet military authority would be worth very little if soldiers were allowed to disobey orders whenever they regarded them (perhaps rightly) as seriously wrong. In this paper, I make no attempt to formulate a theory of legitimacy, one that would enable us to draw the line between this confiscation and some atrocity. Rather I’m concerned with what we should say about responsibility for obedience where wrongful orders do bind.⁴⁸

Any vindication of directive authority must also deal with cases in which obedience to a law or an order seems pedantic rather than atrocious. The hackneyed example is respecting the traffic laws when there is no traffic visible for miles. I agree that there are situations in which it is not the case that we should obey because it is clear to all concerned that there is no rationale for obedience, as there are situations in which it would be silly to stick to my holiday plans. On the other hand, since people are often mistaken about whether compliance is silly, directive authority would be worth little if people felt free to disobey whenever they couldn’t see any grounds for compliance. Our conscientious taxpayer pays up even where obedience has no obvious rationale and surely this is often what they should be doing.

Might the informal social authority exemplified by our charismatic stranger suffice to ensure social coordination without any recourse to directive authority? Maybe, at least among small groups of people of good will but once we leave this face-to-face society behind, a reliable consensus based on common knowledge of relevant considerations will likely elude us. The justification for a system of taxation is intricate and intangible. To cite just one relevant consideration, it will be hard for people to attain common knowledge of a body of information which will settle whether the tax we are being asked to pay is appropriate and, if not, how seriously misguided it is.⁴⁹ And even if we could all know that we all had that information, it would be easy for each party’s self-interest to bias their assessment of it in their own favour. To ensure that tax is paid where practical deliberation based on common knowledge can’t be relied on, we need the qualified form of blind obedience I have been describing.⁵⁰

Organs of adjudication settle disagreements about the application of certain social norms by letting a third party decide. This is often the best way to proceed and yet some social norms resist such institutionalisation. For example, the value of being subject to the norms of friendship would likely be undermined by attempts to establish courts of friendship to resolve disputes over what constitutes disloyalty between friends. Whether institutionalisation is appropriate will depend on the interests served by the norms in question, interests which explain their bindingness. More particularly, it depends on whether the value of compliance with the relevant norm will be lost if you comply despite believing (and perhaps knowing) that the relevant norm does not really apply. If I stay off a stranger’s land to obey a court’s injunction, the owner has got most of what they want from me though I still maintain that I am not trespassing. By contrast, help to a friend provided on the orders of a court of friendship would no longer be an act done out of loyalty to a friend. Norms of neighbourliness hover between friendship and property in this regard; something is lost once we resort to the manorial court to settle who should have the tool but not everything. And, as we’ll see in the final section, the case in favour of institutionalising a given system of norms by recognising organs of adjudication must be strong enough to compensate for any loss of liberty involved.

IV. COMMANDS, COERCION AND LIBERTY

Our discussion has left two questions from earlier sections hanging:

1. Why does personal obedience compromise our liberty in a way that compliance with a binding social rule does not?

2. Why is there any entitlement to enforce commands?

Both questions are answered once we conceive of liberty as a social status, namely the status of being deemed responsible for your own actions.⁵¹

In saying ‘your own actions’ I’m presupposing that there is a sense in which actions are allocated to people prior to any conventional apparatus of imputation. For example, I have both direct knowledge of and direct control over the movements of my body; that body’s movements are, in this sense, my actions. Furthermore, when there are no physical obstacles to my moving my body as I intend, I have what might be called ‘natural liberty’, a form of liberty also possessed by non-rational agents like spiders.⁵² Finally, if I’m under no obligation of obedience then I also possess what Hobbes called ‘the liberty of subjects’⁵³ and my actions are imputable to myself.

Now my actions are not imputable to myself when I perform them in obedience to a binding order. Where this is so, though I may retain my natural liberty, I lack something which usually matters to me. We feel constrained by binding requests and may well resent them but there is a special presumption in ordering someone to do something and a special affront when that presumption is not well grounded. This is most obviously so with momentous decisions like whom to marry or what career to pursue, especially when the agent has a definite opinion on the matter, but it is often true even when the issue is relatively trivial or the decision a hard one to make; people object to being someone else’s charge like a child. Adult responsibility can be a burden and occasionally we are content to lose it – I feel relieved when I am ordered to hand over the tool to my neighbour – but I also lose something even if avoiding a row with my brother is the greater gain. That is because command undermines our liberty, as other interpersonal demands do not.

Describing how we are constrained by the obligations of friendship, Scheffler says that

If I have a valuable friendship with someone, then I may sometimes find myself with reasons to provide assistance to my friend even if I regard what he is doing as silly or ill-judged or unwise. As a general matter, to participate in valuable forms of social interaction is to incur new limits to one’s authority to determine the content of one’s reasons for action; some of that authority is now shared with one’s intimates and associates.⁵⁴

Here it is not just that I have reason to act in my friend’s interest; I am obliged to defer to my friend’s judgement as to what would be in their interests, at least within certain limits.⁵⁵ Scheffler’s observation is true and important but the form of authority he describes is not directive authority and can’t account for obligations of obedience. Contemporary friendship excludes the type of hierarchy that command presupposes, a hierarchy which involves not just binding demands but the imposition of an obligation by declaration together with an assumption of responsibility for compliance. Confronted by my friend’s demand for help, I may decide that I should facilitate their foolishness. Still, I am not relieved of the responsibility for complying with their demand and my liberty is not curbed.

Moving onto question (2), for many liberals, what makes commands, laws and judicial orders a threat to our liberty is their association with enforcement.⁵⁶ I doubt that binding commands are always enforceable. What is true is that directive authorities often claim the right to use force to get us to do something because they claim authority over the matter, a fact more puzzling than it is usually taken to be. After all, command and coercion are completely different ways of persuading someone to do something. Command is speech that communicates the intention of obliging us to comply; coercion compels us to comply by main force or the threat of a sanction. It is often said that both command and coercion undermine our liberty but what can that mean when they operate on us so differently? Any account of the oppressive character of command should explain all of this and not take the link between command, coercion and the deprivation of liberty for granted.⁵⁷

In the last section I suggested that the commands of an authority bind because they are needed to deal with disagreement. Such an authority will be an effective social coordinator only if most people conform to its edicts and it is likely that a significant number of people will remain unmoved for one reason or another. These rebels need some other incentive to conform and the rest of us need an assurance that the rebels have this incentive before we feel obliged to conform ourselves. For Hobbes and many others, this supplementary reason is supplied by giving the authority the power to enforce its commands.⁵⁸ Perhaps so but bribery and shaming also provide us with (perhaps insufficient) incentives to comply with the law without threatening our liberty in the special way that coercive enforcement does. We need to explain firstly why coercion poses a special threat to our liberty and secondly how a rationale which legitimates directive authority might deflect such liberty-based objections to coercion.

I’ve said that liberty is the status of being responsible for what you do. If what is distinctively problematic about coercion is that it threatens our liberty, that must be because it deprives us of responsibility for compliance. This is indeed the case at least when enforcement takes the form of brute force. Once my tool is confiscated, the loan of the tool is no longer imputable to me. Given this, it becomes clear how possession of directive authority might help to explain why one is entitled to enforce one’s commands. To possess such authority is to be entitled to get someone to act by assuming responsibility for their compliance with your orders (i.e. by depriving them of their liberty) and that is exactly what you do in bypassing their agency with brute force. We can’t conclude from this that you ought to use force where there may be other objections to be overcome – coercion is often harmful to its object and terrifying to others – but these observations are enough to establish that your commands are pro tanto enforceable. By contrast, I take it that one is never entitled to enforce a binding request simply because the request is binding. The considerations which render a particular request binding (its great urgency etc.) might also establish that one is entitled to seize the thing one has asked for should it be refused but there is no general expectation that this will be so. Where the bindingness of my request turns on say the depth of our friendship, it is not pro tanto enforceable.

Cases in which enforcement takes the form of coercive threats are more tricky. The manorial court may get you to hand over the tool by threatening a fine or imprisonment. Many writers regard such threats as influencing what you do by providing you with an extra reason for compliance. If so, such threats do not compromise your liberty (as I am understanding it) since you still get to decide what to do by making a judgement about what you ought to do, no less than when you are offered a bribe. I suspect that this justificational model of coercion can’t explain coercion’s problematic character and would argue for a model on which coercion is meant to excuse compliance with the threat. Thus, someone who gives in to coercion can deflect blame onto the coercer even if they are unable to justify their compliance.⁵⁹ Incentive threats work by compromising our liberty conceived of as our responsibility for whether we act on our own judgement about what we ought to do; like commands they require our obedience.

To sum up, I have indicated how one might explain the connection between command and enforcement in its various forms without downplaying their differences. Each undermine our liberty conceived of as our responsibility for whether we act on our own judgement about what we ought to do. Brute force altogether deprives us of agency, command renders our acts no longer imputable to ourselves, coercive threats mean that it is not up to us to justify our compliance, at least in the eyes of others. Each involves the loss of a status we have reason to value but, by the same token, a story which explains why someone can deprive us of our liberty by command will help to explain why they can do so by coercion also.

V. CONCLUSION

Our discussion has been largely diagnostic. I floated a familiar rationale for directive authority, namely the need to deal with disagreement. But that does not tell us when it should be recognised, for command in all its forms threatens our liberty and thereby creates a problem of legitimacy. Some liberals react by insisting that directive authority requires the consent of each of the individuals living under it, others that its scope must be restricted in various ways to protect their liberty and yet others that it must be exercised in the name of those subject to it, furthering their interests and/or under their democratic control. Some anarchists maintain that because directive authority, however constrained, compromises our liberty, it can never be legitimate.⁶⁰ One might seek to defuse this problem of legitimacy by replacing the apparatus of command and obedience with something more like the binding requests of neighbours or the urgent suggestions of a charismatic stranger, substituting the demands of our fellows for the edicts of our superiors.⁶¹ Directive authority as we have it can’t be reconstrued in these terms.

ACKNOWLEDGEMENTS

I am grateful for the comments of Ezequiel Monti, Daniel Viehoff, Kieran Setiya, David Enoch, Ralph Bader, Miranda Fricker, Crescente Molina, Samuel Scheffler, Liam Murphy and to audiences at Kings College London, Oxford University, NYU Law, NYU Abu Dhabi, the University of Fribourg and MIT. Thanks also to three anonymous referees for this journal.

COMPETING INTERESTS

The author has no competing interests to declare.