Skip to main content Accessibility help
Hostname: page-component-55597f9d44-ms7nj Total loading time: 0.558 Render date: 2022-08-13T04:10:59.645Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "useRatesEcommerce": false, "useNewApi": true } hasContentIssue true


Published online by Cambridge University Press:  07 October 2021

Bas van der Vossen*
Chapman University, Orange, California, United States
Rights & Permissions[Opens in a new window]


John Rawls wrote that people can voluntarily acquire political obligations to institutions only on the condition that those institutions are at least reasonably just. When an institution is seriously unjust, by contrast, attempts to create political obligation are “void ab initio.” However, Rawls's own explanation for this thought was deeply problematic, as are the standard alternatives. In this paper, I offer an argument for why Rawls's intuition was right and trace its implications for theories of authority and political obligation. These, I claim, are more radical than is often thought.

Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (, which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright © The Author(s), 2021. Published by Cambridge University Press

In A Theory of Justice, John Rawls writes: “it is not possible to have an obligation to autocratic and arbitrary forms of government.” In his view, it's not possible to bind oneself to obey such institutions. Unless institutions are reasonably just, attempts to incur political obligations are “void ab initio.” Try as we might, we cannot legitimize unjust governments through consent.

If Rawls is right, it would follow that, at least insofar as the consent theory of legitimate authority is concerned, justice limits legitimacy. The demands of justice circumscribe what a theory like consent theory can establish. If indeed “it is not possible to be bound to unjust institutions,”Footnote 1 there are distinct limits to the roles legitimate institutions can play.

Rawls's theory did not ground legitimate authority in the consent of subjects.Footnote 2 Nevertheless, his observation is important. Not only is it important to know what obligations we may be under when we face unjust governments, it's also important to know what kinds of obligations we can create in such contexts. Consent remains a straightforward and plausible way (perhaps the most straightforward and plausible way) in which people can obligate themselves. And the consent theory of political obligation gives clear expression to the idea that the legitimate authority of a state or government is conferred upon it by its subjects. Through their consent, subjects give or transfer to institutions the right to rule. This is another reason to investigate Rawls's thought: what's true of consent theory may well be true of other theories that see authority as given to institutions by subjects.


Consider three possible explanations for Rawls's thought that consent cannot legitimize unjust institutions.Footnote 3 The first, suggested by Rawls himself, holds that it's impossible to bind oneself to unjust institutions because these are excluded from the set of things toward which one can voluntarily incur obligations. Call this the Exclusion Thesis.

Exclusion Thesis: It's impossible to voluntarily incur obligations to unjust institutions.

Others demur. For example, A. John Simmons suggests that the Exclusion Thesis is false. People can incur obligations of obedience to unjust institutions, he argues, just like they can incur obligations to unjust people; however, these obligations must be balanced against other moral considerations. Rawls's thought is then explained by the obligation to obey being outweighed on balance by the institution's injustice. Call this the Balancing Thesis.

Balancing Thesis: It's possible to incur obligations to unjust institutions, but these must be balanced against other moral considerations.

A third explanation holds that we can successfully consent to obey unjust institutions (contrary to the Exclusion Thesis), but that there are moral limits to the obligations consent can generate. Because of these limits, whatever one becomes obligated to obey does not include unjust laws (contrary to the Balancing Thesis). Call this the Limitation Thesis.

Limitation Thesis: Incurring obligations to unjust institutions is possible, but there are moral limits to what such obligations can require.

In what follows, I defend (a refined version of) the Limitation Thesis: consent-based obligations have a built-in justice limitation.

The Limitation Thesis has some startling implications. Its truth may alter how we think about the duty to obey the law. In particular, if the Limitation Thesis is true, it follows that, insofar as a state or government is authorized through the consent of the people, its authority is circumscribed morally. That is, just or unjust, no institution is entitled to obedience to all its laws, including unjust ones, because it governs with our consent.

The logic of the Limitation Thesis does not just apply to consent theory, but to all theories of legitimacy that understand a state's right to rule as something that's transferred to it by its subjects. For such accounts, the Limitation Thesis holds.


Let's take the three theses in turn. Consider first the Exclusion Thesis. Rawls's argument for why consent to unjust institutions fails goes as follows:

It is generally agreed that extorted promises are void ab initio. But similarly, unjust social arrangements are themselves a kind of extortion, even violence, and consent to them does not bind.Footnote 4

Rawls was surely right that unjust institutions constitute a kind of violence. While all political institutions use force in some way or other, unjust institutions rely on force much more frequently, much more severely, and in ways that negate people's basic moral standing. The same is not true of just institutions. Just institutions can reasonably hope for more voluntary support.

Nevertheless, as an argument for the Exclusion Thesis, Rawls's argument fails. As Simmons points out, extorted promises fail not because of the nature of the agent to whom they are made, but because only voluntary promises have morally binding power. And extorted promises are not made voluntarily.

Even if an institution's injustice is akin to violence, consent to it need not be involuntary. It's perfectly possible for people to freely choose, out of their own free will, to consent to such institutions. They might be part of a group that's not subject to the institution's violence, say. Such consent seems binding irrespective of the moral quality of the recipient. The same seems true with promises made to unjust people. If I promise a villain to meet him for lunch, I become obligated to show up.Footnote 5

Of course, the truth of the Exclusion Thesis doesn't depend on this defense. And other defenses have been attempted. Chaim Gans agrees that “it is the institution's unjustness, rather than the involuntary nature of the consent to it that renders such obligations void ab initio.”Footnote 6 But the reason, Gans argues, is that political obligations to such institutions would grant them far too unwieldy a control over our actions. This distinguishes unjust institutions from unjust persons, rendering Simmons's critique misguided.Footnote 7 Writes Gans:

The duty to obey unjust institutions is . . . a duty to perform acts because the institution to which the duty is owed demands their performance. In acknowledging such a duty to a given institution, we grant it the moral power to assign us particular duties, without retaining any control of their creation or contents. Thus, in failing to exclude unjust institutions from the scope of the general duty to obey the law, we will be granting unjust institutions moral power which they may abuse directly. Such a move seems far more drastic than acknowledging the duty to keep promises to villains.Footnote 8

Gans's thought is that, since political obligations authorize institutions to determine what people may do, we can only entrust institutions with that power if it cannot be used in morally dangerous ways. Unjust institutions cannot be trusted in this way and are therefore excluded from the set of potential recipients of political authority.

The key assumption for this argument is that if a person P incurs a political obligation to some agent A, then A becomes able to tell P what to do in a wide range of matters.Footnote 9 For Gans's argument to work, this range has to include committing or supporting injustice. If the duty to obey were limited to exclude such actions, the stated problem with political obligations to unjust institutions could not arise.

Gans's argument, in other words, presupposes that the Limitation Thesis is false. According to Gans, such views “involve a confusion as to the very concept of the duty to obey the law.”Footnote 10 The thought is that justice-limits are conceptually incompatible with political obligations. If institutions have authority, it's up to them what subjects may or may not do.

The idea that, as a conceptual matter, political obligations are unbounded in this way has a long pedigree. Call it, following H.L.A. Hart, the idea that political obligations are content-independent.Footnote 11 The motivation behind the desideratum is that it explains why subjects to a legitimate authority might be obligated to do something “because it's the law” (not because of the law's content).

As a conceptual point, Gans's point is impotent. Whether or not people have content-independent political obligations isn't a conceptual matter. It depends on the nature of the obligations they actually have. Perhaps, if their obligations to obey the law are limited in ways that disqualify them as content-independent, and all political obligations are content-independent, then the obligations they actually have are not properly “political obligations.” But they would still be obligations to obey the law. By what name we refer to this doesn't matter much.Footnote 12

Setting aside the conceptual point, a more significant problem is that Gans is mistaken to assert that content-independence implies the denial of the Limitation Thesis. The binding force of obligations that are clearly content-independent can (nevertheless) be conditional on a requirement's moral quality. What's required for obligations to count as content-independent is that their binding force is not due to their moral quality. But that leaves open that moral quality might disqualify requirements as morally binding.Footnote 13

Plainly, promises and acts of consent are content-limited in everyday life.Footnote 14 If I promise you to help you fix your car, for example, there are generally understood limits to what I can become required to do. If you ask me to hand you a wrench, I become obligated to do so. When you tell me to buy an expensive new part, I can refuse without breaking my promise. Such limitations are often implicit, but clearly my consent did not include that kind of help. Nevertheless, my obligations of consent are content-independent. When you ask for the wrench, I become obligated to hand it over “because I promised.”Footnote 15

There is, then, no problem with consent-based obligations being content-limited in general. And that includes the kind of obligations that give others the power to tell us what to do. If that's true, it's unclear why moral limitations would be a problem for consent-based political obligations in particular.

We can distinguish between two versions of content-independence, then. On the first, when a state has legitimate authority, it can make actions obligatory because they're legally required, even if they were not already required because of their content, and the scope of this power is limited by justice. On the second, when a state has legitimate authority, it can make actions obligatory because they're legally required, even if they were not already required because of their content, and this power is not limited by justice. Affirming the former version, the Limitation Thesis is consistent with content-independence.

Gans's argument for rejecting the Limitation Thesis is mistaken, then. And the Exclusion Thesis itself remains problematic. It either implies, as Simmons argued, that we cannot successfully make promises to unjust persons. And that remains plainly false. Or it must insist on some kind of distinction between obligations to unjust institutions and unjust persons, so that obligations to the latter can be created while obligations to the former cannot.

Gans claims we should accept such a distinction. The argument he offers is that institutions have only instrumental moral value, whereas individuals have intrinsic moral value. And this is said to imply that obligations to institutions, in contrast to obligations to persons, are possible only if those institutions are also instrumentally valuable.Footnote 16

But this looks in the wrong place. The power of consent is primarily grounded in facts about the consenting agent. And, as such, the nature of this power will be sensitive primarily to facts about the agent, not the recipient. The key question here is whether agents ought to be able to bind themselves to such recipients (whether persons or institutions). And just as we can have reason to want the power to consent to things that are morally trivial, or even to things that are regrettable, we can also have reason to want the power to bind ourselves to recipients that lack intrinsic or instrumental value. When I blow all of my family's savings in a casino, indulging in a vice with a company that makes money from vice, I am still obligated to pay the bill.

We can successfully consent to institutions, as to persons, irrespective of their moral quality, then. The Exclusion Thesis is false.


Simmons proposes to replace the Exclusion Thesis with the Balancing Thesis:

[S]urely we ought not to support unjust institutions. But it seems more natural to allow that we can sometimes succeed in obligating ourselves both by promises to villains and by consent to “autocratic and arbitrary forms of government” (to borrow Rawls's phrase). In addition, however, we have a clear duty both to help confound villainy and to fight injustice. Thus, it will be a matter for decision in individual cases whether, e.g., the harm done by supporting an unjust institution and our duty to fight injustice outweigh any obligation we may have to respect its authority (deriving from our consent to it).Footnote 17

In Simmons's view, consent is always a reason for obedience, irrespective of the institution's moral quality. And injustice is always a reason for disobedience. Whether one ought to obey laws, all things considered, depends on the relative weight of these considerations.

The Balancing Thesis seems initially plausible. Many think that moral considerations can clash, and that the right thing to do overall depends on their relative weight. Nevertheless, in terms of capturing Rawls's thought that consent to unjust institutions is void ab initio, the thesis does not perform well. On Simmons's view, such consent is not void, but outweighed.

Of course, some may find this attractive, for example because they think obedience can be required when the injustice isn't very serious. In those cases, the obligation from consent may win the day. However, this implication is difficult to contain. For the Balancing Thesis can also imply an obligation to obey in cases of more serious or weighty injustices. In particular, it can do so when injustices that are small at the individual level aggregate to large injustices overall. Suppose a government institutes a $1 per capita tax to line the president's pockets. This is an injustice, to be sure, but it will not be very weighty for any particular subject. After all, each of the taxed persons is harmed very little. It seems, then, that each of them (assuming they consented) is obligated to pay the tax. Yet the result is an obligation to support a very serious abuse of power.

One might object here that the Balancing Thesis can avoid this implication. After all, if minor abuses of power add up to something very serious would that not outweigh the obligation to obey? But no subject is asked to commit or support a serious injustice. They are asked only to perform a very small, and not very harmful, wrong. The moral weight of this will be accordingly small as well. If the obligation to obey the law is to be effective, it cannot be outweighed by things of such small weight.

Kent Greenawalt suggests a similar view, holding that promises to do unjust things are morally binding.Footnote 18 Greenawalt offers the example of promising to tell a white lie for a friend. Such a promise, he claims, is binding, even though it's wrong to tell lies. But this cannot be right. Suppose Andy plans to rob a bank and promises to give Ben half the loot. After the robbery, Andy takes off with all the money. Supporters of the Balancing Thesis must say that Andy violated his obligation to Ben. After all, if Andy's promise went through, he would have become morally obligated to give part of the money to Ben.

But this is just false. To see this, consider that successful promises create rights in their recipients—rights correlative to the promissory obligation. Thus, if Andy's promise succeeded, Ben would have gained a right to the money. But clearly, Ben has no claim to the money whatsoever—only the money's rightful owner does. If Andy wanted to do the right thing, he need not think about whether Ben's claim or the bank's claim is weightier. Ben has no claim, there's no balancing to be done. Andy's promise to give Ben the money was indeed void.

By contrast, if Ben really did have a claim, outweighed by the imperative to return the money, there would be at least some pressure against returning the money. After all, when rights are overridden, this represents something regrettable, even if doing so is overall best. And, again, this is simply not the case. There is no sense in which it's regrettable if Andy returns the money. Ben's being promised a cut provides no pressure at all.

To see this, consider what Judith Jarvis Thomson has labeled the “Aggravation Principle.” According to this principle: “If X has a claim against Y that Y do alpha, then the worse Y makes things for X if Y fails to do alpha, the more stringent X's claim against Y that Y do alpha.”Footnote 19 While formulated in a slightly different context, the principle can be used to test whether a party has a claim. Imagine, then, that it would be extremely bad for Ben to not receive the money. At some level of badness, if the Balancing Thesis were true, the stringency of Ben's claim should become such as to override the rights of the bank (flush with cash). But this, again, clearly isn't true. We don't have to look into how Ben would be affected to know that returning the money is the right thing to do.

Greenawalt might object, of course. Suppose it's impossible for Andy to return the money to the bank. Surely in that case, even if not in the one above, Andy ought to give Ben the money? And if that's true, doesn't that mean that Andy's wrongful promise went through? But this again is mistaken. As long as it's possible for Andy to do something better with the money than give it to Ben, it seems plain that Andy ought to do that. If Andy can't return the money, surely he ought to try and give it to the bank's depositors. And barring that, Andy should probably just give it to charity.Footnote 20

Still, surely there is something wrong here? Doesn't Ben have a complaint against Andy? This strikes me as true, but this fact can be explained without the Balancing Thesis. Perhaps Andy wronged Ben by first creating and then frustrating expectations. There is definite deception going on. And perhaps there's some other kind of unfairness involved.Footnote 21 The point that matters for now, however, is that whatever the nature of this wrong might be, it cannot be the violation of an obligation to give Ben the money. The Balancing Thesis is false.


The reason Ben didn't get a right to the money is obvious, of course. The money was never Andy's to dispose of. And we can't give away what we don't have. Since Andy lacked all right to the money, he cannot transfer to Ben any claim to it. It's not that easy for robbers to undercut the claims of rightful owners.

Call this thought the Limitation Principle.

Limitation Principle: One cannot transfer to others a right one doesn't have in the first place.

The Limitation Principle applies to the consent theory of authority. According to that theory, a government's or institution's authority is the result of rights transferred to it by its subjects, via their consent. The transferred rights include the right to make certain decisions for subjects, decisions that subjects become obligated to follow as if they authored these themselves.Footnote 22 The Limitation Principle holds that when A consensually transfers some right R to B, the set of rights B can acquire is no greater than the set of rights A had in the first place.

The Limitation Principle identifies what we may call, using J.L. Austin's analysis of illocutionary acts, a “felicity condition.” Illocutionary acts like consent must satisfy these conditions in order to be successful. When illocutionary acts violate felicity conditions, they “misfire.” Such misfires can occur when illocutionary acts involve a misapplication of the correct procedure for their performance, for example. Or they can occur when the agent performing the illocutionary act is not appropriate for the procedure. If I try to consent on your behalf (without prior authorization), my act will fail. Similarly, consent that is not freely given misfires as well.Footnote 23

Misfires are distinctive because they represent cases in which an attempted illocutionary act (here: consent) does not succeed. Austin contrasts misfires with other issues, like breaches, misunderstandings, or infractions. The difference, in his view, is that misfired illocutionary acts are, like Rawls said, void. The speaker attempts to apply a mechanism in circumstances where it does not apply. The result is, quite simply, nothing.Footnote 24

Rawls's comparison of consent to unjust institutions with coercion was apt, then. When consent is coerced, it's void because the correct application of the procedure requires the absence of duress. Only when consent is offered voluntarily does the illocutionary act succeed. Coerced consent is thus void ab initio. It does not signify the right-holder's decision to transfer rights.Footnote 25

The Limitation Principle operates in the same manner. Only successful acts of consent can transfer the rights necessary to authorize institutions. But people generally lack the right to commit acts that are unjust. And since the people lack this right, they cannot transfer it to institutions. This is true no matter how qualified or unqualified their attempted consent. One cannot transfer to an institution the right to command something one lacks the right to do oneself in the first place.Footnote 26

The Limitation Principle focuses on cases of rights transfer. It's true that sometimes B can obtain a right R as a result of A's actions, even though A never had R in the first place. Judges might award damages to plaintiffs, the executor of a will might transfer possessions to beneficiaries, and so on. The Limitation Principle does not deny that such powers exist. But these are not cases where people are empowering others by transferring rights. For those cases, where B gains R as a result of A's transfer, the Limitation Principle holds.

The Limitation Principle explains why Andy cannot give Ben a right to the stolen money. Andy failed to confer such a right upon Ben because Andy lacked a right to the money in the first place. Since the consent theory of authority uses the same framework, considering the powers of legitimate government as the result of a consensual transfer by its subjects, the Limitation Principle implies the same result.Footnote 27 Since subjects lack the right to commit or support injustice, they cannot bestow upon institutions the right to require such things.

We can summarize this argument as follows:

  1. (1) An institution has legitimate authority only if subjects have successfully given it the right that they obey the law through their consent.

  2. (2) One cannot give to others a right one doesn't have in the first place.

  3. (3) Therefore, a legitimate institution has a right that subjects do what the law requires only if those subjects had a right to do those things in the first place.

  4. (4) No one has the right to commit or support injustice.

  5. (5) Therefore, a legitimate institution has a right that subjects do what the law requires only if what the law requires is not unjust.

Premise 1 of this argument expresses the consent theory of legitimate authority. Premise 2 is the Limitation Principle defended above. Premise 4 states the truism that injustices are impermissible. The argument's conclusion implies the Limitation Thesis.

Even the most capacious or unlimited acts of consent, then, cannot give a state the right that we obey the law if its requirements are unjust. Nor can any state credibly claim to be authorized to commit an injustice because it's acting in our name. Obligations to obey the law that are established in this manner come with justice-exceptions attached. In this sense, justice precedes legitimacy.Footnote 28


The truth of the Limitation Thesis can be seen in various contexts. Consider empowering agents to act on our behalf. Suppose I hire you to manage my retirement account. By signing our contract, among other things, I empower you to make investment decisions on my behalf. Being granted that power, you gain some authority over my (financial) life.

At the same time, it's plain that your authority is limited by the act through which I empowered you. Financial agents, like political authorities, typically obtain such powers only as a result of their being transferred by principals. Despite being empowered to manage my retirement account, you may not decide how my other savings are allocated, say. Nor may you allocate them in a manner that's riskier than is allowed by the investment profile I selected. Nor, obviously, can you manage somebody else's retirement savings in virtue of my consent. The first two acts would be a breach of contract. The third, a simple case of theft or fraud. Your authority is limited by the rights I transferred to you through my consent.

Of course, it's usually understood (and contractually stated) that when I empower you to arrange my finances, my act is clearly limited in these ways. And political consent may be different. But this doesn't seem a relevant difference. Suppose that, when I hire you, we write into the contract that my consent also empowers you to manage my neighbor's account. This gives you no more right to do so. Such consent misfires, and for the same reason—I can only empower you in this transaction to do things I have the right to do myself. And I lack the right to manage my neighbor's money.

Perhaps one might think the relation between me and my financial manager isn't one strictly of authority. While financial managers might act on my behalf (and make decisions in my name), they don't quite issue commands. But the Limitation Thesis remains plausible when we add this element. Suppose my local fire brigade is engaged in a range of activities. These include, admirably, combatting wildfires, but also less admirable activities, such as posting videos on social media in which they kick cats. Suppose the group is up-front about this. If you join their group, you consent to follow orders, including the cat-kicking ones you might consider wrong.

Suppose now that I know about this, and voluntarily consent to join the brigade. As we're fighting a fire, I follow their orders, saving people and their homes. Now the order comes down to kick a cat while others record it. Surely, I am in no way morally obligated to kick the cat. And this is true despite my consent. While my consent wasn't void entirely—I remain obligated to obey the life-saving and home-saving orders—it's limited by justice. As the Limitation Thesis suggests, I cannot bind myself to kick the poor cat.

Certainly, there are further important differences between giving power of attorney to financial agents or joining a group of firefighters on the one hand, and life as a citizen on the other. But the cases are the same in all morally relevant respects. The persons or institutions in question receive the right to tell one what to do, and they receive this as a result of a transfer effected by one's consent. Absent one's consent, they would lack this right, since their actions infringe upon what (absent consent or something similar) is one's rights-protected sphere of action. And the moment of consent was the moment of waiving or transferring those rights, thus empowering the authority in question. The extent of the power or authority gained as a result of consent, therefore, is determined by the extent of the rights that are transferred.


At this point, one might wonder whether this argument moves too quickly. Even if we cannot bind ourselves to do things we don't have a right to do, is it really true that people always lack the right to commit injustice? Aren't we sometimes permitted to do things that would otherwise be unjust? Perhaps we may commit a small injustice in order to avoid a very great injustice, for example.

I agree that sometimes injustices might be justifiable all things considered. But that final qualification is crucial. Things that are pro tanto unjust might be permissible all things considered. And those things, all things considered, one has the right to do. However, when things are all things considered unjust, it remains true that one lacks, all things considered, the right to do them.

This qualification applies to the Limitation Thesis, too. The thesis holds only that one cannot obligate oneself to obey in cases where, all things considered, one lacks the right to perform the required act in the first place. Suppose that some agent A has a pro tanto obligation to ~R (where ~R means not violate someone's rights). A thus lacks a pro tanto liberty-right to R. However, if all things considered A does have the liberty-right to R, the Limitation Thesis does not preclude A successfully consenting or promising to R (as long as R remains all things considered permissible). After all, since it's no longer true that, all things considered, A is obligated to ~R, A does have the requisite right to transfer.

This qualification does not make the Limitation Thesis equivalent to the Balancing Thesis. It's true that, in cases where one has a consent-based obligation to obey, this obligation must be balanced against other moral requirements. (This is the truth that gave the Balancing Thesis its plausibility.) But that does not mean one can obligate oneself to obey when one lacks, all things considered, the right to do the legally required thing. In those cases, there is no obligation to be outweighed. Our consent misfired.


While the Limitation Thesis is plausible in a variety of contexts, there are also cases in which, at least initially, an objector might think the thesis must be false. In particular, one might doubt the Limitation Thesis in contexts where the need for authority is very important. Consider, for example, the military. While the authority of a soldier's superiors may rest on the soldier's consent to obey, the Limitation Thesis might seem false. Indeed, one might think the thesis must be false in a context like the military. For the military to work, soldiers cannot be allowed to second-guess the moral quality of their superiors’ decisions.

Something similar might be said about the obligation to obey the law. Political authorities are needed precisely because people disagree about questions of justice. And authorities cannot succeed at this unless citizens are morally required to obey unjust laws. Absent that, it becomes difficult to see how states could help people cooperate and live together peacefully. The workings of a state, then, might be thought to equally require that the Limitation Thesis be false.

On this view, it's a desideratum of a theory of legitimacy that the Limitation Thesis does not apply to it. This leaves us with a choice. Either the Limitation Thesis is false after all, or consent theory is a faulty theory of legitimacy because the Limitation Thesis applies to it. Either way, the arguments here don't establish what they set out to demonstrate.

Consider first the case of military obedience. It's clearly true that no military accepts soldiers making up their own mind about the validity of each and every order. Nor, as a rule, do they accept perceived immorality as an excuse for disobedience. Still, there are generally recognized exceptions to the duty of obedience. Soldiers in the US military vow to obey all lawful orders, in accordance with the Uniform Code of Military Justice. But obedience to unlawful orders is punishable. The same is true in many other countries.Footnote 29

The standard for permissible disobedience, of course, typically refers to an order's illegality, not its immorality or injustice. But the two are related. At least in extreme cases, the duty of obedience is often recognized as limited by morality. Canada's Code of Conduct (2001) includes an exception if “the order is manifestly unlawful,” where this is clarified as follows: “A manifestly unlawful order is one which shocks the conscience of every reasonable, right-thinking person.” Germany's Military Manual (1992) exempts soldiers if an order “violates the human dignity of the third party concerned or the recipient of the order.”

The exemption to obedience in US law is called the “Medina Standard,” named after the captain who ordered the March 1968 My Lai massacre during the Vietnam war. The massacre was famously put to an end by Hugh Thompson, a then twenty-five-year-old helicopter pilot. Thompson saw the massacre, landed his helicopter between the villagers and the soldiers, and ordered his crew to use their machine gun against the American troops if they continued firing on the villagers.Footnote 30

Initial reactions to Thompson's intervention matched the thrust of the objection we're considering here. After an attempted cover-up, Thompson was widely criticized within the US military, as well as by the government and the general public. But later reactions have been markedly different. Thirty years after the massacre, Thompson and his crew received the Soldier's Medal, the United States Army's highest award for bravery not involving direct contact with the enemy. At the ceremony, Major General Michael Ackerman praised Thompson and his men for doing the right thing, having “set the standard for all soldiers to follow.”Footnote 31

Still, one might think all this shows that soldiers are not obligated to obey extremely unjust orders. And to be sure, that is all the jurisprudence supports. So the thought that, for the military to work effectively, soldiers must be required to obey (not extremely) unjust orders remains intuitive.

However, this more moderate objection poses no conflict with the Limitation Thesis. The key lies, again, in the distinction between pro tanto and all things considered obligations discussed above. The objection we are considering holds that the demands of military action create serious moral pressure for soldiers to do things that are unjust. That is, actions that are pro tanto impermissible might nevertheless be permissible all things considered, given the military context. But if that's the case, it's no longer true that soldiers lack the right, all things considered, to perform acts that are pro tanto unjust (at least in the military context). As a result, the Limitation Thesis doesn't rule out that soldiers might empower military authorities to, all things considered, bindingly command such acts.

It's worth noting that this explanation is consistent with the phrasings of the exemptions to obedience in many military documents. Typically, that exemption is limited to laws and commands that are clearly or extremely unjust or criminal. As the US Court of Military Appeals has ruled, “the justification for acts done pursuant to orders does not exist if the order was of such a nature that a man of ordinary sense and understanding would know it to be illegal.”Footnote 32

In cases like these, it's highly unlikely that soldiers have an all things considered right to perform the pro tanto unjust act. After all, when (even) the law doesn't allow a certain act to be commanded, it's hard to think it may be overall justifiable. The same is true when acts are clearly or extremely wrong. Actions that violate human dignity or shock the conscience are plainly beyond what we have, all things considered, the right to do.

By extension, the same can be said of political institutions. Even if it's true that just and stable societies are possible only if citizens obey unjust laws—an assumption some might question—the Limitation Thesis does not rule this out. As above, this objection assumes that there are times at which citizens might be all things considered obligated to do things that are pro tanto unjust. But the Limitation Thesis doesn't rule out successfully consenting to obey, all things considered, those kinds of laws. It precludes an obligation to obey only when the need to cooperate and live together does not outweigh a law's injustice, all things considered. This, again, is as it should be.Footnote 33


The Limitation Thesis is true, then. Even subjects who give their most unqualified consent will not have an obligation to obey the law if its requirements are all things considered unjust.

Consent, in other words, needs to be disaggregated. When we consent to things, we never consent absolutely. We cannot consent absolutely. We can only give our consent to those things that we have a right to do in the first place. And, of course, we only consent to the things to which we actually give our consent. Even the most unqualified consent given to the most just institution thus binds citizens only to obey laws that do not require things that are, all things considered, unjust.

Laws are unjust when they require things beyond the limits of the consent received. And, likely, the consent people actually give is limited because they only intend to legitimize whatever powers are needed for proper government. In that case, laws that go beyond those limits (such as laws aimed at private enrichment) are unjust. Or perhaps there are rights one cannot waive or transfer. In that case, certain laws would remain unjust even if subjects gave their most unqualified consent possible. And in any case one cannot successfully consent to laws that violate the rights of others.

Many things that would be unjust if done without consent can be just when done with consent. Volenti non fit iniuria. Thus, the practical upshot of the argument here depends on the rights we have, our powers to transfer those rights, and the actual consent we give. These are complicated questions, involving the nature or justice, autonomy, and more. The argument here does not answer them. All it establishes is that, whatever the truth about justice might be, the legitimate authority of institutions is limited by it.

In the end, this view strikes me as attractive. Rawls famously said that “justice is the first virtue of social institutions.”Footnote 34 The Limitation Thesis lives up to this statement. After all, both the Exclusion and Balancing Theses allow for unjust laws to carry the moral imprimatur of legitimate authority. If an institution's authority is circumscribed to rule out unjust but authoritative directives, no such thing is possible. According to the Limitation Thesis, justice does precede legitimacy.


1. The quotes are from John Rawls, A Theory of Justice (Belknapp Press rev. ed. 1999), at 96. The context here is Rawls's discussion of the principle of fairness. To Rawls, consent and promises are special cases of the principle of fairness. Id. at 303.

2. Nor did Rawls think legitimacy circumscribed by justice in the way (I argue) consent theory implies. As I understand it, the position Rawls developed (especially in his later work) holds that laws are binding if passed in accordance with a legitimate constitution, where a legitimate constitution is based on an appropriately acceptable political conception of justice. In this theory, public reason might operate as an analogue to consent, but remains importantly different. Thus, laws might be considered unjust yet legitimate or binding to Rawls as long as the institution (or constitution) is not so unjust as to become illegitimate in light of the relevant political conception of justice. See John Rawls, Political Liberalism (Columbia University Press expanded ed. 2005), at 137, 428.

3. For ease, I use the terms “legitimacy” and “legitimate authority” interchangeably. I don't deny that these terms can be used in ways that are not equivalent. Further, I assume that people subject to legitimate authorities have political obligations, and that they can confer legitimate authority upon institutions by incurring political obligations toward them.

4. Rawls, supra note 1, at 302.

5. Simmons, A. John, Tacit Consent and Political Obligations, 5 Phil. & Pub. Affs. 274, 277 (1976)Google Scholar.

6. See Chaim Gans, Philosophical Anarchism and Political Disobedience (1992), at 98 (emphasis in original).

7. Strictly speaking, then, Gans defends a more limited Exclusion Thesis according to which it's impossible to incur political obligations to unjust institutions. The difference is immaterial here.

8. Gans, supra note 6, at 98–99 (emphasis in original).

9. “In acknowledging [a duty to keep promises to villains], we are not granting the villains the power to create the promises and determine their contents. In acknowledging the duty to obey unjust institutions, we are doing just this.” Id. at 99.

10. Id. at 103.

11. See H.L.A. Hart, Legal Duty and Obligation and Commands and Authoritative Legal Reasons, in Essays on Bentham 127, 243 (1982). See also Leslie Green, The Authority of the State (1988), especially 226, 239.

12. For a rejection of content-independence as essential to political obligation, see Klosko, George, Are Political Obligations Content Independent?, 39 Pol. Theory 498, 504–507 (2011)CrossRefGoogle Scholar. The arguments below do not imply this sweeping rejection.

13. Laura Valentini helpfully distinguishes between content-independence and content-insensitivity. The former may be a property of political obligation, but only the latter implies that one cannot look at the law's content to assess whether one is obligated to obey. See Valentini, Laura, The Content-Independence of Political Obligation: What It Is and How to Test It, 24 Legal Theory 135 (2018)CrossRefGoogle Scholar.

14. See W.D. Ross, Foundations of Ethics (1939), at 98–99.

15. For a similar point, see Christopher H. Wellman, Rights Forfeiture and Punishment (2017), at 160ff.

16. See Gans, supra note 6, at 99–100. This is a puzzling argument since even very unjust institutions can presumably do instrumentally valuable things. Thus, the Exclusion Thesis (“[I]t is the general duty to obey this system that is void, not only its individual applications.” Id. at 108) itself seems instrumentally unjustified.

17. Simmons, supra note 5, at 277–278. For the same view, see Zofia Stemplowska & Adam Swift, Dethroning Democratic Legitimacy, in Oxford Studies in Political Philosophy 1, 8 (David Sobel, Peter Vallentyne & Steven Wall eds., 2018).

18. See Kent Greenawalt, Conflicts of Law and Morality (1987), at 84–85.

19. Judith Jarvis Thomson, The Realm of Rights (1990), at 154.

20. Of course, if Andy cannot do anything better, the objection loses its force.

21. David Miller sees an element of desert: “there seems nothing incoherent or bizarre in saying that the man who masterminded the bank robbery deserves a larger share of the loot than the guy who merely drove the getaway car.” See David Miller, Principles of Social Justice (1999), at 135.

22. As John Locke put it: “Men, when they enter into Society, give up the Equality, Liberty, and Executive Power they had in the State of Nature, into the hands of the Society.” Similarly: “Where-ever therefore any number of Men are so united into one Society, as to quit every one his Executive Power of the Law of Nature, and to resign it to the publick, there and there only is a Political, or Civil Society.” See John Locke, Two Treatises of Government (Peter Laslett ed., Cambridge University Press 1988) (1689), Second Treatise, sections 131 and 89, respectively.

23. J.L. Austin, How to Do Things with Words (1975), at 34.

24. Id. at 39.

25. For a similar analysis concerning consent given as a result of deception, see Tom Dougherty, Sex, Lies and Consent, 123 Ethics 717, 732–733 (2013). Dougherty claims that if A consents to B as a result of B's deceiving A, then A's consent was invalid and thus void.

26. The Limitation Principle expresses a necessary condition for successful consent, not a sufficient one. It may be that one has rights that one cannot transfer to or use to authorize others.

27. While it's not clear if he would have accepted the Limitation Thesis defended here, John Locke seems to endorse the logic of the Limitation Principle. See Locke, supra note 22, Second Treatise, sections 135 and 168.

28. It might be said that obeying unjust laws need not involve committing or supporting injustice. Perhaps if a law is created in an illegitimate manner, it might constitute an unjust law—albeit unjust in a manner different (procedural, rather than substantive) from the sense of unjust used in the text here. If such a law requires subjects to do things that are (substantively) morally permissible, the Limitation Principle is not violated. Perhaps our consent does not bind us to obey such law either. I think it doesn't. But the reasons for this would be different from what's expressed here.

29. For an overview, see International Committee of the Red Cross, Practice Relating to Rule 154. Obedience to Superior Orders, IHL Database, (last visited Aug. 27, 2021).

30. Jonathan Glover, Humanity: A Moral History of the Twentieth Century (2001), at 62–63.

31. See Associated Press, My Lai Rescuer Hugh Thompson Jr., Washington Post (Jan. 7, 2006),

32. United States v. Keenan, 39 C.M.R. 108, 110 (1969). Keenan was found guilty of murder after obeying an order to shoot an elderly Vietnamese citizen.

33. Does time make a difference? After all, the limits of consent seem, by this argument, to be set at the time consent is given. And it may be possible that something that was at the time of consent (t1) all things considered unjust becomes all things considered not unjust at the time an order is issued (t2). In that case, the plausible idea that obedience may be obligated may seem to contradict the Limitation Thesis. This objection is avoided, however, if one's consent includes certain conditional permissions. Thus, a soldier might consent at t1 to obey an order at t2 that, at t1, is all things considered unjust, on the condition that it is no longer unjust at t2. Such conditional authorizations are clearly possible, such as when we authorize financial managers to buy stock up to a certain price. It's not implausible that military or political consent would generally be understood to contain such implied permissions.

34. Rawls, supra note 1, at 3.

You have Access Open access

Save article to Kindle

To save this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the or variations. ‘’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Available formats

Save article to Dropbox

To save this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Dropbox account. Find out more about saving content to Dropbox.

Available formats

Save article to Google Drive

To save this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you used this feature, you will be asked to authorise Cambridge Core to connect with your Google Drive account. Find out more about saving content to Google Drive.

Available formats

Reply to: Submit a response

Please enter your response.

Your details

Please enter a valid email address.

Conflicting interests

Do you have any conflicting interests? *