Is There Ever a Duty to Obey Orders to Wage an Unjust War?
Laurance S. Rockefeller Visiting Fellow
University Center for Human Values
University of North Carolina at Greensboro
One gap between the law’s demands and those of morality that combatants confront all too frequently is the order to fight in an unjust war. I will refer to such soldiers as unjust combatants, by which I mean those ordered to fight for a political community without a moral right to wage the particular war in question.<![if !supportFootnotes]><![endif]> Most contemporary discussions of the rights and responsibilities of unjust combatants defend one of the following three positions. The first, familiar to most from Michael Walzer’s account of just warfare, claims that all but the highest ranking combatants lack the ability to assess the justice or injustice of a given war. The reason for this is that they are subject to coercion (or the threat thereof) and manipulation which renders them incapable of forming a judgment on the matter.<![if !supportFootnotes]><![endif]> Since combatants cannot be expected to act as responsible agents when it comes to assessing the justice of a war, it follows that they have no duty to do so, and that they should not be held responsible for the fact that they unjustly wage war.
The second position, defended by Jeff McMahan among others, rejects Walzer’s depiction of combatants and argues instead that all combatants do have at least some ability to assess the justice or injustice of a given war.<![if !supportFootnotes]><![endif]> Since they can form a judgment on the matter, which is obviously one of great moral importance, it follows that they have a duty to do so. Moreover, McMahan argues that they have a duty to act on that judgment, which distinguishes this approach from the third one I describe below. Whether, and to what degree, unjust combatants should be held morally responsible for failing to recognize the impermissibility of the war in which they fight, or for their failure to act on the judgment that the war is unjust, depends on a range of factors that will vary from case to case. There may also be good reasons not to hold unjust combatants legally responsible for participating in an unjust war, even if they bear full moral responsibility for having done so. Still, unlike Walzer, who fully excuses combatants for the injustice of their war by arguing that they should not be viewed as responsible agents when it comes to assessing the justice or injustice of going to war, McMahan excuses combatants (when he does so) because of various factors that limited or impaired their ability to reason well.
The third approach, advocated by writers such as David Estlund, shares with the second the claim that unjust combatants have at least some ability to assess the justice or injustice of a given war, but it departs from that approach in denying that combatants should always act on their own judgment.<![if !supportFootnotes]><![endif]> Rather, if they are subjects of a state with a morally justifiable claim to political authority (henceforth, a legitimate state), then they have a moral duty to defer to the state’s judgment of the war’s justice, even if they believe (rightly) that the state’s judgment is in error.<![if !supportFootnotes]><![endif]> Given such a duty, it follows that unjust combatants should not be held morally responsible for their participation in an unjust war.
Most of this paper consists of a sympathetic critique of David Estlund’s defense of this third approach: in short, how the legal command of a legitimate state can make it not only morally permissible but morally obligatory to act in ways that one (rightly) believes wrongs others. But since talk of a moral duty to wrong others will likely strike many as a clear contradiction – a logical impossibility – I will begin with a brief attempt to clarify the position to be defended.
Estlund writes that
under the right conditions, even though the victim is wronged by the unjustly warring side, the soldier on that side is nevertheless morally obligated (and so morally permitted) to follow all normally binding orders – those that would be binding at least if the war were just. …Even when the killing wrongs the person [the innocent victim], the person doing the killing is not always acting wrongly.<![if !supportFootnotes]><![endif]>
The key to squaring this circle lies in the attribution of the wrong to the state. If an unjust combatant waging war on behalf of a legitimate state kills his victim as part of an effort to comply with his legal superiors’ command, then he does not act wrongly. Rather, the state acts wrongly in commanding him to fight in what is, in fact, an unjust war. Thus it is the state that is the proper object of judgments of responsibility for the wrongful killing; i.e. the proper object of blame, punishment, and claims for reparations. Needless to say, this depiction of who bears responsibility for wrongdoing stands in need of a lengthy defense, and though some of the arguments in the remainder of this paper contribute to such an exercise, other questions remain unaddressed. For example, I make no effort to investigate the ideas of collective agency and collective responsibility that seem to be entailed by the claim that a legitimate state, but not its military personnel, act wrongly when they wage (what is in fact) an unjust war. In suggesting that the state, but not the soldier, wrongs the innocent victim, I only aim to enable the reader to see past what may at first seem a clear and insurmountable barrier to an argument for a moral duty to follow orders in an unjust war.
Estlund’s Defense of a Duty to Follow Orders in an Unjust War
Estlund argues that officials have a moral duty to obey a legal order, even if they think (rightly) that the commanded act is an unjust one, if (and only if?) the following conditions are met:
<![if !supportLists]>1) <![endif]>The particular act (act-token) is of a type (act-type) that is justifiable under certain conditions, even if those conditions are not actually met in the case at hand.
<![if !supportLists]>2) <![endif]>The order is the result of a process that those subject to it have good reason to believe tends to track the truth. That is, the process is one that tends to result in commands to perform particular acts only in cases where the conditions sufficient (and perhaps necessary) to justify those acts actually obtain.
<![if !supportLists]>3) <![endif]>Those subject to the order have good reason to think that those issuing the order genuinely believe that the conditions sufficient (and perhaps necessary) to justify the commanded act do obtain in this case.
<![if !supportLists]>4) <![endif]>Given the fact, or the possibility of, reasonable disagreement over what the truth is, and who knows it (i.e. who can claim to be a moral expert), the procedure leading ultimately to the command to perform such acts must be one that can be defended to all reasonable points of view.
The following example, taken from Estlund, illustrates his claim and perhaps generates some intuitive support for it.
Consider a criminal trial that results in the conviction of an innocent man, despite the fact that the trial procedure is a fair one, and that all those who participate in it make a conscientious effort to properly discharge their roles in that procedure. Or, in a different but equally relevant version of the example, the trial results in an unjustly harsh sentence, one disproportionate to the moral seriousness of the crime for which the prisoner has been convicted, or to the good to be achieved by imprisoning the convicted person. In both cases, the jailer (rightly) believes that compliance with the jury’s (or the Court’s) order entails wronging the prisoner.<![if !supportFootnotes]><![endif]> Assuming that the jailer could free the prisoner, and that he has no other moral reasons not to do so, is the jailer permitted, or even obligated, to act on his own judgment in this case or must he instead act as the Court commands?
Estlund maintains the latter, and offers the following points in defense of his claim. First, the legally required act in question is not a token of a type that is never morally justifiable. Rather, it is just that in this case the jailer believes that one or another of the conditions that must be met in order for that act (imprisonment for a certain number of years) to be justified is not met. Second, the jailer has good reason to believe that the procedure resulting in the prisoner’s conviction and sentencing have a tendency to track the truth; that is, to convict all and only the guilty, and to give them an appropriate punishment. Finally, given the possibility of reasonable disagreement over the guilt or innocence of the accused, and the morally proper response to him should he be found guilty, trial by a jury drawn from his fellow citizens is the only procedure for settling these disputes that can be defended to all reasonable points of view.<![if !supportFootnotes]><![endif]> Against such a background, were the jailer to act on his own judgment rather than deferring to the Court, he would be implicitly asserting his moral superiority to his fellow citizens. That is, embedded in the jailer’s judgment that the Court has erred is a claim to moral expertise (or at least relative superiority) which, Estlund claims, the jailer cannot defend. Put the other way around, the jurors can reasonably contest the jailer’s claim to know better than they do what morality requires or forbids in this case.<![if !supportFootnotes]><![endif]> Given that, in circumstances characterized by reasonable disagreement over what morality requires, the jailor cannot reasonably challenge the use of a fair and responsibly conducted jury trial to determine what the state ought to do to a person accused of a crime, it follows that the jailer has a duty to obey the Court’s order even if he (rightly) believes it to be mistaken. In short, if an agent (or official) is given a command arrived at via a process with a tendency to “get it right” morally speaking, that responds appropriately to the fact of reasonable (moral) disagreement, and that commands an act that would be just were things as those giving the command genuinely believe them to be, then that agent has a duty to carry out the command.
Estlund contends that the same argument applies even when the legally ordered act in question involves killing, be it the execution of a person (wrongly) sentenced to death, or a lethal attack on an enemy combatant in an unjust war.<![if !supportFootnotes]><![endif]> Of course, commands to go to war are not issued by a jury. Estlund maintains, however, that there is a suitable analog in the guise of an adequately democratic state, one whose institutions constitute a process of political deliberation and decision-making that has both some tendency to “get it right” morally speaking and that responds appropriately to the fact of reasonable disagreement. Given the role that the institutions of a democratic state play, directly and indirectly, in the process that leads ultimately to the issuance of a judgment that a war is justifiable, those the democratic state commands to fight in the war have a duty to do so, even if they think (rightly) that the war is, in fact, unjust.
With a clear understanding of Estlund’s argument for a duty to follow unjust legal orders hopefully in hand, I want to examine in greater detail the conditions under which such a duty obtains, the relationship between those conditions, and the contribution that each makes to justifying a duty to defer to the judgment of a legitimate (democratic) state.
Consider the first condition: the particular act must be of a type that is justifiable under certain conditions, even if those conditions are not actually met in the case at hand. A soldier might disagree with the state’s judgment that a given war is just on three different grounds. He might dispute the state’s factual claim that the conditions sufficient (and perhaps necessary) to justify going to war have been met. Alternatively, or in addition, he might dispute the state’s understanding of one or more of those normative conditions. For instance, he may think the state’s interpretation of just cause mistaken, rejecting its claim that a future but not imminent threat, or an imminent threat, or even massive human rights violations perpetrated against a population with no special relationship to the state he serves, can contribute to the moral case for going to war. Finally (or, again, in addition), the soldier may accept the state’s factual claims and share its understanding of the criteria that justify going to war, yet disagree with the state’s application of those criteria to those facts. He may dispute the state’s claim that the war will be proportional, or that sufficient effort has been put into alternative responses (i.e. that the criterion of last resort has been met), or that an internal armed conflict in some other state has risen to the level of genocide.
The first kind of dispute is clearly addressed by Estlund’s requirement that the commanded act (token) be of a type it can be morally justifiable to perform. After all, the soldier accepts that if the world is as the state claims it is, then going to war is justifiable; he merely rejects the state’s claim that, as a matter of fact, the conditions sufficient to wage just war are met. The first condition in Estlund’s argument also appears to address the third kind of dispute described above. In this kind of case, the soldier grants the justice of acts of a particular type, such as wars that are proportional and for which there is a just cause, but disputes the state’s judgment that the particular war it wishes to wage will be a token of this type; for instance, that the war will be proportional.
With regard to the second type of dispute, however, the requirement that the act be a token of a type that can be justifiable contributes nothing to the argument that the soldier has a duty to obey the lawful commands of a legitimate state, even if he thinks (rightly) that the commanded act is unjust. The reason for this, obviously, is that the soldier believes that the commanded act is one that can never be morally justifiable. Estlund clearly seems aware of this third possibility; indeed, he explicitly formulates his thesis so that the duty to obey is conditional on the commanded act being of a type that can be morally justifiable, and the only just cause for war he mentions is one in response to aggression (presumably because he thinks the justice of a war is uncontroversial in such cases). Nevertheless, the possibility, indeed, probability, of disagreement over the very criteria that suffice to justify going to war suggest that if this first condition really does constrain the scope of the duty to fight in (what is in fact) an unjust war, the scope of that duty may be quite narrow. These disputes center not only on what counts as a just cause for war, but also the sort of factors that should and should not figure in calculations of proportionality, and who has a claim to the authority to engage in a particular war (e.g. individual states, multi-lateral institutions, or the United Nations). Recent academic literature, as well as that written for a broader audience (e.g. in newspaper opinion columns), is rife with evidence of such disagreements. In practical terms, then, Estlund’s first condition may well contribute very little to the case for a duty to obey orders to fight in (what is actually) an unjust war.
One possible response to this challenge is to claim that Estlund’s argument only establishes in the abstract when it is the case that a soldier has such a duty, and not how a soldier (or anyone else) can determine whether he has a duty to obey an unjust command in a particular case. If the act type is truly just (though the act token will not be), the soldier has a duty to obey commands to perform it, even if he mistakenly believes the contrary. To make this move, however, is to give up the aim of offering soldiers practical guidance; that is, to help them determine whether they ought to obey a command to fight in a war whose justice they question.
A second possible response would be to replace the claim that the legally required act be a token of a type the soldier believes to be just with the weaker claim that it be one he thinks it reasonable to believe is just.<![if !supportFootnotes]><![endif]> On the one hand, this response significantly broadens the range or number of cases in which the first condition for a duty to obey an unjust command will be met. For example, a soldier might find reasonable the state’s judgment that preemptive wars are justifiable, even if he also thinks it false. Consider, too, the following case of an order to join in an armed humanitarian intervention. A soldier believes that the widespread and massive violation of non-citizens’ basic human rights does not provide a just cause for war, but he admits (at least to himself) that he is not particularly confident in that judgment. In addition, though he shares his state’s understanding of what the ad bellum criterion of proportionality requires, he also believes that the intervention in question will not be proportionate, a judgment in which he is quite confident. On Estlund’s account, it appears that the first judgment (that wars of the type in question can never be just) liberates him from a duty to follow orders, while the second judgment does not (since he disputes the token but not the type), even though he is much more confident in the second judgment than the first. This conclusion feels somewhat odd: one can imagine the soldier saying “I have my doubts about the justice of humanitarian intervention, but the real reason I think this war will be unjust is because it will be disproportionate.” Better, perhaps, to argue that the soldier should obey unless he finds the state’s judgment unreasonable, while allowing that the object of the judgment can be either the state’s judgment of the facts of the case, or its judgment of the conditions sufficient to justify going to war, or its application of those conditions to those facts.
On the other hand, the replacement of the belief that the commanded act is of a type that is just with the belief that it is one the state reasonably but mistakenly thinks just may raise difficulties for the epistemic condition Estlund claims must be met in order to justify a duty to follow unjust commands. The state might consistently draw reasonable conclusions when it comes to going or not going to war without ever getting it right (either objectively, or in the soldier’s eyes). How, then, should we interpret the condition that the order to go to war be issued by a process (or an institution) with a tendency to get it right? If we alter the epistemic condition as well, so that the process must have only a tendency to reach reasonable (but not necessarily correct) conclusions, then while it may not become entirely superfluous, the epistemic condition will likely contribute very little to the case for the soldier’s duty to defer to the state’s judgment. Or at least that is so given a capacious understanding of what counts as a reasonable judgment, which Estlund defends elsewhere.<![if !supportFootnotes]><![endif]>
Interestingly, the substitution of ‘reasonable’ for ‘just’ in Estlund’s fist condition makes no difference to the democratic condition for a duty to follow unjust commands. Estlund claims that mutual subjection to an adequately democratic decision procedure is morally required in cases where people disagree over which person’s moral judgment is (more likely to be) the correct one. That issue remains even among agents that all grant the reasonableness of their opponents’ views, though they also think them false.
The point bears repeating in a bit more detail. Soldiers may have good reason to think that (some of) their superiors are better able to reason through the moral and factual issues involved in determining whether a given war is just than are the soldiers themselves. Their officers and elected leaders have greater access to, and appreciation for, relevant facts, a more detailed and subtle understanding of the moral criteria for waging just war, and a more developed and refined capacity for moral judgment (e.g. greater sensitivity to the various considerations in play, and how much weight each should carry), or so the soldiers may (rightly) believe. Nevertheless, even if some of the soldier’s legal superiors have greater expertise than others with respect to determining the justice of a proposed war, soldiers may have a difficult time identifying who the experts are.<![if !supportFootnotes]><![endif]> Second, even if they (rightly) feel confident that they have identified those with the greatest claim to expertise, Estlund claims that others can (and likely will) reasonably disagree with that judgment. In light of such reasonable disagreement, no person or group can sustain the claim to an exclusive right to rule on the basis of the fact (if it is one) that they are more likely to get it right, morally speaking, than are any others. The inability to make such an argument entails that all enjoy a claim to equal authority, one recognized (only?) in adequately democratic political institutions. It is the need to respect others’ claim to equal authority, if their views are reasonable, that does (nearly) all the work in justifying soldiers’ duty to fight in a war they (rightly) believe to be unjust.
The analysis of Estlund’s argument thus far supports the following conclusion: a soldier should obey an order he believes to be unjust if and only if (a) he thinks the (implicit) claim that the order is just a reasonable one, even if he also thinks it mistaken, and (b) the order arises from a procedure that constitutes a (or perhaps the only) morally justifiable response to reasonable disagreement over which agent’s moral judgment is most likely to be correct. Though this argument does not entirely dispense with truth, since some beliefs will be not only false but unreasonable, it does not make much of democratic institutions capacity to identify it.
Epistemic Arguments for Democratic Authority and the Particularity Challenge
Despite the reservations expressed in the previous section, suppose we grant that the tendency of democratic decision procedures to get it right when it comes to waging only just wars makes a significant contribution to the case for soldiers’ duty to fight in what they (rightly) believe to be an unjust war. It does not follow necessarily that a soldier should defer to the judgment of the democratic state that employs him regarding the justice of a particular war, rather than that of some other democratic state. This is particularly so if when it comes to assessing the justifiability of going to war, the soldier has reason to think that: (a) the democratic decision procedures employed by another state do a better job of tracking the truth than do those of the soldier’s state, and/or (b) that in this particular case officials of the soldier’s state are more likely to suffer from bias and conflicts of interest than are the officials of some other adequately democratic state. So for example, suppose that in 2003 the U.S., France and Sweden were all governed via adequately democratic decision procedures with a tendency to go to war only when they were justified in doing so. Why should a U.S. soldier defer to the judgment of the U.S. government regarding the (moral and legal) justifiability of invading Iraq, rather than that of France or Sweden? In light of the history of Iraqi-U.S. relations since the early 1990s, the emotionally charged atmosphere following the 9/11 terrorist attacks, and the long-lived and close connections between some very high ranking officials in the U.S. government and oil companies that stood to profit enormously from a friendlier Iraqi government, he would have had good reason to worry about the impact of various biases and conflicts of interest on the judgment of the U.S. government regarding the justifiability of war with Iraq. This soldier would also have had some reason to suspect French political officials of bias, given France’s perennial resentment of the world’s sole hyperpower. But he might also have had reason to believe that with respect to the morality and legality of invading Iraq, the French government’s judgment was less likely to be biased than was the judgment of the U.S. government. To my knowledge, the soldier would have had no reason to suspect Swedish officials of suffering from any bias or conflict of interest, nor any reason to think that the truth-tracking quality of Swedish political institutions was far inferior to that of U.S. political institutions. It appears, then, that on epistemic grounds the soldier should have deferred to Sweden’s judgment that an American-led invasion of Iraq would be unjust, and perhaps France’s as well, rather than obeying an order issued by his legal superiors to join in that invasion.
The reader might protest that the U.S. government had access to intelligence that the Swedes lacked regarding the Iraqi military and its biological, chemical, and nuclear weapons activities and plans. Even if that is true, and even if the intelligence had turned out to be far more accurate than it was, it would not necessarily follow that the soldier had more reason to defer to the judgment of the U.S. than to that of Sweden. This is so because the justifiability of invading Iraq turned not just on factual questions (e.g. did the Iraqi’s have an on-going nuclear weapons research program?) but also on normative ones, such as whether preventive war, or even preemptive war, is morally (or legally) justifiable, and whether (morally and/or legally) the invasion required a new authorization by the U.N. With respect to these questions, I suggest that a U.S. soldier in 2003 would have had reason to think the Swedes more likely to get it right than the Americans (even if, in fact, the Swedes got it wrong and the Americans got it right).
As I noted above, it is not clear how much work the epistemic criterion – that is, the tendency of a democratic decision procedure to track moral truth – does in justifying a state’s claim to authority. Rather, most of the justificatory work seems to be done by the requirement that political justification be acceptable to all reasonable points of view. As the discussion of the imprisonment example made clear, were the jailer to act on his own (private) judgment rather than deferring to the Court’s, he would be implicitly making a claim to moral expertise that he could not justify to all reasonable points of view. Or in other words, given that the Court’s judgment was a reasonable one, the jailer could not justify acting on his own judgment rather than deferring to the Court. Might the necessity of being able to publicly justify his action to all reasonable points of view provide the basis for arguing that a soldier in the U.S. military ought to defer to the U.S. government’s judgment regarding the justice of war with Iraq, rather than to the Swedish government’s judgment? I do not see how it does so. The soldier does not privilege his own judgment regarding the justice of invading Iraq over that of all other agents, nor does he implicitly claim any moral expertise. Rather, he defers to the judgment of a democratic decision procedure, which is precisely what Estlund argues agents ought to do in cases where there is reasonable disagreement over the morally proper course of action. Of course, the soldier does act on his own judgment regarding which democratic decision procedure is most likely to get it right on this particular issue. But Estlund already acknowledges that each individual must judge for himself or herself whether the state meets those criteria necessary to justify its claim to authority. It is not clear why that inquiry should not also include determining whether some democratic state other than the one that claims the agent as a subject is more likely to track the moral truth (generally, or in this particular case).
It might be thought that this line of argument is beside the point; the U.S. soldier should obey the U.S. government’s command to fight in (what is actually) an unjust war because he has a legal obligation to do so, and no legal obligation to guide his conduct according to the Swedish government’s judgment that the war is unjust. This won’t do, however, since the question at issue is whether the soldier has a moral duty to do what he has a legal duty to do.
Perhaps the most promising basis on which to argue that a U.S. soldier has a duty to obey the U.S. government, but not the Swedish government, is that he has voluntarily joined the U.S. armed forces and not the Swedish ones. By agreeing to obey all lawful orders issued by his superiors, the soldier places himself in a special relationship vis-à-vis the U.S. government, one that he does not have to Sweden. Note, first, that this argument establishes a duty to obey a state’s command to wage (what is actually) an unjust war only for those who voluntarily join that state’s armed forces. Some further argument will be necessary to show that conscripts also have such a duty, and even among putative volunteers it may be that only those whose decision to enlist is informed and made against a background of reasonable options actually acquire a duty to obey the lawful orders of the state in whose armed forces they serve. Second, consent may obviate the requirement that a state be democratic in order to enjoy legitimate authority. It is not typically the case that one agent’s consent to do as another directs depends on that person’s command being one that is justifiable to all reasonable points of view. Perhaps it is different in the case of consent to rule by a state or political community. But if not, and if the case for democratic authority rests almost entirely on the requirement of public justification, then it follows that consent can generate a particularized duty of obedience regardless of a state’s democratic credentials.<![if !supportFootnotes]><![endif]>
Third, most theorists maintain that one cannot acquire a duty to act immorally. Given a duty not to facilitate murder, a voluntary agreement to lend a criminal assassin the use of your gun is void ab initio, and so creates no duty to lend the gun or to compensate the assassin for failing to do so. Of course, our question concerns what a soldier ought to do when he draws a different conclusion regarding the justice of a war from the one reached by his state. Recognizing that he can have no duty grounded in consent to act immorally, the soldier should act on whatever judgment he has reason to believe is more likely to be correct. If that is the case, however, then it seems that a U.S. soldier ordered to join in the invasion of Iraq ought to act on the Swedish government’s judgment that such a war would be unjust, at least if the foregoing argument regarding the likelihood of bias and conflicts of interest has merit. Even consent, then, may not provide a solution to the problem of particularity. This point should not be overdone, however. It is possible, and perhaps even likely, that in some cases the soldier’s own state will be the one he has most reason to believe will judge correctly the justice or injustice of a given war. In those cases the soldier does have a consent-based duty to obey his legal superior’s commands, even if (a) his state’s judgment that war is justified in this case turns out to be mistaken, and (b) the soldier believed this to be the case (though he also believed that his own judgment on the matter was more likely to be mistaken than his state’s judgment). Notice, however, that the combatant’s duty is not grounded in any special relationship he has to his state; rather, it follows from a contingent fact about the state’s relative epistemic superiority.
On Orders Not to Fight in a Just War
Thus far I have assumed that what is at issue is a combatant’s duty to obey a legal order to fight in what he rightly believes to be an unjust war. I want to bring this discussion to a close by considering the opposite case; that is, a situation in which a combatant is ordered not to engage in what would be a just war. If states never have a duty to go to war, but only in some cases a right to do so, then a situation like this is of considerably less moral interest than the one to which I have devoted the rest of this paper. This is so because in refusing to fight a war that he has a right, but not a duty, to wage, an agent does not wrong anyone. Suppose, however, that states have a Samaritan duty to go to war if there is good reason to believe that it is practically necessary to prevent a great injustice. By ‘a Samaritan duty’ I mean one that an agent must discharge only if he can do so at a reasonable cost to his own interests, while by ‘a great injustice’ I mean to indicate only the gravest of those injustices that provide a just cause for war. For my purposes, large-scale genocide and ethnic cleansing count as great injustices, while the seizure of a portion of a state’s territory, with relatively little harm done to its inhabitants or its natural and built environments, does not. If such a duty exists, what should a combatant do in a situation where he and a certain number of his comrades in arms are in a position to initiate an armed humanitarian intervention, and he believes (rightly) that his state has a moral duty to carry out that intervention, but his state orders him not to do so?
On the one hand, Estlund’s argument may have the same implication in this case as it does in a case where the state orders a soldier to fight in (what is actually) an unjust war. If the state’s judgment has been reached via a process with some tendency to track the moral truth, while also recognizing the equal right of all its citizens to determine (indirectly) what the state ought to do, then the combatant ought to comply with that judgment even if he (rightly) believes it to be mistaken. On the other hand, Estlund does acknowledge the possibility of an exception to the principle that an official in the service of a legitimate state has a duty to obey its commands. Specifically, if the legal official has first-hand knowledge of the order’s injustice, if he knows “with as much certainty as life allows” that obeying the order will result in the wrongful treatment of the victim, then perhaps he is permitted (or even obligated) to disobey it.<![if !supportFootnotes]><![endif]> Estlund seems to suggest that no such situation could arise with respect to the decision to initiate a war, but I want to suggest one case that might prove him wrong, namely that of the United Nations Assistance Mission for Rwanda (UNAMIR) in the weeks and days leading up to the Rwandan Genocide.
On January 11th, 1994, General Romeo Dallaire, the military commander of U.N. forces in Rwanda, sent a fax to his superiors at the United Nations.<![if !supportFootnotes]><![endif]> In it he described in some detail the preparations that those aligned with the Hutu Power movement were making for an extermination campaign aimed at Tutsis and tens of thousands of Hutus judged to be opposed to, or simply insufficiently supportive of, the movement’s political goals. He also reported his intention to carry out a raid on an arms cache members of the Hutu Power movement had created in anticipation of their genocidal attack. Dallaire believed that he was already authorized to carry out this raid under his existing rules of engagement, and remarks made by Kofi Annan, then the head of all U.N. peacekeeping operations, before a U.S. Congressional committee only a few months later would seem to corroborate Dallaire’s belief. Nevertheless, that same day (?) Dallaire received a fax from U.N. headquarters, sent under the name of Kofi Annan and signed by his deputy Iqbal Riza. In it they forbade him from carrying out the raid, and more generally, from taking any action to prevent what Dallaire believed to be a looming genocide, other than sharing what he knew with the Hutu president of Rwanda, Habyarimana, which Dallaire (and, probably, his superiors) knew would be a wholly ineffective response.<![if !supportFootnotes]><![endif]> On many subsequent occasions, Dallaire has stated that with the troops at his disposal at that time he could have done much to mitigate the subsequent deaths of hundreds of thousands of Rwandans, especially had he been permitted to move against agents of the Hutu Power movement months before the genocide began. Our question, then, is this: was General Dallaire morally permitted, and perhaps even obligated, to disobey his legal superiors’ order not to initiate an armed intervention, or was he instead morally required to obey this command even though he (rightly) thought it required him to act unjustly?
What would General Dallaire have needed to know if he were to be justified in initiating an armed intervention to prevent (or at least mitigate) the ensuing Rwandan genocide, in contravention to his legal orders? Various facts about the organization and intentions of those that instigated the genocide, of course, as well as facts about the ability of the troops at his disposal to prevent or at least contain it, and the risks they were likely to confront if they attempted to intervene. But as I argued earlier, normative judgments play at least as important a role in settling whether an agent ought to go to war as do factual judgments. Thus General Dallaire would have had to make a judgment regarding the existence of a Samaritan duty to use force (if necessary) to forestall a genocide. Perhaps he should have concluded that the existence of such a duty was a matter of reasonable disagreement, with the implication that he ought to defer to the judgment of the political body employing him (or at least its adequately democratic member states, especially Belgium, France and the U.S., who were the primary democratic states concerned with events in Rwanda). Yet on the basis of hundreds of public statements that the world should never again stand-by while genocide occurs, the existence of the Convention on the Prevention and Punishment of Genocide, and other such evidence, he might just as well have reasoned that no reasonable agent would deny the existence of a duty to prevent or limit genocide if it could be done at a reasonable cost.<![if !supportFootnotes]><![endif]> Certainly many events after the Rwandan genocide began, such as the efforts by a number of states including the U.S. to deny that what was happening in Rwanda was a genocide, for fear of the legal and moral consequences that would follow subsequent to such an admission, and later remarks by many officials echoing Madeline Albright’s claim (as U.S. Secretary of State) that “we – the international community – should have been more active in the early stages of the atrocities in Rwanda,” suggest that Dallaire could have drawn such a conclusion.<![if !supportFootnotes]><![endif]> Note the crucial point here, namely the assertion that Dallaire would have been right to conclude not only that there exists a Samaritan duty to prevent genocide, but also that it would be unreasonable for anyone to deny it. Without this last claim, on Estlund’s account of legitimate authority Dallaire would not have been morally justified in acting on his own judgment, rather than deferring to the U.N. and/or its member states (or at least its adequately democratic member states). If true, however, then given the facts as he knew them, Dallaire would have been justified in disregarding the command to refrain from any act of armed intervention aimed at preventing genocide.
The foregoing argument seeks to provide a justification for Dallaire privileging his own judgment over that of the political body he serves on the grounds that agents occasionally enjoy such insight into the truth on a given matter that they may justifiably act on that insight rather than defer to the contrary judgment of a legitimate authority. It may be possible to draw the same conclusion by arguing that Dallaire had good reason to think that those states most closely involved in Rwandan affairs leading up to the genocide were severely biased (e.g. the U.S. after the Somalia fiasco) or suffered from conflicts of interest (e.g. France and Belgium), so that the orders he was given did not issue from a procedure with a (sufficient) tendency to track the truth.<![if !supportFootnotes]><![endif]> This argument, too, nullifies legal orders for purposes of an agent’s deliberation.
I want to emphasize that even if Dallaire had no reason to obey U.N., Belgian, etc., orders not to intervene in Rwanda, it does not follow that, morally, he ought to have given a command to intervene. Various other considerations might have entailed the immorality of such a command; for example, the belief that many of the U.N. troops in Rwanda would not have complied with it (even though, morally, they should have), rendering his illegal conduct nearly pointless. Moreover, it is possible that even if the risks of harm arising directly from intervention would have been reasonable, the addition of the risk of (morally unjust) punishment for having acted illegally would have rendered intervention too costly for Dallaire and the other soldiers in Rwanda – too costly, that is, for them to have a Samaritan duty to intervene. My point is not to make an all-things-considered argument for what Dallaire should have done. Rather, my aim is only to raise the possibility that in answering this question, Dallaire should not have given any weight to the mere fact that intervention was legally forbidden. If true, then discussions of the duty to follow unjust orders pertaining to war must pay some (though perhaps not as much) attention to orders not to go to war, as well as orders to go to war.
<![if !supportFootnotes]><![endif]> [See McMahan on the term ‘unjust combatants’ – he may restrict it to those who lack a just cause for going to war, as opposed to the larger group defined by a failure to meet the just cause criterion or any of the other jus ad bellum criteria.] For simplicity’s sake, I refer throughout this paper to ‘soldiers,’ though obviously my remarks apply as well to sailors, airmen, and other types of military personnel.
<![if !supportFootnotes]><![endif]>Michael Walzer, Just and Unjust Wars, ---. Note that Walzer’s claim is not simply that the typical soldier will lack the information she needs, or the skills necessary to evaluate it. Rather, his claim is that modern states (often deliberately) adopt policies that undermine their subjects’ ability to reason well, so much so that we should not view them as responsible agents when it comes to joining in a war. Of course, Walzer maintains that all combatants remain responsible agents when it comes to the actual conduct of the war.
<![if !supportFootnotes]><![endif]> Jeff McMahan, ‘The Ethics of Killing in War,’ Ethics 114 (July 2004): 693-733.
<![if !supportFootnotes]><![endif]> David Estlund, ‘On Following Orders in an Unjust War,’ Journal of Political Philosophy 15:2 (2007): 213-34.
<![if !supportFootnotes]><![endif]> Estlund speaks of knowingly wronging someone, but I find this way of speaking distracting. The question at issue is what one should do when acting under significant, but not altogether debilitating, uncertainty. Thus our concern is with the combatant who has been ordered to fight in what he believes to be an unjust war, but who also recognizes that there is more than a trivial chance that his judgment on this matter is mistaken. For that reason, I place the qualification ‘rightly’ in parentheses when speaking of the soldier’s belief, since what matters is not whether his belief is in fact correct, but whether he has reason to think that his judgment is more likely to be correct than is the state’s.
<![if !supportFootnotes]><![endif]> Estlund, ‘Following Orders,’ 215-16.
<![if !supportFootnotes]><![endif]> In the disproportionate sentence variation on this example, we can imagine that the prisoner has already served what the jailer (rightly) believes to be a period of incarceration proportionate to his offense.
<![if !supportFootnotes]><![endif]> Estlund mentions one other consideration that may increase the appeal of his overall argument for obedience to unjust orders, namely that there exist means by which the jailer can contest the Court’s decision, and perhaps also opportunities for the jailer to “recuse” himself from particular cases or resign his position altogether rather than carry out what he believes to be an unjust order. I set these considerations aside here, since the first one only postpones the issue of what to do when confronted with what one believes to be an unjust order, at least on the assumption that the contestation will sometimes fail, and the second one because I believe that citizens and officials of a legitimate state have no moral right to conscientious objection (selective or otherwise), though a compassionate and merciful state may exempt them from legal obligations they perceive to be unjust.
<![if !supportFootnotes]><![endif]> In some rare cases, it may be that the jailer’s disagreement with the jury is a purely factual one, and that he does have a justifiable claim to know better than the jury what the facts of the case are (meaning that they would be unreasonable to deny his epistemic superiority on this matter). Estlund grants that in such cases the jailer may be permitted, or even obligated, to disregard the Court’s order, but he also claims that no analogous case is ever likely to arise with respect to initiating a war. To this I would add that reasonable disagreement over factual claims are quite common, especially in cases where conclusions must be drawn on the basis of (sometimes seriously) incomplete and even conflicting information, drawn from sources of varying and disputable reliability.
<![if !supportFootnotes]><![endif]> Note that Estlund’s argument is conditional; if execution for a crime is ever a morally justifiable punishment, then the executioner has a duty to carry out such a sentence even if he thinks it unjustified in the particular case at hand. Estlund also argues compellingly against those who would distinguish between punishments that should be meted out even when the official charged with doing so believes it to be unjust and those that should not.
<![if !supportFootnotes]><![endif]> The notion of reasonableness I employ here involves judgments made under what Rawls labels the burdens of judgment, “where it is not to be expected that conscientious persons with full powers of reason, even after free discussion, will arrive at the same conclusion” (John Rawls, Political Liberalism, (New York: Columbia University Press, 1995): 58).
<![if !supportFootnotes]><![endif]> David Estlund, Democratic Authority (Princeton: Princeton University Press, 2008): Chapter 3.
<![if !supportFootnotes]><![endif]> Moreover, as I argue below, it is by no means clear why soldiers should confine themselves to the judgments issued by their legal superiors.
<![if !supportFootnotes]><![endif]> Estlund might not be bothered by this implication if he did not claim that only democratic states can enjoy legitimate authority, which he at least appears to do. See Estlund, ‘Following Orders,’ 224.
<![if !supportFootnotes]><![endif]> Estlund, ‘Following Orders,’ 218.
<![if !supportFootnotes]><![endif]> This depiction of events draws on Romeo Dallaire, Shaking Hands with the Devil: The Failure of Humanity in Rwanda (Toronto: Random House Canada, 2003), especially pp. 141-51, and Philip Gourevitch, ‘The Genocide Fax,’ The New Yorker, May 11th 1998, p.42-6.
<![if !supportFootnotes]><![endif]> Dallaire was also instructed to share his beliefs regarding an impending genocide with the Belgian, French, and U.S. ambassadors to Rwanda, a point I return to below.
<![if !supportFootnotes]><![endif]> My claim is that the facts cited in the text would have provided General Dallaire with evidence to support the judgment that no one could reasonably deny the existence of a Samaritan duty to prevent genocide, not that these facts (in international law terms, practice and opinio juris) made it the case that there is a Samaritan duty to prevent genocide.
<![if !supportFootnotes]><![endif]> Albright’s remark is quoted in Gourevitch, ‘Genocide Fax.’
<![if !supportFootnotes]><![endif]> In discussing why the U.N. and its member states did so little to prevent the Rwandan genocide, or to stop it once it began, Iqbal Riza claims that all discussions of it took place under the shadow of Somalia – meaning, especially, the U.S. experience in Somalia and the commitment of the Clinton administration at that time not to get involved in any interventions or peacekeeping (at least in Africa).