The Ethics of Arming Rebels

Abstract

Despite the popularity of arming rebels as a foreign policy option, there is very little, if any, detailed engagement with the ethical issues surrounding the practice. There is a growing literature on the ethical issues surrounding civil wars and, more specifically, the conditions for engaging in just rebellion; but the focus of this literature is largely on the question of the justifiability of the rebels themselves in engaging in civil war and their conduct when doing so, rather than the permissibility of the arming of rebels by other agents. It is precisely this issue that I want to address here. Overall, I argue that the process should be generally eschewed. More specifically, this article seeks to establish that arming rebels is generally impermissible and only exceptionally morally permissible (even, as I will argue, when rebels are engaged in unjust wars). The former, far more restrictive claim will be established in the first part of the article. The latter, more permissive claim will be established in the second part of the article.

One of the most notable ways that states attempt to realize their foreign policy goals is the arming of rebels in states where there is severe oppression, low-level conflict, or civil war. This was particularly patent during the Arab Spring. Following the uprising in Libya in 2011, various states provided lethal and nonlethal arms to the forces opposed to Qaddafi's regime, including material and financial support from the Libya Contact Group as well as arms from France, Qatar, and the United Kingdom. ¹ In Syria the arming of the various parties by external actors became one of the key elements of the conflict. It was reported that, on the one hand, Qatar and Saudi Arabia (and others) supplied arms to the Free Syrian Army (FSA), the United Kingdom and France provided it with supposed nonlethal support, and the United States and Turkey facilitated and coordinated the supply. ² On the other hand, Russia and Iran supplied weapons such as missile systems, mortars, and rockets to the Assad regime. In 2012 the United Kingdom even secretly drew up plans to train and equip a 100,000-strong Syrian rebel army abroad, which would then strike at Assad in a manner similar to the “shock and awe” strikes on Iraq in 2003. ³ And in September 2014 the U.S. Congress approved President Obama's plan to train and equip “moderate” Syrian rebels. ⁴

For states, there are two main reasons why arming rebels may be preferable to direct military action. First, it is often far less costly for the sending state, both in terms of the lives of military personnel and financial resources. ⁵ (This has led some to refer to the arming of rebels as “intervention-lite.”) ⁶ Second, the arming of rebels can more easily be carried out covertly, that is, out of the public gaze and without the widespread knowledge of the international community. The upshot is that states have often supplied arms to rebels, most infamously under the Reagan Doctrine, which aimed to support anti-Communist insurgencies. Notorious examples include the U.S. arming, training, and financing of the Nicaraguan Contras against the Sandinista government (including the covert funding of the Contras without congressional approval in the Iran-Contra affair); the British arming, in contravention of a UN arms embargo, of Sierra Leone through Sandline, a UK-based private military and security company (the Arms-to-Africa affair); and the Russian supplying of arms to various pro-Russian separatist rebels in former Soviet states (including, at the time of writing, Ukraine). There have also been numerous cases of the arming of rebels in potentially more morally justifiable cases. These include the supply of arms by the United States and some Islamic groups to the Bosnian Muslims during the breakup of the former Yugoslavia, and Soviet and Cuban military support for anti-apartheid forces in South Africa.

Despite the popularity of arming rebels as a foreign policy option, there is very little, if any, detailed engagement with the ethical issues surrounding the practice. ⁷ There is a growing literature on the ethical issues surrounding civil wars and, more specifically, the conditions for engaging in just rebellion ⁸ ; but the focus of this literature is largely on the question of the justifiability of the rebels themselves in engaging in civil war and their conduct when doing so, rather than the permissibility of the arming of rebels by other agents. It is precisely this issue that I want to address here. Overall, I argue that arming rebels should be generally eschewed. More specifically, this article seeks to establish that arming rebels is generally impermissible and only exceptionally morally permissible (even, as I will argue, when rebels are engaged in unjust wars). The former, far more restrictive claim will be established in the first part of the article. The latter, more permissive claim will be established in the second part of the article.

Before I begin, some clarifications are necessary. First, by “arming” rebels, I mean the provision of military equipment. There may be different issues when states merely train and finance rebels, but space precludes considering these.

Second, by “rebels” I mean those who are in armed opposition to their established government and possess political objectives. These objectives may be to supplant the current government (as in the various uprisings during the Arab Spring) or to secede (such as the Kosovo Liberation Army's attempt to secede from Serbia). Other objectives, which all take the form of resisting or challenging the government, include securing a (potentially greater) role in a power-sharing agreement; achieving changes to a particular government policy; and improving the human rights situation (such as the Zapatista rebellion in Mexico). Accordingly, rebels may sometimes seek self-determination, but this is not a necessary feature of rebellion. Rebels may, for instance, attempt to secure their or others’ protection against threats, improve their states’ compliance with international human rights standards, promote their material self-interest, or simply gain power. I will not consider, however, the case of resistance to foreign occupation, which is best seen as a war of liberation, rather than rebellion. ⁹

The General Impermissibility of Arming Rebels

In this section, I will argue that arming rebels is generally morally impermissible, and to do this I will present three objections to arming such groups.

The Rebel-Risk Objection

The first objection—what I call the Rebel-Risk Objection—raises an epistemic problem: it can be very difficult to determine precisely who the rebels are (especially when there is more than one rebellious faction, as has often been the case) and, related to this, it can be difficult or even impossible to determine whether they are fighting with regard to the principles of jus ad bellum and jus in bello and, if they were to get in power, whether they would be likely to be better than the current government. Indeed, these risks were frequently cited and debated over the past two years regarding the Syrian rebels. ¹⁰

Another important concern of the Rebel-Risk Objection is that it is very difficult to determine whether arming rebels will ensure that only those who are liable to harm will be subject to it. For instance, the supplying agent may not possess sufficient information about the liability of those against whom the rebels will use force (or the liability of the rebels themselves). And even if it did have this information, it may lack sufficient control over the rebels to ensure force is only used against those who are liable. Thus, the Rebel-Risk Objection also posits that there are major risks concerning those subject to force by the rebels.

To help flesh out this point, we can frame it in terms of the contingent pacifist objection to revisionist just war theory. ¹¹ This objection runs as follows. The waging of any war is likely to involve harms to those who are not liable to such harms, such as innocent civilians. ¹² On a revisionist approach to just war theory, there are likely to be many nonliable parties. This is because, on this view, not only is it wrong to target innocent noncombatants, it is also wrong to target innocent combatants, such as those who fight justly on the just side and even those who make just contributions to otherwise unjust wars. It is very difficult to ensure that those waging war will target only those who are liable, since it can be difficult both to determine who is innocent and to ensure that there is no harm to innocents. Given the worries about doing harm to innocents (rather than simply allowing it), the risks of waging war are too great.

Such risks are even higher in the case of arming rebels because those supplying arms will typically be in a worse position to determine who is liable, and must rely on those to whom they supply arms to make these judgments. Furthermore, whereas a state can exert some degree of control over its armed forces and have a series of mechanisms to try to ensure that only liable agents are subject to harms (for example, training in international humanitarian law and court-martials for soldiers), it typically lacks such control over rebels. The suppliers rely on a different actor, with its own set of interests and agenda, which they often have little means of influencing, bar the ceasing of funding or military support. Another way of putting this is in terms of the principal-agent problem: the principal (the supplier) has relatively little control over the agent (the rebels) in this context and few mechanisms to avoid, for instance, shirking international humanitarian law. ¹³ Moreover, the agents supplying arms are still doing harm (and not simply allowing it) since their supply of arms helps to produce the harm—it is an action, not an inaction. Hence, there should be a stricter presumption against supplying arms than against waging war.

The Escalation Objection

The second major objection is what I call the Escalation Objection—that is, that the arming of rebels could significantly escalate hostilities if, in response, other parties in the conflict also seek arms or intensify their response by further clamping down on any opposition groups. In this scenario, many more innocents are likely to be subject to serious harms. ¹⁴

Indeed, there is plenty of evidence that shows that such escalation is likely. For instance, during the Sri Lankan civil war the flow of arms prolonged the war and led to an escalation in fighting, as both the Tamil Tigers (LTTE) and the government imported arms in large quantities. ¹⁵ The likelihood of escalation is also documented in a recent study by Matthew Moore. ¹⁶ He finds that in the 114 civil wars between 1946 and 2002 where at least 900 people were killed, no rebel group was transferred major conventional weapons without the government also receiving arms from another source. In general, he finds that arming rebels (and governments) creates more bloodshed and does not end a conflict any quicker. The Escalation Objection, therefore, is a major concern when making proportionality calculations about whether arming rebels will do more harm than good.

One complicating issue, however, concerns what we can call the “interceding agency” (or “intervening agency”) of other actors in arming the opposing groups. ¹⁷ One might hold that the interceding agency of others in a causal chain should be viewed as reducing or removing the responsibility of agents earlier in that causal chain. Thus, it might be held that the state supplying the rebels should not be viewed as responsible if escalation occurs (or at least that this state has a lesser responsibility) because it is the other party that is responsible for the escalation, since it provides arms to the statist forces. On this view, the state supplying the rebels should still be viewed as acting permissibly and, in particular, in conformity with the requirements of proportionality, since the harms caused by other parties are discounted. Suppose, for instance, that the United Kingdom is considering arming a just rebel movement in Sudan. In response, it is reasonably foreseeable that the statist forces will obtain arms from China, which will significantly escalate the conflict, with the result of a much bloodier conflict in which many more innocents will die. The United Kingdom may still be viewed as acting permissibly since it is, on this view, China—and not the United Kingdom—that is responsible for the escalation of the conflict; the United Kingdom has done nothing wrong.

Consequently, considerations of interceding agency might be thought to reduce the force of the Escalation Objection. But how much do they discount the worries of escalation? An extreme view gives the interceding agency of others huge weight. On this view, in the example above the United Kingdom would not have any responsibility for the harms caused by escalation in Sudan. In this vein, Ned Dobos argues that mediated consequences do not matter for rebels and, further, that there is no good reason for thinking that they should matter for intervening parties either. ¹⁸ They do not matter for the rebels, he suggests, since those “facing violations of their basic human rights still retain their right of self-defense, irrespective of how others will react to the exercise of that right.” ¹⁹

However, this is too strong. Agents should still be generally held responsible for the foreseeable consequences of their actions at least to some extent. It seems plausible that interceding agency does not matter that much when assessing the case for the permissibility of states providing arms to rebels, and therefore the Escalation Objection still largely applies. In the hypothetical case of the United Kingdom and Sudan, we should consider the United Kingdom as acting wrongly because of the foreseeable negative effects. Moreover, there is an obvious explanation for this intuition: reasonably foreseeable consequences matter morally when we are assessing what agents should do. Agents should, as far as is reasonably possible, attempt to promote good consequences and avoid bad ones, although certain, nonabsolute deontological constraints (such as against doing harm and the import of having right motives) act as a check on this. This is not to deny that the permissibility of arming rebels is somewhat affected by the fact that the agency of others intercedes when they supply arms to opposing forces. If one holds that the difference between doing and allowing harm has some moral force, it also matters when another's agency intercedes. Thus, if you are not directly causing the harm, your action—like when allowing harm—is more causally remote. ²⁰

In this way, the foreseeable harms of escalation caused by the interceding agency of others seem to be somewhat—but not fully—discounted in the assessments of proportionality. If this is correct, this would reduce some of the force of the Escalation Objection, particularly in marginal cases where arming rebels before the interceding agency of others does much good but overall does marginally more harm, largely due to the reactions of others. Of course, most cases are unlikely to be marginal; consequently, the Escalation Objection will still tend to apply. That is, in most cases, it is likely that the arming of rebels will cause significantly more harm than good, even if due to the interceding agency of others. In such cases, it is impermissible to arm rebels.

The Diffusion Objection

A third major objection—what I call the Diffusion Objection—is that it is difficult to ensure not only that just rebels are armed, as in the Rebel-Risk Objection, but also to ensure that the weapons are delivered to or stay in the possession of the just rebels. ²¹ The weapons may become seized by other, unjust groups as the just rebels lose battles or their stockpiles are raided. More broadly, the weapons—particularly small arms—are unlikely to be destroyed after the conflict, and may be sold on the black market or retained by former combatants. ²² The worry is that there will be a proliferation of arms, which would lead to significant problems, most obviously higher homicide rates, more violent crime, and further conflict. Thus, even if the supply of arms to rebels may appear to be justified in terms of tackling the immediate crisis, it is possible and indeed likely that such arms will enable significant wrongdoing in the longer term as arms are diffused in the affected society and beyond. ²³ Of course, the black market may mean that rebels and other groups are able to obtain arms anyway (especially small arms), and that this is a key means of diffusion. However, this does not repudiate the objection, since the worry is that the supplying of arms will lead to even greater diffusion, that is, provide another means in addition to the black market supply of arms. ²⁴

The upshot of the three objections is that arming rebels should generally be eschewed. Together, these objections show that arming rebels is likely to lead to significant harms to innocents, and therefore will typically be disproportionate—that is, cause more harm than good. ²⁵

Two Implications

The three central objections have two further, notable implications. First, they provide some (if not indefeasible) reason to be skeptical about the general permissibility of arming states. As for the Rebel-Risk Objection, there is a risk that governments may misuse the weapons supplied, it can sometimes be hard to judge the justifiability of the state and its various organizations, and there may be a lack of control over how government forces will use the military resources supplied. As for the Escalation Objection, there is sometimes a worry that arming a state can foreseeably escalate the conflict because other parties will then arm the rebels and because there will be increased incentives for the government not to settle. ²⁶ As for the Diffusion Objection, the arming of even legitimate states can be very difficult to control. For instance, Thomas Jackson finds that one of the central ways for rebels to obtain their weapons is to steal them from government stockpiles. ²⁷

Second, the three main objections also apply to some extent to wars and direct military interventions, and therefore show that there is not significant reason to favor these options over arming rebels. The Rebel-Risk Objection is often likely to apply to wars and direct military interventions, given that intervening militarily in favor of rebels obviously poses notable epistemic risks in terms of who are the rebels that are being supported by direct military intervention. The Escalation Objection may also apply since other parties may intervene militarily to support other sides or may provide arms to them. ²⁸ Finally, the Diffusion Objection may apply as well because the obtaining of arms from external actors is commonplace, whether purchased, stolen, or won on the battlefield. ²⁹

To be sure, these objections are more likely to apply to the arming of rebels than to direct military intervention, largely because of the issue of control considered above. Intervening forces can often, but not always, exert greater control over their armed forces than when relying on other groups, and this can minimize some of the risks of intervening and the potential diffusion of weapons. Overall, arming rebels is likely to be somewhat worse than direct military intervention, given this point, but not much worse, given that the three objections may still apply (to a lesser extent) to war and direct military intervention. ³⁰ Thus, alternatives should be sought to both direct military intervention and arming rebels.

The Exceptional Permissibility of Arming rebels

Given the general problems of providing weapons to rebels, should there be an absolute, blanket ban on arming rebels? In what follows, I argue that, despite the general presumption, there are instances where arming rebels is permissible.

To establish the case for the exceptional permissibility of arming rebels, we first need to note that the arming of rebels might exceptionally be proportionate, that is, achieve more good than harm. This is because, first, the three key objections might not apply, or might not apply to a significant degree. For instance, the Rebel-Risk Objection might not be serious when much is known about the rebels, when the rebels are highly likely to fight justly for just causes, and when there is significant control over them. Second, even when the objections do apply, arming rebels might still be proportionate because the situation is so awful that the risks are worth taking, such as potentially in the Bosnian and South African cases noted above. Furthermore, when the arming of rebels is the only feasible means of responding to mass killings, the risk may also be worth taking. Although this might result in numerous harms to innocents, the alternative is certain: it will result in numerous harms to many innocents. Therefore, there may sometimes be a lesser-evil justification for providing arms, even when there is a notable risk to innocents. ³¹

To be sure, it can be expected that these instances will be rare and that arming rebels will generally be disproportionate. ³² First, it seems that the three objections will generally apply: it will generally be difficult to determine who the rebels are, arming rebels will generally escalate the conflict, and arms will generally diffuse. Second, it is unlikely that arming rebels will do enough good to outweigh these likely harms. For instance, any increase in the fighting capacity of the rebels, and their potential ability to achieve a just end, is likely to be matched by an increased fighting capacity of the government to oppose the rebels, and there will be further problems of rebel risk and the diffusion of weapons after the conflict.

The second step in establishing the exceptional permissibility of arming rebels is to consider exactly why arming rebels might be justified. The case is somewhat analogous to the case for the exceptional permissibility of humanitarian intervention. As with humanitarian intervention, what justifies the (exceptional) right to arm the rebels is a broad and serious threat to the enjoyment of basic human rights. Most clearly, this might consist of helping rebels to overthrow an oppressive regime, to secede from a repressive majority, or to reduce the extent of government repression. (In addition, I will argue below that arming rebels could be a legitimate means for putting an end to human rights violations even when the rebels are pursuing an unjust war.) Like the arming of rebels, humanitarian intervention sometimes bolsters a particular side, albeit with the ultimate, predominant aim of promoting basic human rights in the affected political community. My point, then, is that if one supports the permissibility of humanitarian intervention, which may require intervening in support of a particular side, then there exists a prima facie case for arming rebels, albeit very rarely.

Indeed, as with humanitarian intervention, the importance of upholding the principle of state sovereignty, as traditionally conceived, does not provide sufficient reason to preclude the permissibility of arming rebels in all situations. One of the central, and largely uncontroversial, claims of the responsibility to protect (RtoP) doctrine is that state sovereignty implies responsibility for its population; if a state is manifestly failing to protect its population from mass atrocities, then the international community has a remedial responsibility to protect this population. As a corollary, humanitarian intervention is widely thought to be sometimes morally permissible—and even sometimes morally obligatory—because of the potential for intervention to do more good than harm, despite the various risks that humanitarian intervention poses to international stability and its seeming contravention of the principle of nonintervention. The same is true of arming rebels. When understood as a responsibility, sovereignty is not a bar to arming rebels if the state in question is manifestly failing to protect its population from mass atrocities. In such cases, worries about the effects of the violation of the principle of nonintervention may be outweighed. Indeed, arming rebels may not lead to much international instability. It may in fact improve stability by helping to put in place a more stable regime, or it may enable the rebels to achieve sufficient good (that is, by tackling the mass violation of basic human rights), thus outweighing any harms.

Again, as with humanitarian intervention, there will need to be various conditions on the permissibility of arming rebels. I will not flesh out all the conditions here, but they are similar to the well-rehearsed just war criteria found for engaging in humanitarian intervention. They include, most notably, just cause (rebels must be supplied arms only in response to situations of the ongoing or impending mass violation of basic human rights) and proportionality (arming rebels must do more good than harm). Indeed, we have already seen that arming rebels is likely to be disproportionate.

Notwithstanding, two nonconsequentialist desiderata are worth highlighting, given the emphasis I have already given to considerations related to proportionality. Although my account gives significant weight to the import of achieving good consequences in assessing the justifiability of arming rebels, it is not consequentialist. ³³ First, it matters that those supplying the arms have predominantly the right motive in their overall balance of reasons for supplying arms. For example, those arming should do so in order to end mass human rights violations, rather than, for instance, to make a profit. This is even the case if insisting on having a predominant right motive means that those supplying arms do not maximize the good. This claim that motives matter is based on the Kantian notion that we should be motivated by the right sort of reasons for our actions to have moral worth. And it applies even when arming rebels with a predominantly wrongful motive would otherwise achieve beneficial consequences, such as overthrowing a brutal dictator and establishing democracy.

Second, arming rebels should (generally) be the last feasible option (apart from direct military intervention). The import of the last resort principle is not simply consequentialist, that is, it does not simply compare, from an agent-neutral perspective, the various goods and bads involved with war, arming rebels, and the other options (for example, as in a principle of necessity). Its import also reflects an agent-relative consideration—the difference between doing and allowing harm. That is, agents should avoid doing harm themselves. Potentially violent options—such as war and arming rebels—are more likely (if not always) to involve the doing of harm than other, less violent, options. ³⁴

The Just Rebellion View and Its Exceptions

While it might be thought that rebels must be fighting a just war in order for it to be permissible to assist them, there may be exceptions when it is permissible to arm unjust rebels. To understand this, it helps to consider two questions. First, is the use of force by the rebels justifiable? Second, is it permissible to arm them? Note that the answer to the second question does not depend on the answer of the first. On the one hand, although the rebels might be fighting a just war, supplying arms to these rebels may not meet the relevant conditions for this supply to be permissible. As already discussed, arming rebels is likely to be disproportionate, even if the rebels are fighting a just war, and so is only exceptionally permissible. On the other hand, although the rebels may not meet requisite conditions for their war to be just, arming them may still be permissible. This contrasts with what I will call the Just Rebellion View, ³⁵ which holds that for it to be permissible to arm rebels they must be fighting a just war—one that meets the conditions of jus ad bellum and jus in bello. If the rebels are fighting an unjust war, then the state supplying the arms acts wrongly. In what follows, I will reject the Just Rebellion View. First, as has been noted in some of the recent literature on just war theory, wars are subject to different phases and may involve different causes, ³⁶ and the same is true of wars fought by rebels. Although a rebellion may start out as a response to a just cause, such as remedying gross violations of human rights, it may later come to support an unjust cause, such as securing access to valuable natural resources. For instance, it might have been permissible to arm the more moderate rebels fighting for freedom in Syria in the early stages of the conflict, before it morphed into a sectarian war with some of these groups changing aims and others being subsumed into groups fighting for more problematic causes. (The opposite, where fighters start out with an unjust cause but change to a just one, is less likely.) ³⁷

In addition to the possibility of there being different phases of a rebellion, there could also be different elements of it. For instance, a civil war may be unjust overall because its main aim is to overthrow a democratically elected leader and impose an authoritarian ruler, but it may have tangential elements that are justified, such as fighting for the greater redistribution of resources in particular areas. The arming of rebels in support of the just element may be permissible, even if the overall war is unjust. For instance, it may be permissible to arm rebels who are justifiably protecting a refugee camp from attack by government forces, even if the main war waged by the rebels is unjust.

Moreover, it may be permissible to arm rebels fighting an unjust war throughout. ³⁸ Perhaps most notably, arming rebels may strengthen them and result in the conflict becoming a stalemate. The provision of arms, most notably major conventional weapons, by an external party, could significantly augment the strength of a rebel group, allowing it to engage directly with governmental forces rather than only with individual governmental personnel on a more ad hoc basis. ³⁹ And, in general, when rebels are strong, conflicts tend to be shorter because governments have incentives to compromise. ⁴⁰ Both sides may give up the hope of victory and negotiate a settlement that ends the brutal conflict. By contrast, when rebel forces are weak, governments are less willing to negotiate both because they do not want to encourage other insurrections and because living with a weak rebel force may be less costly than making concessions or trying to eradicate them. ⁴¹ Indeed, equalizing the sides was one of the rationales offered in the public debate about arming rebels in Syria and thus forcing Assad to negotiate. ⁴²

Accordingly, even though one may arm unjust rebels, this is not necessarily in order to assist them to achieve their (unjust) aims. Rather, the goal may be to frustrate the opposing side. For example, the rebels may have little or no hope of victory; indeed, the state supplying arms would not be aiming to help the rebels secure victory or realize their political vision. On the contrary, the final peace agreement may be very different from that advocated by any of the belligerents. Hence, what matters is not the justifiability of the rebels’ war, but rather the justifiability of the state's arming of the rebels, which will be determined in part by the foreseeable consequences of providing such arms. Of course, it is likely that the foreseeable consequences of arming just rebels will generally be much better than arming unjust rebels.

While the justice of the rebels’ war is not a necessary condition for arming rebels, it is important to emphasize again that this is only in exceptional cases, ⁴³ and that potential benefits are often unlikely to be realized. For instance, if one accepts Mary Kaldor's “new wars” thesis, victory is not always sought by belligerents since they may profit from the conflict itself. ⁴⁴ On this view, belligerents are not striving to defeat a clear opponent, but fight for a variety of complex—and seemingly contradictory—political and economic reasons. Therefore, the acceptance of a stalemate by such belligerents will often be unlikely. Moreover, even when stalemate is likely, the probability of arming (unjust) rebels leading to a stalemate that could do enough good—such as securing peace and ending the mass violation of basic human rights—to outweigh the harms noted in the first part of the article seems small. Although there may be exceptions, in most cases the conflict can be expected to last longer as the rebels’ incentives to back down are reduced.

Concluding Thoughts: Exceptions And Just War Theorizing

Why then note the exception? It might be argued that when examining the ethics of war and political philosophy more generally, focusing on exceptional permissibility of a practice, such as the (unjust) arming of rebels, is problematic in that it is unduly conjectural. The worry is that it involves pontificating about hypothetical or extraordinary cases, producing little moral guidance that can be used in the real world. ⁴⁵ In reply to such objections, I believe there can be significant practical merit in considering feasible exceptions. First, we need to know if the practice under consideration, such as arming rebels, is always impermissible so that, when the exception applies, states and other actors can permissibly arm (even unjust) rebels, and those who do so are not subject to wrongful opprobrium and other costs by the international community.

Second, considering exceptional cases helps us determine why exactly the arming of rebels is morally impermissible (when it is impermissible). We have seen that the impermissibility of arming rebels is not determined simply by the justifiability of the rebels’ war. Rather, I have suggested that arming rebels will generally be wrong because of the foreseeable consequences of doing so (the Rebel-Risk Objection, the Escalation Objection, and the Diffusion Objection). This can sharpen the critique of actors that wrongly arm rebels and help to anticipate potential responses by them. For instance, rather than focusing on whether the rebels are fighting a just war, the critique can focus on the likely problematic consequences of arming rebels.

Third, considering exceptions can be important argumentatively, that is, in order to persuade that there is a general case for a presumption against arming rebels. It can be expected that, against a general presumption, some will emphasize that arming rebels should still be a policy option because it is conceivably permissible. Rather than simply ignoring or denying the argument that there are feasible exceptions, a better strategy would be to acknowledge their point, but to highlight its exceptional nature and to show that there should be a general presumption against arming rebels.

Having defended the case for considering exceptions, however, I now want to suggest that focusing only on exceptions is problematic. This is the case for much recent work in just war theory, which adopts a highly idealized, abstract, and reductionist framing, often based on an analogy to cases of individual self-defense. Such accounts focus on necessary features, exceptional cases, and/or unduly idealized situations. In doing so, they overlook the general, contingent (that is, nonnecessary), but still very morally significant nonideal considerations related to war and conflict, such as epistemic shortfalls due to the fog of war, noncompliance with the ideal rules governing war, the likely abuse of ideal norms, and other unintended consequences (such as potential escalation and arms diffusion).

Why is this a problem? To start with, if we want the ethics of war and peace—including just war theory—to be politically relevant, and ultimately to influence decision-makers in the right way, we need an account of the general, nonideal case, as well as an account of the exceptional cases. That is, we need a fully thought-through assessment based on an account of the general permissibility of the practice, such as of arming rebels. For instance, if we focus only on exceptions and the alleged necessary features of the wrongness of arming rebels, our conclusion might be that arming rebels is not morally problematic, given that it is exceptionally permissible. But this would be to misjudge massively the general moral justifiability of arming rebels and would give a radically incomplete and potentially dangerous assessment of the ethics of arming rebels.

In addition, policymakers may wrongfully, and perhaps mendaciously, draw on the case for exceptional permissibility to legitimize their impermissible actions. For instance, the occasional permissibility of arming rebels may be used to legitimize the general policy of doing so. ⁴⁶ To make this harder, the general case should be clearly spelled out. Accordingly, too much of an emphasis on the exceptions, if not heavily (and very explicitly) caveated as incomplete or accompanied by an analysis of the general assessment, could have the (unintended) consequence of giving the impression that the practice in question is typically permissible. Thus, focusing only on exceptional cases offers incomplete guidance and can be more easily open to misinterpretation or abuse. ⁴⁷

Thus, despite establishing the exceptional permissibility of arming rebels, it is important to highlight that this analysis is incomplete. As the first part of this article makes clear, once we have a general assessment of the practice, arming rebels should generally be eschewed.