Intervention in foreign conflicts is, undoubtedly, one of the most urgent and controversial issues in modern politics. After NATO interventions in Bosnia and Herzegovina and Kosovo during the 1990s, which were widely regarded as reasonably successful, states were comparatively receptive to the idea of using military force abroad for humanitarian ends. There followed a flurry of academic literature on the normative implications of intervention, with two noticeable features. First, the literature is predominantly legal in character, focusing on how international law should respond to widespread human rights abuses. Even now, comparatively little philosophical work has been carried out on the morality of intervention, in contrast to the significant amount of work on just causes for war more generally.^1-5-^1-51 Fabre, Cécile. 2011. Cosmopolitan War. Oxford: OUP.
2 McMahan, Jeff. 2010. “Humanitarian Intervention, Consent, and Proportionality.” In Ethics and Humanity: Themes from the Philosophy of Jonathan Glover, edited by N. Ann Davis, Richard Keshen, and Jeff McMahan. New York: Oxford University Press.
3 Oberman, Kieran. 2015. “The Myth of the Optional War.” Philosophy and Public Affairs 43(4): 255–86.
4 Parry, Jonathan. 2017. “Defensive Harm, Consent, and Intervention.” Philosophy and Public Affairs 45(4): 356–96.
5 Renzo, Massimo. 2020. “Revolution and Intervention.“ Noûs 54: 1233–53.
Second, this literature almost exclusively addresses the permissibility of direct humanitarian intervention – that is, of states' deploying their own armed forces abroad, without the consent of the target state, in order to protect foreign citizens from harm.^6-9-^6-96 Buchanan, Allen. 2013. “The Ethics of Revolution and Its Implications for the Ethics of Intervention.“ Philosophy and Public Affairs 41(4): 291–332.
7 Fabre, Cécile. 2018. Economic Statecraft. Cambridge, MA: Harvard University Press.
8 Pattison, James. 2015. “The Ethics of Arming Rebels.” Ethics and International Affairs 29(4): 455–71.
9 Renzo, Massimo. 2018. “Helping the Rebels.” Journal of Ethics and Social Philosophy 13(3): 222–39.

In more recent years, political debates about how to respond to humanitarian crises abroad have shifted away from direct intervention. In the aftermath of disastrous wars in Iraq and Afghanistan, governments are increasingly reluctant to deploy their armed forces to fight in foreign domestic conflicts, and there is little public appetite for large-scale interventionist wars. Limited NATO intervention in Libya in 2011 was followed by a lack of direct Western intervention in, for example, Syria and Yemen, despite the undoubted gravity of the harms facing the Syrian and Yemeni populations.

Instead, several countries are increasingly advocating and employing indirect forms of intervention, such as funding, arming, or training foreign rebel groups. France, Turkey, the United Kingdom, and the United States have armed and trained rebels in Syria, for example. Of course, there is nothing new about states indirectly intervening in foreign conflicts. But, historically, such intervention has been conducted largely covertly, and primarily or solely as a means of furthering the interveners' own political interests. Now, in contrast, indirect intervention is often carried out overtly and presented as a laudable means of discharging one's duties to aid without incurring either the physical or moral risks of waging war. But indirectly contributing to war is not without moral risks, as a growing body of literature in the ethics of war attests. And, of course, not all indirect contributions to foreign conflicts fall under the description of aiding foreign citizens. On the contrary: governments sell, or facilitate the selling of, weapons and equipment to authoritarian states that use those weapons to harm their citizens. Just as we ought to be concerned about vaunted attempts to influence foreign conflicts by supporting rebels, we ought to be concerned about third parties' rather less publicised roles in suppressing resistance abroad.

In ‘Selling Arms and Expressing Harm’, James Christensen argues that we should reject what he calls the inconsequence argument about arms trading. According to this argument, it is permissible to sell arms to oppressive governments – or ‘outlaw states’, as he calls them – because the victims of such governments are thereby made no worse off than they otherwise would be, given that those oppressive governments would just source arms elsewhere.¹⁰10 Christensen, James. 2022. “Selling Arms and Expressing Harm.“ Journal of Applied Philosophy 39(1): 6–22.
Because the harm is substitutive, it is claimed to be morally irrelevant. Christensen's goal is to undermine the inconsequence argument on grounds amenable to consequentialists. Importantly, his consequentialist-amenable response draws inspiration from a deontological reason to reject the argument – namely, that selling arms to an outlaw state disrespects the outlaw state's victims. Christensen argues that although many common ways of invoking this argument in both its prospective and retrospective forms fail, the prospective form is more promising. On this form, it matters what evidence arms exporters have when they make the sale and what procedure they go through to come to their decision to make the sale. Indeed, Christensen's main argument focuses on cases in which ‘arms sellers can reasonably infer from the available evidence that an arms transfer will not increase the harms that an outlaw state is capable of inflicting (relative to a scenario in which that transfer is not made)’.¹¹11 Christensen op. cit., p. 13.

Arms sales can be either conditional or unconditional. Conditional sales are sales enacted on the basis that they will not make a difference to the outlaw state's victims. Unconditional sales are those sales enacted regardless of whether they will make such a difference. The difference between these sales consists in the difference in exporters' intentions. For some consequentialists, intentions are irrelevant to permissibility. Christensen focuses his argument on perceived intentions. He distinguishes between an act's subject-centred meaning and its object-centred meaning. The former is the meaning as determined by the intentions of the agent. The latter is ‘is determined by how the subject's intentions are perceived by the person acted upon (the object)’.¹²12 Christensen op. cit., p. 15.
These can coincide. A person might intend to be rude and be perceived to be rude. But they can also come apart. A person might intend to be rude but not be perceived as being rude. Or a person might not intend to be rude but be perceived as being rude. According to Christensen, an act that lacks a bad subject-centred meaning can nevertheless be potentially objectionable if it has a bad object-centred meaning, and if it is reasonable to attribute such a meaning to the agent's intentions. For example, the person who does not intend to be rude might nonetheless be acting objectionably if it is reasonable for the object of her actions to perceive her actions as rude.

Such acts inflict what Christensen calls expressed harms, which differ from expressive harms as expressed harms do not need to reflect any attitude actually held by the agent. Because expressed harms do not involve intentions, but do reduce the welfare of their victims, they can be recognised as relevant to permissibility by all consequentialists. Arms sales can thus be thought to inflict expressed harms on citizens of the outlaw regime.

Of course, the proponent of the inconsequence argument could try to expand their argument to cover all harms, including expressed harms, and not just material harms. But, Christensen argues, ‘once it is conceded that arms transfers can cause expressed harms, the inconsequence argument begins to lose purchase … because the concession prompts the realization that relevant expressed harms can be inflicted not only by sales, but also by offers, and by other related acts such as invitations to arms fairs’.¹³13 Christensen op. cit., p. 18.
He then points out that the expressed harms of offers, unlike sales, are additive rather than substitutive: ‘Offers from multiple states can be, and are, made simultaneously’.¹⁴14 Christensen op. cit., p. 19.
Christensen then argues that there will often be good reason for victims of outlaw states to believe that arms exporters who make offers to these states do not care about the victims. While it is not impossible for conditional arms exporters to have their intentions perceived accurately, Christensen contends that it is going to be difficult given the circumstances around and consequences of these sales. Because even arms offers will often have a bad object-centred meaning, even arms offers make a difference. So, we have reason to reject the revised inconsequence argument as well.

In ‘Liability for Wrongful Assistance: On Causing Unjust Harm in the Course of Suboptimal Rescue’, Helen Frowe challenges the claim that assisting rebels to defeat oppressive regimes is a morally safe option that avoids the worries arising from full-fledged direct military intervention. Rebellions typically cause unjust harms, such as the killing of innocent people. Assisting rebels enables these unjust harms. Thus, Frowe argues, providing such assistance can be unjustified. Providing assistance is unjustified when the assistors could have prevented more harm without enabling comparable unjust harm, and without incurring a (significantly greater) supererogatory burden. For example, suppose that a government can, at equal cost to itself, either save a certain number of lives by funding a rebellion, thereby enabling a certain amount of unjust harm, or save more lives by treating a preventable disease. Assume that treating the disease will cause less unjust harm than funding the rebellion. According to Frowe, if a government assists rebels under these circumstances, the government incurs liability to bear costs for the sake of those who are unjustly harmed as a result of their assistance. Frowe argues that assistance that deliberately prevents the lesser harm can be wrongful, and ground liability, regardless of whether it enables others to engage in wrongdoing. On her view, assisting rebels can be unjustified even if the rebels are justified in rebelling – that is, even if, for them, the harming of innocent people is justified as the lesser evil.

An important claim for Frowe's argument, then, is that lesser-evil justifications are ‘indexed to agents’.¹⁵15 Frowe, H. 2022. “Liability for Wrongful Assistance: On Causing Unjust Harm in the Course of Suboptimal Rescue.” Journal of Applied Philosophy 39(1): 23–37.
This is because they depend on the alternatives available to an agent and the cost of taking those alternatives for that agent. Suppose Jane can save five lives only by pulling a switch that will divert a runaway trolley away from the five to where it will kill one. Jane has a lesser-evil justification for killing the one, given the options available to her. But suppose that John, standing at a different switch, can save the five either by diverting towards the one, or by diverting down an empty track, killing nobody. John does not have a lesser-evil justification for killing the one, given the options available to him. Jane's lesser-evil justification does not transfer to John. In the same way, it can be permissible for rebels to fight against their oppressive government, even if they foresee that they will cause unjust harms. Given their options, the rebels might have lesser-evil justifications for inflicting these harms. But this does not entail that third parties are similarly justified in enabling these harms and may assist the rebels. On the contrary: assistors wrong those who are harmed as a result of their assistance, when the assistors could have done something morally superior without thereby incurring any (greater) supererogatory cost.

Frowe identifies two ways in which those who assist in the prevention of a lesser harm, and thereby enable unjust harm, act wrongly. First, she argues, causing such unjust harm violates the necessity constraint on justified harming. A rescuer cannot claim that it is necessary to cause an unjust harm if she can prevent an equal or greater harm without causing comparable unjust harm. Second, agents ought to satisfy all the claims they can as a result of the necessary unjust harms they cause. If a rescuer can either save five lives by killing Kim, or save ten lives by killing Kate, she will wrong Kim if she kills her, given that imposing the same degree of harm on Kate will prevent more harm. Frowe argues that both constraints apply even when preventing either the lesser or greater harm is supererogatory, provided that preventing the greater harm does not impose a significantly greater cost on the rescuer.

Christopher Finlay also explores the ethical implications of assisting rebels in ‘Assisting Rebels Abroad: The Ethics of Violence at the Limits of the Defensive Paradigm’. His central claim is that such indirect interventions give rise to complications not typically present in direct interventions. While direct intervention involves a country using its military to overthrow a foreign government, Finlay focuses on cases where rebels (‘armed, nonstate actors aiming at revolutionary change with potential to secure self-determination through a human-rights-respecting democracy’) are given military assistance, including air support, guidance from military experts, training, arms, and funding.¹⁶16 Finlay, Christopher. 2022. “Assisting Rebels Abroad: The Ethics of Violence at the Limits of the Defensive Paradigm.” Journal of Applied Philosophy 39(1): 38–55.
Finlay explores the ethics of assisting rebels abroad by considering two practical dilemmas for would-be interveners.

The first dilemma for interveners is that they can either fully support the rebels' ends, adopting them as their own, or offer support only as a means of redressing an imbalance of power, in order to give the rebels more scope for self-determination. The first horn of the dilemma might be preferable on some occasions, but there are lots of risks associated with expediting victory. Finlay appeals to different analyses that suggest that fast victories often fail to lead to free, democratic successor states, because fast victories often come without the deeper changes required to turn authoritarian states into democratic ones. The problem, though, is that civil wars can be very destructive, and so it is not clear the second horn is any better. He writes, ‘a sensible theory ought to give due attention to all three ends to which intervention might contribute: first, limiting military destruction; second, helping revolutionaries secure their goals; and third, facilitating political transition through Mill's ‘school’ of struggle (understood […] as a process of defection and conversion) by rebalancing domestic military forces’.¹⁷17 Finlay op. cit., p. 48.
There is a dilemma because it is often unclear which ends should be prioritised. That is, it is often unclear whether interveners should help the rebels defeat the oppressive government quickly or offer only more limited help so that the victory over the oppressive government belongs more to the rebels, thereby increasing the chance of a better long-term outcome.

Finlay's second dilemma concerns rebellions that have rival rebel groups. Because there are often rival rebel leaders, a choice must sometimes be made about whom to support. However, this choice means the interveners may have a creative role in the war rather than just a passive one. In order to perform this creative role, Finlay outlines a series of legitimacy conditions that rebel leaders ought to satisfy to some extent in order to be permissibly supported by interveners. When there is established and legitimate rebel leadership, interveners can avoid an overly creative role. But this isn't always the case. Sometimes there are multiple rebel leaders, sometimes they do not satisfy many of the legitimacy conditions, and sometimes they act wrongly in terrorizing their own people and attacking other rebel groups in order to gain support (and thus violating a key legitimacy condition). This leads to further problems. If interveners wait to see how things play out, then while a strong rebel group may emerge, they may not be ‘the most promising morally’.¹⁸18 Finlay op. cit., p. 51.
Often rebels will go towards the strongest rebel group, even if they do not share their moral values, as happened in Syria. Al Qaeda, through the Jabhat al-Nusra Front, managed to recruit many combatants from more secular and nationalist groups because it was more militarily successful than those groups. This gives interveners reasons to cultivate more morally promising rebel groups early on in a domestic conflict. There is a tension, then, between respecting homegrown rebel leadership and acting such to counterbalance the artificial effects of greater military capacity.

The overall upshot is that indirect intervention does not have a purely defensive justification, nor is it a straightforward matter. With respect to direct humanitarian interventions, there is a clear trade-off. If an intervention will cost more lives than it saves, then it is unjustified. But, according to Finlay, it is less clear how to compare the value of a possible future with democratic governance and the value of the number of lives that might be lost in the fight for that possible future. He concludes that, ‘indirect military intervention seems to demand practical moral and political judgements without the bannisters of determinate numerical figures’.¹⁹19 Finlay op. cit., p. 53.

As the articles in this symposium attest, assisting others to inflict harm in pursuit of political ends raises distinctive moral problems that are overlooked in the broader literature on intervention. Exploring these problems sheds light not only on the ethics of indirect intervention, but also on the nature and scope of our duties to rescue and our duties not to harm.