Complicity and ‘stay or go’ questions

Debates around complicity can be fraught for INGOs because they arise in situations where the complicit action or omission is taken in pursuit of a substantive good—and any chance of that good being realised will be lost if the complicit act, or omission, is foregone. The wrongs arising from the complicit action are in tension with the goods arising from it.

It is worth setting out how arguments from complicity are usually constructed. The argument is typically of the following kind. A principal agent, such as a state, is directly involved in wrongdoing. A second agent, such as an INGO, intervenes, ordinarily to assist those subject to the principal agent’s wrongdoing. Despite the intentions of the INGO, and their ordinary prudential actions aimed at delivering benefits, they make some causal contribution to the principal agent’s wrongdoing. A typical example is where an INGO’s humanitarian resources are co-opted by an armed agent. The INGO starts helping the guilty.

Ordinarily, those who make accusations of complicity seek to do several things. First, they identify that the INGO is—inadvertently—making a causal contribution to wrongdoing: in the scenario above it is the INGO’s resources that contribute. Second, because of the causal contribution, they ascribe some degree of moral responsibility to the complicit agent—they are at least partly guilty of the crime and their actions therefore fall under moral judgement; and finally, the charge of complicity is usually presumed sufficient, by itself, for the INGO to withdraw. Arguments based in complicity are frequently regarded as moral trump cards. In the discussion that follows, it is with step 2 and particularly step 3 that we are concerned.

What do we mean by complicity?

Complicity emerges from ‘accomplice’. Legally, it refers to joint enterprise or common purpose: an accomplice shares the goals, and the intentions—the mens rea—of the primary perpetrator. In this sense, accusations of complicity levelled at INGOs are absurd. It is difficult to take seriously the idea that INGOs share the intentions of the perpetrators of wrongdoing—ordinarily the wrongdoing that triggered the INGO’s intervention. We set this version of complicity aside.

More recently, a more nuanced understanding of complicity has emerged. Lepora and Goodin describe a sliding scale of degrees of complicity, with joint enterprise at one end and a far more attenuated connection to wrongdoing at the other—so attenuated that decision makers may have little or no idea their choices are contributing to the wrongdoing of others.1 In between, they identify a range of what they call complicity’s ‘conceptual cousins’. These include conspiracy, co-operation, collusion, connivance, condoning, consorting and contiguity. They divide this cluster of concepts into three groups: acts involving coprincipals or joint wrongdoing; acts involving contributors, whether intentional or not and those acts, such as connivance, which involve non-contributors, such as physical proximity or ‘contiguity’.

Among these conceptual cousins, the focus of our concern are those acts involving contributors. These acts causally contribute to the wrongdoing of others, ‘without … in any way constituting part of that principal wrongdoing’.2

There are many strengths to Lepora and Goodin’s analysis. They recognise the moral messiness of the contexts in which INGO’s work. They also recognise the enormous variation—and the enormously various moral shades—of inadvertent wrongdoing by INGOs. They identify that INGOs can be involved in harms they neither intend nor foresee and recognise that complicity is sometimes used lazily, as a catch-all term, when what is at a stake is a more nuanced range of moral concerns. They also recognise that complicity may be morally required. They argue that Oskar Schindler’s complicity with the Nazi’s was justified by the good he could do in keeping large numbers of Jewish people from transportation and death.3

Problems arise with Lepora and Goodin’s ascription of moral responsibility for acts involving contributors. In extending the reach of complicity and its conceptual cousins, they risk overextending the scope of moral responsibility. At times, Lepora and Goodin promote a familiar account of moral responsibility involving voluntariness, intention and knowledge of likely outcomes.4 At others, they ascribe moral responsibility to decisions that contribute to the wrongdoing of others in ways the contributor did not intend, had no control over and could not possibly foresee. As French argues, this takes us outside any plausibly defensible idea of moral responsibility and accountability for actions. In doing so, they can make INGOs seem worse than they are and risk a retreat into moral narcissism, with associated problems for decision making.5

Complicity and moral responsibility

As we have seen, accusations of complicity carry a strong charge—have a particularly bad moral odour—and are sometimes presented as trump cards over stay or go decisions. The reasons for this are not transparent and it is worth asking why it might be so. One straightforward (consequentialist) reason may be that complicity suggests that instead of doing good, INGOs are doing harm and should desist. The difficulty is that complicit decisions are seldom of this straightforward kind. Good and bad outcomes are often inextricably linked: the good of the intervention cannot be achieved without the bad. Nor do concerns about complicity seem directly linked to overall sums of good and ill—the charge of complicity does not seem to rest on crudely consequentialist concerns. Another possibility, one that may account for some of the psychological charge of complicity, is that proximity to wrongdoers leads to feelings of compromise and contamination: guilt by association. This is a close ally of what Lepora and Goodin call ‘contiguity’. They highlight two possible sources of wrongness here. The first has to do with character: contiguity should repulse a good person and to remain in close proximity is to raise concerns about character. The second is that those who remain close to wrongdoing are more likely to end up contributing to it. (Interestingly, Lepora and Goodin identify an important problem for INGOs when they recognise that contiguity can at times be interpreted as implicit approval. Many INGOs ask themselves whether their presence by itself can add legitimacy to a regime perpetrating wrongdoing, and therefore unwillingly contribute to it.)6 Another plausible contribution to the malodour attached to complicity is its capacity to confuse the scope of individual moral responsibility—that some responsibility for the most serious wrongdoing is transferred to humanitarian agents. We now turn to questions of moral responsibility.

We have seen that a complicit action is one in which a secondary agent, the complicitor, makes a causal contribution to a wrongdoing perpetrated by another. As Peter French has convincingly demonstrated, causal contribution is not, by itself, sufficient to accrue moral responsibility.5 A simple example makes this clear. Consider an INGO deciding whether to provide significant humanitarian aid to a large population displaced into a neighbouring country by war. Following a needs and risk assessment, the INGO decides at a specific time (Tx) to commit significant resources. Sometime later (Ty), a warlord invades the neighbouring country and makes off with those resources, thus enhancing his ability to wage the war that displaced the population. Although at (Ty) the INGO has now made a causal contribution to the conflict, it is absurd to suggest that an ordinarily prudent INGO was morally responsible at (Tx). Causal contribution is not synonymous with moral responsibility.

More importantly for our purposes, if the INGO makes a similar decision at (Ty), knowing that there is now a significant likelihood that its resources might be confiscated for nefarious purposes, the scope of its moral accountability shifts. A decision made knowing that there is a reasonable likelihood that it will causally contribute to wrongdoing does accrue some moral responsibility. This is probably closer to reality for most INGOs when making decisions in circumstances that give rise to accusations of complicity. Several questions remain though: how much moral responsibility, and for what?

Without wanting to enter a long discussion concerning the conditions for ascribing moral responsibility, it is reasonable to suggest that someone who kills, and someone who makes an unintended causal contribution to that killing are not equally morally responsible. The one who intends and carries out the killing is guilty of the crime. At (Ty), our INGO does not take on responsibility for the waging of the war, or anything like it. It neither intends nor perpetrates the primary wrongdoing. But given that it is within the scope of the INGO’s agency to pull its resources, and cease to contribute, it seems fair to ascribe some form of moral responsibility. But for what? If, at (Ty), the INGO decided to continue contributing, but the warlord is deposed and peace ensues, the INGO is no longer contributing and ceases to be complicit. As French argues, there is no plausible way that the INGO, as a complicitor, can predict an outcome that it neither intends nor works towards—the outcome is too contingent on other actors. What it must accept, if it continues to supply resources, is responsibility for a decision at (Ty) that involved a risk that its resources would be co-opted: a lesser evil by some degree. Given that these decisions are ordinarily made in pursuit of some other good, moral responsibility for the principal wrongdoing seems more attenuated still.

When thinking about complicity in relation to inadvertent wrongdoing by INGOs, as a first step we need to distinguish those consequences for which moral responsibility can reasonably be imputed to the INGO. As more information about unintended consequences emerges, our moral responsibility can change. If, despite our best intentions, we have a reasonable belief that our decisions will have undesirable consequences, then they will fall to some degree within the scope of our moral accountability. Our retrospective assessment at (Ty) of our decision at (Tx), is different to our assessment of likely prospective moral responsibility for the consequences of the decision we make at (Ty). We do not however take moral responsibility for the primary wrongdoing that remains with the perpetrator.

Motivation, humanitarian agents and moral narcissism

The fact that we may have some moral responsibility for the ills arising from our contribution to another’s wrongdoing—that we are complicit—does not bring our decision making to a close. It is not decisive. As Lepora and Goodin make clear, in some circumstances the complicit decision may, as in the case of Schindler, be morally preferable. There will be times when we decide that, despite an action being complicit, it may, nonetheless, based on an ‘all things considered’ judgement, be the right or necessary choice. Questions of complicity give way to a familiar—although by no means easy—decision-making process, whereby the overall balance of good and ill, based on an assessment of all morally relevant reasons, will need to be made. It is at this point—where we are involved in the difficult decision as to whether to stay or go in morally compromising circumstances—that moral narcissism can threaten judgement.

The academic literature on moral narcissism is not extensive, nor are we aware of its use in academic reflection on humanitarian engagement. Kundnani, writing in a foreign policy context, describes it as “the tendency to think about morality in terms of how your actions make you feel about yourself rather than in terms of their consequences for others”.7 More recently, it has been used to criticise American liberals who regard moral goodness as being dependent on the content of their beliefs rather than the consequences of their choices.8 From our perspective, moral narcissism refers to the tendency of some humanitarian actors to think more in terms of how their actions are perceived, and their image as a ‘good’ person or agent, than the effect of their actions. The fear of being perceived as morally contaminated involves the risk that humanitarian actors will withdraw from circumstances in which they could, on balance, do far more good than harm. Concerns about the moral self-image of the individual or organisation can usurp the benefits that could be provided to those in dire need.

Although the phrase moral narcissism is—deliberately—pejorative, it emerges from a problem first identified by Bernard Williams in his critique of consequentialism—what is sometimes called the integrity objection. Consequentialism is the straightforward view that an act is only right or wrong in so far as its consequences are right or wrong. And a consequence is right or wrong dependent on whether it increases or reduces overall utility, well-being or happiness. It follows therefore that, to be a good person, we are under a moral injunction to live in such a way as to maximise good consequences. Although the consequences for the agent matter morally, they count no more than the consequences for all other individuals whose well-being is affected by the decision. As good consequentialists, we are under an injunction to step out of our own perspective and view the world objectively, ‘from the viewpoint of the universe’. Our decisions should be in accordance with what, from this perspective, will lead to the highest overall satisfaction. Williams’ main objection is that this misconceives what we are like as moral agents:

The point is that (the agent) is identified with his actions as flowing from projects or attitudes which… he takes seriously at the deepest level, as what his life is about… It is absurd to demand of such a man, when the sums come in from the utility network which the projects of others have in part determined, that he should just step aside from his own project and decision and acknowledge the decision which utilitarian calculation requires.9

For Williams, moral agents have good reasons to seek to realise their (good) personal projects and, at times, their commitment to these projects will be in tension with the demands of utility. There can be something incompatible between the demands of personal integrity—our virtuous desire to realise our self in the world—and the demands of consequentialism.

Some humanitarians, faced with a conflict between a strongly felt personal injunction of a ‘do no harm’ type, and a decision based on utilitarian calculus that would involve them in identifiable wrongdoing, may understandably seek to prioritise the former. Some may also be troubled by the presence of conflicting—and possibly incommensurable—moral reasons, both of which seem to have a claim on them and that cannot be reconciled by reference to an external standard.

Although we are wary of generalising about motives, it is plausible to suggest that humanitarian work attracts (some of) those strongly motivated to do good and who are therefore likely to be particularly troubled by conflicts of these kinds. We recognise the importance of these issues, and in what follows, do not wish to trivialise them. We seek instead to identify a problem for INGOs arising from these contradictory demands. Williams recognises the difficulty. What happens, he asks, where an agent ‘is forced with a reliable choice between a detestable action and an outcome which will be utilitarianly worse?’10 And if I decline the detestable action, am I “open to the charge of being concerned with (my) own integrity or purity or virtue at others’ expense”?11 Am I guilty of what Williams calls moral self-indulgence? Barry and Fried make a similar point. In criticising the Kantian injunction to refrain from lying, even where lying could save someone’s life, Barry and Fried suggest that maintaining personal integrity in such a context ‘becomes a form of narcissism’.12

The point at which an appropriate concern for one’s own moral integrity turns into moral self-indulgence or narcissism is not easy to identify. For Williams, it arrives when motivation shifts—assuming the agent is not a thoroughgoing moral narcissist and no shift is possible—from concern for the interests of others to concern for the preservation of the agent’s self-image.13 It is initially tempting to see this as cousin to an issue in political philosophy, what Michael Walzer has termed the problem of ‘dirty hands’. Put simply, this involves the question of whether political leaders might be called on to violate deep moral prohibitions to bring enormous benefits—or avoid catastrophic harms—to their communities.14 Does the pursuit of some great, even overwhelming good justify malicious means?

Fortunately, few INGOs, and fewer of their staff, face questions on this scale. Dirty hands problems also involve deliberate, eyes-wide-open wrongdoing, whereas complicity here involves non-intended causal contributions to wrongdoing. Rubenstein has recently introduced the helpful concept of ‘spattered hands’.15 Considering ‘stay or go’ type questions, Rubenstein argues that complicity is unhelpful, running the risk that it will ‘characterise INGO’s intentions as worse than they are’, lead INGOs to ‘commit sins of omission (ie, to withdraw)’ and also lead to ‘unhelpful defensiveness on the part of aid practitioners’.16 Complicity, she argues, should be replaced by the clear-eyed, if rueful acknowledgement that INGOs may get drawn into wrongdoing, and that this may be necessary to achieve the goods at which they aim. Accusations of complicity, and fears that their hands may be spattered by the wrongdoing of others, may precipitate an unhelpful retreat into moral narcissism and deflect the proper objects of INGOs concern away from beneficiaries towards self-image.

Complicity in context: Myanmar, Libya and ‘the berm’ refugee camp on the Syria-Jordan border

In this section, we look at three humanitarian interventions that gave rise to concerns about complicity. Although complicity was not determinative in decision making, and was not the only moral issue engaged, we identify where accusations of complicity arose and examine how useful they were in decision making. These decisions were challenging and controversial. Reasonable people took opposing views. The purpose of this section is not to explore the full range of moral questions that arose but to look at those that gave rise to questions of complicity.