2016 Joint Series on International Law And Armed Conflict: Janina Dill On Assessing Proportionality

In the final installment of our Transatlantic Dialogue Series, Janina Dill discusses assessing proportionality. Ms. Dill is an Assistant Professor at the Department of International Relations of the London School of Economics and a Research Associate of the Oxford Institute for Ethics, Law and Armed Conflict at the University of Oxford. The views expressed in this article are the author's own and do not necessarily reflect those of the ICRC or the other blogs taking part in this series.

Assessing Proportionality: An Unreasonable Demand on the Reasonable Commander?

Proportionality in International Humanitarian Law (IHL) demands that the attacker weighs incommensurate values: the concrete and direct military advantage anticipated to arise from an attack against the expected incidental harm to civilians and damage to civilian objects. It is common place that for that reason (amongst others) it is difficult to apply Article 51(5)b of the First Additional Protocol to the Geneva Conventions and the corresponding principle of customary law to real world cases (here, here, here, here, here). The legal rule seemingly bends to endorse diametrically opposed interpretations of the same attacks; salient examples include some Israeli air strikes in the 2014 campaign in Gaza (here and here). References to proportionality in the court of public opinion therefore often fan the flames of discord rather than adjudicate between diverging views. In the court of law, specifically in the chambers of the International Criminal Tribunal for the former Yugoslavia, proportionality has largely failed to add to the justiciability of unlawful attacks.

At the same time, proportionality – and indeed the task of comparing seemingly incommensurate values – are not unusual in law. What then is the problem with the principle of proportionality in IHL?

Proportionality according to the reasonable observer

A common approach to assessing an agent’s judgment of excessiveness is to look at it from the point of view of a “reasonable observer”. However, an empirical investigation of attitudes towards collateral damage yields anything but a concretization of what proportionate incidental harm looks like. When asked to put themselves in the place of a commander partaking in a mission to clear an Afghan village of Taliban fighters, 27% of British respondents and 20% of American participants in a survey I conducted in 2015 said they would not accept any foreseen civilian deaths as a side-effect of an attack meant to kill a group of Taliban fighters. At the same time, 17% of British and 21% of American respondents said they would accept however many casualties the attack would cause. 44% and 41% of the populations respectively hence rejected the very premise of proportionality in war: the prospect of a military advantage warrants a positive, but limited number of unintended, yet foreseen civilian casualties.

Even more striking than this division of over 40% of respondents into either no or unlimited civilian casualty tolerance is the fact that a roughly equal share of respondents refused to answer the question or said they were not sure how many civilian deaths they would accept. Only 13% and 14% of the survey participants were ready to specify a maximum number of incidental casualties that they would allow in pursuit of the military advantage. That share fell to 9% when respondents were asked how many deaths they thought were a morally acceptable side-effect of the same attack. The range of numbers they suggested was widely distributed between 1 and 10,000 foreseen civilian casualties. In other words, those respondents who did entertain the possibility that the prospect of a military advantage could be meaningfully connected to a limited, but greater than zero number of expected civilian casualties, either refused to specify (43% and 45%) or radically diverged (13%) in their conception of how much was too much incidental harm.

These numbers naturally depend on the scenario in the survey. For instance, if an attack were framed as crucial in a belligerent’s quest for victory, the share of people who reject any foreseen civilian deaths might be smaller, and possibly even more respondents might accept unlimited incidental harm. Of course, few individual attacks in real wars are ever truly pivotal or without alternative. At the same time, there is no reason to believe that the prospect of a greater or smaller anticipated military advantage would make people more confident in judging the magnitude of acceptable civilian harm. It seems observers do not so much agree or disagree on when incidental civilian harm is excessive in relation to a military advantage, but they converge on the rejection of the notion that there is a knowable “balance” between human life and military gain with neither value simply overriding the other.

Proportionality according to the reasonable commander

Does this finding not suggest that it is a mistake to consider the attitudes of lay people and that the observer whose views should be standard-setting is the reasonable military commander? I have argued elsewhere that for political, strategic and moral reasons we should care about what civilians think. But naturally the views of military experts also matter greatly if we want to move past our diagnosis of “incommensurate values”, to figure out what exactly is wrong with the principle of proportionality in war and how we may measure it. Interviews I conducted over the last eight years with current and former members of three state militaries (the Armed Forces of the United States, the Afghan National Army and the Israel Defence Forces (IDF)) have, however, led to the inescapable conclusion that even individuals with military expertise have widely diverging reactions to most attacks that cause some incidental harm and yield some military advantage. Emerging research relying on surveys has reported similar findings for legal experts.

Most military commanders I have encountered over the years admitted that in other than atypical “easy cases,” two commanders with the same knowledge of fact and of law may reach diverging conclusions about the projected excessiveness of an attack. Although some interviewees found this lack of the principle’s action guidance problematic, others considered the resulting flexibility and reduced justiciability proportionality’s greatest assets. A third and more surprising narrative centered on the denial that the rule even required balancing military gain and loss of civilian life. Instead, some commanders argued that proportionality demanded that the attacker made sure they had chosen a truly valuable target and did everything possible to minimize expected incidental harm. Adequacy (that the attack will likely produce a military advantage) and necessity (that it is the mildest means to achieve this particular advantage) are indeed conditions of an attack’s legality. Of course, they are not sufficient to make an attack proportionate under international humanitarian law.

The collapsing of proportionality into necessity or the occasional denial of the principle’s importance in the targeting process tend to go hand in hand with a strong emphasis on the multiple stages of target verification and the processes of weaponeering and mitigation followed in branches of the US military and the IDF. These procedures (corresponding to the obligation to take precautions in attack) clearly work to make expected incidental harm adequate and necessary. But does conformity with procedural obligations ensure that collateral damage is also proportionate strictu sensu to the military advantage?

Proportionality and procedural measures

A common institutionalization of proportionality judgements I encountered is a system whereby, depending on the collateral damage estimate, the authorization of an attack is pushed further up the chain of command. Of course, this displaces rather than solves the problem of how to weigh loss of civilian life against progress towards military victory. In some contexts, armed forces in addition use variations on a matrix according to which incidental harm and military advantage are independently sorted into scales (for instance, small/medium/large). Which box is ticked may likewise determine the level of authorization required for an attack. Such a matrix is certainly a valuable addition. Senior decision-makers called upon to approve an attack may be put on notice more effectively if the reason for their involvement is not simply the crossing of a collateral damage threshold, but a potentially problematic disconnect between the expected incidental harm and the military value of a proposed attack.

But does the procedural response of involving more senior personnel solve or even just lessen the problem that reasonable observers with and without military expertise have divergent or no intuitions about how to balance loss of civilian life and military progress? It is often argued that commanders in the higher echelons have access to additional intelligence and generally a better understanding of an attack’s strategic and political context. While this enhances their grasp of the military advantage and possibly of the implications of incidental harm, it leaves unaffected their ability to weigh them. The abilities to judge the soundness of a collateral damage estimate, to gauge the operational significance of a military advantage, to assess the feasibility of an additional step of target verification or yet another change in the angle of attack are all plausibly improved by additional military expertise and combat experience. But after all these judgements have been made proportionality has yet to be established.

The conceptual problem with proportionality

If not military seniority, what would make an individual systematically better at determining whether the death of another is a valid means to a military end? To answer this question, we need to return to the puzzle of why proportionality in war seems to be resistant to the usual ways of concretization. An alternative to taking the view of the reasonable observer in law is to convert ostensibly dissimilar values into a common metric, for instance monetary value. The more appropriate metric to measure loss of life and injury to the human person is moral value/disvalue. Can we determine the moral value of a military advantage in order to judge how many civilian lives lost it “outweighs”?

If a belligerent is fighting for a morally just cause, a military advantage may be valuable in that it contributes to the achievement of a moral good. Yet, IHL envisages that an individual fighting for an unjust aggression also makes proportionality judgements. In those cases, when measured by reference to moral value/disvalue, the balance between loss of life and military gain is not merely hard to find, metaphysically speaking there is no such thing. As I have argued elsewhere, this does not mean that we should therefore question the independence of the legal permissibility of conduct in war from questions of resort or indeed mess with the equal treatment of belligerents under IHL. But maybe we can admit that in most wars, at least on one side, the notion that the inevitable loss of civilian life is “outweighed” by military advantage is a fiction as both incidental civilian harm and military advantage are not prima facie morally valuable.

Even if military advantage can in principle be expressed in terms of progress towards a moral good, making a proportionality judgement then requires that an attacker gauges the moral import of her ultimate war aims, makes multiple future oriented judgements about how a particular attack furthers their achievement, and assesses the wrongness of harming civilians with presumably varying moral status. Every single one of these steps poses an enormous, possibly prohibitive epistemic challenge. In this light, it appears eminently reasonable that 87% of lay observers in one way or another reject identifying the number of civilian lives a military advantage is “worth” taking. Is it reasonable that we expect military commanders to nonetheless try?


Schedule of blog posts:

The joint blog series arising from the workshop follows on from our collaboration in hosting a similar series last year (see here, here and here). The Transatlantic Workshop is organized and sponsored by the Oxford Institute for Ethics, Law and Armed Conflict, the Oxford Martin Programme on Human Rights for Future Generations (both directed by Dapo Akande), the Washington DC & London delegations of  International Committee of the Red Cross, the Houston College of Law (through the good offices of Geoff Corn), and the Robert S. Strauss Center for International Security and Law at the University of Texas (directed by Bobby Chesney).

Click here for the 2015 Series.

Click here for the 2014 Series. 

For ICRC's Customary IHL Database on Proportionality in Attack, click here.