1. Introduction
The precipitant increase in civilian casualties from organised violence in the twentieth and twenty-first centuries can be traced to a variety of political, social, and technological factors. Among these is the increased prevalence of asymmetrical and non-international armed conflicts. It is a platitude that the laws of war in force today are not well adapted to such conflicts. Although the modern Hague Convention and Regulations,Footnote 1 Geneva Conventions and their protocols,Footnote 2 and other conventional sources of jus in bello Footnote 3 were adopted in the twentieth century, these treaties are essentially designed for the circumstances of late nineteenth century and early twentieth century warfare: two regular armies meeting on a battlefield far removed from civilian populations, or defending cities or towns from which civilians have either fled or are otherwise rarely in the direct line of fire. Only Additional Protocol I to the 1949 Geneva Conventions promulgates rules limiting combat techniques for the protection of civilians in substantial detail,Footnote 4 and its rules do not account adequately for the indirect effects of armed conflict on such facilities as transportation infrastructure, drinking water facilities,Footnote 5 power plants, telecommunications, and other civilian or dual-use property necessary for the survival and health of the resident population. Nor do they include especially practical guidance for asymmetrical warfare with an enemy that hides among, and indeed may disguise itself as, the civilian population – a scenario that has become disturbingly common.
More troubling still – with the notable exceptions of common Article 3 of the 1949 Geneva Conventions and Additional Protocol II to the same – most rules of jus in bello adopted for the protection of civilians do not clearly apply in non-international armed conflicts. The relevant provisions are framed in the vaguest manner and require merely that civilians be treated ‘humanely’ (common Article 3) and afforded ‘general protection’ from the effects of the hostilities (Additional Protocol II).Footnote 6 They superficially prohibit little more than direct targeting of civilians and indiscriminate attacks.
The tragic reality of civilian casualties in modern armed conflicts calls into question the effectiveness of two core legal protections for civilians: the principles of discrimination and proportionality. The lack of interpretive guidance and enforcement precedents with respect to the proportionality principle especially, combined with frequent accusations of its violation,Footnote 7 have fuelled pessimism as to whether it even has meaningful content beyond a general injunction not to ignore the consequences to civilians when making targeting decisions. The devotion of international and military lawyers to the concept of lex lata, far from reassuring critics of the value of proportionality as a restraint on organised violence, merely highlights the dissonance between the prevailing international law of war doctrine and its operational code. Adhering to a conception of proportionality unsupported by state practice does not merely promote a delusive view of international law; it threatens to discredit the law of armed conflicts as a functional mediator between community values and state behaviour.
Consider the relatively bright-line rule of civilian immunity. Throughout the twentieth century, there was no reliable international system to threaten accountability for the ubiquitous dictators, commanders, and soldiers who intentionally targeted civilians. Following the formation of two ad hoc international criminal courtsFootnote 8 and several hybrid national–international criminal tribunals,Footnote 9 civilian immunity received a boost in 2002 from the entry into force of the Rome Statute of the International Criminal Court (ICC).Footnote 10 A formal mechanism is now in place to (potentially) hold accountable individuals responsible for murdering civilians on a sufficient scale. While realistically there is little reason to expect courts of this kind to deter civilian devastation in many armed conflicts,Footnote 11 the clarity of the discrimination rule and the symbolic power of a permanent institution for prosecuting war crimes together proclaim the international community's commitment to the rule of law on this question.
Its commitment to the principle of proportionality is not quite so certain. The principle, first codified in Additional Protocol I of 1977 to the Geneva Conventions, prohibits attacks likely to cause ‘incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated’.Footnote 12 The principle is formulated in intentionally general terms to protect the discretion of military commanders to achieve legitimate strategic and tactical objectives.Footnote 13 The subjectivity of the principle, its vague formulation, and the incommensurability of the values (civilian lives and property versus military advantage) render the principle troublesome in its application.Footnote 14 The standard is so difficult to apply that there is no widely accepted metric for judging state compliance, and there are no well-known precedents in which a state, military commander or soldier was clearly sanctioned for a violation of the principle.Footnote 15 Further confusing the matter is the disjunction between Additional Protocol I and the ICC Statute. Although an attack causing ‘excessive’ anticipated collateral damage violates Additional Protocol I, it may be prosecuted as a war crime before the ICC only if it is ‘clearly excessive’.Footnote 16
Two examples of criminal prosecutions for disproportionate attack that are commonly cited by commentators – the Blaškić case and the Galić case – do not, in fact, relate to Article 51(5)(b) of Additional Protocol I at all. Blaškić concerned three artillery attacks on Vitez and Stari Vitez that resulted in the killing of some one hundred civilians and a smaller number of combatants. The ICTY Trial Chamber found that this ratio of civilian casualties to military advantage was excessive.Footnote 17 Dinstein and others characterise this case as an application of the proportionality principle,Footnote 18 and the tribunal did discuss proportionality at some length in obiter dicta. However, the ruling of the case deals solely with the related but distinguishable rule of discrimination.Footnote 19 The Trial Chamber held that the attack proceeded ‘without distinction’ among adults and children; that the attackers had not targeted combatants specifically with their artillery fire; that the attacks ‘targeted the Muslim civilian population’; and that Vitez and Stari Vitez lacked any ‘strategic or military’ value.Footnote 20 In other words, the attack went beyond disproportionality and violated the principles of necessity and discrimination.Footnote 21
The Appeals Chamber, while accepting that a majority of victims of the attacks were civilians, specifically rejected the conclusion that the civilian casualty figures could be ‘relied on in determining the nature of that attack’Footnote 22 and accordingly exonerated the commander from ordering an illegal attack.Footnote 23 Indeed, the Appeals Chamber's analysis seems to imply the position that an attack targeting civilians would constitute a war crime, but that a wilful attack resulting in more than 50 per cent civilian casualties and ‘lacking sound military judgment’ would not.Footnote 24 If the Blaškić case tells us anything about proportionality, it is to call into question whether a disproportionate attack can ever rise to the level of a war crime unless the attack qualifies as wholly indiscriminate.
In the second example sometimes cited, the Galić case, the ICTY found a series of attacks on random locations in a populated town which resulted in 50 per cent civilian and 50 per cent combatant deaths to be disproportionate or ‘indiscriminate’.Footnote 25 One commentator described this case as a ‘successful proportionality prosecution’.Footnote 26 This, too, is a false positive identification. As the Appeals Chamber observed, the Trial Chamber found the disproportionate effects of the attacks to merely constitute evidence that the defendant was intentionally targeting civilians: ‘No mention is made of indiscriminate or disproportionate attacks as the basis for conviction.’Footnote 27
At the level of treaty obligations, some functional guidance is available. First, there are blanket prohibitions on the manufacture and use of certain highly indiscriminate weapons.Footnote 28 Additional Protocol I also includes some specific operational rules to deter indiscriminate and highly disproportionate attacks. Article 51 prohibits attacks that treat ‘as a single military objective a number of clearly separated and distinct military objectives located in a city, town, village or other area containing a similar concentration of civilians or civilian objects’.Footnote 29 Such attacks are ‘considered indiscriminate’, which, again, differs from proportionality analysis while achieving a similar objective. Article 57 requires military commanders to choose ‘means and methods of attack’ that minimise civilian casualties and dictates that, when confronting multiple military objectives conferring a ‘similar military advantage’, commanders must choose the objective posing the least danger to civilian lives and property.Footnote 30 Attacks otherwise violate the proportionality principle, although Article 57 does not explicitly relate these concepts to ‘excessive’ collateral damage.
These rules are intended to guide military judgments to protect civilians from the worst effects of armed attacks. Unfortunately, they do not apply equally in non-international armed conflicts; instead, Additional Protocol II merely provides: ‘The civilian population and individual civilians shall enjoy general protection against the dangers arising from military operations.’Footnote 31 This amounts to only the slightest enhancement of the principle of discrimination.
In the absence of conventional sources of more detailed guidance for the interpretation of the proportionality principle, and of any guidance whatsoever on the question of whether and how the proportionality norm applies in non-international armed conflicts, scholars,Footnote 32 international tribunals,Footnote 33 the International Committee of the Red Cross (ICRC),Footnote 34 and even some statesFootnote 35 have turned to customary international law. Customary international law embodies the long-standing, widespread, and consistent practice of states undertaken from an expectation of legally binding obligation (opinio juris sive necessitatis). Those looking for guidance on the application of the principle of proportionality must study state practice with a view to identifying consistent patterns of behaviour, characterised by those marks of respect for the expectations of the international community that are sufficiently unambiguous to establish recognition of a legal rule. Producing decisive evidence of a customary rule of international law is notoriously tricky and laborious, and all the more so in the study of the customary law of armed conflict. Studies of international custom have so far proven impervious not only to all attempts to identify the content of the rule of proportionality in any detail, but to assess the status of the principle itself as a binding custom in both international and non-international armed conflicts.
If proportionality is to play a role as a standard of behaviour rather than an abstract ideal then international lawyers, military commanders, and other relevant actors need better metrics of how states interpret the rule of proportionality to accurately assess the content of custom as well as the incidence and methods of state compliance. The purpose of this article is to analyse the challenges in developing a methodology to study these questions, and to describe one methodology designed to study empirically state practice of proportionality. In the course of the analysis, the limitations of any private research methodology will become clear, and ideas for making proportionality analysis more approachable will be explored.
In Section 2 of this article, the practical obstacles to understanding how (and whether) states interpret and apply the proportionality principle in practice are discussed. I then analyse the strengths and weaknesses of some potential methods for conducting empirical research on proportionality, followed by a description of the method settled upon by the author and colleagues for conducting a long-term, multinational empirical study of state practice of proportionality.
In Section 3, the limitations of empirical research will be discussed, and some proposals for overcoming the obstacles to understanding how states implement proportionality will be explored. Specifically, in Section 3, I discuss the merits of adopting a transparency norm with respect to state implementation of proportionality obligations, including training, monitoring and enforcement measures. In Section 3, I then propose a new legal duty for the gathering and dissemination of information on post-engagement civilian and combatant casualties and property damage, and discuss the kinds of disclosure obligation most likely to be acceptable to states and their military organisations.
2. Empirically Assessing Proportionality in Operation
2.1 Obstacles to Recognition of the Proportionality Principle as International Custom and to Assessment of its Content
Inferring the existence and content of the principle of proportionality in customary international law is obstructed by several factors. First, until relatively recently, states did not consider themselves legally bound to afford significant protection to the civilian population under international law. The unnecessary civilian deaths caused by carpet bombing, firebombing, and the use of indiscriminate weapons such as toxic gases, landmines, drifting contact sea mines, unguided rockets, and atomic bombs in the Second World WarFootnote 36 – not to mention the intentional murder of millions of civilians – demonstrated a general ruthlessness and disregard for the safety of civilians during that period. The establishment of a long-standing, widespread, and consistent custom of taking special precautions to protect civilians from the effects of hostilities, if it has arisen, must have done so in the last 60 years or so.Footnote 37 Yet, as already noted,Footnote 38 the ratio of civilian to combatant casualties does not seem to have abated since the Second World War. Civilian casualties continue to overshadow military casualties in most armed conflicts globally. The facts do not seem to support, at least superficially, the many assertionsFootnote 39 that the principle of proportionality has yet crystallised into customary international law.
Second, military planning and operations are generally opaque. The planning processes in which decisions are made about targeting and attack methods, in particular, are highly secretive. To expose the decision process to public view could enable current or future adversaries to predict the military organisation's strategy and tactics, undermining its effectiveness and exposing its personnel to danger. For similar reasons, military discipline enforced against the organisation's own commanders and soldiers is rarely conducted with total transparency (although several military organisations do conduct public trialsFootnote 40 or at least publish the judgments). Nor are military organisations usually eager to allow the media free access to their battlefield operations. As a consequence, experienced students of the laws of armed conflict are quick to acknowledge the nearly insuperable difficulty of assessing state practice and opinio juris based on actual battlefield events.Footnote 41 The ICTY itself observed in the Tadić case:Footnote 42
When attempting to ascertain State practice with a view to establishing the existence of a customary rule or a general principle, it is difficult, if not impossible, to pinpoint the actual behaviour of the troops in the field for the purpose of establishing whether they in fact comply with, or disregard, certain standards of behaviour.
Given the chaos typically governing armed hostilities, even military commanders themselves may not always be able to ascertain with certainty ex post facto whether any specific attack complied with the applicable rules of engagement, much less the principle of proportionality.
Third, it is possible for a sufficiently serious violation of the principle of proportionality to be considered a war crime,Footnote 43 and few states are eager to publicise crimes committed by their own military organisations. Even those that privately recognise that an attack violated the jus in bello will rarely concede the fact openly. Admission of such violations brings disrepute to the military organisation and, by extension, the government. It also may invite intrusive scrutiny into military planning and operations by civilian political authorities or the international community. As a result, breaches of the laws of war may be denied, intentionally overlooked or, at best, handled privately and informally with no public acknowledgement of the violation. Often, official information about a breach of the jus in bello becomes available only after the rare international outcry or investigation by an international human rights authority.
Finally, most of the armed conflicts since 1945 have been, or at least started out as, non-international. There are several reasons why states have shown special reluctance to treat the targeting rules applicable in international armed conflicts as equally binding in non-international conflicts. Traditionally, a state's treatment of its own civilians was considered a matter of sovereign internal control, beyond the purview of international law.Footnote 44 This attitude has since become more enlightened, but the Geneva Conventions and their additional protocols do not reflect the full extent of change. As noted, Additional Protocol II omits any reference to a proportionality principle in non-international conflicts. While some authorities have asserted or implied that the principle of proportionality applies equally in non-international armed conflicts under customary international law or otherwise,Footnote 45 the accuracy of the claim is difficult to test. The opacity of military decision-making is, if anything, redoubled in armed conflicts confined to a single state. There, the international media and non-governmental organisations may have less interest and be excluded more easily. Moreover, when the opposition group must remain clandestine in order to survive, there is limited access to viewpoints opposing the government's official line. When anti-government forces can manage to communicate, the secrecy of the organisation may cast doubt on its veracity; it seems psychologically difficult to attribute honesty and candour to a violent group that lurks in the shadows. A single state government firmly in control of the ostensibly credible information about traumatic events in a dangerous and often remote location multiplies the already substantial obstacles to acquiring objective information.Footnote 46
2.2 Methodological Approach to the Study of the Customary Law of War
The most thorough analysis of the state practice of proportionality currently available is the monumental study, Customary International Humanitarian Law (‘ICRC Study’), conducted under the auspices of the ICRC by Jean-Marie Henckaerts and Louise Doswald-Beck.Footnote 47 The ICRC Study assembles valuable guidance on how states have interpreted their obligations through treaty negotiations, national legislation, military manuals, and public statements. Through this guidance, the ICRC Study identifies proportionality as a principle of customary international law.
The ICRC Study's methodology and the content of the rules it identifies as customary have not, however, achieved universal acceptance. The legal advisers to the US Department of State and Department of Defense, for example, characterised the study as excessively reliant on documentary sources and lacking verification by actual state practice on the battlefield.Footnote 48 They further claimed that the study failed to adduce specific evidence of opinio juris.Footnote 49 Academic commentators have echoed these concerns.Footnote 50
The criticisms of the approach by Henckaerts and Doswald-Beck to opinio juris take issue with the inference that states adopting legal constraints on attack modalities do so from a sense of international legal obligation. Technically, this is indeed a mere inference and not indisputable proof of opinio juris. Yet the relevance of the ICRC Study's sources to the assessment of customary international law is beyond doubt. States do not promulgate limitations on attack methods in their military manuals, or create criminal liability in national legislation, for the satisfaction of arbitrarily hampering the effectiveness of their military commanders. When states voluntarily limit their own military capabilities as a general rule applicable in all future conflicts it is because they consider such tactics to be either generally ineffective (for example, wasteful of firepower), immoral, or prohibited by international law.Footnote 51 It may be difficult to separate pragmatic motivations for adopting a rule of proportionality from the opinio juris motivation, but practical and humanitarian concerns need not be mutually exclusive.
In any case, it would be impossible to present the irrefutable proof of widespread, long-standing, and consistent state practice of proportionality accompanied by opinio juris desired by the ICRC Study's critics, unless states were to adopt the habit, which they have very rarely done,Footnote 52 of publicly and formally acknowledging customary rules of international law that they consider to be legally binding. Given that opinio juris can be inferred only indirectly in almost every case, the chief methodological difficulty is to gather evidence of battlefield behaviour during the confusion of combat and, afterwards, subject it to intentional obfuscation by the relevant actors. In adversarial legal practice, this problem may be overcome through the allocation of burdens of proof, ethical rules, and rules requiring cooperative evidentiary discovery. In international law, the door to such evidence is closed. Given the real-world constraints on evidence gathering, the value of the ICRC Study for assessing opinio juris is beyond doubt, even if its conclusions are presented with greater confidence than the evidence presented in the study can justify.
The ICRC Study is, in fact, more useful to that end than it is for assessing state practice. The main limitation on its assessment of state practice is its failure to reflect the potential gulf between two concepts described by Pound as ‘law on the books’ and ‘law in action’.Footnote 53 Battlefield behaviour does not always comply with the ideals set forth in military manuals or national legislation. It is true that the ICRC Study omitted any serious evaluation of actual state military practice. The question remains open whether states in fact observe a proportionality principle with any consistency and, if so, how they implement the principle in battlefield situations. The ICRC Study establishes that the principle of proportionality in its general formulation has achieved significant influence in state attitudes towards the conduct of hostilities. The next step – a method for assessing how states interpret and operationalise the proportionality principle – is sorely needed.
2.2.1 Journalistic methodology
None of the critics of the ICRC Study has thus far offered any concrete suggestion as to how it might supplement the Study with a more accurate and complete study of practice. The four factors discussed earlier in Section 2.1 serve to deter the deeper investigation of state practice. While they do not pose insuperable obstacles to supplementing the ICRC Study with an alternative methodology, they offer challenges that no study has yet answered.
One option would be to supplement the ICRC Study's sources using a journalistic methodology. This approach involves four general steps:
(1) defining the scope of the research and identifying sources of information, including legislation, media reports, reports and press releases of non-governmental organisations, public announcements of national and international elites, witness statements, and acts and reports of intergovernmental organisations;
(2) gathering the available information through documentary research and interviews;
(3) evaluating the validity of each source by confirming it against independent sources of evidence; and
(4) combining the facts into a narrative – with appropriate qualification for, or discounting of data falling towards, the less reliable end of the spectrum.Footnote 54
From this narrative, it may be possible to draw tentative conclusions about elite expectations for the actors involved in the incident under study, or at least their public positions on these expectations.
The strengths of this method include its openness to a variety of non-documentary data sources and its ability to cast doubt on some unreliable facts by seeking independent confirmation. On the other hand, journalistic methods for the evaluation of the reliability of evidence are less refined than legal evidentiary principles. For example, they do not necessarily account for the unreliability of second-hand evidence or the self-interest of the information source. The result is usually a neutral report of the statements of relevant actors or the journalist's own anecdotal observations rather than a thorough and objective analysis of the events. Moreover, the journalistic method is typically used for gathering and evaluating facts based on an incident limited both temporally and in subject matter (Step 1). Journalistic methodology is not adapted to the open-ended study of long-standing, consistent, and widespread international custom.
2.2.2 Historiographical methodology
Historians, in contrast, are accustomed to using such methods to survey developments occurring through a large number of incidents over a long span of time. From a thorough study of state practices through all available and reliable resources, historiography may generalise with appropriate nuance. Such a study will inevitably involve the use of numerous secondary sources dealing with specific incidents or trends in incidents. The resulting research is highly resource-intensive, involving the winnowing of large quantities of information in memoirs, biographies, or government records to isolate a few legally relevant details in the mass of data.Footnote 55 Because an accurate depiction of historical trends requires both a long view of the subject and a firm grasp of important details, a thorough study typically takes many years to complete. Such a study of international practice in the application and interpretation of the proportionality principle would be theoretically possible with sufficient resources, but the researcher then faces the obstacles discussed earlier in Section 2.1. Moreover, historians rarely gather information personally from direct sources except in distant retrospective interviews, and secondary sources relating to armed conflicts are usually written by journalists and historians with no training in international law. They rarely include the kind of detail about targeting decisions that are necessary to draw reliable legal conclusions about how states apply and interpret the proportionality principle, even if such information were available to researchers.
2.3 The Proportionality Multinational Study
The present author and a colleague, Professor Douglas Sylvester of the Arizona State University College of Law, have designed, and are in the process of conducting, a long-term, multinational study of state practice in observing and interpreting the proportionality principle (‘the Proportionality Study’). This project aggregates data gathered through individual ‘incident studies’. The incident study is a method of identifying and evaluating a specific event or connected series of events as a legal precedent. The methodology has been fully described in the writings of Andrew Willard and W Michael ReismanFootnote 56 and combines legal, journalistic, and historiographical research techniques. In the evaluation of each source of data, the observational standpoint of the source must be weighed to assess the reliability and bias of the source, and whenever possible such data is corroborated independently. The claims of all politically relevant actors are evaluated in light of the facts to ascertain how the operative legal norms affect the outcome of the dispute or the international community's appraisal of the event. Incident studies use the journalistic and historiographical methods described above, but supplement them with the international lawyer's understanding of the political relevance and viewpoints of actors in the international community and of legal principles of evidence. The Proportionality Study departs from the incident study methodology in some ways to be described in this section, but the most important departure is that the Proportionality Study is not limited to a single incident, but rather aggregates data from numerous incident studies. Specifically, our purpose in organising the Proportionality Study was to gain a more factually accurate and complete understanding of state practice of proportionality by cumulating incident studies in a wide range of major conflicts, both international and non-international, since 1945. The Study's key goal is to move beyond formal, documentary evidence of customary law and to analyse real battlefield practice.
The first step in each incident study is to identify and circumscribe the incident to be examined. We began with a set of armed conflicts, the selection of which was based on intensity, geographic and temporal diversity, and relative geopolitical importance. In the selection process, we have tried to obtain a representative sample of both international and non-international conflicts. With the parties to the conflict identified, we search through conventional sources of information on each party's general military legal training, law compliance monitoring, and enforcement techniques, through military manuals, public documents, and internet resources. This includes training techniques, deployment and the accessibility of legal personnel,Footnote 57 target selection and approval procedures, procedures for investigating and punishing commanders allegedly responsible for disproportionate attacks, and procedures for deterring or correcting dysfunctional attack methods. This general information will later be supplemented through interviews with military personnel.
With a relevant conflict identified, we select incidents of potentially disproportionate attack based on news reports, NGO reports, or allegations by opposing military forces or civilian populations. Given the Study's goals, at no stage do we judge whether such allegations are accurate. However, to determine the credibility of allegations that significant civilian casualties or property damage resulted from an attack, we perform documentary research relying primarily on reports of NGOs, the news media, or intergovernmental organisations such as the UN, Human Rights Committee, Inter-American Commission on Human Rights, or European Court of Human Rights. Based on this data, we determine whether a credible question of disproportionate attack has been raised. This research always requires assistants who are fluent in the languages spoken by the participants in the conflict.
Once we have identified a potentially precedential incident, our next step is to gather as much information as possible about the circumstances and consequences of the engagement. We attempt to overcome the difficulties described in Section 2.1 by examining each incident of allegedly disproportionate attack to determine:
• the actors involved, including their characteristics and relative advantages;
• the strategic goals of the participants;
• the immediate goal of the attack;
• the weapons used;
• the forces, threats, and defences faced by the attacker;
• the environment and context in which the engagement took place (for example, geographic, temporal, institutional);
• the political and logistical circumstances of the attack (for example, crisis, terrorism, war of attrition);
• the nature and strategic value of the military targets chosen in light of the range of targets reasonably available for attack;
• what method of attack was used of those reasonably available (for example, in terms of timing, approach, or distance);
• the military advantage expected from the attack;
• the expected consequences for civilian lives and property at the planning stage; and
• the claims and strategies of the relevant actors as to the legality of their actions, including their strategies for defending their actions.
Any clearly intentional killings of civilians, such as execution-style killings or attacks distant from any military objective, are factored out as being irrelevant to the question of proportionality.Footnote 58
If the facts are insufficiently developed to determine with confidence whether significant civilian casualties or property damage resulted from the engagement, we discontinue study of the incident and move on to the next. With sufficient facts, we then seek further information on the consequences of the incident from representatives of the armed forces involved. Our general interview protocol, which must be translated into each relevant foreign language, is set out below as an Appendix, although it is subject to customisation depending on the facts of the conflict and the information already publicly available. In addition, the research assistants are instructed to follow up with detailed questions about any vague or ambiguous practice. We also attempt to interview any available witnesses, field reporters, NGO and IGO personnel, and government sources to corroborate facts and supplement the record with unpublished information about the course of events and ultimate consequences for the actors.
In those cases in which, after extensive investigation, we are unable to identify any remedial or punitive consequences that were at least arguably connected to the allegedly disproportionate attack, we note the apparent absence of enforcement action. In the event that disciplinary or corrective action was taken, or a formal investigation was undertaken, we record the manner of investigation, the corrective measure, or the punishment and its result. Finally, we survey the reaction of the international community to the incident and its outcome (if any) through public statements of the government and international community, actions by intergovernmental organisations, national and international lawsuits, and NGO reports or press releases.
When a representative sample of countries and incidents have been studied, we will synthesise the findings and present an analysis of state practice in terms of general practices for ensuring compliance with the proportionality principle and specific corrective or enforcement measures in cases of credible allegations of violation. From our study of specific incidents, we will be able to describe the circumstances that provoked an investigative or other response, the range and incidence of responses (such as compensation to victims, implementation of new procedural protections, and punishment of individual combatants), and the expressed views of the international community on what kinds of measures are expected to satisfy proportionality obligations.
To conduct a sufficient quantity of incident studies to permit a competent analysis of the existence and content of a general customary norm of proportionality requires the evaluation of numerous military conflicts in diverse geographic locations over an unusually long time span. Each incident represents a single data point; it tells us about one event in one conflict only. To extrapolate a general rule or trend from such incidents requires a representative sample of incidents in a representative sample of armed conflicts. A systematic study of state practice in this manner demands patience, extensive resources, and assistants knowledgeable in over a dozen languages.
Our findings based on this methodology are subject to several important limitations. The first and most consequential is the absence of direct access to the tactical planning process of the military commanders involved. As noted above, military planning is by nature secretive, and decisions not to attack, or to change mode of attack, to reduce civilian casualties are not susceptible to discovery through conventional fact gathering techniques. The influence of the rule of proportionality, though sometimes invisible, nonetheless may have important consequences in practice. Not all legal rules operate by threat of sanction – indeed, very few do. Many legal rules become integrated into cultural expectations of rectitude and so operate psychologically with greater effect and universality than would be possible through threatened coercion.Footnote 59 But the difficulty of demonstrating empirically the operation of such influences in those countries with closed military cultures leaves them vulnerable to underestimation. Because there is often no real possibility of gaining access to this data, our findings should be qualified by the understanding that they may understate the role of the proportionality principle in protecting civilian populations. The main value of the incidents – as opposed to our more general research on military training, target planning and approval processes – is in assessing whether and how states investigate allegations of disproportionate attack and respond to discovered violations.
Gathering information about the consequences (or absence thereof) of an alleged violation of jus in bello is especially difficult. It is possible that, in some instances, disciplinary action was taken against a commander who directed a disproportionate attack without the action becoming publicly known. There are sanctions that fall short of court martial or criminal prosecution that could nonetheless punish and deter effectively. Corrective measures could result from disproportionate attacks without the military organisation publicly drawing a connection between the two. Whenever possible, we question our informants about whether any given incident could have resulted in corrective measures or informal sanctions,Footnote 60 but the frequent lack of direct access to military officers who would know best impedes their discovery.
The distortive effect of secrecy should not be overstated, however. A military organisation that refrains from publicising the consequences of a violation of jus in bello mutes the deterrent effect on its military commanders. This is not to say that an unpublicised sanction is equivalent to no sanction at all, but silent sanctions are much less conducive to the development of a military culture respectful of the laws of war. A military organisation's commitment to observing international law, to be effective, should be a public one.
In addition, in each case, whether a given attack was disproportionate in the first place is a question that can never be finally settled by our Study. No information dragged kicking and screaming from the din and mayhem of war will ever be totally beyond impeachment. Many military organisations do not even collect data on civilian or enemy military casualties. This is especially true of irregular armed forces, but even describes the practices of highly professional military organisations. Reliance on NGO and witness reports should be qualified by the understanding that, between the fog of war, the danger and difficulty of surveying a battlefield during or immediately after an armed engagement, and the traumatic effects of witnessing human devastation, the data will inevitably be partial and subject to some uncertainty. The news media and NGOs may fail to verify allegations of disproportionate attacks, may misreport accidental deaths as intentional killings of civilians (or vice versa), or may misstate the number of civilian or military casualties. The headline-grabbing effect of intentional executions of civilians also tends to overshadow reports of accidental, even if disproportionate, civilian casualties. Consequently, an investigation by the self-interested military organisation that ultimately exonerates the commander concerned cannot be disbelieved out of hand. At the same time, the organisation's failure to seriously investigate credible allegations of disproportionate attack, or consistent exonerations following credible reports of disproportionality, speaks forcefully about the regard in which the organisation holds the jus in bello.
The Proportionality Study has already begun to uncover empirical evidence of how states react to allegedly disproportionate attacks. Although we consider the Study to be an important advance over currently available information on international custom, our methodology does not touch the ideal. Because we perceive no readily available remedy for the deficiencies of this methodology at the present time, empirical researchers face the alternatives of a flawed study or none at all. To prevent the best from becoming the enemy of the better, we prefer the former to the extent that it produces at least some useful and reliable data, if regarded with sufficient caution.
3. The Need for Procedural Reform: Towards Greater Protection of Civilians in Practice
3.1 The Benefits of a Basic Transparency Obligation
The obstacles to a more reliable and complete picture of international custom draw attention to a critical gap in the jus in bello that, if filled, could increase our understanding of how the proportionality principle operates and, if the principle offers insufficient protection to civilians in practice, to improve compliance or develop alternatives. Given our findings in the five conflicts we have studied to date, it seems very unlikely that the Proportionality Study will be able to conclude that the principle is consistently implemented with great deference to the protection of civilians in most international armed conflicts. It clearly receives still less deference in non-international conflicts. It would be surprising indeed if the sanguinary civilian casualty statistics in modern wars could support a sanguine conclusion.
Without prejudging the effectiveness of the proportionality principle, one might hypothesise that the perceived failure of a generally phrased legal principle to achieve its policy objective should prompt proposals for the adoption of treaties, regulations, or other guidance with clearer or stricter obligations. The intuitive cure for vague law is more precise or detailed law. There are, no doubt, improvements to the phrasing of the proportionality principle that could render it more effective in protecting civilians, or at least make it easier to judge compliance with the principle.Footnote 61 In general, however, stricter rules are a double-edged sword. If not sufficiently restrained, they could upset the balance between military flexibility and humanitarian concerns struck in the framing of Additional Protocol I. The proportionality principle was intended specifically to provide a certain degree of flexibility to military commanders. An interpretation of the principle that hamstrings a state's ability to achieve its critical military objectives would not attract a general consensus or serve the international community's interests. The reforms that I propose here – not exclusive of other proposals for the improved protection of civilians – do not involve upsetting that balance.
Additional Protocol I mandates that the proportionality principle should be implemented through an ex ante assessment of the consequences of an attack on civilians. Both the Inter-American Court of Human Rights and European Court of Human Rights have also held that the human right against arbitrary deprivation of life requires an effective, public investigation of alleged killings of civilians during an armed conflict.Footnote 62 Amichai Cohen has argued that such ex post review by the military organisation or national government of commander decisions resulting in civilian deaths should also be implied as a requirement of the proportionality principle.Footnote 63 That a legal norm only becomes operative with institutional detection of violations and enforcement against violators is definitional, and Cohen's observation that investigations may yield useful information about the consequences of reckless or negligent targeting decisions is an insightful one.
Ex post investigation should be facilitated and supplemented, however, with other measures. As an enforcement technique, any investigation requires a standard of sufficient clarity and precision by which to appraise the commander's targeting decisions. It is impossible to evaluate a state's compliance with the proportionality principle, much less to understand how military commanders engage in the proportionality calculus in battlefield situations, without access to information about training in the law of war, tactical planning procedures, and enforcement practices. A few states have been forthcoming with some information about training, procedural safeguards, and enforcement of jus in bello. The US Army, for example, has publicly released summaries of its standing rules of engagement that address proportionalityFootnote 64 and has even adopted a general methodology for estimating collateral damage prior to attack.Footnote 65 Excerpts or summaries of both have been declassified and can be accessed by the interested public. On the level of enforcement, the US Army conducts public criminal trials of officers and soldiers accused of violating the jus in bello. However, few military organisations have shown comparable candour.
What I propose is that the first step toward the greater protection of civilians should be the adoption of regulations for the gathering and dissemination of information about the military organisation's compliance procedures and the consequences of its operations on civilians. This could be accomplished through the adoption of a new information-sharing treaty for the protection of civilians in armed conflicts or by the creation of an intergovernmental agency charged with monitoring compliance with existing laws of war, or both.
Although the words ‘transparency’ and ‘military’ are rarely found as neighbours, a greater if limited degree of transparency promises a significant increase in civilian protection. As noted, most military organisations do not keep track of civilian deaths or property damage.Footnote 66 Some do not even keep an accurate account of enemy combatants killed. National legislation and most military manuals are a matter of public record, but many states have extremely secretive and closed military cultures that inhibit the gathering of information about their methods for ensuring compliance with the laws of war. This is especially so if, when the proportionality rule has been violated, the practical consequences take the form of informal or sub rosa sanctions or corrective measures, such as demotion, unpaid leave, reforms to classified procedures, or simply a private reprimand.
My proposed approach – which would apply in the case of any significant use of force whether international or non-international – would not impose any new limitations on the ability of states to use military force. Its sole purpose would be the promotion of a limited kind of transparency. Specifically, what I propose are commitments by states, either directly or through a neutral clearinghouse:
(1) to track all direct and indirect civilian and combatant casualties on all sides of the conflict to the extent possible without endangering combatant lives or important military objectives, and to make this information publicly available;Footnote 67
(2) to make publicly available the military organisation's training methods for compliance with the laws of war;
(3) to make publicly available the methods adopted for ensuring compliance with the laws of war by units in the field (for example, through the appointment of legal advisers, detailed rules of engagement, no-strike lists, limitations on the use of relatively indiscriminate weapons, a collateral damage estimate methodology, or clearance by commanders of targeting decisions that potentially endanger civilians);
(4) as suggested by Cohen and in the decisions of the regional human rights courts,Footnote 68 to require investigation of all credible allegations of disproportionate attacks and to publish the findings, with redactions if necessary for national security purposes; and
(5) to make publicly available the results of any formal charges brought against a military commander for indiscriminate or disproportionate attacks resulting in civilian casualties or extensive property damage.
3.2 Limitations and Objections to a Transparency Obligation: Beyond Criminalisation of Disproportionate Attacks
The reason for the inclusion of combatant casualties in the first obligation may not be obvious, considering the emphasis of humanitarian law – and the proportionality principle especially – on civilian protection. Aside from honouring the intrinsic value of combatant lives, tracking combatant as well as civilian casualties serves a dual function that would indirectly support the goal of protecting civilians. With regard to a state's own combatant casualties, the proposal would encourage states to exercise greater candour about the human consequences of their military operations. The international community and NGOs show special solicitude for civilian lives (and especially the lives of women and childrenFootnote 69), implicitly if unintentionally minimising the fact that combatant lives have moral weight as well. More to the present point, knowledge of the number of combatant casualties may facilitate an assessment of compliance with and interpretation of the proportionality principle by emphasising the importance of post-engagement information gathering.Footnote 70
With regard to enemy combatant casualties, in the era of classical warfare such a reporting obligation would usually have been superfluous because the enemy belligerent could be expected to assess its own military casualties. Today, when asymmetric and non-international armed conflicts predominate, enemy belligerents are often either failed states or are not states at all but rather insurgent groups. The organisation and the information gathering and disseminating capabilities of such groups are typically greatly inferior to those of opposing state military organisations. Unless the latter attempts to track enemy combatant casualties, the consequences of armed conflicts for enemy combatants may remain unknown in asymmetrical engagements. Enemy combatant casualties are not decisive in proportionality analysis because prior expectations matter, not post hoc consequences, and enemy casualties cannot be equated with military advantage by any means. Nonetheless, such statistics may partially inform the military advantage expected from the attack during the planning stage.
The first obligation listed above also raises the inevitable question of which deaths and injuries should be included in the calculus. Military organisations reporting on civilian deaths will naturally tend to count civilian deaths resulting from direct attack only. Marco Roscini has proposed to include indirect casualties in any assessment of proportionality.Footnote 71 Depending on how ‘indirect’ casualties are defined,Footnote 72 this practice could greatly increase the number of civilian casualties reported in protracted – where the diversion of resources from civilian to military uses on a significant scale tends over time to reduce civilian essential services and ultimately their life expectancy – or intense armed conflicts near urban areas, where damage to civilian infrastructure may endanger resources necessary for the survival of concentrated populations. The necessity for including such indirect but foreseeable civilian casualties in a reporting obligation may deter unnecessary attacks on important civilian or dual-use infrastructure.
The first and last listed obligations could be objectionable if construed to require military organisations to disclose sensitive national security information. No state would consent to be bound by a transparency obligation that compromised national security. The transparency commitments would necessarily stop short of obligating states to reveal sensitive military information, such as specific procedures for selecting targets for attack. As noted, the United States has revealed declassified versions of its training, collateral damage estimates, and targeting rules of engagement without suffering any noticeable military disadvantage.
Of course, allowing secrecy creates the potential moral hazard of exempting states from the consequences of disproportionate attack by overclassifying conflict information as being necessary for national security. However, there are some kinds of relevant and important data that can be uniformly qualified as non-sensitive. This includes the number of enemy combatant deaths, the number of civilian casualties, and the civilian property damage following an armed attack. Such information cannot reasonably be categorised as militarily sensitive. In addition, the identity and general nature of the military target being attacked, and the general method of attack and weapons used, will generally not be militarily sensitive information after the attack is completed. In those presumably rare cases in which such information can be used to predict future methods of attack, the disclosure obligation can wait until the engagement has terminated. While an immediate disclosure obligation might be ideal, even a more realistic duty of delayed disclosure would improve accountability. Moreover, outside sources of information such as NGOs and opposing parties will help to correct or at least moderate distortions, as currently occurs for human rights obligations through the UN Human Rights Council's Universal Periodic Review.Footnote 73 Furthermore, states that repeatedly fail to comply with their reporting obligations by claiming security privilege with respect to violations of the jus in bello or distorting information will be subjected to reputational pressures to reform, including possible sanctions or expulsion from the transparency regime.
Given the widespread acknowledgement of the subjectivity and indeed potential arbitrariness of proportionality judgments, a criminal conviction in the case of a borderline disproportionate attack may be viewed as inequitable. Articles 51 and 57 of Additional Protocol I are designed to restrain violations of the law of war prospectively rather than to create a framework for criminal punishment. Neither the Fourth Geneva Convention nor Additional Protocol I defines a disproportionate attack as a grave breach. The principle as stated in Additional Protocol I includes no detailed enumeration of elements of the crime, it does not clearly specify a mens rea requirement, and it proposes no specific punishment for individual violators. Nor does it specify whether individual combatants can be found liable for violating the principle, or whether the principle applies only to commanders who plan attacks.
The language of the principle is, in short, more characteristic of a general legal obligation incumbent upon the state than a specific criminal prohibition applicable to individuals. Perhaps some attacks, falling short of indiscriminate, may be so evidently disproportionate that no reasonable person could disagree as to their character as prohibited, but many others will inevitably fall within the twilight zone between sufficiently restrained and wildly reckless. The ICC Statute is consistent with this reasoning in authorising international criminal prosecutions for ‘clearly excessive’ attacks only.Footnote 74 However, for attacks not so ‘clearly’ disproportionate, an ordinary military commander cannot be expected to understand what the principle prohibits with great confidence and precision. Such uncertainty opens the door to arbitrary or discriminatory enforcement. Overly vague criminal laws are inherently unfair, a point international law reflects through the principle nullum crimen sine lege. Wherever there is uncertainty as to whether a military commander's conduct falls within the scope of a criminal prohibition, all doubt should be resolved in favour of exoneration.Footnote 75 The inappropriateness of a criminal penalty for most violations of the proportionality principle does not, however, doom it to irrelevance. A more nuanced approach could preserve the principle while fostering transparency and observing the basic human rights of military commanders to due process of law.
Additional Protocol I directly binds state parties, not individuals. It does not specify that its provisions must be enforced by criminal prosecution; it simply prohibits specified acts and leaves the means of ensuring compliance to state parties. At least in its manifestation in Additional Protocol I, then, the proportionality principle imposes an obligation on states to deter and punish any kind of excessive attack, but not necessarily through criminal penalties. This approach is both appropriate and consistent with the policy goals of international humanitarian law. Criminal prosecution is a fortiori less appropriate for attacks other than those that threaten civilian lives or property in a manner clearly in excess of the direct and concrete military advantage anticipated. In such cases, the state satisfies its international legal obligation through measures sufficient to deter future disproportionate attacks accompanied by compensation to the civilian victims and their families.Footnote 76 With criminal penalties off the table for all but the most plainly disproportionate attacks, no serious obstacle to the adoption of a basic transparency obligation should remain.
4. Conclusion
Widespread adoption of the proposed commitment to the collection and disclosure of information on compliance with the rules of discrimination and proportionality would have two salutary consequences for the protection of civilians. First, the publicising of casualties caused by the armed conflict would place reputational and other political pressures on states to minimise such casualties whenever possible and, by motivating domestic and international opposition to the irresponsible use of armed force, potentially increase the threshold of provocation before a state will resort to armed coercion. Opacity and secrecy promote irresponsibility and unaccountability.Footnote 77 As US Supreme Court Justice Louis Brandeis once wrote, ‘sunlight is the best disinfectant’.Footnote 78 In the context of proportionality, where assessing compliance is complicated by the obstacles discussed earlier, transparency would promote the necessary accountability now lacking.
Transparency would also lead to a better understanding of how proportionality operates in practice. The proposal would make possible a more reliable evaluation of state compliance with the jus in bello most affecting civilians. At present, an accurate assessment of the operation of the proportionality principle is hobbled by an information vacuum that, when completed, the Proportionality Study will only partially fill. The systematic collection of more comprehensive information will require cooperation by states in sharing relevant compliance data. To the extent that the proportionality principle could be refined, supplemented, or replaced to achieve more effectively the objectives of civilian protection consistent with legitimate military operations, transparency would facilitate the analysis greatly.
Some states may prove reluctant to agree to these requirements. Publicity about civilian and combatant deaths is likely in many cases to stimulate public opposition to the armed conflict at issue. This is one reason why states often avoid gathering casualty statistics in the first place. Moreover, because it will not be possible to enforce comparable requirements in asymmetrical conflicts on terrorist organisations or irregular armed opposition groups, some states may object to reporting burdens that do not encumber their opponents (who may already disregard the most fundamental laws of war).
These are legitimate concerns, but they are also reasons for building into the transparency regime effective institutional deterrents to non-compliance with the transparency obligations, not for adhering to the status quo. The measures proposed here are far from drastic. It is possible to define them in such a way so as not to require an extensive allocation of resources by states, nor to put any state's critical defence needs at risk. Further development of a transparency regime would require more extensive discussion than is possible here. The point of raising it is to emphasise the limitations of empirical research in state interpretation of, and compliance with, the proportionality principle, and to suggest the aptitude of transparency for improving our understanding of, and potentially the very operation of, this aspect of the law of armed conflict.
In the meantime, the Proportionality Study continues to plod along. Whatever the outcome of that research, the utility of our methodology is not limited to studying the proportionality principle itself or even jus in bello more generally. It is adapted to the exigencies of research on secretive military organisations, but it is easily customisable. The investment of resources required for thorough empirical research is substantial, and the payoff uncertain and glacially slow. But for those international lawyers not content to rely for their understanding of customary international law on the ex cathedra pronouncements of international tribunals or each other's ipse dixit, there are few good substitutes.
Appendix
Interview Protocol for Foreign Military Representatives
If we send our questions in writing, or pre-send them in preparation for an oral interview, we begin with the following introductory comments:
Dear Sir/Madam:
I am writing to request your advice and assistance with a long-term, multinational study on the international law of war, specifically on the concept of ‘proportionality’ as defined in Protocol I Additional to the 1949 Geneva Conventions, Article 51. I am a graduate law student working with the authors of the study, Professors Aaron Fellmeth and Douglas Sylvester of the Arizona State University College of Law.
The purpose of the study is to document state practice in interpreting and enforcing the proportionality principle in battlefield practice, in both international and non-international armed conflicts. To that end, the study will describe in detail some twenty armed conflicts in which allegations of disproportionate attack were made. The purpose of the study is not to accuse any state, military organisation, or person of violating international law, but instead to deepen our understanding of how the proportionality principle works. This includes information about how military organisations such as your own generally train commanders and combatants to respect the proportionality principle; the role of military lawyers in battlefield practice; any interpretive guidance given to commanders and combatants; and monitoring and enforcement mechanisms and practices. We are also studying specific military engagements involving civilian casualties in which disproportionate attack was alleged, and how any investigation and enforcement was handled.
To be clear, we guarantee that we will not use any person's name without explicit permission. We will not purport to arrive at any conclusions about the criminality of any action or omission by any state or person. The purpose of our project is solely educational.
Thank you in advance for your kind assistance.
General questions
1. Do military units in the field always or generally include a legal adviser trained in the laws of armed conflict?
(a) If so, is the legal adviser a military officer? Of what rank?
(b) What is the minimum size of the military unit for which a legal adviser is required or responsible?
(c) Is the adviser automatically involved in targeting decisions, or only when requested by a military commander? If the latter, are there published guidelines on when military commanders should consult legal counsel during active military operations?
(d) Does the legal adviser have power to block targeting decisions, or is his advice optional?
2. How are military commanders, soldiers, and pilots trained in the laws of war?
(a) Does initial training involve instruction in the laws of war?
(b) Are soldiers or commanders issued a military manual describing their rights and obligations under the Hague and Geneva Conventions and Additional Protocols?
(c) If commanders have questions about their obligations under international law, are there clear channels through which they can question legal advisers?
3. What is the protocol for commanders to assess the probability and number of civilian casualties involved in an attack?
(a) Before commencing military operations, does the organisation always or usually create a no-strike list?
(b) Must targeting decisions and attack methods be approved by a high-level officer? If so, what level?
(c) Does the military organisation have standing rules of engagement in addition to general rules on laws of war and proportionality in particular?
(i) If so, are they classified or public?
(ii) If so and they are public, where can we find them?
(d) Does the organisation have an established methodology for estimating collateral damage?
(i) If so, is it classified or public?
(ii) If so and it is public, where can we find it?
(iii) If so and it is classified, is there a summary available? Where can we find it?
4. Does your military organisation keep track of civilian deaths and property damage caused by its operations? If so, how?
(a) What are the usual consequences of a military commander making a tactical decision that results in excessive civilian deaths?
(b) Does your military organisation include an ombudsman, military lawyer, or other independent person to evaluate the legality of battlefield behaviour?
5. Does your military organisation have its own system of courts and judge advocates general?
(a) If so, are its proceedings open to the public in criminal cases?
(i) Are its judgments published and publicly available?
(ii) If not, are summaries or outcomes of the judgments made public?
(b) If not, how does the organisation handle violations of the laws of war? By civilian criminal courts or through some other means?
6. To whom do you recommend we talk to get more information about your training, monitoring, and enforcement practices? How can we get in touch with that person?
Questions specific to the incident under study
1. How did the military organisation respond to allegations of disproportionate attack in the incident under study?
(a) If there was an investigation, who conducted it?
(i) What was the procedure followed?
(ii) Were the results of the investigation made public? If so, how?
(b) Was the commander found to have followed the protocol for assessing danger to civilians?
(c) What was the judgment of liability?
(i) If someone was found liable, at what rank?
(ii) What were the consequences for that individual?
2. Were there some other, non-criminal consequences? For example:
(a) Was anyone demoted as a result of the investigation?
(b) Was anyone dishonorably discharged or fired?
(c) Was anyone put on mandatory leave? If so, for how long? Paid or unpaid?
(d) Was the unit responsible reorganised? If so, when and how?
(e) Was any military commander transferred to another unit? For what official reason? What are the consequences of the transfer for that commander?
(f) Was anyone given a letter of reprimand or otherwise informally cautioned or sanctioned?
(g) Were any of the above consequences made public? If so, how?
3. Regardless of the outcome, did the incident prompt any change to your methods of training or supervising commanders or soldiers for compliance with the principle of proportionality? Any change to your target selection, weapon selection, precautionary measures, or collateral damage estimation methods?
4. To whom do you recommend that we talk for more information or other perspectives on this incident? How can we get in touch with that person?
Follow up on all answers as appropriate, with special attention to vague or ambiguous responses.
