Rethinking direct participation in hostilities and continuous combat function in light of targeting members of terrorist non-State armed groups

Abstract

Endless armed conflicts against terrorist groups put civilian populations at risk. Since France has been involved in the Sahel from 2013 onwards, transnational non-international armed conflicts (NIACs) of extended geographical and temporal scope against groups designated as terrorists are not a US exception anymore. NIACs against terrorist groups, conducted not only by the United States but also by France, persist and have been reconfigured around threat anticipation. How can anticipatory warfare be best constrained? This article argues that it can be best done through more constraining rules regulating target selection in NIACs and, in particular, by redefining the notion of continuous combat function (CCF). Many elements explored in this article indicate that the United States and France select targets that they pre-designate as terrorists, before these targets are engaged in hostilities. Instead of responding to the observed participation of these individuals in hostilities, strikes are based on contextual and behavioural elements ahead or outside of such moments. This paper argues that when war consists of threat anticipation, it becomes very extensive and particularly risky for civilians. Furthermore, recent State practice in the counterterrorism context reveals the pitfalls of the notions of direct participation in hostilities and CCF as defined in the 2009 International Committee of the Red Cross Interpretive Guidance. Outside this context, the interpretations proposed in the Interpretive Guidance might seem sufficient to constrain target selection processes and to protect civilian populations. However, when applied to armed conflicts that are driven by threat anticipation, the pitfalls of these interpretations emerge. I formulate a critique of these interpretations as being partly responsible for anticipatory warfare and propose an alternative theory for the CCF test.

Introduction

In ten years, Europe will still be deployed in the Sahel, and probably more so than it is today.

Général François Lecointre, French Military Forces Chief of Staff¹

How can endless and anywhere wars in the counterterrorism context be constrained? Non-international armed conflicts (NIACs) against terrorist groups, conducted by the United States,² but also by France,³ persist and have been reconfigured around threat anticipation. The United States and France are the two case studies of this paper as they both happen to select targets that they designate as terrorists outside moments and areas of hostilities. Because these interventions against members of designated terrorist groups⁴ aim to anticipate threats, they are extended in time. For this reason, the terms “US forever war”⁵ and “French endless war”⁶ populate our imaginaries about armed conflicts against non-State armed groups designated as terrorists. Even if underexplored by legal scholars to date, France has, for the past eight years, been leading counterterrorism operations against various non-State armed groups in the Sahel and excludes the prospect of terminating its intervention in the region.⁷ While the French intervention in the Sahel is being reconfigured at the moment of finalizing this paper,⁸ the government does not intend to put an end to the military intervention but simply to reshape it and increasingly rely on air power.⁹ This is in line with the French Law on Military Programming for 2019–2025¹⁰ which provides a military strategy for France's “war against jihadist groups” – thus entering the international humanitarian law (IHL) paradigm – and suggests that France's counterterrorism operations abroad will persist for as long as these groups exist, re-emerge and recompose themselves.¹¹ This law, together with other sources, including the testimony of a French drone operator,¹² indicates that France's military strategy in the counterterrorism context, just as that of the United States, rests on the objective of pre-empting threats: instead of targeting individuals at the very moment when they participate in hostilities, their aim is to pre-emptively identify possible targets before hostile acts are conducted.¹³ Given this similarity in the military strategy of the two countries and the extensive temporal scope of their military engagement against designated terrorist groups, France and the United States serve as case studies for this article. This identification of targets is sometimes based on contextual and behavioural elements that do not include the witnessed participation in hostilities and leads, as cases studied in this paper show, to civilian casualties.¹⁴

I argue in this paper that when armed conflicts consist of threat anticipation, they become not only endless,¹⁵ but also risky for civilian populations. Recent State practice in the counterterrorism context reveals the pitfalls of the interpretations of direct participation in hostilities (DPH) and continuous combat function (CCF) crystallized in the International Committee of the Red Cross (ICRC) Interpretive Guidance of 2009. Outside this context of anticipatory warfare, it was not necessarily possible to foresee that these interpretations would leave unconstrained target selection processes and civilian populations insufficiently protected. I formulate a critique of these interpretations as failing to constrain anticipatory warfare and propose an alternative definition of the CCF test. The point is not to say that the targeting practices were explicitly justified on the basis of CCF as interpreted by the ICRC in its Interpretive Guidance – in fact, as shown elsewhere the US rationale for these practices is an admixture of ad bellum and in bello justifications,¹⁶ and France mainly presents vague justifications for its military operations in the Sahel.¹⁷ The point of this paper is to argue that not only practically speaking but also on the advocacy front, the mainstream interpretation of CCF provided in the ICRC Interpretive Guidance does not allow for successful and robust pushback against pre-emptive targeting practices that put civilian populations at risk.

The persisting threat of anticipatory (and thus endless) warfare resulting in civilian harm should stimulate renewed interest in scrutinizing States’ rules of engagement and related interpretations of the law. It requires asking what open-textured norms of IHL, when left undiscussed, allow for the practice of targeting individuals who have never been witnessed actively participating in hostilities and thus create a high error risk for civilians. The paper identifies the IHL concepts that States use in targeting members of designated terrorist groups who are enemy parties in a NIAC, and discusses how the interpretation of these legal concepts has evolved over the past two decades. All norms present a certain degree of uncertainty and the rules regulating target selection in NIACs – DPH and CCF – are no exception. While these norms have been the object of lively and highly important scholarly debate since the publication of the ICRC's Interpretive Guidance on the notion of DPH, this debate has not much focused on the counterterrorism context. Yet, elements of State practice highlight the pitfalls and strengths of alternative interpretations of these norms. To the question of how to constrain anticipatory warfare, I argue that we need to depart from the current definition of CCF.

This contribution is first descriptive and consists of the study of the notions of DPH and of CCF. In the first section, the article gives a brief account of the evolution of DPH and the notion of CCF, as well as an explanation of the intervention of the ICRC Interpretive Guidance in this evolution. In the second section, the article unpacks what the test of the notion of CCF, as defined by the Interpretive Guidance and as commented in scholarly debates, involves. In other words, it concentrates on CCF as functional membership derived from evidence not limited to witnessing actual involvement of the target in hostilities In the third section, the paper shows that the CCF test proposed by the ICRC Interpretive Guidance is very close to how some States target individuals designated as terrorists beyond areas and actual moments of participation in hostilities. It is argued that the CCF test, unless it is revisited, puts civilians at great risk. In this, the paper takes issue with the functional approach to membership as it does not satisfactorily constrain targeting.¹⁸ Finally, in conclusion, I briefly undertake the more normative task of identifying an alternative, more restrictive interpretation of CCF that would change the CCF test, and considerably reduce the targeting opportunities of States, thus lessening the risk posed to civilians.

Methodologically speaking, this paper intends to show that looking at IHL both from an internal and external point of view is necessary to explain in detail how law is conceptually and practically rearranged by States around pre-emptive warfare. I hope to highlight the value of abandoning the divide between studying the norms of IHL in a doctrinal manner, on the one hand, and studying IHL “in action”, on the other hand. This field is one deprived of routine judicial review, so one cannot wait for court decisions to analyse facts and realities of contemporary warfare. Similarly, it concerns military strategies developed by States that do not necessarily have constitutional traditions where the executive branch is vocal about defence policy choices – this is the case of France, for instance. Here again, IHL scholarship cannot wait for official declarations to be made to try and trace rules of engagement and suggest related interpretations of the norms. Therefore, studying the facts of reported cases is sometimes the only way to formulate hypotheses about the legal interpretations chosen by operating States for the norms of IHL. By analysing the practice and rhetoric of certain States active in armed conflicts against parties that they simultaneously label as “terrorist” in nature, I intend by no means to draw conclusions about their interpretation of the law, nor to describe as de lege lata their interpretation of these notions. Given the importance of the studied States’ participation in these armed conflicts (time, space and intensity), as well as the laconic information about their legal and operational framework, an active inquiry of State practice that moves beyond the narrow focus on what the States voluntarily share seems timely and needed.

The concept of direct participation in hostilities in context

The DPH test serves to determine whether a civilian's protection should be temporarily suspended on account of his participation in hostilities.¹⁹ While Article 3 common to the four Geneva Conventions merely refers to “persons taking no active part in the hostilities” without defining the category, Article 51 of Additional Protocol I and Article 13(3) of Additional Protocol II are more specific. Besides, according to Rule 6 of the ICRC Study on Customary International Humanitarian Law, this is also a customary law rule in both international armed conflicts (IACs) and NIACs.²⁰ These articles provide that in both IACs and NIACs, civilians enjoy protection “unless and for such time as they take a direct part in hostilities”. Because treaty rules remain open-textured, case law and doctrinal debates have served to establish that in addition to a limited temporal scope (i.e. suspending protection only for such time as the hostile acts are perpetrated), DPH has a limited material scope (i.e. it only covers acts that are intended to cause and do cause actual harm).²¹ The three cumulative constitutive elements of direct participation as consolidated by the ICRC guidelines are that: (i) the person must engage in an act that reaches a certain threshold of harm; (ii) a direct causal link exists between the act and the harm caused; and (iii) the act is precisely designed to directly cause harm, in support of a party to the conflict and to the detriment of another.²² The notion of DPH so conceived practically restrains targeting in warfare to individuals when they directly commit harmful acts.

The extension of targeting in warfare to individuals recognized as members of the enemy group when hostilities – understood here as acts of combat and military operations – are not ongoing would thus be unlawful under this restrictive interpretation of DPH. However, the notion of CCF as crystallized in the Interpretive Guidance opens up targeting possibilities beyond actual moments of hostilities. In 2009, the ICRC Interpretive Guidance introduced the notion of CCF as follows:

For the purposes of the principle of distinction in non-international armed conflict, all persons who are not members of state armed forces or organized armed groups of a party to the conflict are civilians and, therefore, entitled to protection against direct attack unless and for such time as they take a direct part in hostilities. In non-international armed conflict, organized armed groups constitute the armed forces of a non-state party to the conflict and consist only of individuals whose continuous function it is to take a direct part in hostilities.

For the sake of clarity, it is useful to note that there are two ways in which CCF can be defined in relation to DPH. CCF can either be considered as a version of DPH, or DPH and CCF can be understood as two separate categories. No matter what option one chooses, CCF is defined as someone's function in the fighting wing of an armed group. To avoid any confusion, it is simply useful to stress that the present paper portrays CCF as one version of DPH as proposed by the ICRC.²³

Aside from this issue, and in any event, the CCF appears in the Guidance as a criterion for membership in a non-State armed group; but it triggers only a quasi-membership status as it does not come with a privilege of combatant status (that is, with a right to participate in hostilities, combatant immunity or a prisoner-of-war status if the person holding that function is captured). Besides, CCF refers to the person whose continuous function involves the “preparation, execution, or command of acts or operations amounting to direct participation in hostilities”.²⁴ Such a function can be assumed “even before he or she first carries out a hostile act”.²⁵ As such, CCF does not define membership on the basis of an official status as a member of a non-State armed group, nor does it necessarily require the witnessing of repeated conduct of participation in hostilities in order to attribute membership (this form of membership is called “continuous direct participation”).²⁶

By introducing the CCF, the ICRC intended to give clarity on the terms “civilian”, “armed forces” and “organized armed group” that treaty-based IHL uses without defining them.²⁷ In doing so, the ICRC wanted to respect the wording and logic of common Article 3 and Article 13 of Additional Protocol II to ensure that civilians, armed forces and organized armed groups remain distinct categories. The goal, beyond terminology, was to concretely distinguish members of the organized fighting forces from civilians who directly participate in hostilities sporadically or support the group with non-combat functions.²⁸

The CCF test: Less is not more

Much scholarly literature on the CCF has focused on discussing the desirability of this concept as defined by the Interpretive Guidance.²⁹ Although it has been extremely useful in many regards, this discussion has rarely included developments on the CCF test, which is of the specific elements taken into consideration and the analysis undertaken to conclude or not that someone holds a CCF. Yet, having a conversation about the evidence and criteria that practically are, or on a normative level should be, used to determine a potential target's CCF seems essential to reach a well-informed opinion on the desirability question. The ICRC Guidance, for its part, did provide preliminary elements to start constructing a CCF test: it says that it requires elements demonstrating “lasting integration into an organized armed group acting as the armed forces of a non-State party to an armed conflict”, such as “the preparation, execution, or command of acts or operations”;³⁰ it adds that “an individual recruited, trained and equipped by such a group to continuously and directly participate in hostilities on its behalf can be considered to assume a continuous combat function even before he or she first carries out a hostile act”;³¹ and finally, it excludes “individuals who continuously accompany or support an organized armed group, but whose function does not involve direct participation in hostilities”.³² As CCF is not status-based and not necessarily conduct-based – an individual's CCF does not need to be derived from repeated participation in hostilities, but can be based on evidentiary elements other than witnessed participation in hostilities. As was just mentioned, apart from this logical definition of the CCF test, the ICRC clearly states in its Interpretive Guidance that an individual “can be considered to assume a continuous combat function even before he or she first carries out a hostile act”.³³ Therefore, someone can be found to hold a CCF even when their engagement has never been witnessed and even if such engagement has never actually taken place (if someone, for instance, has only been trained but has not yet actively engaged in hostile acts).

However, the purpose of the Guidance is to “facilitate these distinctions [between armed forces of parties to an armed conflict and civilians, and between civilians who never take a direct part in hostilities and those who do so on an individual, sporadic or unorganized basis only] by providing guidance”.³⁴ As such, it does not pretend to solve every issue and only “summarizes the ICRC's position on the interpretation of IHL on a particular legal question”.³⁵ No further details or examples are given in the Interpretive Guidance on what type of elements can be taken into account to demonstrate lasting integration into an organized armed group acting as the armed forces of a non-State party to a NIAC.³⁶

Nonetheless, the CCF is found to be a reasonable way of distinguishing members of non-State armed forces that can be targeted from civilians punctually participating in hostilities, while simultaneously extending the targeting possibilities of States against members of non-State armed groups.³⁷ Indeed, these scholars often share the ICRC's (most certainly laudable) concern that evolving interpretations of the law should maintain a high level of protection for civilians. While they might have criticized the CCF for other reasons, they supported it for being constraining and for protecting the principle of distinction. Notwithstanding the first impression that CCF strikes the right balance within the above-mentioned typology of membership definitions, an analysis of the notion in light of the actual test it implies, and the type of evidence used to identify a CCF, will invite the reader to nuance and even rebut the presumption that the CCF test as currently defined extends the scope for targeting without putting civilians at greater risk. First, there is the idea that CCF should be applauded for constraining targeting options as it only concerns members of an armed group's fighting wing and not of the broad organization, e.g. the members of its “political” wing.³⁸ Yet, approving the functional nature of CCF membership, without trying to build on the foundations provided by the ICRC Guidance and specifying which elements should be expected, neglects the risks it poses to civilians. Not only is there no clarity to date on the specific elements that permit the identification of someone as holding such function, but also its explicit distinction with a conduct-based membership inherently tells us that CCF is not solely derived from repeated engagement in hostile acts. What are, therefore, the specific elements that lead to identifying someone's “lasting integration into an organized armed group”, i.e. as holding a CCF? Also, what is the effect of targeting individuals based on elements other than their repeated engagement in hostile acts? Without answering these questions in detail just yet, it is already possible to say that the error risk inherent to targeting on the basis of evidence other than witnessed engagement in violent acts necessarily puts civilians at greater risk.

Some authors have also applauded the CCF as constraining in comparison to status-based membership.³⁹ According to this view, membership in an armed group derived from an official status for the purpose of targeting could be too extensive because “as indicated by ex post acknowledgement of attacks by groups such as the Islamic State, for instance, armed groups might have various incentives to exaggerate their membership”.⁴⁰ Again, basis for appraisal of CCF is vulnerable to a very concrete counterargument related to the test it implies. It may be true that individuals and groups claim membership sometimes in the absence of very close ties between the individual and the group; it would actually be very interesting, as a socio-legal endeavour, to complement this analysis with empirical evidence. This point thus definitely deserves to be made especially if the discussion is placed at the descriptive level. If the discussion is a normative one, though, I am not convinced that this is a good reason to prefer the alternative option of function-based membership, which consists of deducing membership from evidence not limited to engagement in violent acts. There is a need to refine and rethink the current mainstream interpretation of the CCF test in order to powerfully push back against situations that we will explore below.

As mentioned above, the third option is to derive membership from conduct: if States witness individuals engaging in a series of offensive violent acts (in support of a non-State armed group party to the conflict), then they can continuously target them. This option is sometimes considered too similar to punctual attribution of DPH and too restrictive: the idea is that “if membership entirely depends on actual conduct, this notion becomes useless and should simply be abandoned. In other words, if targeting based on ‘membership’ has to mean something, it has to be more permissive than just direct participation in hostilities”.⁴¹ This is not the reason, however, or at least not the apparent or only reason why the ICRC chose to create a membership status based on a function rather than a conduct, at least when looking at the explicit rationale of the Guidance. The ICRC Guidance's goal, by excluding that people showing a continuous combat conduct and not function be targeted, was to preserve “the conceptual integrity of the categories of persons underlying the principle of distinction” and protect civilians.⁴²

The truth is that both reasons for choosing a conduct-based membership appear to be misguided, for different reasons, and this will clearly emerge from the last section. Indeed, the last section shows that the preferred version of the CCF test is too permissive in the context of States’ targeting members of groups designated as terrorist. Dissatisfaction with the conduct-based definition of membership might stem from underestimating how extensive targeted practices can be when States derive membership solely from elements other than engagement in hostile acts. Yet, a test based on evidence that is not limited to conduct (engagement in hostile acts), and can even exclude conduct (remember the inclusion in CCF of individuals who have been trained but who have not yet engaged in hostile acts) puts civilians at greater risk than a test based on engagement in hostile acts.

Another viewpoint is that CCF is in fact much closer to continuous direct participation (conduct-based membership) than meets the eye.⁴³ Continuous direct participation describes the situation where an individual engages in a series of hostile acts and is identified, for this reason, as holding a CCF and thus as continuously targetable. Some have considered the ICRC's rejection of the conduct-based membership to be only “apparent” as “adopting either the ICRC approach or the continuous direct participation approach would probably mean that a person who takes part in a series of hostile acts would lose immunity from targeting, even beyond the specific occasions when he acts”.⁴⁴ While it is true that a CCF test can lead to identifying a person who engages in a series of hostile acts as a (continuous) legitimate target, one cannot ignore all the other scenarios where no direct engagement in hostile acts is witnessed and the CCF test is positive. These scenarios will appear legally justifiable (even if debatable from a normative standpoint) for as long as the CCF test can be positive without evidence of direct participation in violent acts. The following section will reveal how problematic this can be for the civilian population.

Continuous combat function in action: Targeting practices

The arrival of this open-texture CCF in the counterterrorism context has marked the individualization and de-materialization of the way States make targeting decision under IHL. Indeed, State targeting practice in this context shows that CCF allows the targeting of individuals when and where hostilities are not ongoing on the basis of information gathered about the target that do not necessarily include their witnessed direct engagement in hostile acts. Targeting practices thereby are not necessarily based on material circumstances (hostile acts) and are focused on personal features and behavioural patterns of the target. In this context, empirical evidence studied here brings to light what is at stake in the definition of DPH as CCF and makes its impact on targeting choices and scope very tangible: first, targeting based on evidence other than (witnessed) participation in hostile acts presents a higher error risk than when an individual is seen participating in combat; second, the spatio-temporal scope for targeting is considerably extended as CCF opens targeting options beyond moments and areas of active hostilities. While this is exactly what the CCF sought to achieve, by mirroring the risk to which States’ armed forces are exposed, there are concrete cases showing the pitfalls of such an extended window for targeting. The point is not to say that these cases reveal what the CCF test necessarily leads to but what it can lead to if CCF is attributed without any witnessed participation in hostilities.

The practices of the United States and increasingly of France, in the context of the armed conflicts they fight against groups they designate as terrorist, show that a CCF rationale practically aligns not only with personality strikes, where the identity of the target is known, but also signature strikes. Signature strikes consist of firing at people whose identities are not known, based on evidence of suspicious behaviour.⁴⁵ To understand the role played by the CCF in framing targeting operations, looking at the US Executive Order of July 2016, which remains in force until the date of writing, is helpful. It explains that:

The United States considers all available information about a potential target's current and historical activities to inform an assessment of whether the individual is a lawful target. For example, an individual may be targetable if the individual is formally or functionally a member of an armed group against which we are engaged in an armed conflict. As Administration officials have stated publicly, to determine if an individual is a member of an armed group, we may look to, among other things: the extent to which the individual performs functions for the benefit of the group that are analogous to those traditionally performed by members of a country's armed forces; whether that person is carrying out or giving orders to others within the group; or whether that person has undertaken certain acts that reliably connote meaningful integration into the group.

This Executive Order, in other words, allows the targeting of individuals who might not have been personally and formally recognized as being members of a non-State armed group party to a conflict, but rather on the basis of behavioural elements (which might or might not include witnessed participation in hostilities). The point here is not so much to affirm that all strikes analysed below were framed by the United States and France as justified under the CCF. Rather, the point is to show that the CCF test as currently defined does not prohibit such practices, which might be legal in some cases, but pose a high risk to civilians when they rest on contextual or behavioural elements other than their witnessed participation in hostilities.

Bringing to light an actual target selection process, a report by the New York Times tells us the story of the Razzo family, whose house was targeted on the evening of 20 September 2015 in Islamic State of Iraq and Syria (ISIS)-occupied Mosul, Iraq.⁴⁶ That day, Basim Razzo, his wife, his brother and brother's wife were heading to bed in the family house. Basim had spent his usual night hours of clicking through car reviews on YouTube. A strike hit them. Only Basim and his sister-in-law survived the strike. Later that same day, the American-led coalition uploaded a video to its YouTube channel titled “Coalition Airstrike Destroys Daesh VBIED Facility Near Mosul, Iraq 20 Sept 2015”. The target, the video explained, was the leader of a car-bomb factory, a hub in a network of “multiple facilities spread across Mosul used to produce VBIEDs [vehicle-borne improvised explosive devices] for ISIL's terrorist activities” posing “a direct threat to both civilian and Iraqi security forces”. The buildings that appeared in the recording were Basim and his brother's houses. That day, the evidence and intelligence gathered on Basim led to his identification as a lawful target. It is likely that his daily YouTube and internet activity, exploring cars, led the US drone authorities to identify him as the leader of a car-bomb factory. It could be argued that this targeting decision is not in flagrant violation of the CCF test as currently framed considering that the type of elements that can be taken to demonstrate “lasting integration into an organized armed group acting as the armed forces of a non-State party to an armed conflict” have not been defined, and as the ICRC Interpretive Guidance includes that a person can be targeted without having been witnessed engaging in hostile acts. Although it can be seriously contested that the evidence used demonstrates lasting integration, this event, as numerous others, should urge us to define the test more narrowly.

Another relevant situation is that of a US drone strike conducted on the evening of 16 March 2017, which targeted the Sayidina Omar Ibn Al-Khattab Mosque in Al-Jinah, in the province of Aleppo, Syria. Around forty dead were reported by the Syrian Observatory for Human Rights.⁴⁷ The Forensic Architecture team reported that after the strike, the US Central Command claimed responsibility for the airstrike on the building, which it had wrongly identified at first as located in the province of Idlib and as a “partially constructed community meeting hall” rather than a mosque.⁴⁸ The US officials further claimed that there were no civilian casualties. They then admitted that the building was in fact a mosque but maintained that they were “confident this was a meeting of al-Qaeda members and leaders; this was not a meeting of civilians”.⁴⁹ However, on 8 August 2017, a United Nations Syria Commission report concluded that US forces “lacked an understanding of the actual target, including that it was part of a mosque where worshippers gathered to pray every Thursday”, “failed to take all feasible precautions to avoid or minimize incidental loss of civilian life”, and was “in violation of international humanitarian law”.⁵⁰ In this case, although it cannot be affirmed with certainty, it is possible (also in light of the 2016 Executive Order) that the regular reunion of a group of people in a building triggered the decision to kill. What matters in this case study is not to conclude with certainty which considerations exactly led to conducting the strike but to emphasize that the CCF test should be better defined to leave no room for doubt that elements such as the regular reunion of a group of people in a building cannot be taken to derive lasting integration into an organized armed group. Of course, it can already be considered far-fetched to deduce lasting integration from such uncertain elements. Such an interpretation also goes against the goal pursued by the ICRC and scholarship when respectively creating and developing the notion. It is also arguably incompatible with other principles of IHL, such as the principle of precaution which requires that, in the conduct of military operations, constant care has to be taken to spare civilians and civilian objects.⁵¹ Still, it is urgent to make some efforts towards discussing and proposing ways to fill the gaps that have been left open.

These two episodes took place in Iraq and Syria, but the United States has also been conducting so-called signature strikes “outside areas of active hostilities”, in other words, where there is no ongoing situation of intense violence in the precise targeting location (justifying this on the basis of a –solid – legal interpretation of the geography of armed conflicts according to which conflicts follow their participants).⁵² On 17 March 2011, in the city of Datta Khel in North Waziristan, Pakistan, a US drone strike killed between forty to fifty people seated in circles.⁵³ US officials claimed that the men were all legitimate targets, stating that “these people weren't gathering for a bake sale. They were terrorists”,⁵⁴ and “not the local men's glee club”.⁵⁵ Yet, it was then established by Forensic Architecture,⁵⁶ whose work was later used by UN Special Rapporteur Ben Emerson,⁵⁷ that the men were in fact participating in a government-sanctioned meeting, and more specifically a jirga, a traditional assembly of leaders that makes decisions by consensus and serves as a dispute-settlement system.⁵⁸ Despite the posthumous identity check of the deaths, US government officials reiterated that no civilian casualties resulted from those strikes, which in their view successfully targeted a large group of heavily armed men.⁵⁹ The fact that a large group of men were seated in circles and appeared to gather and discuss together for two days and carried weapons led the United States to reach the conclusion that these men were legitimate targets. It is regrettable and dangerous, as it can lead to violations of IHL, including of the principles of distinction and precaution, that cultural and political local specificities are not systematically acknowledged when searching for CCF indicators: in some areas of high insecurity, open carrying of weapons can be quite common including for self-defence purposes.⁶⁰ Here again, one should ask if States should be able to strike non-State actors on the basis of elements other than witnessed participation in hostile acts, or if, on the contrary, the risk is too high for civilians.

On 3 January 2021, the Barkhane forces conducted a strike on a group of individuals located in the village of Bounty, in Mali.⁶¹ In the aftermath of the strike, locals and journalists reported that the operation resulted in numerous civilian casualties. France, for its part, declared that its troops had targeted “members of a terrorist armed group” in full compliance with IHL.⁶² This event, and the divergent views on how to characterize what happened, led the United Nations Multidimensional Integrated Stabilization Mission in Mali (MINUSMA) to open an investigation, which resulted in a report published on 30 March 2020.⁶³ The report concludes that twenty-two casualties resulted from the strike, among which nineteen civilian deaths and three members of the terrorist group Katiba Serma, a semi-autonomous arm of the jihadist group the Macina Liberation Front.

The casualties, including those identified as civilians by the MINUSMA, were the intended targets of the strike: France, both before⁶⁴ and after⁶⁵ the report, insists that only legitimate targets were hit. On 7 January 2021, the Ministry shared the following detailed information on what triggered the decision to strike (translated from French):

On Sunday January 3, in the afternoon, after several days of intelligence gathering, the Barkhane Force conducted an operation in the Douentza region, an area characterized by the presence and action of terrorist armed groups such as the Katiba Serma. […]

In this area, more than an hour before the strike, a REAPER drone had identified a motorcycle with two individuals at the north of RN16. The vehicle joined a group of about forty adult men in an isolated zone. All the intelligence gathered, including real time information, allowed us to identify this group as belonging to an armed terrorist group. The observation of the zone for more than an hour and a half also allowed to exclude the presence of women and children. Considering the behavior of the individuals, of the tools identified, and combining all collected intelligence, [the strike was ordered].⁶⁶

The MINUSMA report gets into interesting detail about the local tradition of having men and women spend part of the wedding celebrations in separate groups and concludes that only three of those men were presumably part of a violent non-State armed group.⁶⁷

Although the targets were not engaged in hostilities at the moment of the strike, it seems like their behaviour, age, sex and location were taken to reveal their role as part of the military arm of a terrorist group. Some elements of evidence might not have been released, and press releases from the French Ministry of the Armed Forces and the MINUSMA report might diverge on some elements, but one thing cannot be contested: the Barkhane forces gathered pieces of evidence other than witnessed participation in hostilities which, taken together, allowed them to identify the individuals as legitimate targets.⁶⁸ These pieces of evidence seem to have included the facts that all individuals were men, apparently military-aged, unusually gathered, in a region known for the presence of terrorist groups, and that some of them carried weapons.⁶⁹

These cases do not only bring to light the lethal consequences of flawed or outdated intelligence and of misperceptions caused by other factors, but also show that the data and evidence used to identify legitimate targets in armed conflicts against a party designated as a terrorist group include criteria such as, but not limited to: geographic location of individuals or buildings in an area of a city or region identified as concentrating terrorist activities; regular male gatherings; group gatherings; group conversations; mobile phone communications or connections (the target does not need to have expressly communicated with an identified member of a terrorist group to trigger suspicion, as attempts of communication are also intercepted); internet activity (e.g. related to items associated with any terrorist group's activities); possession of weapons. In an era where knowledge data collection capacities are unprecedented and constantly growing, from current drone surveillance to the promise of their algorithmic capacities in the near future, the CCF test that is not conditioned upon witnessing someone engage in hostile acts exacerbates the risks posed by anticipatory warfare to civilian populations.

Concluding remarks and proposition

Rethinking the CCF notion and test in light of French and US strikes against members of terrorist groups shows that extensive practices cannot be robustly constrained by the CCF test as currently defined. However, this practice puts civilians at great risk and completely deviates from the goals pursued by the drafters of the Interpretive Guidance and those cherished by the IHL scholars whose work has been discussed here. As a result, this paper hopefully showed that the CCF test needs to be collectively reframed and further developed if we want it to perform a constraining function on target selection, including against members of terrorist groups in NIACs. Because the CCF test proposed in 2009 builds on evidence other than witnessed engagement in combat and thus sources of information that could be partial or ambiguous, the CCF architecture creates a particularly high risk of error when mobilized against transnational non-State armed groups which are dispersed and whose modes of action make the identification of members of the group more difficult than in traditional non-State actors. It also arguably results in heavier consequences in such conflicts than it would in State-centred non-State armed groups, as it opens the door to geographically unlimited conflict.

A clearer and narrower version of the CCF test should be identified to avoid such extensive targeting practices. The above developments show that the only solution to do so is by including in the CCF test evidence of repeated direct engagement in violent acts that cause harm (in support of a party to the conflict and to the detriment of another). The test could also include, but only in addition to evidence of repeated engagement, behavioural indications of CCF. Such a CCF test would strike a good balance between the willingness to extend the scope for targeting members of terrorist groups beyond moments of active hostilities, while simultaneously not lowering the threshold of civilian protection. To the concern that this version of CCF might put civilians more at risk than the punctual attribution of DPH, two elements should be shared. First, the present paper hopefully shows that there is a less dangerous version of CCF than that which does not require repeated participation. Second, these concerns can be addressed by supplementing the CCF test with additional requirements. One could think of the following formula: according to a reframed CCF test, not only would participation need to be identified, but State forces would have to corroborate this information with additional evidence to support the apparent function of continuous combatant. Such a change would have the added value of consistency by requiring a higher evidentiary threshold when targeting spatio-temporal possibilities are extended (beyond “unless and for such time as” someone is witnessed as engaging in combat), instead of setting a lower threshold when extending the targeting window. In any event, I am certain that other scholars will have sharper ideas to revisit the CCF test. My hope is that this contribution raises awareness about the need to do so in light of existing targeting practices against members of transnational non-State armed groups designed as terrorists.