What type and degree of organization is required? The need for fighting forces and accountability structures as minimal requirements

IHL treaties do not establish the threshold for a NIAC in the sense of Article 3 common to the Geneva Conventions, nor do they define the notion of an “organized armed group”. Some elements of the concept may however be inferred from the treaties, namely from common Article 3 itself and from Additional Protocol II to the Geneva Conventions (AP II).

Common Article 3, which is applicable to all NIACs,Footnote ²¹ does not employ the term “organized armed group”. It simply states that “[i]n the case of armed conflict not of an international character occurring in the territory of one of the High Contracting Parties, each Party to the conflict shall be bound to apply, as a minimum, the following provisions”.Footnote ²² The term “Party” has been interpreted as requiring the existence of organized actors, be they States or non-State actors.Footnote ²³ The ICRC Commentary on Geneva Convention I noted already in 1952 that one possible criterion for determining the existence of a NIAC is that “the Party in revolt against the de jure Government possesses an organized military force, an authority responsible for its acts, acting within a determinate territory and having the means of respecting and ensuring respect for the Convention”.Footnote ²⁴

In its case law,Footnote ²⁵ the ICTY further clarified the content of the criterion of organization. The Tribunal never developed strict standards as to what elements a group would have to meet to be considered an OAG, but simply used a list of indicative factors to that end.

These factors could be classified into two broad categories. The first category, which could be called structural elements, relates to a group's internal structure. These elements include indicators such as the existence of a responsible command capable of exercising authority and direction over the forces in the accomplishment of the mission; of an organizational chart; and of internal rules, including disciplinary rules and standards. The second category – let us call them operational elements – is important for the group's operations: its ability to recruit and train new fighters, to control territory, to set up headquarters, and to launch operations bringing together different units. Some indicators, such as a group's ability to “speak with one voice”, may be both structural and operational.

The ICTY has highlighted that these factors are neither cumulative criteria nor a checklist for demonstrating the existence of the organizational threshold – they are simply indicators.Footnote ²⁶ Thus, organization cannot be determined in abstracto but requires analysis on a case-by-case basis.

It has rightly been argued that this approach provides little added value for determining the existence of a NIAC. The jurisprudence of the international criminal tribunals is of course fact-driven, but which, or how many, factors need to be present in order to classify a situation of armed conflict remains totally open and unspecified.Footnote ²⁷ In jurisprudence subsequent to the Tadić case, the ICTY never made an effort to clarify its underlying “classification logic” – namely, why it considered specific factors as relevant and others as not, or whether all of those used carried the same weight in the Tribunal's deliberation. For example, the Trial Chamber in the Haradinaj case made no reference at all to principles underlying its classification of the conflict between Yugoslav forces and the Kosovo Liberation Army (KLA) in the southern Serbian province of Kosovo in 1998–99: it simply regarded the situation ex post facto and took up certain elements of that situation in order to say that it had indeed amounted to a NIAC.Footnote ²⁸ These elements, the Tribunal found, were also present in the conflicts in Bosnia and Herzegovina and Croatia, and finally in Kosovo itself.Footnote ²⁹ Therefore, because certain indicators existed in one situation which was described as an armed conflict, their presence elsewhere implied that armed conflicts existed in those territories as well. Such a methodology – evidently based on a comparison between various contexts – does not provide clear guidance for classification purposes in contexts which are very different from the ones examined by the ICTY, but which may nevertheless amount to NIACs. Particularly compelling in this regard is the argument that the approach used by the ICTY to classify “conflicts of high intensity … provides no meaningful guidance in assessment of low-intensity conflicts”.Footnote ³⁰

Another unknown element is the extent/degree of organization required. Should the threshold of organization be high (as for the threshold of intensity), or could armed actors with some rudimentary organization be considered as OAGs for the purposes of IHL? The phrase “minimum” or “minimal level of organization” is occasionally employed, for example in the Boškoski case,Footnote ³¹ although it does not seem to have ever become a proper term of art. Even if this concept was taken as a yardstick, it may be given two different meanings: it may set a low threshold, accepting for an OAG any level of organization higher than that of a mob; or it may be seen, to the contrary, as setting a higher standard, namely if “minimum” were interpreted as referring to a certain level of organization, which is not necessarily low.

In its early jurisprudence, the ICTY seemed to adopt a reasoning that was congruent with the latter meaning. In the Appeals Chamber judgment in the Tadić case, the Tribunal evoked the notion of “an organised and hierarchically structured group, such as a military unit or … armed bands of irregulars or rebels”, that has “a structure, a chain of command and a set of rules as well as the outward symbols of authority”.Footnote ³² In later jurisprudence, though, the ICTY seemed to be less demanding. In the Limaj case, for instance, the ICTY adopted a very flexible approach to qualify the KLA as an OAG despite the fact that the “organisational structure and the hierarchy of the KLA were confusing”.Footnote ³³ Similarly, in the Haradinaj case, the Tribunal found that the KLA did not possess disciplinary rules and mechanisms,Footnote ³⁴ and that it had developed “a mainly spontaneous and rudimentary military organization at the village level”, but that it nevertheless satisfied the requirement of an OAG for the purposes of IHL.Footnote ³⁵

Of related note is the ICTY's reluctance to cast the criterion of organization in light of the capacity to respect IHL; this is particularly visible in its handling of the KLA's lack of disciplinary mechanisms. In fact, when the defence in the Limaj case raised that “a party to a conflict must be able to implement international humanitarian law and, at the bare minimum, must possess: a basic understanding of the principles laid down in Common Article 3, a capacity to disseminate rules, and a method of sanctioning breaches”, the Chamber explicitly rejected this argument, finding that “some degree of organisation by the parties will suffice to establish the existence of an armed conflict” and that “[t]his degree need not be the same as that required for establishing the responsibility of superiors for the acts of their subordinates within the organisation”.Footnote ³⁶ A similar approach to the one taken by the Limaj and Haradinaj chambers was subsequently applied by the chambers of the International Criminal Court (ICC) in the Lubanga case, when the Trial Chamber rejected the argument that an armed group needed to be under “responsible command” (no mention is otherwise made of the group's capacity to respect IHL)Footnote ³⁷ – all that was necessary was for the group to be “sufficiently organized” in light of factors that were “potentially relevant”.Footnote ³⁸

The “minimal level of organization”, thus understood, does not seem to have much precedent in earlier practice and discussions, nor is it supported by the International Committee of the Red Cross's (ICRC) Commentaries to the Geneva Conventions. According to the latter, the organizational element of any non-State armed group, in order for that group to be considered a party to an armed conflict, must be such as to allow it to implement IHL, a “fundamental criterion which justifies the other elements” of the NIAC definition.Footnote ³⁹ This understanding is commonly restated in doctrine, and is logically necessary for the application of common Article 3.Footnote ⁴⁰ Thus, “[w]hile the group does not need to have the level of organisation of state armed forces, it must possess a certain level of hierarchy and discipline and the ability to implement the basic obligations of IHL”.Footnote ⁴¹

We therefore disagree with the ICTY's approach in Limaj and Haradinaj, which in our view does not reflect older and contemporary doctrine, nor indeed the Tribunal's own earlier position in Tadić. We are not saying that the Kosovo War did not amount to a NIAC: we simply disagree with the ICTY's methodology. This methodology is contrasted by the one chosen by the ICC Trial Chamber in Katanga, which found that “‘organised armed groups’ must therefore have a sufficient degree of organisation to enable them to carry out protracted armed violence and to implement the provisions of humanitarian law applicable to that type of conflict”.Footnote ⁴² This is the better view, one which is also supported by doctrine.Footnote ⁴³

We submit that any analysis of whether a particular group meets the criterion of organization must be teleological – namely, looking at the object and purpose of IHL rules – and not simply comparative. There are in fact two main reasons why the “organization” of armed groups is key from an IHL perspective.

1. 1. The first reason is to distinguish proper armed conflicts from “isolated”, “sporadic” or otherwise less serious acts of violence performed by unorganized criminal bands.Footnote ⁴⁴ The armed group needs to be organized with a view to conducting hostilities or a continuum of attacks (the ability to launch a “continuum of attacks” seems to us the most precise understanding of the capacity to engage in hostilities, as the group should be able to engage in more than just an isolated incident of violence and should rather be capable of engaging in recurrent and prolonged violence). This is supported by the ICRC's recent study on The Roots of Behaviour in War, which notes that “[a]ll armed groups capable of launching operations with some semblance of a military character have structures of one kind or other – one or more leaders and degrees of organization which, though they may vary, exist and need to be identified”.Footnote ⁴⁵ To engage in a continuum of attacks, the armed group must possess the ability to launch military operations bringing together different units and to confront enemy armed forces in the context of proper armed clashes, and this ability is illusory without proper organization. In other words, there must exist an armed wing whose function is to regularly conduct hostilities. The criterion of “continuous combat function” developed by the ICRC to determine individual membership in an OAG is thus also useful to determine whether an OAG exists in the first place.Footnote ⁴⁶

2. 2. The second reason is to make sure that the group would be able to respect IHL.Footnote ⁴⁷ The existence of an accountability mechanism allows the group to respect IHL. It has been argued by some authors that the existence of a disciplinary mechanism and the ability to implement IHL is merely an indicative element of organization, and not a necessary aspect of an OAG.Footnote ⁴⁸ We disagree with this view. While it is not required of a group to actually respect IHL in order for that body of law to apply, what is required is the existence of some accountability in the group for the acts of its members in general. Such “accountability” could exist even if the group's fighting forces are given a great deal of initiative in their accomplishment of various missions, but at the very least the armed group, under responsible command, exercises sufficient control over the acts of its members for those acts to be considered as its own. Such an accountability mechanism need not be exclusive to ensuring accountability for IHL violations: indeed, the group might rather be interested in punishing disloyalty or desertion, but the same mechanism could in principle be applied for breaches of IHL. This goes hand in hand with the previous requirement for the existence of sufficiently regular/stable fighting forces (an “armed wing”). Volatile groups whose members participate in hostilities in a sporadic manner, and which possess no accountability mechanisms at all, should not be considered OAGs.

In brief, it is only if a group is capable both of launching a continuum of attacks and ensuring the accountability of its members that it may be considered sufficiently organized for the purposes of IHL. These are in our view minimum requirements and not mere indicative factors.

What lessons can be derived for classifying groups designated as terrorist? The relevance of the dichotomy between the networked and hierarchical models of terrorist organizations

Linking these theoretical considerations more specifically to designated terrorist groups, it is first pertinent to recall that scholarship commonly distinguishes between “networked” and “hierarchical” models of terrorist organization. Thus, Gunaratna and Oreg (as well as a number of other authorsFootnote ⁴⁹) distinguish between groups with a “command-cadre (hierarchical) structure” and those with a “network structure”:

Hierarchical groups’ functioning may more closely resemble that of armed forces,Footnote ⁵¹ while networks employ “intermediaries to keep cells isolated from each other but in communication with leadership”.Footnote ⁵² Networked organization allows for greater group resilience to detection and elimination by the authorities, but only a hierarchically organized group is capable of carrying out “complex attacks such as the 11 September attack on New York and Washington, DC”.Footnote ⁵³ Although there are different types of terrorist (and, in general, criminal) networking, networks typically have a core (hub) “characterized by dense connections among individuals who, in the case of a directed network, provide the steering mechanism for the network as a whole”, and a periphery featuring “less dense patterns of interaction and looser relationships than the core”.Footnote ⁵⁴

We submit that this dichotomy is pertinent when classifying armed conflicts as well, although the structural type (hierarchical or networked) should be seen as indicative of a group's organizational capacity rather than as dictating it. Thus, hierarchical organizations will sooner conform to the notion of “organized armed group” than networks because, essentially, they will be better equipped to launch a continuum of attacks and to ensure the accountability of their members. This dichotomy is similar to the one identified by the ICRC between centralized and decentralized groups in its study The Roots of Restraint in War. Footnote ⁵⁵ This study further demonstrates that decentralized groups have looser coordination in military planning and operations and few observable signs of military discipline.Footnote ⁵⁶ In other words, they are less capable of launching a continuum of attacks and of ensuring the accountability of their members – both of which are necessary to fulfil the two minimum requirements of organization described above.

The hierarchical–networked dichotomy is not a strict categorization but rather a spectrum: some groups may lean more to the poles, while others may be in between, or their internal factions/branches may all levitate between the two models. Some groups exhibit characteristics that do not properly conform to either model, but this is the exception rather than the rule. Crossover types of groups are not exclusive to designated terrorist groups; this is the case, for instance, for the Free Syrian Army (FSA), now known as the Syrian National Army (SNA). The UN Independent International Commission of Inquiry on Syria found that a NIAC erupted in Syria between November 2011 and August 2012, but at the time of its initial report, when some OAGs such as the FSA had already been founded, the Commission was not able to conclude that the intensity and organization criteria had been met as of March 2011.Footnote ⁵⁷ Some months later, it determined that a NIAC had come into existence on the basis that Syrian opposition forces were now engaged in sustained military operations against the government (including in the capital), had organized local military councils claiming leadership over fighters in an area, and had access to sufficient weapons, funding and logistics.Footnote ⁵⁸ And yet, it does not appear that the FSA/SNA ever achieved a strong hierarchical model or managed to establish a command structure, despite attempts to that effect.Footnote ⁵⁹ The opposition remains “fractious and deeply divided”,Footnote ⁶⁰ and the group itself is described by one commentator as having “no rules and no military or religious order. Everything happens chaotically. … The FSA lacks the ability to plan and lacks military experience.”Footnote ⁶¹ Assuming that the FSA/SNA is indeed a single armed group (as most media tend to do), then it must be seen as a decentralized one made of several structured factions.Footnote ⁶² In such a case, the question of whether such a group is an “organized armed group” for the purpose of IHL may be disconcerting given the absence of an internal disciplinary system.Footnote ⁶³ The better view, however, is perhaps to consider the FSA/SNA as a network of smaller groups which are merely nominally affiliated to the FSA/SNA and are fighting for a common cause.Footnote ⁶⁴

In any case, the existence of groups which are difficult to classify on the networked/hierarchical spectrum does not detract from the latter's usefulness as a tool for conflict classification. It would therefore be wise to develop tools – with the help of social scientists – to better understand the structure and functioning of various armed non-State actors and to classify them along the lines of the hierarchical and networked models. This would ensure a more scientific approach to conflict classification while keeping the criterion of organization as the focal point of the exercise. Such classification would need to be conducted over time for each group, as a group may (d)evolve from a hierarchical to a networked model, typically to avoid detection and to act more clandestinely. Obviously, the characterization of such a group as an OAG can and should equally evolve overtime.

Although understanding the structure of clandestine organizations designated as terrorist groups is difficult, a convincing IHL analysis cannot be done without adequate consideration of the structure and functioning of armed non-State actors and their evolution over time. A combined use of minimal requirements for determining the existence of the organization criterion and of typologies of various armed non-State actors would facilitate legal classifications and help predict their evolution.

Consider the structure of Al-Qaeda in the Arabian Peninsula (AQAP) or IS. In its heyday, the latter even claimed statehood and worked hard, in addition to its military efforts, to provide services to the local population in lieu of the Syrian and Iraqi authorities; it began organizing to this effect at least as early as 2006.Footnote ⁶⁵ The expansion of IS and the establishment of its authority over a number of settlements in Iraq allowed the group to build up functioning public services that included oversight of electricity supply, health spending, education, street cleaning and food provision.Footnote ⁶⁶ Its military wing was organized by experienced foreign jihadists who had fought in other conflicts, such as in Chechnya, and who had perhaps even served in State armed forces before joining IS; these individuals created a centralized chain of command to manage the IS paramilitary force,Footnote ⁶⁷ which in 2014 may have numbered as many as 31,000 fighters in Iraq and Syria.Footnote ⁶⁸ The organization kept detailed information on its incoming and permanent fighters, their salaries and duties, and their permission and reasons for leaving, leading Shapiro and Siegel to conclude that “[a]ny human resources manager would want to capture that information”.Footnote ⁶⁹ There can be no doubt that IS, a designated terrorist group,Footnote ⁷⁰ met the requirements for an OAG under IHL in 2015 (and allegedly well before thatFootnote ⁷¹): this is because it was able to mount sustained attacks and engage in concerted military action guided by a clear strategy. The IS military wing possessed an internal structure and disciplinary system which would allow it to ensure respect for the rules of military discipline – but also IHL, had the responsible command so desired.

This structure of IS is thus very different from that of other jihadist groups, such as the Pakistan-based Al-Qaeda in the Indian Subcontinent (AQIS). AQIS is a smaller group, a loose collection of mujahedeen across the Indian subcontinent which is affiliated to Al-Qaeda Central and which has also pledged allegiance to Mullah Omar, former head of the Afghan Taliban. It does not engage in sustained military action but rather in more “traditional” terrorist conduct such as assassinations and hijackings.Footnote ⁷² The violence projected by such a group can be devastating, but this alone does not make it an OAG under IHL. Although information on the internal functioning of AQIS is admittedly scarce, it seems to resemble that of older terrorist groups such as those functioning in Europe in the second half of the twentieth century (e.g., the Red Army Faction), which demonstrated a complete lack of a military (hierarchical) structure and were not considered armed groups even though they projected high degrees of violence.Footnote ⁷³

Designated terrorist groups also tend to evolve very quickly, and such a group may still be generally perceived as a single OAG even after its structure has changed considerably. For instance, in Nigeria, after the death of Boko Haram's founder Mohammed Yusuf in 2009, the group split into different factions, of which there were as many as six in 2014, answering to different leaders with limited control over their many cellsFootnote ⁷⁴ (some of these later pledged allegiance to IS while others did not, and Boko Haram as a whole seems to have lost its ability to “speak with one voice”).Footnote ⁷⁵ It is difficult, in our view, to speak of a single “organized” armed group if the group breaks up into different factions acting separately and using different means.Footnote ⁷⁶

There are thus various reasons why the ideological, strategic, tactical and organizational workings of designated terrorist groups in principle do not correspond to the classical hierarchical OAG model. IS and other terrorist OAGs such as the group's predecessor, Al-Qaeda in Iraq, are therefore quite exceptional in the milieu of groups labelled as jihadist and/or terrorist. Most of them are either decentralized or should be considered as an assemblage of various smaller armed groups.

Moreover, no sufficient evidence has been produced to demonstrate that either Al-Qaeda or IS are worldwide, unitary, transnational OAGs. Should terrorist attacks conducted by IS cells or adherents around the world be considered part of the same “continuum of attacks” as the conflicts involving the group in Iraq or Syria? Do their perpetrators find themselves within the same chain of command as the IS fighters in the Middle East? The answer is probably no, and simply sharing an ideology is not enough to replace proper organizational links as required by IHL.

It may be tempting for States to consider jihadist networks such as Al-Qaeda or IS as forming part of the same OAG, but a deeper analysis shows this to be little more than a fiction, one which allows the use of more permissive rules on targeting than those governing simple law enforcement operations.Footnote ⁷⁷ Some OAGs may also pledge allegiance to – or pretend to belong to – a broader terrorist network, such as IS, in order to attract financial resources and manpower, even though such links are not (yet) in existence. Such a phenomenon has been recorded in the Philippines in relation to the Maute group pretending to belong to the Islamic State of Iraq and Syria (ISIS) without any actual linkages to this terrorist network.Footnote ⁷⁸ In any case, the mere fact that various jihadist groups share a common ideology and have loose connections with one another is not enough to consider that they together constitute an OAG for the purposes of IHL. Certainly, there is nothing in IHL to limit OAGs to a particular State or geographical area,Footnote ⁷⁹ and to the extent that IS operates as a single entity in the territories of Iraq and Syria, notably, it clearly qualifies as a “transnational” armed group.Footnote ⁸⁰ What matters is that the group is able to perform attacks of a sufficient magnitude and which are linked with one another in a single continuum, and that the responsible command/management exercises sufficient control over the acts performed by subordinates. Only such parts of a broader organization/network that correspond to a group which can perform attacks with sufficient magnitude and in a single continuum can be equated with OAGs for the purposes of IHL. In other words, the ISIS/Al-Qaeda networks are to be understood as ideological/political organizations that contain various distinct groups, some of which correspond to OAGs for the purposes of IHL.

What degree and type of violence is required? The existence of proper armed clashes and the necessity of moving away from law enforcement means as a benchmark

Besides organization, the other criterion used to establish the existence of a NIAC is that of intensity. Violence between a State and individuals under its jurisdiction is a matter of domestic law and IHRL “as long as such violence does not rise to a certain level”.Footnote ⁸¹ The intensity requirement should therefore be understood as an indicator that a certain situation can no longer be resolved through law enforcement means, allowing governments to use force under the conduct of hostilities paradigm.

The ICTY has considered a number of factors to be indicative of the existence of a level of violence sufficiently intense for fighting involving non-State armed groups to be described as a NIAC. For instance, in the Haradinaj case, it first recalled that the criterion of “protracted armed violence” as discussed in the Tadić case referred “more to the intensity of the armed violence than to its duration”Footnote ⁸² (the duration of a conflict may, in any event, only be gleaned retrospectively,Footnote ⁸³ and international bodies have not given it much relevance in their practice),Footnote ⁸⁴ before providing a non-exhaustive list of indicators to be taken into consideration when evaluating the intensity of armed violence. These included

As mentioned earlier, these indicators are depicted as indicative and not determinative of a NIAC. Contrary to the organization criterion, there appears to be consensus that the level of intensity needs to be high. In terms of classification, this is one of the distinguishing features between the IHL of international armed conflicts (IACs) (where there is no threshold of violence, at least according to the “Pictet theory”Footnote ⁸⁶) and the IHL of NIACs (where the level of intensity needs to be highFootnote ⁸⁷).

While the focus is generally put on degree, more thought should be given to the type of violence required. This matter is not discussed in IHL treaties, other than to exclude internal tensions or disturbances from the category of NIACs.Footnote ⁸⁸ However, the fact that the violence should be of a specific nature is implicit in the ICTY's list of factors.

All of the ICTY's indicators focus on the military, and not law enforcement, nature of the State's response, implying that the State in casu is no longer capable of containing the situation, which has consequently escalated into an armed conflict. This view is echoed in the ICRC's position on the scope of application of common Article 3 – namely, that the threshold of intensity may have been reached “when the hostilities are of a collective character or when the government is obliged to use military force against the insurgents, instead of mere police forces”.Footnote ⁸⁹

Terrorist violence can doubtless leave a horrifying number of casualties, but this does not in itself qualify it for regulation by IHL: the violence must be of a specific nature and not just of a specific degree. We propose that two aspects be taken into consideration when classifying a situation from an intensity perspective: whether the violence under examination amounts to armed clashes, and whether those armed clashes reach such a degree that they may no longer be resolved through law enforcement means.

1. 1. The “armed clashes” requirement has been raised several times in doctrine.Footnote ⁹⁰ The existence of such clashes is a necessary requirement for a NIAC to arise. We understand “armed clashes” to mean an exchange of violence or individual confrontations between organized parties. The violence cannot be committed only by one side; it must be projected mutually, which requires an exchange of violence between at least two sides. The ICRC's 2019 Challenges Report states that “confrontations must take place between at least two organized parties” for a NIAC to arise.Footnote ⁹¹ Unilateral violence, however shocking, does not amount to an armed conflict. Although arguments have been made in favour of a unilateral trigger for a NIAC,Footnote ⁹² this does not conform to the traditional understanding of an armed conflict and could lead to an over-classification of such situations.

2. 2. The armed clashes (and related violence) must achieve a sufficient level to amount to an armed conflict. In cases of violence between a State and a non-State actor, which are primarily governed by the State's human rights obligations, this means that a resort to military force and conduct of hostilities rules must be justified by the State's law enforcement mechanisms no longer being able to maintain order in the relevant territory.Footnote ⁹³ This analysis presupposes an underlying principle of necessity, which is both objective and situational. It is objective in the sense that it depends not on the State's will but on whether, as a matter of fact, its law enforcement authorities can be expected to cope with the matter. It is situational as it involves an in situ analysis, and not just a theoretical capacity-based analysis. This means that even States with very powerful law enforcement and police forces can become a party to a NIAC if in a part of their territory (or when acting abroad) they have no choice but to take more severe measures, such as deploying the military.Footnote ⁹⁴

The lack of a central monitoring body responsible for independent conflict classification becomes particularly problematic with respect to internal tensions and disturbances, namely “other situations of violence” (OSVs) that could evolve into NIACs. It is often practically impossible to pinpoint the exact moment when OSVs become NIACs, even though this point entails a crucial shift of legal frameworks. This problem is compounded by the fact that international jurisprudence typically does not take into consideration the reasons for which armed forces and military-grade weapons are deployed to contain an OSV, and simply takes these as retrospectively indicative of an armed conflict. Yet States should not resort to such measures unless it seems impossible to cope with the situation by means of law enforcement – otherwise, the door could be left wide open to abuse as a State could take wartime measures and claim that the conduct of hostilities paradigm applies as a direct result of its own excessive measures. In other words, a State cannot “create” the required level of intensity if the situation does not require it. It is therefore crucial to consider whether such measures were necessary in the first place, rather than simply being taken in order to determine the beginning of a NIAC. It is only when the bilateral violence has reached such an intensity that law enforcement measures have proven inadequate, or would be inadequate, to maintain or restore order that a situation of violence may be considered a NIAC and may allow the State to resort to the conduct of hostilities.

Specific problems may arise in relation to NIACs taking place between OAGs with no State involvement (“non-State NIACs”). States, and not armed groups, are obliged by international law to ensure law and order in a given territory, and the principle of necessity related to the maintenance of order discussed above would therefore not make sense when applied to a non-State actor directly. However, NIACs usually take place within the territory of a State,Footnote ⁹⁵ and regardless of whether that State is a party to the armed clashes in question, the conflict classification must take place on the basis of its own capacity to maintain order. Gasser writes that “[a]nother case [of NIAC] is the crumbling of all government authority in the country, as a result of which various groups fight each other in the struggle for power”Footnote ⁹⁶ (the implication is that the “non-state NIAC” emerges in the absence of State authority, not in spite of it). Thus, whether armed clashes between rival drug cartels in Mexico or Colombia have reached the threshold of an armed conflict cannot be considered in isolation from those States’ ability to contain the clashes. Non-State, “private” violence is normally much better regulated by domestic law and IHRL than by IHL.

In the counterterrorism context, a specific problem may arise when States act against terrorist groups abroad. Unless the intervening State is joining a pre-existing NIAC on the side of the territorial State (in which case the criterion of intensity will have already been fulfilled for the “domestic” NIAC), the requirement to exhaust possibilities for maintaining order may be problematic for the foreign State. For example, State A can only be in a NIAC against armed group X in the territory of State B (which is itself a non-belligerent) if the armed clashes are of a sufficient degree for law enforcement means to no longer be able to resolve the matter. The fact that X is not situated in A's territory will have to be taken into account for establishing whether the criterion of necessity – as defined above – is met (namely, A is probably not in a position to maintain order in B).Footnote ⁹⁷ However, there still need to be armed clashes between A and X, which is to say bilateral violence. If A were to simply launch air strikes or drone strikes against X, which is not capable of defending itself against such strikes, such unilateral force will be inadequate to be considered a NIAC. The situation may be different if B invites A to take care of X in its stead: assuming that B is not already involved in a NIAC, it would be absurd not to consider B's capacity to maintain order in its own territory when determining the intensity of the conflict between A and X. This would create a loophole wherein it would be possible for intervening States to engage in conduct otherwise prohibited to the territorial State.Footnote ⁹⁸ Nevertheless, the existence of such an invitation should probably be taken as indicative that the necessity criterion has been met (although it should not be definitive).

We recall that the above arguments are not imported from jus ad bellum but are rather derived from the necessity of demonstrating that law enforcement measures are inadequate or unavailable before considering IHL applicable. If States were to wage war against armed non-State actors without demonstrating the necessity of doing so (e.g., when the group is disorganized or does not project a sufficient level of violence), such use of force would almost certainly amount to a serious violation of human rights law. IHL should not be considered applicable before there exist violent armed clashed between two or more parties. The application of IHL should never become a reward for States opting to employ extreme measures, but can only ensue from an objective state of necessity in a given moment.

To summarize, the simple enumeration of indicative factors by the jurisprudence is not sufficient to ascertain whether the criterion of intensity has been met in each situation. The intensity criterion should be proven considering the existence of armed clashes and the related necessity of resorting to hostilities because law enforcement measures would be inadequate or have proven to be inadequate.

Who should project violence, and can intensity be accumulated in cases of alliances or coalitions between armed groups labelled as terrorist? Assessing theories pertaining to “co-belligerency”, the “support-based approach” and “cumulated/aggregated intensity”

Numerous jihadist groups around the world function with a certain degree of cooperation and mutual assistance. This does not necessarily make them part of the same supranational group, but it does raise the problem of whether such cooperation can impact the classification of a particular situation of violence. Jihadist groups are often perceived as entertaining various levels of interaction with other factions espousing a common ideology. These dynamics can become very intricate. For example, the Mali-based Nusrat Al-Islam, described as an Al-Qaeda “affiliate”, is apparently a merger of several other jihadist groups in the region, most of which are also former individual “affiliates”; the group's leaders apparently swore oaths of fealty to Al-Qaeda leadership following the merger.Footnote ⁹⁹ Depending on the relationships between its constituent elements, Nusrat Al-Islam may be either a unitary group or a coalition of different groups (as understood by Nikolić, de Saint Maurice and Ferraro, who describe it as an “umbrella organization”: a coalition of “various armed groups operat[ing] with different structures and motives”Footnote ¹⁰⁰). Nusrat Al-Islam has also been active in Burkina Faso, where it has allegedly been “cooperating” with the Islamic State in the Greater Sahara (ISGS) even though IS and Nusrat Al-Islam are rivals elsewhere.Footnote ¹⁰¹ In Libya, various jihadist and Islamist militias banded together in the Shura Council of Benghazi Revolutionaries, a coalition of separate groups with a common purpose of defending themselves against anti-Islamist operations by other groups.Footnote ¹⁰² It is less common for non-jihadist groups to be perceived as working in a coalition, but such examples do exist, for example in Mali, where a coalition of nationalist groups works under the aegis of the Coordination of Azawad Movements.Footnote ¹⁰³ Therefore, although there is a tendency to link designated terrorist groups into coalitions (a term that is seldom defined by authors), inter-group cooperation in any form is not restricted to such groups. Our analysis here, therefore, has broader implications than just counterterrorism.

Legally, the question boils down to the following: if State A is in a NIAC with armed group B, and B cooperates with group C, could there ipso facto be a NIAC between A and C (and in the absence of direct clashes between A and C)? If our answer to this question is yes, then we can only conclude that the intensity of violence between A and B may be extended to cover C as well. There are several strains of thought arguing for such a possibility in doctrine.

We may call the oldest such position the co-belligerency approach. Scholars of the law of armed conflict rarely feel compelled to define “co-belligerency”, which is certainly an old IHL concept. Traditionally, “[c]o-belligerency is a concept that applies to international armed conflicts and entails a sovereign State becoming a party to a conflict, either through formal or informal processes”.Footnote ¹⁰⁴ The concept has been used historically to describe members of a common alliance (which may be formal or informal) vis-à-vis a specific armed conflict wherein States could be considered belligerent parties regardless of whether or not they actually took part in hostilities (for example, one or more co-belligerent States may simply have declared war against the common enemy without themselves engaging in battle). Geneva Convention IV refers to the concept when establishing who may become a protected person when in the hands of a belligerent party.Footnote ¹⁰⁵ As a relation of co-belligerency between States may usually be inferred from formal treaties, declarations and other public sources, international law does not spell out the conditions for co-belligerency.

It is very controversial whether the concept can be applied by analogy in a NIAC, in particular as regards relationships between different armed groups. The US judiciary has notably understood the term “associated forces” to be synonymous with co-belligerency and has applied it to OAGs. In Hamlily v. Obama, the Columbia District Court accepted that “[a] ‘co-belligerent’ in an international armed conflict context is a state that has become a ‘fully fledged belligerent fighting in association with one or more belligerent powers’”, and proceeded to apply the concept to associated forces of Al-Qaeda.Footnote ¹⁰⁶ This interpretation of co-belligerency would require the co-belligerent to actually fight alongside one of the principal belligerents, excluding situations where no such fighting takes place.Footnote ¹⁰⁷ As the concept is here used with respect to OAGs and not States, it makes sense that the US government and judiciary are referring to co-belligerency in this more narrow sense. Therefore, the co-belligerency approach as advanced by some governments and scholars in relation to OAG coalitions requires, perforce, that each “associated force” is itself engaged in hostilities, and does not simply profess a common worldview or pay lip service to the same ideological crusade – an OAG may not simply “declare war” on a State and thereby trigger a NIAC or become a co-belligerent in the same way as this may happen in an IAC.

We are not convinced that co-belligerency, a concept imported from the law of IACs and the law of neutrality, may be applied by analogy in a NIAC, for two main reasons: (1) unlike IACs, NIACs require that a threshold of intensity be reached, and there is no reason why violence perpetrated by a third OAG in a NIAC should not first be evaluated separately when classifying the conflict;Footnote ¹⁰⁸ and (2) even when it comes to States, not every violation of the law of neutrality would result in the third State becoming a party to the conflictFootnote ¹⁰⁹ (as wrongly suggested by the Court in Hamlily Footnote ¹¹⁰). It also goes without saying that the notion of “association” or “support” required for co-belligerent status is completely undefined for a NIAC. Even authors arguing in favour of the co-belligerency approach (under one name or another) have trouble ascertaining the requisite standard to be applied in such cases. Thus Koutroulis, who argues that non-State armed group (NSAG) “coalitions” should be treated as a single party to a NIAC if they have a person or a group (e.g., a joint council) that exercises the overall leadership of the groups in terms of operational coordination and strategic authority,Footnote ¹¹¹ admits that it would be difficult to distinguish such coalitions from decentralized NSAGs.Footnote ¹¹² If the existence of a coalition depends on the existence of common leadership, even on a purely strategic level, then we do not see the added value compared to analyzing the coalition as a single OAG (other than, perhaps, lowering the organization threshold, which we disagree with). Finally, the co-belligerency approach does not resolve the problem of a higher threshold for the application of AP II (namely, it may apply to some groups, but not to others). Although a “cumulative approach” has been suggested to apply AP II to all groups involved in the coalition if just one of them fulfils its criteria, this does not seem to correspond at all with positive law.Footnote ¹¹³ It is also problematic in practice because an OAG which does not have territorial control will simply not be able to fulfil all the obligations foreseen in AP II.

A second position regarding OAG coalitions/alliances relies on the ICRC's support-based approach. This approach was developed by the ICRC with regard to NIACs involving foreign intervention, in order to determine whether “the nature of the intervening power's involvement in the pre-existing NIAC could mean that it is considered a ‘co-belligerent’, making it a party to the conflict”.Footnote ¹¹⁴ Because a NIAC already exists, a third party's involvement would need to be evaluated not against the general NIAC criteria, but according to its relationship with one of the belligerent parties. Under this approach,

The main difference between the co-belligerency and support-based approaches is that the latter does not require potential coalition members to engage in hostilities themselves.Footnote ¹¹⁶ In this respect, it sets a lower threshold for an external party to become a belligerent party to a conflict. Zwanenburg summarizes the criteria under the support-based approach as follows:

The criticism levelled at the co-belligerency approach may equally be applied to the support-based approach, at least when applied to OAG coalitions. We may add for good measure that the ICRC examines the intervention and involvement of foreign States, not armed groups.Footnote ¹¹⁸ However, some international lawyers, such as van Steenberghe and Lesaffre, argue that “no logical reason prevents the ICRC's ‘support-based approach’ from applying to other types of interventions in pre-existing NIACs, in particular to the support provided by armed groups to one party to a pre-existing NIAC”.Footnote ¹¹⁹ These authors submit that the requirements of the support-based approach are better defined than those of co-belligerency, and are thus a viable alternative to the broad interpretation given by the United States to the concept of “associated forces”. If one group, however well-organized, does not engage in acts of violence (a traditional prerequisite for determining the existence of a NIAC), but simply provides material and logistical support to another, why would we accept ipso facto that it has become a party to an armed conflict and that lethal force may be employed against the group according to the conduct of hostilities paradigm? Doesn't this approach unduly expand the categories of targetable individuals? Isn't the concept of “support” too elastic, and what would be its source in the context of conflict classification? Isn't the notion of direct participation in hostilities enough to deal with individuals who support one party to the conflict to the detriment of another's military operations? It is worth recalling that the Columbia District Court in Hamlily has already dismissed the US government's similar arguments with respect to individuals providing “substantial support” to the Taliban or Al-Qaeda.Footnote ¹²⁰ Is the support-based approach workable in practice for OAGs? For instance, the last criterion, the existence of an official decision by the intervening “power”, would in practice be quite difficult to establish for fluid OAGs. The idea of applying the support-based approach to OAGs therefore causes more problems than it could ever solve.

Several authorsFootnote ¹²¹ also contend that NSAG coalitions may be inferred on the basis of the ICRC's approach to direct participation in hostilities. This position produces similar results to those of the support-based approach, but the criteria are rather based on the ones for direct participation in hostilities as developed by the ICRC (namely, the existence of a threshold of harm, direct causation and a belligerent nexus).Footnote ¹²² It is debatable whether this interpretation is more restrictive or more permissive than the support-based one, but as direct participation in hostilities normally pertains to individual rather than group conduct, the fact that it finds a place in discussions of NSAG coalitions is somewhat perplexing. Deeks writes:

If the nature of Group Y's support to Group X in its hypothetical NIAC with State A meets certain cumulative criteria, then Y is a “functional co-belligerent” of X. The test of functional co-belligerency requires positive answers to three questions, cumulatively: “Was Group Y's participation or action likely to adversely affect the military options of State A? Did Group Y's participation directly lead to tangible harm to State A? Was Group Y's participation intended to benefit Group X and harm State A?”Footnote ¹²⁴ This approach raises many challenges, not least because of the difficulty inherent in determining whether any of these criteria have been met by a group (Zwanenburg remarks that the question of whether the act was intended to harm State A will be very difficult to answer in practiceFootnote ¹²⁵), but also because direct participation in hostilities, developed primarily to determine whether civilians lose protection from direct attack at a given moment, is always provisional: when a civilian is no longer directly participating in hostilities, he or she may no longer be targeted. By implication, this would mean that Group Y as a “functional co-belligerent” could only be targeted when it is directly participating in hostilities, but not afterwards or in between such acts, which is counterintuitive.Footnote ¹²⁶

The last approach we examine here may be labelled as cumulative or aggregated intensity. This theory takes several different forms, all of which have in common the notion that violence projected by separate OAGs could be evaluated jointly for the purposes of conflict classification. For example, Nikolić, de Saint Maurice and Ferraro observe that, in some complex situations of violence,

These authors invoke in this regard the ICRC's 2019 Challenges Report, which finds that “[w]hen several organized armed groups display a form of coordination and cooperation, it might be more realistic to examine the intensity criterion collectively by considering the sum of the military actions carried out by all of them fighting together”.Footnote ¹²⁸ The prerequisite for aggregating intensity would be to have a number of groups that fulfil the criterion of organization separately but work together to form a “coalition”. Clearly the problem remains of how to define such a coalition and the nature of inter-group relations necessary for it to arise: Nikolić et al. offer a number of indicative factors in this regard, but beyond excluding merely ideological and political similarities between coalition members, they ultimately do not provide a clear instruction on how to identify them. They also refer to notions such as “coordination and cooperation” almost interchangeably with the notion of coalition, even though the former conveys the idea that the relationship between the groups is even more fluid and unstable than in a coalition. Writing separately on the same issue a few years earlier, Kleffner had developed the “cumulative intensity” theory, but divorced from the notion of coalition. He suggested that the violence projected by all armed forces in a “geographical and temporal continuum” should be added to the equation of determining whether a NIAC exists, regardless of whether various OAGs are fighting together or against each otherFootnote ¹²⁹ (this is the key difference between Kleffner's view and the aggregated intensity approach suggested by Nikolić et al., which does not require the existence of such a continuum, merely a “coalition”). A modification of Kleffner's view is also advocated by Redaelli, who finds that the coalition ought to be fighting against a “common enemy” as an additional requirement.Footnote ¹³⁰ Again, these interpretations are based on theoretical constructions driven by the desire to facilitate conflict classification in situations where the facts are difficult to ascertain.

The cumulative/aggregated intensity approach presents an easy way out of persistent and frustrating legal headaches, but is it justifiable to loosen legal standards simply because this is more easily done than collecting and evaluating data from the field? Is it even legally sound to do so?

Ultimately, like the other approaches we examine here, cumulative/aggregated intensity suffers from the same problem of potentially leading to an over-classification of situations of violence. An armed conflict, whether international or non-international, is an objective, factual state, one which in principle exists independently of the perception of external observers – and it is also a legal state, in the sense that it triggers the applicability of a certain body of law. While the question of whether IHL or IHRL provides greater protection to human beings is more grey than black and white, when it comes to fundamental issues such as the use of force or deprivation of liberty, the latter is clearly preferable to IHL. By simplifying conflict classification, each of the above theories effectively creates a presumption (in case of doubt) in favour of IHL, when there should rather be a presumption against it. As aptly noted by Sivakumaran, “[t]he default presumption for international human rights law is that it applies, and for international humanitarian law that it does not”.Footnote ¹³¹

It may be submitted that there is a humanitarian interest in facilitating conflict classification in order to bind an armed non-State actor by rules of international law (specifically, IHL). We do not find this argument convincing for several reasons. As stated above, IHL contains not only protective rules but also a “license to kill” (or at least an absence of prohibition to intentionally kill legitimate targets) in a manner which is very foreign to human rights law – it is extremely difficult to conclude whether the interest of binding a non-State actor by IHL outweighs the loss of human rights protection vis-à-vis the State. The growing body of doctrine holding that non-State actors, including OAGs, have human rights obligations should also not be disregarded.Footnote ¹³² At any rate, if an armed conflict exists, then every individual (whether civilian or combatant/fighter) within its geographical scope may violate IHL if they engage in prohibited or criminalized conduct with a nexus to it – it is not necessary for them to first belong to an OAG in order for them to be bound by IHL. Above all, the interests underlying the aggregating/cumulating intensity cannot make up for any organizational deficiencies in the groups involved: to combine the violence projected by such groups could even lead to a situation where all the violence projected in a territory, including isolated acts of criminal violence, might be considered for conflict classification. If that were the case, the distinction between other situations of violence and proper armed conflicts would tend to disappear.

Accordingly, we submit that less is sometimes better than more, and that when classifying violence projected by and against armed groups, a bilateral approach should be taken to determine which groups are belligerent parties. This remains without prejudice to the possibility that one or more individuals who do not clearly belong to one of the belligerent parties may be considered as direct participants in hostilities and thereby lose their protection from attack for such time as they directly participate in hostilities.

An analogy may be made here with the “internationalization” of NIACs through third-State intervention. Both the International Court of Justice (ICJ)Footnote ¹³³ and the ICTYFootnote ¹³⁴ have treated the acts of an OAG and a third State intervening on its behalf separately for purposes of conflict classification and determination of the applicable law – Robert Kolb refers to this as the theory of “bilateral bundles” (faisceaux bilatéraux).Footnote ¹³⁵ It has never been suggested that the acts of violence projected by the OAG could be cumulated with the acts of violence projected by the intervening State; even the view that the conflict should be considered “internationalized” as a whole has been abandoned despite its practicality and humanitarian value.Footnote ¹³⁶ It is therefore all the more disconcerting that a cumulative approach should be advanced for relations between OAGs (which would result in the extension to one or more of them of the law of NIACs, instead of considering human rights law applicable to them), when the notion of internationalization (which would merely transform a NIAC into an IAC) is itself losing favour. If a bilateral approach is adopted for a foreign State intervention on behalf of an OAG, there is no good reason why the same should be abandoned when two groups operate in the same area or when they loosely coordinate their actions.

Although a situation of violence may be colloquially described as a single “war”, there may be many armed conflicts, all of which represent a bilateral legal relationship between individual belligerent parties. Only on a bilateral level may the content of these parties’ legal obligations be fairly ascertained. Of course, such bilateral legal relations do not exist in a factual vacuum, and an ongoing NIAC with one group may influence the evaluation of the NIAC criteria in relation to fighting against another (for example, the State's capacity to maintain law and order could be impacted by the ongoing NIAC and thus create a situation where the criterion of intensity for another NIAC may be more easily met), but the classification analysis itself should always be done on a bilateral level. Otherwise, if the facts are unclear or insufficient to establish the existence of an armed conflict between specific belligerent parties, then human rights law is the applicable legal framework, not IHL.