The ICC’s Al Hassan Case: A Rejection of the Bilateral Approach to Conflict Classification?
On June 26, 2024, Trial Chamber X of the International Criminal Court (ICC) finally delivered its judgment in the Al Hassan case. The defendant, Mr. Al Hassan Ag Abdoul Aziz—now convicted of eight counts of war crimes and crimes against humanity—was a member of the Islamist militia Ansar Dine and chief of its Islamic police when the group controlled Timbuktu in Mali from April 2012 until January 2013. The case drew interest for some of the complex questions of law in armed conflict that international tribunals had not previously addressed. Some of these questions were already explored in other posts by Katharine Fortin, Sean Watts, and Diletta Marchesi; again, separately, by Katharine Fortin; and by Ezequiel Heffes after the parties had delivered their closing statements in May 2023.
Now that the trial judgment has been delivered, I would like to take another look at the court’s approach to an aspect of the case that may have the most far-reaching implications, namely, its logic in classifying the conflict. As we will see, the court has explicitly abandoned the classical “bilateral” or “fragmented” approach to conflict classification without, in my view, adequately elaborating a new model to be used in its place.
The Bilateral Approach and Its Contenders
The application of the principal rules of international humanitarian law (IHL) requires the existence of an armed conflict, whether of an international (IAC) or non-international character (NIAC). The existence of an armed conflict and whether it is an IAC or NIAC is ascertained as part of a legal analysis referred to as conflict classification. In a typical classification exercise, those performing the analysis will try to identify the various parties involved in a situation of violence and determine if their mutual relations amount to one or the other type of armed conflict. The approach whereby conflict classification is done by separately examining the relationship between belligerent parties A and B, A and C, B and C, and so on, is known as the bilateral or fragmented approach.
The bilateral approach has two main advantages. From a practical perspective, understanding bilateral legal relations is crucial to determine the applicable law. Despite expansive interpretations of customary law, differences in the framework governing international and non-international armed conflicts remain significant. It is therefore necessary to understand which rules apply to a given military operation or govern the treatment of a particular individual affected by armed conflict. The bilateral approach also arguably reflects the state of positive or existing international law, having been implicitly used by the International Court of Justice as early as the 1986 Nicaragua judgment (para. 219). In a case of parallel NIACs, the bilateral approach means that the traditional requirements for the application of Common Article 3 to the Geneva Conventions, those of organisation—whereby the parties need to possess a minimum level of organisation to engage in hostilities and implement IHL—and intensity—whereby the fighting needs to reach a certain threshold to warrant the label of an armed conflict—have to be ascertained separately for each conflict.
However, it has been argued the bilateral approach is inappropriate or even impossible to implement in complex situations involving a multitude of armed non-State actors with unclear or shifting alliances. Such a “messy” conflict environment can make it difficult to ascertain the scope of geographical presence and control over territory by individual groups or even to attribute specific attacks. It may even come across as counterintuitive and formalistic to insist that State authorities conduct a separate legal analysis and determination when encountering different groups of insurgents within the same territory.
Therefore, several authors have suggested alternatives to the bilateral approach. The most prominent alternative models are the cumulative intensity model, put forward by Jann Kleffner, who argues in favour of cumulating the intensity of fighting projected by multiple armed groups in the same geographical and temporal continuum, and the aggregated intensity model, advocated in various forms by: Jelena Nikolić, Tristan Ferraro, and Thomas de Saint Maurice; Marten Zwanenburg; and Chiara Redaelli. This model was essentially accepted in the International Committee of the Red Cross (ICRC) 2019 Challenges Report (p. 50-51). The aggregated intensity model adds up the intensity of violence projected by groups working together in some sort of “coalition” or at least fighting a common enemy (in the Redaelli model). Others have proposed more nuanced or eclectic models. While all these models recognise the bilateral approach is inadequate for situations involving multiple non-State actors, they exhibit significant and even irreconcilable differences regarding when intensity may be aggregated or cumulated.
It is important to note these authors agree the bilateral approach is the “basic” model and their proposals are designed to address its deficiencies in complex conflict settings. I respectfully understand this to signify the existence of consensus regarding the status of the bilateral approach in positive law; by implication, the new models are all advocated de lege ferenda. To the best of my knowledge, the aggregated intensity model has not been employed by any judicial body at the international level prior to the Al Hassan judgment. Conversely, Professor Kleffner points to earlier jurisprudence from the ICC (specifically, the Katanga trial judgment, paras. 1207-1211) as supporting his cumulative intensity model. As we will see, the ICC has not been consistent in its classification logic, and it is difficult to speak of a uniform ICC approach.
The Trial Chamber’s Classification Logic in Al Hassan
The facts underlying the Al Hassan case are a poignant example of a situation that would merit, in the eyes of advocates, a certain deviation from the bilateral approach. Focusing on events preceding and during the capture of Timbuktu by jihadist militants in 2012 and 2013, the Trial Chamber had to examine the nature, status, and activities of not just one, but as many as four distinct groups, each of which was said to separately meet the criterion of organisation (which the advocates of alternative models do not bring into question). These are: Ansar Dine, a “homegrown” Tuareg Islamist militia counting Mr. Al Hassan as a member; Al Qaeda in the Lands of the Islamic Maghreb (AQIM), another jihadist group consisting primarily of foreign fighters that entered an “alliance” with Ansar Dine; the Movement for Oneness and Jihad in West Africa (MUJAO), which split from AQIM in 2011 but rejoined the Ansar Dine/AQIM coalition in November 2012; and the predominantly secular nationalist National Movement for the Liberation of Azawad (MNLA), which coordinated its attacks against government forces in Northern Mali with Ansar Dine and AQIM in early 2012 before breaking off with them in June of that year. The jihadist coalition took over Timbuktu in April 2012 and held it until January 2013 when it was ousted from the city by Malian and French forces (an overview of these groups’ organisation is provided in paras. 430-43 of the judgment).
As could be expected, the complicated, opaque, and changing nature of relations between these groups makes it difficult to precisely apply the bilateral approach when classifying the conflict. The Defence was certainly aware of this fact when it argued (correctly, in my view) that the “[i]ntensity level is normally ‘assessed for each bilateral situation on its own, meaning that’ violence ‘between an organized armed actor and its adversary would need to satisfy the criterion of intensity, regardless and separately from the actions of other organized armed actors that might be fighting the same adversary’” (para. 102 of the Defence Final Brief; the quote is from the above-mentioned piece by Nikolić, Ferraro, and de Saint Maurice). The court rejected this view, finding that, “Given the holistic nature of the assessment of the facts that is required, the Chamber also finds unpersuasive the Defence’s arguments relating to individual confrontations taken in isolation, to show that they were not ‘intense’ enough by themselves” (para. 1266).
It is not readily apparent that this statement concerns the bilateral approach, as the Chamber is really dismissing the Defence’s argument that armed confrontations between January and April 2012 were too sporadic to be considered an armed conflict, and that accordingly there was no conflict in April 2012 when Timbuktu was captured. However, in the relevant footnote (n. 4064) the judges were quite explicit, observing,
The Chamber also finds unconvincing the Defence’s argument that it is necessary in the present case to assess each bilateral situation on its own (the intensity between each non-state armed group and the Malian army). . . . In this regard, the Chamber observes that the reference cited by the Defence envisages an assessment of the aggregated military actions carried out between all the non-state armed groups fighting together and their common enemy, ‘rather than requiring that each bilateral relationship of violence meets the criterion on its own’. . . In any case, the evidence makes it clear in the present case that Ansar Dine, AQIM, the MUJAO (and the MNLA to a certain extent and at a certain point in time) undertook military actions in a coordinated manner.
Though it may be hidden away in a footnote, this finding may be the most legally impactful part of the entire eight-hundred-page judgment. It amounts to an explicit rejection—and a rejection in full—of the bilateral approach, pronounced so casually and with so little reverence for settled legal doctrine that one would think the Defence had desperately grasped for some sort of fringe theory. Even if there is much to criticise in the arguments of the Defence, on this point their argument reflects positive law.
My concern with the Chamber’s logic is not merely formalistic. In rejecting the bilateral approach, the Chamber offers no alternative model for classification. While it seemingly agrees with the model argued by Nikolić, Ferraro, and de Saint Maurice (requiring the existence of a “coalition” whereby the groups in question exhibit a degree of cooperation and coordination that goes beyond simply having a common enemy), it is actually unclear what type and degree of inter-group cooperation it takes as relevant to aggregate intensity. On the one hand, it refers to the advanced and sophisticated level of cooperation between Ansar Dine, AQIM, and later MUJAO as “the Coalition.” But in the very next sentence it also takes into account their temporary coordination with the MNLA, which was far less extensive and highly opportunistic. Furthermore, the Chamber’s reasoning contradicts the ICC’s own earlier jurisprudence in Katanga, resembling Professor Kleffner’s model, and of course Lubanga, where the bilateral approach was used (at least implicitly; see para. 543 of the trial judgment).
To be clear: I do not disagree with the court’s conclusion in ascertaining the existence of a NIAC in Mali at the relevant time; I only disagree with its approach. The relationship between Ansar Dine and AQIM, at least, appears to have been so elaborate that an entirely different and much more traditional logic could have been applied, eschewing any sort of novel classification model.
Al Hassan and Armed Group Coalitions
As noted earlier, several aggregated intensity models allow to jointly consider the violence projected by separate armed groups acting as part of a coalition. While the term “coalition” is catchy, it is absent from treaty law and there is also no proof that a useful concept of armed group coalitions has emerged in customary law. Consequently, while different authors may agree on the relevance of the empirical phenomenon of armed group coalitions, they usually have considerably divergent views of how to define and identify them.
For example, the ICRC Challenges Report seems to understand an armed group coalition as possessing a certain degree of inter-group cooperation and coordination (in a 2024 opinion paper on the notion of armed conflict, the ICRC only required “a sufficient level of coordination” between groups to aggregate intensity, at p. 17). In contrast, Professor Zwanenburg’s model, a reaction to the ICRC position, explicitly rejects this in favour of cooperation or coordination. Meanwhile, the Nikolić, Ferraro, and de Saint Maurice proposal provides several indicative elements that may be used, on a case-by-case basis, to determine the existence of a coalition. These indicators range from something as simple as coordinating attacks against a common enemy to much higher levels of cooperation including the establishment of a centralised joint command and the sharing of operational tasks (my understanding is that none of these elements is constitutive of a coalition on its own). Where such a sophisticated level of cooperation is present, I wonder whether it is still justified to regard the “coalition members” as truly separate groups.
Take the example of the relationship between Ansar Dine and AQIM as discussed in Al Hassan. These two groups, while initially separate, formed an “alliance” to achieve their common goals. Although they retained separate group identities—presumably to attract more locals to join Ansar Dine, perceived as less extreme than Al Qaeda (as discussed in paras. 437 and 465 of the trial judgment, where Ansar Dine is even described as a “cover” for Al Qaeda)—they seemingly formed a single chain of command, with the Ansar Dine leader hierarchically subordinate to the leader of AQIM (para. 468), at least until their roles were reversed in October 2012. They maintained unified communication with the local population (paras. 469, et seq.) and jointly established institutions in Timbuktu, including training centres, Islamic courts, and a police force (paras. 516, et seq.).
With such a high degree of cooperation between Ansar Dine and AQIM, is it truly still justified to regard them as two separate groups instead of a single armed group, even if it was somewhat decentralised? The Trial Chamber itself employs the undifferentiated term “Ansar Dine/AQIM” more than 1,500 times in the judgment. Indeed—and while this only came to pass several years after the events discussed in Al Hassan—it should be borne in mind that both groups eventually formed, along with several others, the notorious Jama’a Nusrat ul-Islam wa al-Muslimin (JNIM), a collection of well-connected jihadist militants that is variously described as a group and coalition. In fact, the elements of extensive cooperation the court identified between Ansar Dine and AQIM recall the indicators of organisation for a single armed group used by the International Criminal Tribunal for the Former Yugoslavia in the Haradinaj case, including those relevant to the presence of a command structure, logistics, and the ability to speak with one voice.
Instead of coming up with new theories of conflict classification, it may be more useful to further clarify the scope and content of existing legal tools and ascertain at which point two groups merge into a single entity. Had such an attempt been made by the Trial Chamber in Al Hassan, it would arguably have been unnecessary to stray from the bilateral approach.
Concluding Thoughts
The Al Hassan trial judgment doubtless constitutes a rejection of the bilateral approach to conflict classification and a doctrinal vindication of scholars advocating alternative approaches to the classification of conflicts involving multiple non-State actors. At the same time, as the Chamber did not adequately explain its own reasoning, the judgment cannot really be invoked in favour of any of the major alternative classification models. It also remains to be seen how the approach will stand on appeal, should it come to that.
For the time being, Al Hassan will likely influence legal doctrine, although it may also suffer the fate of Katanga which does not appear to have had a significant influence on the views of most scholars. Some stakeholders will certainly pick up on the judges’ reasoning and use it in other situations, possibly those with little resemblance to Al Hassan.
In my view, more flexible approaches to conflict classification will necessarily lead to the phenomenon of over-classification of armed conflicts, of which I have written with Professor Gloria Gaggioli. Over-classification means treating as armed conflicts situations which are better dealt with as law enforcement matters, governed by more restrictive human rights standards. This may not be true of Mali, but it could easily be the case elsewhere. In the elaboration of new doctrinal models, it therefore behooves humanitarian-minded lawyers and scholars to take good care not to throw out the champagne with the cork.
***
Pavle Kilibarda is a Postdoctoral Researcher at the Faculty of Law of the University of Geneva.
Photo credit: Pexels
