Ntaganda Appeals Chamber Judgment Divided on Meaning of “Attack”

by | May 12, 2021

Ntaganda. ICC. 2021

 On 30 March 2021, the International Criminal Court Appeals Chamber issued its judgment in the Ntaganda case. The judgment, which primarily responds to the appeal submitted by Bosco Ntaganda against his conviction for numerous counts of crimes against humanity and war crimes, raises a number of interesting international criminal law issues. The most significant in terms of the interaction between international humanitarian law (IHL) and international criminal law is actually in response to the Prosecutor’s appeal on the question of “attack.”

Strikingly, the judgment findings did not so much expound on the definition of “attack” in the context of the Rome Statute’s war crimes provisions. In fact, there was a stalemate of sorts between the judges in their opinions regarding the definition of “attack.” This point is particularly notable since the Appeals Chamber had identified the issue last year as one warranting extra attention. As readers of Articles of War will likely remember, the Appeals Chamber agreed with the Prosecutor that “the meaning of the term ‘attack’ in Article 8(2)(e)(iv) of the Statute … may have implications beyond the present case” (para 11). It therefore solicited opinions from amici curiae—a process in which numerous scholars participated. A fascinating symposium summarized these proceedings on Articles of War

On the heels of those posts, and the Appeals Chamber judgment itself, the present post will summarize the Appeals Chamber judges’ legal findings and offer some critical insights. In particular, I shall observe that what ultimately led to the discord in the respective judges’ opinions was the different methodologies they adopted. Moreover, it will be argued that the correct methodology is one which interprets the term “attack” in accordance with its meaning in IHL. Yet, only some of the judges adopted this approach. Before undertaking the analysis, I provide a brief background on the relevant factual circumstances and findings of the Trial Chamber’s judgment, which was the subject of the Prosecutor’s appeal.


Ronald Alcala and Sasha Radin’s introduction to the Articles of War symposium summarized the factual background to the legal issue of “attack” in the Ntaganda case. I therefore shall only very briefly mention the main facts of relevance.

Two incidents involving the Union des patriotes congolais/Patriotiques pour la libération du Congo (UPC/FPLC) of which Bosco Ntaganda was a senior member, implicated the definition of “attack,” as found in Article 8(2)(e)(iv) of the Rome Statute. That Rome Statute provision stipulates that the following act constitutes a war crime in non-international armed conflicts:

Intentionally directing attacks against buildings dedicated to religion, education, art, science or charitable purposes, historic monuments, hospitals and places where the sick and wounded are collected, provided they are not military objectives ….

The first incident concerned UPC/FPLC looting during a ratissage (sweep-up) operation, following its takeover of the town of Mongbwalu, in Ituri, Democratic Republic of the Congo. The purported Article 8(2)(e)(iv) violation related to the looting of medical equipment from the town’s hospital (Trial Chamber Judgment, para 514). The Trial Chamber considered that the looting did not constitute an “act of violence against the adversary”—referring to the definition of “attack” in Article 49(1) of Additional Protocol 1 (API)—and therefore decided no “attack” had occurred (para 1141).

The second incident concerned further ratissage operations in the town of Sayo, where “the UPC/FPLC had set up a base inside the church in Sayo.” During this period “they broke the doors of the church, removed the furniture, dug trenches around the church, and started a fire inside to prepare their food” (Trial Chamber Judgment, para 526). Here, the Trial Chamber determined that the incident did not amount to a crime under Article 8(2)(e)(iv). It opined that because “the attack on the church in Sayo took place sometime after the assault, and therefore not during the actual conduct of hostilities” (para 1142), it thus did not amount to an “attack.”

The Prosecutor appealed both these conclusions, arguing that the Trial Chamber’s conclusion that certain acts of the UPC/FPLC  did not amount to “attacks” for the purposes of Article 8(2)(e)(iv) of the Rome Statute was in error.

The Appeals Chamber Judgment and Individual Opinions on “Attack”

In its judgment, the Appeals Chamber rejected the Prosecutor’s appeal by a 4-1 margin, with Judge Luz del Carmen Ibáñez Carranza dissenting. However, the majority were divided in their reasoning. Actually, the most striking aspect of the Appeals Chamber judgment and the annexed individual opinions is the manifest lack of agreement among the judges over what constitutes an “attack” for the purposes of Article 8.

Indeed, as will be seen, Judges Ibáñez Carranza and Eboe-Osuji’s conception of “attack” radically differs from that of Judges Morrison and Hofmański. That such a gulf exists in the context of defining criminal conduct in regard to the precise definition of “attack” is clearly an unhappy state of affairs (see Abhimanyu George Jain’s analysis).

It should be added that their analyses of the term “attack” were not confined to Article 8(2)(e)(iv), but extended to its meaning under the war crimes provisions of Article 8 more generally. This was contrary to the Prosecution’s contention that “attack” in Article 8(2)(e)(iv) has a different meaning than normally used in IHL (Hearing 12.10.2020, 11-13, see also Al Mahdi, paras 14-15).

Summary of Judges’ Opinions on “Attack”

Judges Howard Morrison and Piotr Hofmański, in their joint separate opinion, considered that the Trial Chamber “committed no error in its interpretation of the term ‘attacks’ in Article 8(2)(e)(iv) of the Statute and its application to the facts of the case” (para 1). They therefore rejected the Prosecutions appeal on this issue.

In defining “attack,” Judges Morrison and Hofmański placed particular emphasis on the 1987 ICRC Commentary to API (para 1882). The Commentary considers that the term “attack” reflects a “combat action” which “refers simply to the use of armed force to carry out a military operation at the beginning or during the course of armed conflict.” The judges also approvingly quoted the statement in Geoffrey Corn et al’s amicus brief (para 15) that “violent acts directed at harming the adversary (including the civilian population and civilian objects) through physical injury or destruction, are ‘attacks’ within the meaning of IHL.”

Judge Solomy Balungi Bossa considered the Trial Chamber’s approach in defining “attack” too narrow—at least according to her interpretation of the latter’s judgment.[1] She explicitly adopted the opinion in ALMA’s (Association for the Promotion of IHL) amicus brief. That brief stated that if a ratissage operation is the same as “a so-called ‘mo[p]-up operation,’ which forms part of the military operation to seize control over a certain location …, it might be considered … an integral part of the conduct of hostilities” (para 13). Judge Balungi Bossa thus considered that when the ratissage operations were occurring, the conduct of hostilities had not ceased. She referred to the combative nature of these operations, their temporal proximity to the use of heavy weaponry, and their characterization as a “military operation” (para 8).

She viewed the ratissage operation in Mongbwalu as part of a single “planned violent attack against the enemy” (para 10). However, in rejecting the Prosecutor’s appeal, Judge Balungi Bossa was of the opinion that the acts in Mongbwalu “should have been more appropriately charged as a crime other than an attack under Article 8(2)(e)(iv)” (para 12).

Regarding the church in Sayo, she deemed the acquittal to be justified. There was insufficient evidence situating the incident during the ratissage operation (para 14). And in any event, here too, she considered that more appropriate charges should have been brought (para 15).

It should be noted that Judge Balungi Bossa appears to have refrained from taking a clear stance on the material elements of “attack.” Instead, she primarily focused specifically on the circumstances in which an “attack” can occur—that is, whether “attacks” may take place in the course of ratissage operations. While she answered this question in the circumstances of the case in the affirmative, she ultimately agreed with Judges Morrison and Hofmański that “attack” can only occur as part of hostilities.

Judge Chile Eboe-Osuji rejected the view that “rear-guard actions, such as ratissage operations, taken to secure or consolidate the capture does not amount to ‘attack’” (para 131), opining that an “attack” need not occur during the actual conduct of hostilities (para 132). Moreover, he appears to have considered a range of coercive acts as capable of amounting to “attacks” (para 111), and saw no need for “a particular motive or purpose” on the part of the “attacker” (para 110). Nevertheless, he ultimately rejected the Prosecutor’s appeal on the basis that “the prosecutor’s charge ought to be brought under any special provision that caters better to the conduct charged, where such a provision is available” (para 136).

Finally, Judge Ibáñez Carranza, quoted approvingly the Pre-Trial Chamber’s confirmation of charges decision regarding the definition of “attack” (para 1166). That Chamber stated that “what matters is the consequences of the act, and particularly whether injury, death, damage or destruction are intended or foreseeable consequences thereof,” and thus extends to acts such as rape, pillage and destruction of property, so long as “there exists a sufficiently close link to the conduct of hostilities” (emphases in original, Pre-Trial Chamber Confirmation Decision, para 46). On this basis, Judge Ibáñez Carranza rejected the Trial Chamber’s decision (para 1168).


If the “attack” aspect of the Appeals Chamber judgment was the subject of a post-mortem, a pathologist would likely be tasked with identifying how such a gulf in interpretation had occurred. From reading the judges’ opinions, it appears that the methodology employed in interpreting the term “attack” was the main division.


Both Judges Eboe-Osuji and Ibáñez Carranza began their analyses by referring to standard—but, curiously, different—dictionary definitions for the term “attack.” Judge Eboe-Osuji referred to a definition from the Shorter Oxford English Dictionary, which defines “attack” as “[t]o go against with violence or force of arms” (para 110). Judge Ibáñez Carranza said that “[t]he term attack is ordinarily defined as ‘an act of using violence to try to hurt or kill somebody’ and as ‘a violent act intended to hurt or damage someone or something’” (para 1166). Admittedly, Judge Eboe-Osuji—and, to a more limited extent, Judge Ibáñez Carranza—listed supporting reasons. These explanations ranged from the use of the term “attack” in Article 7 concerning crimes against humanity (para 112), to improving the intelligibility of the law to “ordinary soldiers” (para 121), to the Martens Clause (para 114) and the prevention of impunity (Judge Eboe-Osuji para 120 and Judge Ibáñez Carranza para 1167).

Conversely, Judges Morrison and Hofmański considered that pursuant to the chapeau of Article 8(2)(e) and the Elements of Crimes, the interpretative exercise should be conducted within the framework of IHL (paras 9-10)—“the established framework of international law.”

Judges Morrison and Hofmański’s approach is the most defensible approach. Indeed, their interpretation of the term “attack,” as found in Article 8 of the Rome Statute, accords with the customary rules of treaty interpretation (Alleged Violations, para 35), codified in Articles 31-33 of the Vienna Convention on the Law of Treaties. Under Article 31(1) of the Vienna Convention, “[a] treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.” The “ordinary meaning” of a term is inseparable from its context (see, e.g., Gardiner, 194).

The war crimes listed in the Rome Statute are defined as “serious violations of the laws and customs” of IHL, existing “within the established framework of international law” (the respective chapeaux of Article 8(2)(b) and Article 8(2)(e)). Therefore, the starting point for a search for the ordinary meaning should not have been standard dictionaries. Rather, it should have begun with identifying whether there exists an ordinary meaning of the term under IHL. In particular, the judges should have adopted the International Court of Justice’s reasoning in one of its cases (Spain v Canada, para 70) and analyzed whether the term has a meaning which has “long been understood by States in the treaties which they conclude” and one which arises from “the practice of States.”

The term “attack” has a particular meaning in the context of IHL. As noted, a specific definition is enshrined in Article 49(1) of API. This meaning was preceded—and has been followed—by a very similar understanding of the term in State discourse and practice. Therefore, these sources should have been used as the starting point for its definition. The following examines in greater detail what arises from these sources.

The Concept of “Attack” in State Practice

The term “attack,” as a concept particular to IHL, was first proposed by the ICRC in its 1956 Draft Rules for the Limitation of the Dangers Incurred by the Civilian Population in Time of War. That instrument defined “attacks” as “acts of violence committed against the adverse Party by force of arms, whether in defence or offence.” The ICRC explained in its commentary that it was concerned with “physical harm” (p 44), conducted “in order to cause harm to the adverse Party” (p 45 emphasis added). Thus, both the physical harm and the motive of harming the adversary elements of “attack” are reflected in the Draft Rules.

At the time, the Draft Rules “elicited virtually no reaction from governments” (Schindler, 718). However, the ICRC later resurrected its concept of “attack” in the lead up to the drafting of API. It did so first implicitly in one of its background documents to the first Conference of Governmental Experts on the Reaffirmation and Development of International Humanitarian Law Applicable in Armed Conflict (p 21, fn 15). Then it explicitly put forth a concept of “attack” in the draft protocol it circulated in the lead up to the second Conference of Governmental Experts (p 90). Although the definition was tweaked in later drafts, there was never any indication that its substance was altered.

More significantly, not a single State—or governmental expert—is recorded as having any reservations to the material aspects of the definition of “attack” as originally presented, despite its centrality in the respective drafts and final version of API. Rather, controversies during API’s drafting appear to have focused on other issues. Notably, there was disagreement around the applicability of provisions relating to “attacks” to different domains of warfare and to acts conducted in a party’s own territory (see, generally, here, here, here, and here).

Subsequent State practice appears to support these understandings relating to physical harm and necessary motive. Acts which are not physically destructive—such as psychological and electronic warfare—are typically not conducted in accordance with rules relating to distinction, to which “attacks” are subject. Yet, these acts are a common part of military operations. Moreover, State practice reflects the understanding that an “attack” is an act conducted to harm the adversary. Various intentionally destructive acts—such as the destruction in the course of building fortifications, breaching through walls, and clearing fields of fire—are typically not conducted in accordance with rules relating to distinction yet are nevertheless conducted by militaries around the world, including during the conduct of hostilities (cf, generally, Neuman).

“Attack” as Interpreted in API Pursuant to the Vienna Convention on the Law of Treaties

A similar picture arises from a Vienna Convention-rules interpretation of Article 49(1) of API. To begin, the occurrence of “attacks” should be confined to when hostilities occur, a reading supported by Article 49’s placement under API Section I Part IV, entitled “General Protection against Effects of Hostilities.” Accordingly, the main point raised by Judge Balungi Bossa should be considered correct. That is, one should not be too quick to rule out the possibility of “attacks” occurring after the taking over of a geographical area—a condition that does not exist in API—but that the existence of “hostilities” are necessary for an “attack” to occur.

Article 49(1) defines attacks as “acts of violence against the adversary, whether in offence or defence.” In this regard, it is a “well-established principle in treaty interpretation that words ought to be given appropriate effect” (effet utile) (Georgia v Russia, ICJ, para 133). As such, the various elements of Article 49(1)’s definition should have meaning: “acts of violence,” “against the adversary,” and “whether in offence or in defence.” The former two elements are of greater importance for the current discussion.[2]

First, an “act of violence” in the context of “hostilities” implies that an “attack” involves intentional physical force. This understanding is confirmed by the fact that an “attack” often carries with it the possibility of causing “incidental loss of civilian life, injury to civilians [or] damage to civilian objects” (API, Article 57(2)(b)).

Second, adding the term “against the adversary” to the definition indicates the element of a military purpose of harming the adversary, or motive thereof. It thus explains how an “attack” is not synonymous with the mere intentional “destruction” of objects (API, Article 54(1)). Destruction of objects is often not conducted with the motive of harming the adversary, but rather purely incidentally or instrumentally, such as fortifying a position (Jachec-Neale, para 17, Hayashi, 112). Indeed, it is not clear how the term “against the adversary” would have the said “appropriate effect” were there, as Judge Eboe-Osuji suggests, no requirement of motivation.

Consequently, Judges Morrison and Hofmański’s views—which endorse the Corn et al’s position, quoted above, that “attacks” under IHL are “violent acts directed at harming the adversary (including the civilian population and civilian objects) through physical injury or destruction”—must be the preferred interpretation.

Concluding Observations

There is a taboo which has not been mentioned, and perchance needs no mentioning. Nevertheless, at the expense of using the international lawyer’s “f” word—and raising a subject considered by Judge Cançado Trindade “at most, a topic for a university thesis (for an LL.M., rather than a Ph.D. degree)”—it would appear that some of the judges’ opinions in the Ntaganda Appeals Chamber judgment are paradigmatic examples of fragmentation in international law. Some of the judges have given the term “attack,” taken from IHL instruments, a meaning divorced from IHL.

It would seem, however, that such a result was precisely the opposite of what they sought to achieve. In fact, Judge Eboe-Osuji asked rhetorically (para 121):

“How does one explain to the average soldier not well versed in the complex amalgam of international humanitarian law instruments ancient and modern—including now the Rome Statute—that precisely the same conduct would expose him to liability for crimes against humanity (in peacetime and in war) but not to war crimes (during war)?”

Of course, one can question whether the “average soldiers” actually read the provisions of the Rome Statute in seeking concrete guidance on how to act in conflict. It is more likely they rely on their respective militaries’ rules of engagement, drafted in a manner so applicable legal rules can be more readily understood for practical purposes. Regardless, even if a degree of coherence in the Rome Statute framework is achieved—a questionable conclusion—the approach posited by Judge Eboe-Osuji would likely have deeper repercussions. Indeed, how does one explain to that same average soldier that intentional conduct which up to today has been practiced by militaries worldwide and been considered lawful and indispensable conduct in armed conflict—such as breaching of walls, fortification, and clearing fields of fire—is now deemed unlawful? How does one explain that that same conduct would now constitute a violation of distinction rules—and even a crime—because the term “attack” must now have a meaning divorced from that traditionally accepted? With a divided bench, such an explanation may ultimately be unnecessary, depending on what future jurisprudence holds.


Ori Pomson is a member of the Israel Bar and currently an LLM candidate at the Hebrew University of Jerusalem. Previously, he served for six years as an officer (rank of captain) in the Israel Defence Forces Military Advocate General’s Corps International Law Department, where he served as Assistant Legal Adviser for Cyber Affairs and Assistant Head of the Legal Development Section.



[1] Though her interpretation of the Trial Chamber’s stance may be disputed in the sense that the Trial Chamber differentiated between the existence of hostilities—when “attacks,” in its opinion, can occur—and post-hostilities—“based merely on the weapons used” (para 10).

[2] The latter element emphasizes that an “attack” can occur both when a party to an armed conflict is on the offense and defense (Rogers, 207), and need not concern us for present purposes.