1. Introduction
Unlike domestic crimes, international crimes typically involve the efforts of many individuals, often acting in concert.Footnote 1 This poses a challenge for those tasked with trying them: namely, how to hold individuals criminally responsible for offences requiring collective action.Footnote 2 Collective conduct produces unique harms and dynamics which make individual conduct within the group difficult to capture and prosecute under classic theories of criminal responsibility. Firstly, the ‘group effect’ often enhances the power of each individual to accomplish a particular criminal objective. Crimes that are difficult to accomplish individually are facilitated when multiple individuals participate.Footnote 3 Secondly, the ‘group effect’ tends to augment the degenerative nature of the criminal conduct. A false sense of impunity derived from the collective presence unleashes ‘the worst’ in the members of the group, leading them to commit acts they would likely not have committed when acting alone.Footnote 4 Thirdly, the group dynamics, whereby many participate and yet few are the physical perpetrators of a crime, allow individuals to plead innocence as regards a particular offence. Participants rely upon the diffuse manner in which criminality spreads across the group – attaching a little to everyone, but not fully to anyone, except, perhaps, the physical perpetrator.Footnote 5 This diffuse sense of criminality is heightened when members of the group are geographically apart, act at different times, or perform different roles in the crime. As a result, virtually all jurisdictions, both civil and common law, have developed at least some principles that specifically address collective criminality.Footnote 6
International criminal tribunals are no different. They have developed a number of legal theories designed to hold individuals responsible for their role in collective criminal conductFootnote 7 in interpreting their Statutes.Footnote 8 These doctrines of criminal participation, known as modes of liability, are the subject of significant scholarly commentary.Footnote 9 Yet missing from much of this debate, particularly as regards the International Criminal Court (ICC), has been an analysis of how current doctrine on modes of liability responds to the need to hold collective perpetrators criminally responsible for crimes of sexual and gender-based violence (SGBV).Footnote 10 Indeed, many writings in this area of the law address perceived shortcomings in the theoretical underpinnings of modes of liability doctrine in the abstract but ignore the application of this doctrine in concreto. As a result, facially neutral writings on modes of liability may, in fact, be gendered in application, either because they fail to account for the specific characteristics of sexual and gender-based violence, or because they are applied in a manner that requires higher thresholds for finding culpability for the commission of SGBV crimes. This article fills the gap between theory and practice, examining past and present doctrine, and suggesting ways in which the treatment of modes of liability by international criminal courts and tribunals can properly respond to both the need for personal culpability and the dangers of collective criminal activity, particularly as regards SGBV crimes.
In particular, we examine the various ways in which the ad hoc criminal tribunals for the former Yugoslavia (ICTY) and Rwanda (ICTR) and the ICC have held individuals liable for their role in collective criminal conduct, including joint criminal enterprise liability (at the ad hoc tribunals) and direct and indirect co-perpetration and common purpose liability (at the ICC). Thus, our general focus is on individual – as opposed to superior or command – responsibility, although the ICC Appeals Chamber’s decision in the Bemba caseFootnote 11 suggests that the difficulties we identify in the Court’s application of these modes of liability also apply to its approach to command responsibility.Footnote 12
As Section 2 describes, the ad hoc international criminal tribunals used a theory of liability, known as joint criminal enterprise (JCE), to prosecute perpetrators who joined together to implement a common criminal objective. First developed in Prosecutor v. Tadić,Footnote 13 the Tadić Appeals Chamber identified three forms of JCE under customary international law, the first two of which, known as JCE I and II, were relatively uncontroversial. An extended form of JCE known as JCE III – described more fully in the next section – was used to successfully convict perpetrators of SGBV crimes, but is contentious, especially in regard to its mens rea, which allows liability to attach if the perpetrator assumes the risk that foreseeable crimes will occur.
The Rome Statute, negotiated in 1998, has much longer and complex provisions on criminal responsibility (and modes of liability) than prior instruments. Article 25 of the Statute addresses individual criminal responsibility generally whereas Article 28 addresses command responsibility. As Section 3 explains, the ICC’s judiciary have taken international criminal law in a new direction; applying Article 25(3) of the Rome Statute, they developed a more rigid approach to modes of liability law, engaging in judicial interpretation no less astonishing than the development of the JCE theory of liability by the ad hoc tribunals. The ICC’s new interpretations have resulted in acquittals of sexual violence counts in every case presented to the Court, except in the recent Prosecutor v. Bosco Ntaganda case.Footnote 14
Our study reveals two divergent patterns emerging from the jurisprudence of the ad hoc tribunals and the ICC, each of which presents challenges for the successful prosecution of SGBV crimes. First, the ad hoc international criminal tribunals’ prosecution of SGBV crimes under the doctrine of JCE III proved successful when evidence of these crimes was examined in a ‘contextually comparative’ manner to evidence of other atrocity crimes. However, this mode of liability was subject to both judicial dissent and scholarly critique. Second, at the ICC, new modes of liability were developed (perhaps as a response to critiques of JCE). Known as direct and indirect co-perpetration – and common purpose liability – these have largely been interpreted restrictively and/or applied differently to acts of sexual violence than to other crimes, leading to acquittals of the accused on the sexual violence counts.
The Nuremberg and Tokyo Tribunals set precedents for SGBV prosecutions,Footnote 15 as did subsequent Second World War proceedings in Europe and Asia.Footnote 16 Nevertheless, much of this jurisprudence was ignored and victims of SGBV committed during periods of conflict or repression were rendered largely invisible in the decades after the war.Footnote 17 Modern international criminal law, thankfully, has increasingly recognized the need to prosecute SGBV crimes for all the same reasons other atrocity crimes are punished: to deter perpetrators, address specific harms, render justice to the victims, and express official condemnation of the behaviour.Footnote 18 Yet there remains a long way to go, even at the ICC, where emphasis on highly technical and scientific explanations of the ICC’s modes of liability provisions have tended to side-line SGBV crimes as mere ‘incidents’ of conflict.Footnote 19
Violence against females and males, whether in detention or during or following attacks, is a pervasive twenty-first century phenomenon. Examples abound: sexual violence committed against prisoners held by the United States at Abu Ghraib following the invasion of Iraq in 2003;Footnote 20 rape as a tool of ethnic cleansing in the Democratic Republic of the Congo;Footnote 21 mass rapes against females in the Central African Republic;Footnote 22 sexual slavery practiced by members of the Islamic State against the Yazidis (and others);Footnote 23 and sexual violence against male and females detainees in Syrian prisons.Footnote 24 These SGBV crimes are both ‘social and systemic’ expressions of domination and power in conflicts where the vulnerable are specifically targeted for violence by those who believe themselves to be beyond the reach of the law, or, worse, within the bounds of conduct tolerated by the law.Footnote 25 Indeed, to the extent the law has failed to sanction this violence when directed against females, or even to recognize how it manifests against males, one wonders whether this is not, yet again, a way in which international law remains a ‘thoroughly gendered system’.Footnote 26
In light of these findings, this article makes two specific recommendations. First, national and international criminal courts and tribunals applying and developing modes of liability jurisprudence under customary international law should accept the customary basis of JCE III and the jurisprudential precedents of the ICTY and ICTR, as well as of other courts that have examined this issue more recently, in particular the Extraordinary African Chambers. Second, the ICC should revisit its restrictive approach to Article 25(3)(a), and interpret it consistent with a plain reading of that provision, while addressing what appears to be a discriminatory application of common plan liability in Article 25(3)(d) by more appropriately situating sexual and gender-based violence within the context of other crimes. While the recent Trial Chamber decision in Ntaganda goes some way toward properly contextualizing sexual violence within other criminal conduct and interpreting Article 25(3)(a) more broadly, the case is still subject to appeal and it is not yet clear that the decision has ushered in a new normative approach to sexual violence crimes at the ICC. Absent a shift in that direction, cases of conflict-based sexual and gender-based sexual violence may continue to suffer significant obstacles, threatening to render such violence as invisible as it was just decades ago,Footnote 27 in spite of provisions in the Rome Statute specifically designed to prevent this result.
2. JCE liability for SGBV crimes: Successful cases, but the doctrine is criticized
2.1 JCE as a mode of liability
JCE liability originates in Article 7(1) of the ICTY Statute.Footnote 28 In the ICTY’s first case, Prosecutor v. Tadić, the Appeals Chamber held that JCE is a form of ‘commission’ that captures instances in which perpetrators implement a common criminal objectiveFootnote 29 to commit an international crime.Footnote 30 As the late Antonio Cassese, a member of the Tadić Appeals bench, explained:
[JCE] is crucial more in international criminal law than at the domestic level. In the world community international crimes such as war crimes, crimes against humanity, genocide, torture and terrorism share a common feature: they tend to be expression of collective criminality, in that they are perpetrated by groups of individuals, military details, paramilitary units or government officials acting in unison or in pursuance of a policy. When such crimes are committed, it is extremely difficult to pinpoint the specific contribution made by each individual participant in the collective criminal enterprise.Footnote 31
Indeed, individual liability may be difficult to assign when crimes are the product of collective criminality, as certain individuals physically perpetrate the crime while others contribute to its commission in other ways, such as through planning or preparation.Footnote 32 JCE liability recognizes that all participants may be equally culpable.Footnote 33 As the Appeals Chamber in Tadić explained:
Although only some members of the group may physically perpetrate the criminal act (murder, extermination, wanton destruction of cities, towns or villages, etc.), the participation and contribution of the other members of the group is often vital in facilitating the commission of the offence in question. It follows that the moral gravity of such participation is often no less – or indeed no different – from that of those actually carrying out the acts in question.Footnote 34
The Tadić Appeals Chamber reasoned that limiting liability to the physical perpetrator would disregard the role of co-perpetrators who enabled the physical perpetrator to carry out the criminal act.Footnote 35 The Chamber considered such liability distinct from aiding and abetting, which, depending on the circumstances, might understate the degree of the co-perpetrators’ criminal responsibility.Footnote 36 Given the often unorganized, chaotic, and decentralized nature of the Yugoslav conflict,Footnote 37 JCE liability was an effective way for the ICTY to capture the collective criminality committed during the conflict.
Although concerns arose about the Tribunal’s methodology in ‘gap-filling’ outside the express terms of the Statute,Footnote 38 there was precedent for the Tribunal’s approach. Historically, organizational liability was recognized in post-Second World War cases,Footnote 39 and forms of liability for participation in group crimes, such as complicity, conspiracy,Footnote 40 ‘common plan’ liability;Footnote 41 and common purpose liability,Footnote 42 were found in other international instruments.
The ad hoc tribunals interpreted JCE liability as consisting of: a plurality of persons;Footnote 43 the existence of a common plan, which amounts to or involves the commission of a crime provided for in the Statute;Footnote 44 and participation of the accused in the execution of the common plan.Footnote 45 The ICTY recognized three categories of JCE, commonly referred to as JCE I,Footnote 46 JCE II,Footnote 47 and JCE III,Footnote 48 which were differentiated by their circumstances and their required mens rea. JCE I’s design or common purpose may take multiple forms, such as a campaign of terror, as long as the participants share the intent to commit the crime.Footnote 49 JCE II circumstances often manifest as a system of ill-treatment, such as a concentration camp or police detention centre.Footnote 50 This form of JCE requires personal knowledge of the system and an intent to further it,Footnote 51 which may be proved directly or indirectly as a matter of inference from an accused’s position, tasks or authority within the system.Footnote 52 JCE III’s mens rea focuses on the foreseeability of a criminal act and the willingness of the accused to assume the risk that it will occur.Footnote 53
Important JCE III cases have resulted in convictions for SGBV crimes, especially for non-physical perpetrators. The use of JCE III to hold SGBV perpetrators accountable challenged the misconception that sexual violence committed in the context of conflict or mass violence was often isolated, unrelated to the other crimes, or the unfortunate acts of lone corrupt soldiers.Footnote 54 JCE III liability furthered a contextualization of crimes that required a non-discriminatory approach to the assessment of evidence of SGBV crimes. Essentially, to establish JCE III crimes as the natural and foreseeable consequence of intended crimes in the JCE I common plan or in the JCE II system of ill-treatment, such crimes had to form part of the context of the initially atrocity-propelled conduct. Thus, evidence of SGBV crimes had to be appraised together with other concurrent criminal conduct rather than be viewed as an abnormality.
2.2 Comparatively contextualizing SGBV crimes: The Krstić case
While rarely noted,Footnote 55 an early JCE III case, based in part on SGBV crimes, was the case of Prosecutor v. Krstić.Footnote 56 Radislav Krstić was charged with genocide for the massacre at Srebrenica as well as other killings and sexual violence as persecution.Footnote 57 These other crimes occurred just prior to the Srebrenica genocide, in pursuit of a JCE comprised of forcibly displacing Bosnian Muslims into the UN compound in Potočari, creating a ‘humanitarian crisis’.Footnote 58 In finding Krstić liable under JCE III for rapes committed in the UN compound, the Trial Chamber focused on Krstić’s awareness of the ‘catastrophic humanitarian situation’,Footnote 59 noting in particular:
the lack of shelter, the density of the crowds, the vulnerable condition of the refugees, the presence of many regular and irregular military and paramilitary units in the area and the sheer lack of sufficient UN soldiers to provide protection.Footnote 60
The crisis environment made it foreseeable to Krstić that other crimes, such as killings and rapes, might occur.Footnote 61 Emphasis was placed on contextualizing the sexual violence not only within the intended JCE I crimes, but also within the foreseeability of the other JCE III crimes of killings. Such non-discriminatory assessment concerning the foreseeability of SGBV crimes became a determinant factor in ICTYFootnote 62 and ICTRFootnote 63 cases using JCE III.
2.3 Šainović and the foreseeability standard: Possibility or probability?
Prosecutor v. Šainović Footnote 64 clarified the mens rea standard of foreseeability, namely whether there needed to be a possibility or probability that further crimes would occur, given the JCE I or JCE II context. Šainović concerned a JCE I common plan to forcibly displace the Kosovo Albanian population through a persecutory campaign of terrorFootnote 65 in which sexual assaults had occurred.Footnote 66 At trial, Nikola Šainović and his co-accused, Sreten Lukić, were acquitted of rapes charged as persecution because the acts were not considered foreseeable as required for JCE III liabilityFootnote 67 and the prosecution failed to prove their knowledge of the specific sexual assaults charged.Footnote 68 On appeal, the prosecution argued that the Trial Chamber had confounded the possibility standard of the JCE foreseeability requirement with a probability standard. The Appeals Chamber agreed, characterizing the forcible displacement of the Kosovo Albanian population as an ‘environment of ethnic animosity’ where ‘aggression and violence prevailed’.Footnote 69 It recognized the accused’s awareness of these circumstances,Footnote 70 and their knowledge of sexual assaults committed during the expulsions, even though these were not the same sexual assaults charged against the accused.Footnote 71 Given these conditions, the Appeals Chambers held that the rapes were clearly foreseeable and found the accused liable under JCE III for those rapes.Footnote 72 Significantly, it recognized that, as members of a JCE I, the accused had a privileged position that permitted them to detect possible crimes triggered by the original common plan. Subsequent appellate jurisprudence affirmed the JCE III mens rea standard. As the Karadžić interlocutory appeal held:
[Foreseeability] is not satisfied by implausibly remote scenarios. Plotted on a spectrum of likelihood, the JCE III mens rea standard does not require an understanding that a deviatory crime would probably be committed; it does, however, require that [whether] a crime could be committed is sufficiently substantial as to be foreseeable to an accused.Footnote 73
2.4 Đorđević and Karemera: JCE III liability for SGBV crimes of specific intent
Whether JCE III liability should apply to specific intent crimes, such as persecution and genocide, has been among the doctrinal issues raised by some, including Judge Antonio Cassese.Footnote 74 In Đorđević,Footnote 75 the Appeals Chamber assessed the contentious issue of whether JCE III liability could be extended to persecutory acts. Vlastimir Đorđević was charged with participating in a JCE I that aimed to change the ethnic balance of Kosovo to ensure Serbian dominance in the territoryFootnote 76 and was convicted of over 700 murders and countless episodes of forced transfer and persecutory acts. He was acquitted of rapes alleged under JCE III liabilityFootnote 77 because the Trial Chamber was not convinced that the SGBV satisfied the discriminatory intent requirement of persecution.Footnote 78 On appeal,Footnote 79 the Appeals Chamber found that the Trial Chamber committed a legal error when it ‘failed to evaluate the surrounding circumstances of Witness K20’s and Witness K14’s sexual assaults and the broader context in which these crimes occurred’.Footnote 80 It meticulously reviewed the specific and contextual circumstances of the five sexual assaults charged as persecution under JCE III. After examining elements such as the ethnicity of the victims and the perpetrators, the status or authority of the perpetrators, the concurrent criminality inflicted on the females and other victims, the verbal abuse, and the location and timing of the physical assaults, the Chamber concluded that the sexual assaults satisfied the discriminatory nature of persecution.Footnote 81 It then held that Đorđević had significantly contributed to the JCE IFootnote 82 and reiterated that JCE III foreseeability entailed the possibility that the rapes would occur.Footnote 83 It found the atmosphere of violence and fearFootnote 84 had left the population ‘highly vulnerable’,Footnote 85 particularly given the separation of the community along gender lines, whereby Serbian forces could act ‘with near impunity’.Footnote 86 It recalled Đorđević’s personal knowledge that Serbian forces reportedly committed crimes during the previous year’s military offensiveFootnote 87 and his knowledge about the dire humanitarian situation through media sources.Footnote 88 It, therefore, concluded that Đorđević ‘willingly took the risk when he participated in the JCE’Footnote 89 and that given the discriminatory nature of the Kosovo campaign, ‘it was foreseeable to Đorđević that such sexual assaults might be carried out with discriminatory intent’.Footnote 90 The Đorđević Appeals Chamber’s contextual examination of the JCE III crimes in relation to the common plan, and its contextual analysis of the infliction of sexual violence – including its discriminatory intent – evinces a non-discriminatory and gender-competent approach to JCE III. Krstić as well as Šainović resonate in Đorđević.
The ICTR case of Karemera et al. reviewed whether JCE III liability could apply to the specific intent crime of genocide.Footnote 91 Two accused were convicted, inter alia, of genocide and crimes against humanity.Footnote 92 Their liability stemmed from their participation in a JCE I to destroy the Tutsis.Footnote 93 Their liability for sexual violenceFootnote 94 as acts of genocide and as crimes against humanity was established via JCE III for rapes committed by the Interahamwe militiaFootnote 95 that were a natural and foreseeable consequence of the JCE I common plan ‘to exterminate Tutsis in Rwanda’.Footnote 96 The accused appealed.Footnote 97 After reiterating the possibility standard for JCE III liability set forth in the Karadžić interlocutory JCE decision,Footnote 98 the Appeals Chamber affirmed the Trial Chamber’s analysis that the rapes had been foreseeable to the accused.Footnote 99 Furthermore, it rejected the accuseds’ argument that the Interahamwe did not possess a genocidal intent when committing the rapes and sexual assault,Footnote 100 explaining:
Bearing in mind that these rapes and sexual assaults were intricately linked to the killing of members of the Tutsi group and intended to inflict further suffering, the Appeals Chamber finds that the Trial Chamber adequately explained and reasonably concluded that the perpetrators possessed genocidal intent.Footnote 101
Thus, the contextual analysis places the rapes on par with the killings in terms of assessing the presence of genocidal intent. Again, rather than extract and mischaracterize the perpetrators’ intent with respect to the rapes as different from their intent with respect to the other acts committed to destroy the Tutsi population, the Chamber understood that the same specific intent permeated the various means to achieve the genocide.
In sum, a mature judicial reasoning regarding JCE III liability has emerged at the ad hoc tribunals, as illustrated by the above jurisprudence, and recently confirmed by the judgment of the Appeals Chamber in the Karadžić case. Case law over two decades has developed distinct contextual analyses of JCE III that examines the circumstances of the JCE I common plan and/or JCE II ill-treatment system; the foreseeability or possibility of JCE III crimes in those circumstances; and JCE III crimes in view of the continued commission of the specific intent crimes of the JCE I and II. Thoughtful, logical, non-discriminatory reasoning about SGBV evidence amid unfolding criminality has significantly guided the development of JCE III.
2.5 The judicial and scholarly critique of JCE III liability
Several prominent commentators have lauded the application of JCE III to atrocity related sexual violence crimes.Footnote 102 Nevertheless, JCE III has been subject to judicial dissent and scholarly critiques, challenging its compatibility with the legality principle and the principle of individual culpability.Footnote 103
With respect to the principle of legality, judicialFootnote 104 and scholarly critiquesFootnote 105 have contested the customary origins of JCE III and the ad hoc tribunals’ analogy to modes of liability in domestic legal systems. For instance, in 2010, a Pre-Trial Chamber decision of the Extraordinary Chambers in the Courts of Cambodia (ECCC) rejected the customary basis of JCE III as of 1975, the beginning date of the ECCC’s temporal jurisdiction.Footnote 106 It found that the war crimes cases decided pursuant to Control Council Law No. 10 relied on by the ICTY did not explicitly identify JCE IIIFootnote 107 as the basis for their decisions and that the national jurisprudence they cited was similarly insufficient to establish JCE III’s customary international law basis.Footnote 108
Critics have also argued that JCE III holds individuals liable for conduct too distant from the actions of the accused, since the perpetrator need not intend the foreseeable crime to be committed nor make any substantial contribution to its commission.Footnote 109 According to these critics, JCE III ‘endangers the principle of individual and culpable responsibility by introducing a form of collective liability, or guilt by association’.Footnote 110 One commentator famously quipped that this extended form of JCE liability might as well be called ‘just convict everyone’.Footnote 111
Thus, notwithstanding the continued use of the JCE III by international tribunals, extensive challenges to aspects of JCE persist.Footnote 112 Although this mode of liability has been successfully used to hold perpetrators of sexual and gender-based violence accountable, and has parallels in many domestic legal systems,Footnote 113 it is not clear whether and to what extent it will continue to be used successfully to prosecute such crimes at international, hybrid or domestic criminal tribunals that rely on customary international law.
3. The impact of the ICC’s modes of liability jurisprudence on the prosecution of SGBV crimes
At the ICC, a troubling pattern is emerging in the manner in which the Court’s modes of liability jurisprudence is intersecting with crimes of sexual and gender-based violence. The ICC Pre-Trial Chambers departed from the jurisprudence of the ad hoc tribunals and adopted interpretations of Article 25(3) that were both novel and complex. This may have been a response to the criticism of the ad hoc tribunals’ approach to JCE, but generally has resulted in an overly restrictive approach that has also been applied differently to acts of sexual violence than to other criminal conduct. This has led to acquittals for those accused of SGVB crimes in all cases, with the exception of Ntaganda. The following section explores this phenomenon.
3.1 The ICC’s restrictive interpretation of Article 25(3) of the Rome Statute
The drafters of the Rome Statute found the provisions on individual criminal responsibility in the statutes of the ad hoc tribunals too laconic.Footnote 114 Thus, during the Statute’s negotiation, states proposed a more comprehensive provision and eventually adopted Article 25, which sets out a more detailed framework of liability than predecessor instruments. Subparagraph (3)(a) makes criminally liable one who commits a crime whether as an individual, jointly with another or through another person, while subparagraphs (3)(b)–(c) set out a variety of other forms of liability including ordering, soliciting, inducing, aiding and abetting.Footnote 115 Subparagraphs (d), (e), and (f) provide liability for contributing to the commission or attempted commission of a crime by a group, incitement to genocide, and attempt, respectively. Per Saland, who chaired the Working Group on General Principles of Criminal Law throughout the ICC negotiations, noted that the provision
… posed great difficulties to negotiate in a number of ways. One problem was that experts from different legal systems took strongly held positions, based on their national laws, as to the exact content of the various concepts involved. They seemed to find it hard to understand that another legal system might approach the issue in another way: e.g., have a different concept, or give the same name to a concept but with a slightly different content.Footnote 116
While Article 25 has its origins in the work of the ILC on the Code of Crimes against the Peace and Security of Mankind, later drafts more closely representing the final version of Article 25 were proposed by ‘an informal group representing various legal systems’.Footnote 117 The final version draws on various sources of national criminal law and international law.Footnote 118 The vast and divergent literature on Article 25 indicates, as one of us has suggested in earlier writings, that it is like a ‘Rorschach blot’, in which jurists tend to see and read into the provision their own experience and understanding of criminal liability, based on their national legal system.Footnote 119
The travaux préparatoires of the Rome Statute say little about the content of Article 25.Footnote 120 Although some scholars have insisted after the fact that Article 25 embodies both a differentiated and hierarchical approach to criminal participation,Footnote 121 others disagree equally vociferously, noting that ‘it is difficult to find an unambiguous answer’ to the question of what model of perpetration and participation it adopts.Footnote 122 In point of fact, Article 25(3) was a consensus provision that does not embody a strong and logically cohesive theoretical underpinning of the kind that can be found in some domestic jurisdictions.Footnote 123 Rather, as William Schabas has suggested, ‘the terms in paragraph (b) seem to be drawn from continental models, whereas those of paragraph (c) belong to the common law’ and ‘should not be viewed as two different or distinct bases of liability, but rather as an effort to codify exhaustively various forms of complicity by drawing upon concepts familiar to jurists from different legal traditions’.Footnote 124
In terms of collective criminal responsibility, or participation in a criminal enterprise, the Rome Statute does not use the term joint criminal enterprise, in part because it was adopted prior to the doctrine’s clear emergence at the ICTY.Footnote 125 As a response to the critique that the ICTY and ICTR Statutes had omitted conspiracy (except as regards the crime of genocide), and in search for a form of ‘common plan’ liability, the Statute incorporated a specific paragraph on common plan liability (paragraph 3(d)), the text of which was drawn from the International Convention for the Suppression of Terrorist Bombings.Footnote 126 According to the chair of the Working Group on General Principles of Criminal Law, this avoided divisive discussions about incorporating the concept of conspiracy, which was rejected by most states parties, although common law countries advocated for it extensively.Footnote 127 While some scholars now argue that Article 25(3)(d) is a ‘residual’ form of liability, that was not at all apparent in 1998, and following the Tadić Appeals Chamber decision, there was at least some thought that this might be a Rome Statute version of JCE, given that the text refers to the individual contributing ‘to the commission or attempted commission of … a crime by a group of persons acting with a common purpose’.Footnote 128 In fact, one leading scholar has recently suggested that JCE I and II may be included in this provision.Footnote 129
No sooner had the Rome Statute become operational, however, than a group of scholars insisted on an interpretation of Article 25 that would cause it to diverge dramatically from the jurisprudence of the ad hoc tribunals and customary international law. In Prosecutor v. Thomas Lubanga Dyilo,Footnote 130 the ICC’s first completed trial and judgment, the majority assigned a complex meaning to Article 25(3)(a), based upon German theorist Claus Roxin’s ‘control of the crime’ theory.Footnote 131 Essentially the judges parsed the meaning of the words ‘commits such a crime, whether as an individual, jointly with another or through another person’ to create three separate forms of ‘commission’ liability: individual perpetration, direct co-perpetration, and indirect co-perpetration. For both direct and indirect co-perpetration, an essential element of the offence would be the establishment of a ‘common plan’,Footnote 132 and the accused’s contribution to the plan had to be ‘essential’.Footnote 133 Indirect co-perpetration added as a requirement (in the Katanga case discussed below)Footnote 134 that the accused also had control over an organization through which he committed the crime.Footnote 135 This, the judges found, was an appropriate use of Roxin’s theory.Footnote 136
Two judges, Adrian Fulford (in the Lubanga case)Footnote 137 and Christine Van Den Wyngaert (in Prosecutor v. Mathieu Ngudjolo Chui),Footnote 138 have taken issue with the elaboration of the Court’s modes of liability doctrine, noting that it had improperly imported the control of the crime theory from national law into Article 25 and that this interpretation was inconsistent with the text of the Statute. Many scholars, including one of us,Footnote 139 have similarly argued that the Court’s interpretation neither comports with the ‘plain meaning’ of the Statute nor is appropriate under the interpretative methods open to the judges under Article 21 of the Statute.
It is ironic that just as the critiques of the ICTY’s JCE doctrine intimated that the Tribunal’s judges had engaged in judicial activism and improperly incorporated common law doctrine into international criminal law,Footnote 140 the ICC’s judges elaborated a stunningly complex interpretation of Article 25, subjecting themselves to the same critique, perhaps even more so.Footnote 141 Although the ICC’s case law may well be more favourable to the accused than the ICTY’s, perhaps in response to the complaint that JCE was overly favourable to the prosecution, it is neither well-grounded in the Statute nor mandated by customary international law. Nor have the critics of JCE established that its application led to any wrongful convictions or a miscarriage of justice,Footnote 142 suggesting that the moniker ‘just convict everyone’ is invalid and unfounded.
Finally, this new doctrine has produced, perhaps unintentionally, an additional problem: by restricting the application of Article 25(3)(a) to principal liability and limiting it to those with control over the crime, the Court has had to rely on Article 25(3)(d) in cases of group liability where the facts do not support such a limited interpretation.Footnote 143 This has led – as the next section describes – to high rates of impunity for SGBV crimes.
3.2 The absence of gender competence in the ICC’s application of Article 25(3)(d) of the Rome Statute to SGBV crimes
As noted above, the additional elements required to establish ‘commission’ of a crime at the ICC are perhaps more favourable to the accused than analogous ICTY case law. Yet, as the discussion of the Katanga case below reveals, this interpretation of ‘commission’ led the Court in that case to reject the application of common purpose liability under Article 25(3)(a) in favour of Article 25(3)(d). At the same time, the Court applied this provision differently to acts of sexual violence than to other crimes, even though the sexual violence occurred simultaneously at the hands of the same group of perpetrators. The result, as Katanga illustrates, has been a high level of impunity for sexual violence crimes.
Germain Katanga, the commander of the Ituri-based Walendu-Bindi collectivité, a Ngiti militia group in the DRC, was accused of co-orchestrating an attack on the village of Bogoro in the region of Ituri on 24 February 2003 and charged with the war crimes of rape, sexual slavery, outrages upon personal dignity, directing an attack against a civilian population, wilful killings, destruction of property and using child soldiers, and the crimes against humanity of rape, sexual slavery, and murder.Footnote 144 He was initially charged under Article 25(3)(a) as a co-perpetrator for the crime of using child soldiers and as an indirect co-perpetrator for all other crimes. However, during the deliberation phase at the end of the trial, the majority of the Trial Chamber notified the parties that they would consider whether the evidence presented during the trial actually satisfied a different mode of liability, namely common purpose liability under Article 25(3)(d)(ii).Footnote 145
The Trial Chamber’s decision to alter the mode of liability from Article 25(3)(a) to Article 25(3)(d) had to do with the Court’s earlier restrictive interpretation of Article 25(3)(a). After reviewing the evidence – in particular evidence relating to Katanga’s contribution as co-ordinator of the preparations for the attack on Bogoro – the Trial Chamber’s majority came to the conclusion that the evidence better supported liability under Article 25(3)(d) than under Article 25(3)(a).Footnote 146 This was later confirmed in the judgement issued by the Trial Chamber, which having chosen to follow Lubanga’s control over the crime approach to distinguish between principals and accessories, found the evidence insufficient to support Article 25(3)(a) liability.Footnote 147 As laid out in Katanga, indirect perpetration under Article 25(3)(a) requires proof of both ‘the existence of an organised and hierarchical apparatus of power, characterised by near-automatic obedience to the orders it hands down’ and the perpetrator’s ‘control and genuine authority over the organization … [such that he is able] to steer it intentionally towards the commission of a crime, without leaving one of the subordinates at liberty to decide whether the crime is to be executed’.Footnote 148 Applying this standard, the Chamber concluded that the evidence was insufficient to establish that:
(1) in February 2003, the Ngiti militia was an organised apparatus of power; and (2) Germain Katanga, at that time, wielded control over the militia such as to exert control over the crimes for the purposes of article 25(3)(a) of the Statute.Footnote 149
The Court then turned its attention to Article 25(3)(d), which attaches criminal liability to anyone who ‘[i]n any other way contributes to the commission or attempted commission of such a crime by a group of persons acting with a common purpose’.Footnote 150 The contribution must ‘be intentional’ and either:
(i) Be made with the aim of furthering the criminal activity or criminal purpose of the group, where such activity or purpose involves the commission of a crime within the jurisdiction of the Court; or (ii) Be made in the knowledge of the intention of the group to commit the crime.Footnote 151
The Trial Chamber interpreted this mode as ‘a residual form of accessory liability, included in the Statute to vest the Court with jurisdiction over accessories whose conduct does not constitute aiding or abetting the commission of a crime within the meaning of article 25(3)(c)’.Footnote 152 According to the Chamber, this mode differs from JCE in that it cannot be used to attach liability to an accused for crimes which form part of the common purpose but to which he or she did not contribute.Footnote 153 According to the Trial Chamber, an accused’s ‘contribution may be connected to either the material elements of the crime’ (for instance, the provision of resources such as weapons) ‘or their subjective elements’ (which may involve tacit or explicit encouragement).Footnote 154 Further, it held that although the group of persons acting with a common purpose must harbour the same intent to commit the crime and the contribution must be made in the knowledge of that intent, intent can be demonstrated either by showing the group meant to cause the consequences that constituted the crime or were aware that the crime would occur in the ordinary course of events.Footnote 155
In March 2014, a majority of the Trial Chamber convicted Germain Katanga as an accessory for all crimes with which he was charged except for rape and sexual slavery as war crimes and crimes against humanity and the war crime of using child soldiers. Specifically, the Chamber convicted Katanga as an accessory to the war crimes of directing an attack against a civilian population, pillaging, and destruction of property, as well as murder as a war crime and a crime against humanity, but unanimously acquitted Katanga as an accessory to rape and sexual slavery as war crimes and crimes against humanity, as well as of the war crime of using child soldiers.Footnote 156 The acquittal on the sexual violence charges was not due to an absence of credible evidence regarding the acts of sexual violence. In fact, the Chamber recognized that rape and sexual slavery had been committed by Katanga’s fighters on the day of the Bogoro attack.Footnote 157 What the Chamber found was that the sexual violence crimes were not part of the common purpose of the attack, unlike the crimes of directing an attack against a civilian population, pillage, murder, and destruction of property.Footnote 158
The Trial Chamber first found that the purpose of Katanga’s militia was not only to reconquer Bogoro from the opposing (UPC) troops but also to ‘wipe out’ the ‘Hema civilians’ from the village.Footnote 159 It looked at the manner of the attack – including that it was initiated in the dark against civilians, and that the crimes were committed systematically, with ‘great violence’ and accompanied by ethnic references – to conclude that the militia did, in fact, have ‘a common purpose of a criminal nature’.Footnote 160
The Trial Chamber then looked at whether the particular charged acts formed part of the common plan. It found that murder, directing an attack against a civilian population, pillaging, and destruction of property were part of this common purpose, in part because of the scale of the crimes and the fact that Katanga’s militia had engaged in these acts prior to the attack on Bogoro, enabling it to conclude that they must have been intended and therefore part of the common purpose.Footnote 161 The Chamber went on to note that, in contrast to these crimes, there was no evidence that rape and sexual slavery had been committed on a wide scale or that these crimes had been committed by Katanga’s militia before the attack on Bogoro.Footnote 162 It found that the ‘obliteration’ of Bogoro did not, therefore, ‘perforce entail … the commission of [rape and sexual enslavement]’.Footnote 163 Oddly, it concluded that ‘although rape and enslavement formed an integral part of the militia’s design to attack the predominantly Hema civilian population of Bogoro’,Footnote 164 these crimes did not form part of the group’s common purpose.Footnote 165
The Trial Chamber’s common purpose analysis is problematic for a number of reasons. First, it seems to imply that for an act to be part of the group’s common purpose – absent direct evidence that the group agreed to commit the act – it has to both have been committed on a wide scale and repeated basis and have occurred prior to the attack in question. Notably, neither of these factors is identified in the Chamber’s own findings on what the law requires or what would be sufficient to show that an act is part of the group’s common purpose in the absence of direct evidence.Footnote 166 In the part of the judgment interpreting the relevant law, the Chamber adopts an approach that excludes from the common purpose only those ‘crimes ensuing … from opportunistic acts by members of the group’.Footnote 167 It is hard to see how the rape and sexual enslavement of Hema women was merely opportunistic when these crimes were committed on the same day as the crimes committed against other members of the Hema community in Bogoro and, by the Chamber’s own assessment, ‘formed an integral part of the militia’s design to attack the predominantly Hema civilian population of Bogoro’.Footnote 168
Second, even if scale were required, it is unclear why the scale of the sexual violence was insufficient here. Only three witnesses testified to the rape, as opposed to 60 people who the Chamber found had been killed in the attack.Footnote 169 However, each of them testified to being raped multiple times, with a total of 17 acts of rape occurring during the attack on Bogoro.Footnote 170 Moreover, all of these witnesses were enslaved at a camp run by Katanga’s militia following the attack, for a period of one to 18 months.Footnote 171 It is hard to see why 17 acts of rape and subsequent enslavement for months do not qualify as crimes committed on a wide scale or repeated basis. Indeed, the Chamber’s failure to consider the number of assaults or the subsequent months of sexual enslavement shows a lack of gender competence in assessing the evidence related to the gender crimes.
Although it did not say so explicitly, perhaps the Chamber had difficulty finding that the sexual violence crimes were crimes the group either intended to commit or would occur within the ordinary course of events, which the accused must have been aware of at the time the crimes were committed.Footnote 172 Yet, in its discussion of the law, the Chamber itself notes that a group’s common purpose need not have been previously agreed upon; it can ‘materialize extemporaneously’ and ‘be inferred from the subsequent concerted action of the group of persons’.Footnote 173 Katanga’s militia not only repeatedly raped the women but enslaved them after the attack for up to 18 months. Even if the militia had not committed prior acts of sexual violence, the repeated acts of rape and subsequent months-long enslavement of Hema victims were evidence from which the Chamber could have inferred that the crimes were either intended or would occur within the ordinary course of events. Again, the failure to do so suggests a lack of gender competence in evaluating the evidence of SGBV crimes.
More significantly, the Chamber’s analysis appears to ignore the way in which sexual violence has been used in other conflicts to achieve the very kind of purpose the Chamber indicates Katanga’s militia wanted to accomplish, i.e., to forcibly displace or wipe out a rival ethnic group. Although the Chamber stated that it would reference the common purpose jurisprudence of the ICTY and the ICTR,Footnote 174 it failed to cite cases from these tribunals which had found that sexual violence had been used effectively to destroy or discriminate against a particular group and/or to drive them out of a particular territory and, therefore, formed part of the perpetrators’ common purpose. For instance, in the Stakić case, the ICTY Appeals Chamber found that the common purpose of Stakić and other JCE members was to deport Bosnian Muslims and Bosnian Croats from the town of Prijedor in Bosnia and to persecute them in order to establish Serbian control over that territory.Footnote 175 Importantly, the Trial Chamber had recognized that forcing people to flee required waging a persecution campaign against Bosnian Muslims and Bosnian Croats, which in turn was accomplished through various acts of violence, including sexual violence.Footnote 176 In confirming Stakić’s conviction for persecution, including through acts of sexual violence, the Appeals Chamber recognized that sexual violence formed part of the group’s common purpose.Footnote 177
In Katanga, the Trial Chamber examined the SGBV crimes apart from the other violent acts, as if they were somehow unconnected to the violence targeted against the very same civilians. The Chamber suggested that because the victims denied their ethnicity (which they did to avoid being targetedFootnote 178 ), the rapes were somehow not connected to the victims’ ethnicity, and therefore not a part of the intended attack against the Hema in Bogoro. Yet, as one author has noted, the transcripts of the witness testimonies indicate that they were all ‘asked about their ethnic identity repeatedly before and after being raped’,Footnote 179 suggesting that the issue of ethnicity was clearly related to the rapes irrespective of whether the victims denied their Hema heritage. As in Stakić, there was ample evidence in the record from which it could be inferred that that sexual violence was used – like the other crimes – to drive the Hema population from Bogoro.
It is hard to avoid the conclusion that the Chamber in Katanga isolated and treated SGBV crimes differently, both factually and legally, from the other charged crimes. Rather than viewing the evidence in the record as linking the sexual violence with the broader context in which it occurred, the Chamber analysed these crimes separately, requiring more concrete evidence than its own findings on common purpose liability suggest is legally required to show that the sexual violence was part of the common plan.
Importantly, even if the Chamber had found that the sexual violence crimes were part of the common plan, it is not clear that the Chamber would have found Katanga liable for these crimes under Article 25(3)(d), as per this provision, Katanga’s contribution would have to be ‘intentional’ and either:
(i) made with the aim of furthering the criminal activity or criminal purpose of the group, where such activity or purpose involves the commission of a crime within the jurisdiction of the Court; or (ii) made in the knowledge of the intention of the group to commit the crime.Footnote 180
According to the Chamber’s interpretation of Article 25(3)(d)(ii):
[k]nowledge must be established for each specific crime and knowledge of a general criminal intention will not suffice to prove … that the accused knew of the group’s intention to commit each of the crimes forming part of the common purpose.Footnote 181
Moreover, the accused must be aware this intention existed ‘when engaging in the conduct which constituted his or her contribution’.Footnote 182 Given the Chamber’s refusal to consider the repeated acts of rape during the attack and the subsequent months-long enslavement of Hema victims in its analysis of the militia’s common purpose, it is unlikely that the Chamber would find these same facts sufficient to show Katanga had contributed to the attack with the knowledge that the group intended to commit these crimes at the time he made his contribution.Footnote 183
The take-away from Katanga is clear. First, a restrictive interpretation of Article 25(3)(a) is likely to serve as a particularly high bar for cases involving sexual violence charges. It will likely be difficult to prove that those accused of SGBV ‘unquestionably … conceived the crime, oversaw its preparation at different hierarchical levels, and controlled its performance and execution,’ as required by the Court’s doctrinal construction of indirect perpetration,Footnote 184 given that sexual violence – even when widespread – often occurs because it is tolerated and permitted rather than explicitly ordered or planned.Footnote 185 Second, even though common purpose liability under Article 25(3)(d) can theoretically be used to hold perpetrators accountable for sexual violence crimes and the provision has not been subject to the same level of critique as JCE III liability, that provision has been applied differently to sexual violence crimes than other crimes, resulting in the acquittal of SGBV charges. Thus, the combination of a restrictive interpretation of Article 23(a) with a lack of gender competence in applying the elements of Article 25(3)(d) to SGBV crimes has led to a high level of impunity for those crimes.Footnote 186
4. Two trains running – Customary international law and the Rome legal regime
The above analysis suggests that accurately capturing and characterizing individual contributions to collective criminal conduct, particularly when that conduct involves sexual and gender-based violence, remains a challenge for international courts and tribunals. Despite significant advances in the development of sophisticated theories designed to hold individuals responsible for their role in atrocity crimes and the recognition of rape and other forms of sexual violence as war crimes, crimes against humanity and even acts of genocide, the cases discussed in the previous section demonstrate that judges struggle with how to understand and situate sexual and gender-based violence within the context of other atrocity crimes.
The solution, moving forward, will undoubtedly need to operate on several tracks. For various reasons, including perhaps the ‘positive complementarity’ incentive on states to investigate and prosecute atrocity crimes within their own domestic systems,Footnote 187 we have seen – and may continue to see – an increasing number of states trying atrocity crimes at home.Footnote 188 Some have incorporated the Rome Statute into their domestic legal systems, while others rely on existing penal codes and customary international law to try international crimes. Moreover, we will likely continue to see the establishment of new hybrid or internationally-supported tribunals with temporal mandates requiring the application of customary international law. Thus, it is important to explore potential solutions to the problems we have identified under both the Rome Statute and customary international law.
4.1 JCE III under customary international law
Our first recommendation – relevant primarily to jurisdictions applying customary international lawFootnote 189 – is to push back against the critique of JCE III. First, we disagree with the scholarly criticism and judicial dissents that reject JCE III’s customary basis. As an initial matter, we find the ad hoc tribunals’ response to these challenges persuasive.
In the Đorđević case, for instance, the accused challenged the bona fides of JCE’s customary origins, arguing that the Appeals Chamber should depart from its previous decisions because the reasoning set out in the Tadić Appeal Judgement is ‘shallow and uncertain’.Footnote 190 In support, he cited, inter alia, the ECCC’s Pre-Trial Chamber decision that concluded that JCE III was not rooted in customary law by the 1970s, the relevant time of the Khmer Rouge atrocities.Footnote 191 The ICTY Appeals Chamber disagreed and affirmed the customary basis of JCE III.Footnote 192 We concur for several reasons.
First, JCE III is rooted in legal instruments and the case law of the post-WWII era. For example, the final clause of Article 6 of the Charter of the International Military Tribunal established that:
Leaders, organisers, instigators and accomplices participating in the formulation or execution of a Common Plan or Conspiracy to commit any of the foregoing crimes [crimes against peace, war crimes, and crimes against humanity] are responsible for all acts performed by any persons in execution of such plan.Footnote 193
In other words, the IMT Charter provided that a person who participated in a common plan or conspiracy to commit any crime under the Charter could be held liable for all acts resulting from the execution of that common plan or conspiracy.Footnote 194 Similarly, Control Council Law No. 10 provided that a person could be held criminally responsible for a crime if he ‘was connected with plans or enterprises involving its commission’,Footnote 195 without any requirement that the crime in question have been intended.
The case law of the post-World War II tribunals further confirms that JCE III was a recognized mode of liability well before the establishment of the ICTY. Although the lack of reasoned judgments in many cases makes it difficult to determine with certainty the exact mode of liability applied, a careful review of the facts of those cases strongly supports the conclusion that they applied a mode of liability akin to JCE III. For example, in the Essen Lynching case, a Germany Army Captain instructed a private to transport three British prisoners of war through the town of Essen, and not to interfere if the civilian crowd were to ‘molest’ the prisoners – an instruction given within hearing of the crowd.Footnote 196 The crowd then proceeded to beat the prisoners, ultimately killing them.Footnote 197 Together, the captain, private, and crowd all implicitly agreed upon a common purpose – to attack the prisoners – thereby causing them injury. Even though it was unclear who struck the fatal blow, the five defendants were found guilty of the war crime of killing – not just assaulting – a prisoner of war, suggesting that the tribunal found them guilty of the foreseeable crime of murder and not just the agreed upon assault.Footnote 198 The Borkum Island case is similar. There, several airmen were killed by a civilian crowd after the military commander ordered his subordinates not to interfere with any attack by the crowd.Footnote 199 Although there was evidence that many of the civilians intended their murder,Footnote 200 there was no evidence that the commander intended that the prisoners be killed, rather than assaulted.Footnote 201 Nonetheless, the commander was found guilty in their deaths.Footnote 202
Third, we find the ECCC decision rejecting the customary status of JCE III to be unpersuasive.Footnote 203 Not only did the ECCC Pre-trial Chamber restrict its examination to sources relied on by Tadić rather than consulting other sources of custom,Footnote 204 but it did not dispute that the Essen Lynching and Borkum Island cases might ‘indeed be directly relevant to JCE III’.Footnote 205 Rather, the ECCC declined to rely on them ‘in the absence of a reasoned judgment’.Footnote 206 Yet other cases, not relied upon by Tadić or considered by the ECCC, support the conclusion that JCE III was a recognized mode of liability in the post-WWII era.
For example, in The Queen v. Ikeda, the Batavia Military Tribunal prosecuted a Japanese army colonel for war crimes for his role in the provision of Dutch ‘comfort women’ to Japanese troops who were raped and subjected to enforced prostitution.Footnote 207 The tribunal found that Ikeda was responsible, along with others, for planning the establishment of brothels filled with girls and women recruited from internment camps – conduct that, of itself, amounted to a crime against humanity.Footnote 208 Although there was no evidence that Ikeda had intended the use of force, the tribunal held that Ikeda was responsible for the additional ‘criminal offences committed in the process’ because they ‘could and should have been anticipated’, i.e., they were foreseeable, by Ikeda.Footnote 209 The Australian Military Tribunal likewise applied a mode of liability similar to JCE III. For example, in the Hatakeyama and Matsumoto cases – which concerned the deaths of individuals who had been tortured – the defendants admitted that they had intended to interrogate and/or torture the victims, but the evidence was ambiguous as to whether there had been any agreement to murder the prisoners. Nonetheless, the defendants were convicted of murder, apparently because it was a foreseeable result of their crimes.Footnote 210 The conclusion that post-WWII decisions were based on JCE III is buttressed by the arguments of the prosecutors and Judge Advocates in these cases. In Hatakeyama, for instance, the Judge Advocate stated before the Australian Military Tribunal that:
[i]f several persons combine for an unlawful purpose, or for a lawful purpose to be effected by unlawful means … and one of them, in the prosecution of it, kills a man, it is murder in all who are present.Footnote 211
Significantly, even the ECCC’s decision appears to leave room for the argument that some form of extended JCE existed under customary law in the post-Second world War period. Although the ECCC’s Supreme Court Chamber found ‘that criminal liability based on making a contribution to the implementation of a common criminal purpose was … limited to crimes that were actually encompassed by the common purpose’,Footnote 212 its discussion of which crimes are encompassed by a common purpose included ‘crimes that are foreseen as means to achieve a given common purpose, even if their commission is not certain’.Footnote 213 The Chamber went on to explain that:
if attaining the objective of the common purpose may bring about the commission of crimes, but it is agreed to pursue this objective regardless, these crimes are encompassed by the common purpose because, even though not directly intended, they are contemplated by it.Footnote 214
In language similar to that used by the ad hoc tribunals when assessing foreseeability for purposes of JCE III liability, the Chamber stressed that:
the common purpose may encompass crimes in which the commission is neither desired nor certain, just as it is sufficient for the commission of certain crimes that the perpetrator acted with dolus eventualis and therefore neither desired that the crime be committed nor was certain that it would happen.Footnote 215
Finally, the Chamber explained that:
[w]hether a crime was contemplated by the common purpose is primarily a question of fact that – absent an express agreement – has to be assessed taking into account all relevant circumstances, including the overall objective of the common purpose and the likelihood that it may be attained only at the cost of the commission of crimes.Footnote 216
Thus, even under the ECCC’s formulation of the JCE standard, a member of a JCE may be held responsible for crimes he or she did not intend and the commission of which was not certain, provided that the crimes were foreseeable as a means to achieve the common purpose.Footnote 217 This formulation appears to extend beyond the usual limits of JCE I and II, even if it does not fully encompass the full breadth of JCE III liability.Footnote 218
Fourth, there have been a number of decisions since the ECCC Pre-Trial decision and Đorđević’s appeal affirming the customary status of JCE III. Most significantly, the Extraordinary African Chambers (EAC) in Senegal convicted former Chadian dictator Hissène Habré of sexual slavery based on JCE III.Footnote 219 The trial chamber painstakingly reviewed the customary basis of JCE, examining the post-World War II jurisprudence, ICTY and ICTR jurisprudence and the ECCC decision.Footnote 220 In regard to JCE III, the chamber concluded that it was part of customary law as of the time the relevant events occurred in 1982Footnote 221 and that it was sufficiently foreseeable and accessible to the accused at that time.Footnote 222 In sum, we find persuasive that, time and again, JCE III has passed the test of judicial scrutiny.Footnote 223
We also find unpersuasive the argument that JCE III liability is tantamount to the imposition of guilt by association or collective guilt. Unlike guilt by association, in which mere membership in an organization is sufficient for liability, a finding that an accused is responsible for crimes committed as part of a joint criminal enterprise requires an intent to commit a crime as well as the participation of the accused in the execution of the JCE’s common purpose.Footnote 224 Indeed, there can be no JCE III liability without participation by the accused in either a JCE I or II, meaning there must first be a finding that the accused intended and participated in the JCE I common plan or JCE II system of ill-treatment. In other words, an individual charged with JCE is ‘charged not for his membership in a joint criminal enterprise but for his part in carrying it out’.Footnote 225 As the Martić Appeals Chamber explained,
when all the elements of JCE are met in a particular case, the accused has done far more than merely associate with criminal persons. He has the intent to commit a crime, he has joined with others to achieve this goal, and he has made a significant contribution to the crime’s commission. Thus, he is appropriately held liable also for those actions of other JCE members, or individuals used by them, that further the common criminal purpose (first category of JCE) or criminal system (second category of JCE), or that are a natural and foreseeable consequence of the carrying out of this crime (third category of JCE).Footnote 226
Further, as Antonio Cassese has observed, an accused’s participation in the original common plan puts him in a privileged position to foresee the further crimes that could arise out of the original JCE.Footnote 227 Thus, imposition of JCE III liability in such cases ‘convey[s] the message that [the accused] should have opposed or impeded the crime of the “primary offender”’.Footnote 228
In addition to the guilt by association critique, some critics contend that JCE III could lead to unfair convictions, in the sense that its application could lead to the unfair labelling of minor participants as principals.Footnote 229 As an initial matter, while theoretically possible, we find it significant that despite decades of jurisprudence by the ad hoc tribunals, critics have failed to establish that, in practice, the use of JCE III has resulted in a serious miscarriage of justice.Footnote 230 Moreover, although the Tadić Appeals Chamber held that JCE – including JCE III – is a form of ‘commission’,Footnote 231 this need not imply that JCE III must always be viewed as a principal rather than accessorial form of liability.Footnote 232 Indeed, domestic or hybrid courts could use JCE III to hold accused liable for crimes that are a natural and foreseeable consequence of a common criminal plan without necessarily characterizing it as principal liability. This would allow courts to use JCE III when other forms of liability, like aiding and abetting or command responsibility, might understate or not fully capture the way in which an accused’s participation in a group’s common criminal plan was inextricably linked with and led to the foreseeable commission of sexual violence crimes.Footnote 233
Significantly, JCE III is critical to capturing how SGBV crimes often unfold during times of mass violence. Sometimes rape is ordered by military or political leaders, but often it is not explicitly planned or orchestrated from the outset. Sexual violence may initially be committed because the atmosphere of war and the violence, lawlessness, and chaos it produces permits it to occur. However, once it becomes clear that superiors do not disapprove of sexual violence, the ‘opportunistic rapes typically then become more public, more frequent, and more violent, growing indistinguishable from and becoming part of the organized rapes committed at least in part to inflict widespread terror and/or harm on the targeted group’.Footnote 234 Thus, while sexual violence may at first appear unintended, it is often connected to the commission of other intended crimes. JCE III captures the reality of the way in which this violence unfolds.Footnote 235
In sum, in addition to agreeing with the decisions affirming the customary status of JCE III, we believe JCE III’s ability to capture the unique contextual circumstances under which sexual violence is often perpetrated in times of conflict or mass violence provides yet another compelling reason to push back against the critique of JCE III. The importance of the JCE III contextual analysis of SGBV crimes is illustrated in the final trial judgment in Prosecutor v. Karadžić.Footnote 236 In convicting Karadžić for a myriad of JCE III crimes, the Trial Chamber held that it:
is convinced beyond reasonable doubt that it was foreseeable to the Accused that persecution through torture, beatings, physical and psychological abuse, rape and other acts of sexual violence, and the establishment and perpetuation of inhumane living conditions in detention facilities as cruel or inhumane treatment, killings, [and] forced labour at the frontline … might be committed by Serb Forces used to carry out the objective of the common plan, during the execution of the common plan, with discriminatory intent.Footnote 237
Thus, we suggest that courts heed customary international law and its attendant liability forms, including JCE III, when assessing liability for SGBV crimes under international criminal law.
4.2 Article 25(3) of the Rome Statute
Our second recommendation relates to the ICC’s interpretation of Article 25(3) of the Rome Statute, in particular the provisions relating to co-perpetration, indirect perpetration, and common purpose liability. As discussed above – and as both scholars and dissenting judges at the ICC have pointed out – the Court’s restrictive interpretation of Article 25(3)(a) is neither well-grounded in the Statute nor mandated by customary international law. Moreover, by limiting Article 25(3)(a) to those with control over the crime and requiring the Court to rely on Article 25(3)(d) in cases of group liability where the facts do not support such a limited interpretation while at the same time applying Article 25(3)(d) differently to crimes of sexual violence than other crimes, the Court has adopted an approach with negative consequences for the prosecution of SGVB crimes. Indeed, although Article 25(3) appears facially neutral in its application to crimes within the jurisdiction of the ICC, the Court’s interpretation of these modes of liability, with the recent exception of Ntaganda, has adversely impacted the prosecution of SGVB crimes. If the ICC is to fulfil its mandate, it must recognize and address these issues. We identify here two ways the Court might do this:
4.2.1 Adopt a plain reading interpretation of Article 25(3)(a)
One possibility is for the Court to revisit its interpretation of commission under Article 25(3)(a). Rather than interpreting Article 25(3) as requiring a distinction between principals and accessories or importing the control over the crime theory to give effect to this distinction, the Court could adopt a plain reading of the provision, as dissenting judges Fulford and Van den Wyngaert have suggested.Footnote 238 There is nothing in the text of Article 25(3) itself that warrants the imposition of the hyper complexity that the Court’s early case law has developed, in which the term ‘indirect co-perpetrator’ is used to describe the responsibility of top-level accused (like Muammar Gaddafi) for crimes committed by their own forces,Footnote 239 and the inclusion of a ‘common plan’ is required for ‘committed jointly’ under Article 25(3)(a), even though ‘common plan’ is absent Article from 25(3)(a) and explicitly referenced in Article 25(3)(d). Under standard canons of treaty interpretation, as well as Article 21 of the Rome Statute, looking to plain language should be the starting point of the Court’s analysis,Footnote 240 as opposed to references to abstract legal theories such as Roxin’s control of the crime theory. This might lead not only to ‘a simpler, more internationally acceptable and predictable understanding of modes of liability at the ICC’Footnote 241 as one of us has argued, but also to greater accountability for SGBV crimes. To the extent, for instance, joint commission is interpreted to require some kind of co-ordinated actionFootnote 242 (rather than a ‘common plan’Footnote 243 ) and a directFootnote 244 (rather than an ‘essential’Footnote 245 ) contribution, it would likely be more possible to hold perpetrators accountable for the way in which collective sexual violence often unfolds in situations of conflict – again, not necessarily orchestrated, yet nevertheless interconnected with the commission of other intended crimes.Footnote 246
This approach is consistent with that of the ad hoc tribunals, which have recognized that perpetrators can be held responsible for sexual violence crimes not only when those crimes clearly form a component of a common plan from the outset, but also when they become part of the group’s action over time. In the Krajišnik case, for instance, while sexual violence was not part of the group’s original plan to forcibly remove Bosnian Muslims and Bosnian Croats from areas of Bosnia, the Trial Chamber found the accused could nevertheless be held liable for the sexual violence, after the prosecution showed that the accused had been informed of the violence, did nothing to prevent its recurrence, and persisted in pursuing the group’s plan.Footnote 247 As the ICTY explained:
Whether other crimes were “original” to the common objective or were added later is of course a matter of evidence, not logical analysis. The Chamber’s preference is for a strictly empirical approach which does not speculate about the crime profile of the original JCE objective, but conceptualizes the common objective as fluid in its criminal means.Footnote 248
While the facts in Krajišnik would likely not have met the strict ‘common plan’ and ‘essential’ contribution standards required by the control over the crime theory, they would likely be sufficient to hold perpetrators responsible for the sexual violence if the Court adopted a plain reading of Article 25(3)(a).
We underscore that Article 25(3)(a)’s restrictive outcomes depend upon the judicial interpretation of factual evidence admitted by the chamber about the common plan. A tilt toward a broader application of the facts to Article 25(3)(a) is tenable, as demonstrated in Ntaganda. There, the Trial Chamber deliberated upon the common plan of the UPC/FPLC, which was to occupy key positions in Ituri through a military operation that involved conducting an attack on Lendu civilians.Footnote 249 To understand the common plan, the Chamber examined the overall objective of the attacks, including the content of military instructions and direct orders.Footnote 250 Likewise, evidence of discussions to plan the operations, months prior to the deployment, and, most notably, a side conversation about infliction of rapes to instil fear on the enemy were contemplated.Footnote 251 Accordingly, the Ntaganda Chamber viewed the attack to destroy, disintegrate and drive out the Lendu community as an agreement that ‘inherently involved the targeting of civilians individuals by way of acts of killing and raping as well as … acts of appropriation and destruction’ of property.Footnote 252 SGBV crimes were integralFootnote 253 to execution of the agreement:
Regarding acts of sexual violence, the Chamber notes that the unfolding of these acts were like the acts of killings and other acts of physical violence, a tool used by the UPC/FPLC soldiers and commanders alike to achieve their objectives to destroy the Lendu community The Chamber notes the words pronounced by a UPC/FPLC soldier during a rape and at a moment when many of the victims were raped: he compared the Lendu to non-human elements to be exterminated. It is significant that the UPC/FPLC soldiers killed or attempted to kill many of the civilians they subjected to sexual violence. A survivor compared her experience to dying.Footnote 254
Further, although the Chamber found certain acts of sexual violence not within the co-perpetrators’ express understanding of the common plan, it nonetheless found their commission a virtual certainty, given the circumstances prevailing at the time, and therefore part of the plan.Footnote 255 As the Chamber stated,
the co-perpetrators were virtually certain that the implementation of their plan to drive out all the Lendu from the localities targeted during the course of their military campaign … would lead to: (i) the recruitment and active use in hostilities of children under the age of 15 within the UPC/FPLC (Counts 14, 15 and 16); and (ii) the rape and sexual slavery of these children (Counts 6 and 9). Indeed, the Chamber finds that, in the circumstances prevailing in Ituri at the time, the occurrence of these crimes was not simply a risk that they accepted, but crimes they foresaw with virtual certainty.Footnote 256
Thus, Ntaganda emphasized the contextualization of all the criminal acts, highlighting the connection between the crimes rather than characterizing the sexual violence as unexpected and unrelated to the other crimes.
Notwithstanding the Chamber’s broad approach to the common plan, Ntaganda’s liability as an indirect co-perpetrator required the Chamber to assess the accused’s control over the crime through his essential contribution to it and resulting ability to frustrate its commission.Footnote 257 While it found such control to be present in this case,Footnote 258 cases in which sexual violence unfolds as part of the plan but evidence of the accused’s control over the crime is less compelling would still likely fail. Moreover, whether Ntaganda represents a one-time exception or a partial attempt to correct the Court’s previous failures to recognize and understand how sexual violence evolves in the context of collective criminal conduct remains to be seen.
4.2.2 Use a gender informed analysis when applying Article 25(3)(d) to SGBV crimes by examining the overall context in which the sexual violence occurred and the role such violence played in the group’s common plan
In Katanga, the Court applied Article 25(3)(d) differently to SGVB crimes than to other crimes. Yet, as the Appeals Chamber underscored in Đorđević, sexual violence crimes ‘must not be treated differently from other violent acts simply because of their sexual component’.Footnote 259 Taking a contextual approach to analysing whether SGBV crimes formed part of the common plan would help the court avoid isolating and treating such crimes differently from other crimes. As the Stakić case discussed aboveFootnote 260 suggests, examining the overall context in which the sexual violence occurred, and the role such violence played in achieving the objectives of the group, would allow the Court to properly understand whether sexual violence crimes were part of the common plan, rather than opportunistic or unrelated crimes. Had the Chamber in Katanga case taken a similar approach, the result might have been different.
Significantly, taking into account context might also help the Court more accurately assess whether an accused’s contribution was made in the knowledge of the intent of the group to commit the crime, an element also required under Article 25(3)(d). According to Katanga, this knowledge requirement can be shown either by proving ‘the group meant to cause the consequences which constituted the crime’ or was ‘aware that the crime would occur in the ordinary course of events’.Footnote 261 Although the latter phrase – as interpreted by the Lubanga Appeals Chamber – requires ‘virtual certainty’ that the crime will occurFootnote 262 (presumably a higher standard than the foreseeability standard in JCE III cases), the factors used by the ad hoc tribunals to assess whether SGBV crimes were connected with other crimes, and therefore foreseeable or even intended, may be equally helpful in meeting this standard.Footnote 263 Indeed, while scale and prior commission of SGBV crimes may help show the accused was aware that sexual violence would occur in the ordinary course of events, other factors such as the overall context in which the violence occurred and the objectives of the group may be equally relevant and sufficient to show the requisite knowledge. For instance, in the Kvočka case, the ICTY found rape in detention was foreseeable – despite the absence of evidence showing that the accused knew that women had been previously raped in that camp – by drawing common sense inferences from the surrounding circumstances.Footnote 264 The Trial Chamber first noted that ‘[a]pproximately 36 women were held in detention, guarded by men with weapons who were often drunk, violent and physically and mentally abusive and who were allowed to act with virtual impunity’. It then concluded that:
it would be unrealistic and contrary to all rational logic to expect that none of the women held in Omarska, placed in circumstances rendering them especially vulnerable, would be subjected to rape or other forms of sexual violence. This is particularly true in light of the clear intent of the criminal enterprise to subject the target group to persecution through such means as violence and humiliation.Footnote 265
This approach could be equally helpful in meeting the knowledge requirement in Article 25(3)(d).
Interestingly, the Trial Chamber’s analysis in Ntaganda, although examined under Article 25(3)(a), could be instructive. As discussed earlier, although the Chamber found that the rape and sexual slavery of child soldiers was not within the co-perpetrators’ express understanding of the common plan, it nonetheless found the commission of these crimes a ‘virtual certainty’, given the circumstances prevailing at the time, and therefore part of the plan.Footnote 266 Indeed, as with Kvočka, the Chamber found the overall context in which the violence occurred and the objectives of the group relevant and sufficient to show that the perpetrators were virtually certain that those crimes would be committed. Clearly, if this was adequate to meet the common plan standard under Article 25(3)(a), it should also suffice to meet the knowledge standard under Article 25(3)(d).
5. Conclusion
We conclude by offering a few thoughts about the benefits of adopting these approaches to international criminal law more broadly. Using a contextual approach to assessing foreseeability under JCE III and common purpose liability under the Rome Statute may help international criminal courts and tribunals more accurately apply other modes of liability where foreseeability or constructive knowledge is a required element, such as command responsibility, which makes liable commanders who knew or should have known that crimes were – or were about to be – committed by his or her subordinates. In our view, a contextual approach may have resulted in a different outcome in the Bemba case, which acquitted Bemba – who had organized, paid and led a mercenary army – of the SGBV crimes committed by his troops because, among other things, he was a ‘remote’ commander.Footnote 267
More generally, acknowledging how sexual violence actually unfolds during times of mass violence and using a contextual approach to assessing common purpose liability or co-perpetration might also help make more visible conduct that is often overlooked in the context of conflict or mass violence, such as sexual violence against males. Although sexual violence against males has been infrequently investigated and prosecuted, it too has played a role in atrocity crimes, especially detention-related violence.Footnote 268
Adopting this approach might also help more accurately explain the motivation of perpetrators, a part of the narrative not often explored by the tribunals. As others have noted, not all perpetrators are deviants. It is sometimes a shockingly ordinary person that participates in atrocity crimes.Footnote 269 Taking the time to understand and recognize the context in which violence occurs might help demonstrate more clearly what motivated the perpetrators to act. Understanding these dynamics more clearly will not only help tribunals more accurately characterize the liability of the accused but may also help the international community develop more effective prevention strategies.
Finally, it is worth observing that the International Criminal Court is the institution charged with the mandate to carry international criminal law forward at the international level. The ICC is embedded more generally in the international legal system, a legal system in which power imbalances between genders, between rich and poor, between strong and weak play out just as they do in national legal systems. Yet we can and should expect better of the Court – an institution charged with the administration of justice not only for the strong, but especially for the weak. A high level of impunity for SGBV crimes persists at the ICC, even though the situations giving rise to acquittals for SGBV crimes are universally acknowledged to be situations in which rape, sexual slavery and other SGBV crimes are and have been endemic. While much criticism has been directed at the Office of the Prosecutor regarding those acquittals,Footnote 270 following the decisions in the Katanga and Bemba cases, it is now clear that the Court’s highly restrictive jurisprudence on modes of liability presents a major obstacle to the successful prosecution of these cases,Footnote 271 at least as regards individuals of high rank who did not themselves perpetrate the crimes. This, of course, undermines the very purpose for which the ICC was established, following the Nuremberg and Tokyo precedents. While the Trial Chamber’s judgment in Ntaganda may represent a partial attempt to right the near-categorical impunity for SGBV crimes seen thus far at the ICC, it is not yet clear if the case will be appealed or what impact it will have on other cases now at trial, such as the Ongwen and Al Hassan cases.Footnote 272 Absent a decisive shift toward – and beyond – the Ntaganda approach, the Court’s otherwise restrictive jurisprudence presents a cautionary tale about how the elaboration of well-meaning intellectual constructs can deprive legal text of practical force and effect; and how – in spite of the advances made – international criminal law remains, like international law itself, a ‘thoroughly gendered system’.Footnote 273
