A Battle of Ideas: Modes of Liability and Mass Atrocities

International criminal law (ICL) is premised upon the principle of individual criminal responsibility—the idea that individual persons, rather than states or other entities, bear criminal responsibility for mass atrocities. This principle has been operationalized in practice through the development of various legal theories, technically known as “modes of liability,” that delineate individual contributions to collective forms of violence. However, the modes of liability have been interpreted and applied differently across international courts and tribunals. Crucially, the International Criminal Court has construed those modes in a significantly more restrained manner than previous international tribunals, essentially limiting the set of situations in which a person could be found guilty of an international crime. This article contributes to the interdisciplinary literature on international criminal justice by exploring the social construction of different conceptualizations of the modes of liability in ICL and their transformations over time. Specifically, the framework focuses on the promotion and contestation of ideas among various actors participating in the international criminal justice field and the power dynamics that govern the field. This is a key issue for ICL and global justice because it demarcates the legal boundaries of “guilt” and “accountability” for international crimes.


INTRODUCTION
In the aftermath of the Second World War, the International Military Tribunal (IMT) in Nuremberg, which was established to prosecute the defeated Nazi major generals, famously proclaimed that international crimes were committed by "men, not abstract entities." 1 Thus, the Nuremberg IMT placed the foundations of a new system of international law that was premised on the idea of holding individuals, rather than states or other collective entities, criminally responsible for conducts such as genocide, crimes against humanity, war crimes and aggression.From this perspective, international criminal law (ICL) was equipped to hold accountable those "individuals who, normally hiding behind the shield of state sovereignty, grossly breached human rights" (Cassese 2011, 272).
The foundational principle of international criminal justice-individual criminal responsibility-has been operationalized in practice through the development of various legal theories that seek to delineate individual contributions to collective forms of violence.Those legal theories have become technically known as "modes of liability," or modes of participation in a crime, and have varied across international courts and tribunals: from the "conspiracy" charge used at the IMTs at Nuremberg and Tokyo, to the famous "joint criminal enterprise" theory developed at the United Nations (UN) tribunal for the former Yugoslavia (ICTY), and, more recently, the "control-over-thecrime" theory adopted at the permanent International Criminal Court (ICC).
The manner in which modes of liability are interpreted and applied is of crucial importance for the operation of international criminal justice because those legal theories elucidate the link between the accused and the crime.The ICC has delivered only nine verdicts as of June 2022.Yet even this modest record illustrates the differences between trial outcomes at the ICC and the UN tribunals-namely, the ICTY and the International Criminal Tribunal for Rwanda (ICTR).While the UN tribunals held an acquittal rate of about 15-20 percent, 2 the ICC's acquittal rate currently stands at 44 percent (five convictions and four acquittals for core international crimes). 3This raises an important question: why has the ICC approached the assessment of criminal responsibility differently from other international tribunals?
A potential explanation might be that ICL has "matured tremendously" over the past few decades (a term borrowed from Combs 2017, 124), leading to greater respect for legal norms at the ICC compared to its predecessors.This idea has not been developed as an independent thesis in ICL scholarship.Rather, it constitutes an implicit assumption among many legal scholars and practitioners that logically follows from the liberal notions of progress, development, and incremental improvement.The issue with this line of reasoning is that it presupposes a single model of the perfect international legal order at which ICL aims.Yet, in practice, it is far from clear what are the requirements of that perfect legal order.As observed by Darryl Robinson (2020, 95), the plurality of plausible moral standards that could underpin such an order, along with the "malleability and imprecision" of those standards, has rendered unsuccessful the search for a single "correct" moral theory from which the ICL principles can be deduced.
This article brings a new point of view to the vital question of the development of modes of liability in ICL by exploring the construction of different conceptualizations of those modes and their transformations over time and, specifically, by looking at the promotion and contestation of ideas among the various actors participating in the international criminal justice field.The article examines why some approaches to assessing criminal responsibility have gained popularity while others have been vigorously contested, and it traces the implications of these developments for the outcomes of international trials.Some might argue that the variety of approaches to assessing criminal 2. Based on data from '"Key Figures of the Cases," International Criminal Tribunal for the former Yugoslavia, https://www.icty.org/en/cases/key-figures-cases;"The ICTR in Brief," International Criminal Tribunal for Rwanda, https://unictr.irmct.org/en/tribunal.
3. Excluding the cases for offences against the administration of justice.See "Defendants," International Criminal Court, https://www.icc-cpi.int/defendants.responsibility is a matter of secondary importance for international criminal justice since what ultimately matters is holding the trial.As Antoine Garapon (2004, 717) observes, the power of international courts and tribunals plays out not merely in the "power to judge" but, more crucially, in the "power to prosecute"-namely, "to designate, isolate and, therefore, disqualify a human being" from the community of mankind. 4 Consequently, the defendants in international trials might appear to be of instrumental value for obtaining the ultimate goal of ICL-namely, to express the utmost form of condemnation concerning international crimes (Devresse and Scalia 2016, 811).
Nevertheless, questions of legal procedure, and, especially, the process of assessing the defendant's criminal responsibility, remain of crucial importance for international criminal justice (Stolk 2018, 156), not least because of its symbolic function (Minkova 2021).The modes of liability demarcate the legal boundaries of "guilt" and "accountability" for international crimes.The rules for assessing criminal responsibility also bear implications for trial outcomes and, ultimately, for the operationalization of the promise codified in the ICC's Rome Statute to end impunity for international crimes. 5It is true that the very existence of the Rome Statute to an extent fulfills this goal because it signals to the international community the wrongfulness of international crimes. 6 It is also true that liberal trials are bound to end up in some acquittals and that the trial record should not bear implications for the court's institutional standing.But, in reality, at least part of the court's audience, including the state parties on which the court depends on for funding, often measure the performance of international courts and tribunals in terms of successful prosecutions (Combs 2010, 358). 7 Consequently, the construction of modes of liability in ICL has not only legal, but also practical, implications for the operation of international criminal justice.This article engages with the construction of modes of liability by borrowing insights from "practice" studies in international relations.Building upon the work of Emanuel Adler and Vincent Pouliot (2011; see also Pouliot 2008), I examine the "background" knowledge, or the "thoughtless and inarticulate" rules (Pouliot 2008, 270), pursuant to which actors within the international legal field operate.These background rules are reproduced and modified through the everyday practices of the actors participating in the field or the so-called "community of practice" (Adler and Pouliot 2011, 17).This process can involve different kinds of actors and different forms of action.Specifically, while states continue to play an important role within the community of international legal practice, other actors such as judges, legal scholars, and non-governmental organizations (NGOs) gain leverage with respect to the interpretation of the laws, which affects trial outcomes.
This article aims to contribute to the literature engaged with questions of modes of liability in ICL by exploring the normative clashes between the interpretative 4. On "symbolism," see also Aksenova 2017. 5. Rome Statute of the International Criminal Court, 1998, 2187 UNTS 90. 6. See the third section of this article.7.For instance, at the 2018 International Criminal Court's Assembly of State Parties, the United Kingdom expressed serious concern about the low level of core crime convictions (three at the time): "After 20 years, and 1.5 billion Euros spent" on the court."UK Statement to ICC Assembly of States Parties 17th Session," Foreign and Commonwealth Office, December 5, 2018, http://www.gov.uk/government/speeches/uk-statement-to-icc-assembly-of-states-parties-17th-session.

Modes of Liability and Mass Atrocities 511
preferences of different actors within the ICL field, including legal academics, and tracing the influence of these preferences on the development of modes of liability.Hence, the article not only engages with legal scholars such as Elies van Sliedregt (2012), Kai Ambos (2013), andRobinson (2020) but also assesses the role of their and other academics' writings in (re)producing particular visions of criminal responsibility.The article reflects upon the findings of a three-year research project on the construction of standards of liability in ICL and a forthcoming book on the subject (Minkova 2023).The project comprises a study of the official records of the ICC's establishment, more than two hundred legal documents from the court's archive, three hundred academic publications, and numerous NGOs' statements.
The primary method for collecting data was discourse analysis, supplemented with interviews with ICL scholars and practitioners.The analyzed documents include the official records of the Preparatory Commission for the establishment of the ICC and legal decisions at the ICC and the UN tribunals (on issuing arrest warrants, confirming charges, trial and appeal judgments), which were available at the court's archive and the Legal Tools database. 8The article further employs discourse analysis of academic legal publications from peer-reviewed journals that publish extensively on ICL topics. 9Academic discourse is explored from a critical perspective-as a means for reproducing, justifying, and contesting legal norms.Finally, the analysis includes statements by NGOs monitoring the work of the ICC, such as Women's initiatives for Gender Justice, and international human rights NGOs, such as Amnesty International.
The empirical analysis shows that earlier tribunals and the ICC 10 have followed different visions of what the principle of individual criminal responsibility entails, which has resulted in different conceptualizations of the modes of liability at those institutions.Former international tribunals, like the ICTY and ICTR, often interpreted broadly the notion of "personal culpability," but this was deemed acceptable within the ICL community of practice at the time due to the perceived necessity to put in motion the wheels of international criminal justice.The tribunals' broad interpretation of modes of liability, such as "joint criminal enterprise" (JCE), reflected the background understandings about the law of those practitioners who sought to satisfy the demands of victims to see justice being served.However, the establishment of the permanent ICC and its founding document-the Rome Statute-opened room for new actors to enter the ICL field, notably criminal law practitioners with a civil law background, which changed the socially meaningful standards of acceptable legal decision making.Starting in the mid-1990s, these actors called for the codification of the general principles of criminal law in the Rome Statute and for strict compliance with those principles at the future court, even if the results were more acquittals (Eser 1993;Ambos 1996) 10.This article focuses on the ICC as the only permanent international criminal law (ICL) court.Future research could examine whether similar trends have occurred at the "hybrid' tribunals, such as the Special Court for Sierra Leone and the Extraordinary Chambers in the Courts of Cambodia, or the trial of Hissène Habré.
the modes of liability in a manner that further restricts the scope of situations in which the defendant could be found guilty, even perhaps going further than what was intended by some of the statute's drafters.From this perspective, fidelity to the legal process upholds the liberal values that underpin international criminal justice.This article is structured as follows: the second section discusses the development, institutionalization, and enforcement of ICL principles; the third section turns specifically to the modes of liability; and the fourth section concludes.

IDEALS, INTERESTS, AND PROFESSIONAL NORMS IN ICL The Socio-political Embeddedness of Legal Norms
The socio-political embeddedness of international criminal justice points to two potential explanations of the question why the ICC and other international tribunals have interpreted and applied the individual criminal responsibility principle differently: that powerful states have sought to constrain the mandate of the ICC and that the differences in the professional backgrounds of ICL academics and lawyers has created clashes in the interpretation of legal principles.To begin with the first explanation, the ICC and the UN tribunals were designed differently.The statutes of the UN tribunals11 were significantly less detailed compared to the ICC's Rome Statute, which left more room for judicial creativity to the judges at the UN tribunals compared to their colleagues at the permanent court.The level of judicial independence that the ICTY was granted has been explained by the seeming confidence of the UN Security Council's member states in the low probability that any of their nationals would be prosecuted by those tribunals (Danner 2006, 19-20).However, the Security Council had a different predisposition toward the ICC.Unlike the ICTY and the ICTR, whose jurisdiction was limited to addressing specific instances of mass atrocities, the ICC constitutes a permanent institution that could potentially prosecute crimes committed worldwide.Consequently, authoritative commentators have noted that, during the establishment of the ICC, states have sought to constrain the future court to a far greater extent than the UN tribunals (Cassese 2008, 526).To protect their nationals from prosecution at the permanent international court, the argument goes, powerful states have curbed the freedom of action of ICC judges by drafting in significant detail the Rome Statute (Hunt 2004).Darren Hawkins and colleagues (2006, 29-30) suggest that, as a result of "fears of a runaway international court," states have limited the discretion of the ICC, for instance, by the Rome Statute's "complementarity" provision, which enables states to challenge the court's jurisdiction by genuinely investigating the crimes at the domestic level.
While there is certainly merit to this observation, it is usually made in a rather generic manner that obscures the complexities of the dynamics taking place inside the ICL field.States remain highly influential actors in international relations, but modern interdisciplinary analyses have highlighted that states are not the only, or perhaps not even the most authoritative, actors in the field of international law (Helfer and Slaughter 2005;Alter 2014).Nicole Deitelhoff's (2009) study highlights the significant role played by NGOs, supported by a group of small-and middle-state powers, in countering the highly constrained project design of the ICC, which was originally supported by the permanent members of the Security Council, and in lobbying for some limited freedoms for the ICC judges and the prosecutor (see also Hall 2004, 125).
This article similarly examines the interplay of interests and ideas in the ICL field by providing two types of further insights.First, the interdisciplinary literature has extensively examined the influence of states and NGOs during the drafting of the Rome Statute's provisions concerning the court's jurisdictional scope, including its "complementarity" mechanism (De Vos 2020) and the powers of the prosecutor (Deitelhoff 2009).However, the role of state and non-state actors in drafting the individual criminal responsibility provisions in the ICC's statute has attracted less scholarly attention among international relations scholars, despite the significance of that part of the statute for fulfilling the court's mandate.Second, Deitelhoff's study as well as other studies depict a continuous struggle between, on the one hand, powerful states seeking to protect their sovereignty and, on the other hand, non-state actors, supported by small-and middle-state powers, promoting "the fundamental principles of law" over state interests (60).While there is merit to this account, it neglects the lack of a consensus within the ICL field over defining the precise requirements of the fundamental principles of law.The question whether state sovereignty or international law should take precedence might be one source of contestation in ICL, but the purpose, requirements, and goals of international law are similarly disputed among legal scholars and practitioners.Assuming there is a dichotomy between "state interests" and "legal principles" in ICL suggests the existence of an objective legal order that cannot be located anywhere and removes international legal principles from the realm of contestation.
The legal scholarship on international criminal justice has provided another insight into the reasons why the individual criminal responsibility principle has been construed differently across time and space-namely, that the ICL field has integrated legal experts with different legal backgrounds, which may sometimes result in clashing interpretations of legal norms.Most prominently, the work of Darryl Robinson (2020, 22-24) and other scholars has highlighted a schism between those ICL practitioners with a background in public international law, such as international humanitarian law and human rights law, and those with a background in domestic criminal law.These branches of law significantly differ in terms of their purpose, guiding principles, and the methods they use to interpret legal provisions.12Scholarly work has also observed that the practices of assessing criminal responsibility somewhat vary across domestic penal systems, such as the Anglo-Saxon system and the continental system (see van Sliedregt 2012).Because the ICL field integrates criminal lawyers with professional background in a variety of national penal systems, it makes reaching a consensus on the appropriate requirements of the individual criminal responsibility principle particularly challenging.
This article seeks to contribute to these scholarly observations by providing new insights into the dynamics taking place in the ICL field.Specifically, this article explores the constant (re)construction of the modes of liability in ICL by taking into consideration two factors.First, I will examine the interaction between different ideas about the law or, in other words, the combined impact of the "common law versus civil law" and the "international law versus criminal law" cleavages on the conceptualization of the modes of liability in ICL.Second, I will explore the institutional context that provides the arena for the competition of ideas to take place.The establishment of the ICC was marked not only by a shift in the position of states-from expressing little interest in the statutes of the UN tribunals to becoming actively involved in the design of the ICC-but also by the unprecedented participation of NGOs and legal experts.
The rest of this section presents a framework of analysis, which borrows from "practice" studies to explain the construction of legal norms and ideas.The following section relies on this framework to analyze the construction and application of the modes of liability in the ICL field.

International Law: A Dynamic Phenomenon
Borrowing insights from Friedrich Kratochwil (1989) and Jutta Brunnée and Stephen Toope (2010, 22), this article approaches international legal rules as "continuing struggles of social practice" (see also Johnstone 2011), where "social practice" means "socially meaningful patterns of action" that guide interaction within epistemic communities (Adler and Pouliot 2011, 6).Inside the legal domain, social practices comprise the application of legal rules to the case at hand and the process of appraising the reasons given for a particular decision before the community of legal practice.This pattern of rule following distinguishes social practice in the form of legal reasoning from other social practices, such as political and moral judgments (Kratochwil 1989, 18).
Practice studies enable the researcher to trace the interaction between the structure of the legal field-the rules of practice-and the actions of the agents within that field.The rules of legal practice are not to be found in a perfect legal code but are instead reproduced in the everyday application of those rules within the epistemic community of international law (see Adler and Pouliot 2011, 17).At any point in time, the participants in that epistemic community are bound by "shared understandings" concerning the appropriate interpretation and application of legal rules, which they use as guidelines to assess the validity of legal reasoning (Brunnée and Toope 2010, 24).In a similar manner, Jürgen Habermas (1996, 226-27; emphasis in original) describes the law as a form of an "argumentative process."This framework of analysis locates the source of validation of legal rules neither in abstract moral philosophy nor in state compliance but, rather, in social practice (see also Johnstone 2011, 35).
The "standards of competence" of the legal field that guide the practice of international law are constantly being "renegotiated" within the legal epistemic community (Adler and Pouliot 2011, 15-17).While individual legal decisions remain bound by the pre-existing standards of competence recognized by the epistemic community, members of that community can nevertheless take action to promote new interpretations of legal rules (Bourdieu 1987, 839).Yet any revisions of the background know-how of the Modes of Liability and Mass Atrocities 515 epistemic community are not random and are unlikely to introduce entirely new rules of practice.Instead, reforms in social practices usually constitute modifications of the already existing standards of competent performances (Fish 1980, 349).
Notably, some actors within the community of legal practice would possess more power over the discourse than others (Zehfuss 2002, 148).Social practice can reproduce different forms of discrimination present in society (Brunnée and Toope 2011, 112), a problem that has long ago been recognized by feminist scholars highlighting the unequal "dialogical capacities" of women and men to influence "public opinion" (Fraser 1985, 116).Similarly, in the international legal arena, some states have been "disproportionately influential" (Brunnée and Toope 2011, 119).States, in general, despite the significant power disparities among them, remain highly influential actors in the field of international law, not only by exerting control over the international courts' budgets and the (re)election of judges but also by their ability to question an international court's reasoning and delegitimize its authority (Helfer and Slaughter 2005).But while state power inequalities certainly play a role in the international legal field, the dynamics taking place in the latter are more complex and involve a greater range of actors.This is because the form of politics that enters the international legal field is "constrained and shaped" by the background rules of the international legal epistemic community (Abbott and Snidal 2013, 35).Because the revision of the background rules of legal practice constitutes an incremental process, even great powers cannot simply replace the law with a new one that serves their interests (Johnstone 2011, 52).Therefore, a different type of power dynamics take place within the international legal field-the "competition for monopoly of the right to determine the law" (Bourdieu 1987, 817).Expert knowledge, or "knowledge of the rules" of practice (Fish 1980, 343), thus, becomes the key source of power and authority within the epistemic community of international law (Bourdieu 1987, 816).
The centrality of expert knowledge inside the international legal field grants significant authority to determine the acceptable interpretations of legal rules to non-state actors, such as judges, legal experts and NGOs.Thirty-five years after Pierre Bourdieu (1987, 834-35) observed the "professional monopoly" over the provision of legal interpretations instituted by lawyers and legal experts, his comments seem more relevant than ever.The "invisible college" of international lawyers has become a well-known term to describe the growing community of academics and practitioners actively voicing their opinions on matters of international law (Schachter 1977-78, 217).To present their ideas nowadays, legal experts rely on publications in leading academic journals, which have become particularly influential in the field (Vasiliev 2015, 703), academic blogs, and even social media.In the ICL field, Mikkel Jarle Christensen (2017, 246) observes that judges often rely on academic writings to validate their interpretation of legal rules in the case at hand.NGOs have also been active participants in the ICL field.Over the past two decades, NGOs have not only championed the establishment of the ICC and promoted ratification among states but also strictly monitored the work of the court and demanded efficient administration of justice (Schiff 2008, 144).
Overall, "practice" studies offer important insights into international legal decision making-namely, that the international legal arena is governed by specific rules, or "standards of competence" of legal reasoning, and involves a variety of actors claiming authority over legal decision making in accordance with those standards.The variety of actors within the international legal field implies a variety of interpretations of international laws, premised on different visions of the purpose of law in society.Nevertheless, novel approaches to international law do not come "out of nowhere"-rather, they are non-deliberatively developed with respect to the background knowledge of the field and need to gain at least some traction, or social validation, among actors within that field.The next section applies this framework to the subfield of ICL and, specifically, to questions pertaining to the development of modes of liability.

Standards of Competence in the ICL Field
The ICL community of practice has coalesced around the idea that the perpetrators of mass atrocities, including genocide, war crimes, crimes against humanity, and aggressive war, bear personal responsibility and should be punished for their conducts (Sikkink 2011, 162-63).But ICL does not just recognize any form of punishment as acceptable-the idea of summary executions was rejected by the United States in the aftermath of the Second World War (Drumbl 2007, 3).Franklin Roosevelt's secretary of war, Henry Stimson, reportedly argued that the defeated Nazi generals had to be put on a legitimate trial to reflect the United States's own domestic recognition of fair legal procedure (Bass 2000, 148).Thus, the Western "liberal" trial came to be recognized as the legitimate form of punishment for international crimes (Bass 2000, 24;Sikkink 2011, 13).As observed by Elizabeth Borgwardt (2007,205), the Nuremberg trial reflected the projection of "peculiarly American, New-Deal-style" ideas of justice onto the international arena.
Regarding the accepted forms of punishment, unlike other branches of international law, which typically concern state behavior, international criminal justice has the authority to hold individuals accountable.Consequently, ICL is premised on the specific legal norms protecting the defendant's rights that are common to domestic penal systems but absent from other branches of international law.The "principle of legality" stipulates that an individual can be punished only for a conduct that has been unambiguously criminalized at the time of its commission (Gallant 2008, 11-12).The principle of "personal culpability" requires that persons can only be punished for their own conduct, and it protects individuals from being punished for mere association with the wrongdoers (Danner and Martinez 2005, 85).The culpability principle essentially requires evidence that the accused had consciously or intentionally made some contribution to the commission of the crime.
These criminal law norms have been applied in a novel context in the international law field, raising questions about their appropriate interpretation and application.The context was novel mainly because of the different nature of international crimes and domestic crimes.International crimes often constitute mass atrocities that involve hundreds or even thousands of perpetrators and are said to be facilitated by a social climate of fear and hatred.This social climate "authorize[s]" conduct that would be considered delinquent behavior under normal circumstances (Kelman 2009, 36;emphasis Modes of Liability and Mass Atrocities 517 omitted).Mass atrocities, thus, are often orchestrated or incited by a higher authority.Consequently, the individuals that are considered the "most responsible" for international crimes are often the leadership figures who have orchestrated their commission from above.But to establish a link between the person at the top and the actual criminal conduct is a challenging task.
The efforts to operationalize the legal norms associated with individual criminal responsibility, such as the modes of liability, in the novel context of international law resulted in an emerging community of ICL practice.This community attracted participants from a variety of legal backgrounds, who brought specific background understandings of what constitutes a competent legal performance.Due to the initial scarcity of "international criminal lawyers," as observed by Robinson (2020, 22-24), the development of ICL at the UN tribunals was largely led by experts in international human rights law and international humanitarian law (IHL).Judges with experience in human rights promotion have played a crucial role in the development of greater protection for the victims of international crimes, such as gender-based crimes, at the ICTY and ICTR.Examples include Judge Elizabeth Odio Benito, Judge Gabrielle Kirk McDonald, Judge Florence Mumba, and Judge Navanethem Pillay.
However, human rights law and IHL have a different purpose and, consequently, hold different understandings of the appropriate legal sources and procedures compared to criminal law.Criminal law places the burden of proof on the prosecution and protects the defendant's rights.The "rule of lenity" in criminal law requires that any ambiguities at trial should be resolved in favor of the defendant (Van Schaack 2008, 121).This defendant-oriented nature of criminal law is justified with the harshness of the punishment imposed-deprivation of personal freedom or even the death penalty in some systems (Danner and Martinez 2005, 86-88).By contrast, as observed by Allison Danner and Jenny Martinez, human rights law and IHL focus on preventing or recognizing the victims' suffering and allowing broader interpretations of legal principles to ensure that harm is addressed.Unlike criminal law, human rights law sanctions do not include incarcerating individuals but, rather, mechanisms such as public shaming and symbolic findings of state wrongdoing (86-89).In those subfields of legal practice, it is acceptable to evoke moral considerations, such as "the laws of humanity, and the dictates of the public conscious" when determining humanitarian obligations.13By contrast, in criminal law, the legality principle-namely, the requirement of providing fair notice of legal prohibitions to society-precludes the reliance on such broad concepts as "morality" and "public consciousness" in assessing criminal responsibility.
Another contributor to the plurality of norms within the ICL field is the difference in practices that can be observed among criminal lawyers.Common law generally relies on pragmatism or common sense, while civil law systems adopt a highly systematized and theoretically developed approach to criminal law (Trechsel 2009, 26-27).The "pragmatic" approach has historically led some common law systems to apply less strictly the culpability principle in cases of group criminality and to grant lesser significance to the legality principle.By contrast, civil law systems tend to codify in detail the fundamental legal principles valid for all crimes into the so-called "general part" of criminal law, which is considered to provide greater clarity about the requirements of the law and, thus, greater compliance with the legality principle (Gallant 2008, 13, 33-35).
As the next subsection discusses, different international tribunals have taken different approaches, with the post-Second World War and post-Cold War tribunals applying less strictly the culpability principle compared to the ICC.Ideas about the moral significance of punishing the perpetrators of mass atrocities, even when the link between those persons and the crimes is tenuous, have influenced the jurisprudence of the 1945 Nuremberg tribunal, the ICTY (established with UN Security Council Resolution 827 in 1993 during the conflicts in the former Yugoslavia), and the ICTR (established with Resolution 955 in 1994, following the Rwandan genocide). 14 But the ideas of criminal lawyers, and especially those of civil law countries who respect the strict application of the legality and culpability principles, have become particularly influential at the ICC.

Nuremberg and the UN Tribunals
Both the Nuremberg and the UN tribunals borrowed from humanitarian ideas of morality and public consciousness to set in motion the project of international criminal justice.More specifically, these tribunals based their practices on the premise that, due to the unique nature of mass atrocities, the criminal law principles used in domestic systems could not be applied in the traditional manner in international law, lest the perpetrators of those atrocities were "allowed to go unpunished." 15This required a special approach to delineating the guilt, or culpability, of the individuals standing trial for mass atrocities.Specifically, to assess the accused's criminal responsibility, the judges at those tribunals often relied on establishing an indirect link between the accused and the crimes-the accused was linked to the collective entity that planed or engaged with criminal activities rather than directly to the commission of those crimes.Thus, the Nuremberg tribunal convicted former Nazi officials for either participating in a "conspiracy" to wage an aggressive war or for being members of a "criminal organisation" (see Yanev 2015).In a similar manner, in the 1990s, the ICTY developed the theory of JCE, which became widely used at the tribunal.Pursuant to JCE, the prosecutor did not have to prove that the accused had personally committed the crime but only that they had provided "assistance in, or contribution to, the execution of the common purpose" of the criminal enterprise, which had resulted in the commission of the crime. 16 One of the variants of JCE set such low requirements for the attribution of criminal responsibility that the accused could be convicted for a crime that fell outside the scope of JCE's common purpose if that crime was committed by one of the group's members 14.This is not to suggest that all cases at the United Nations (UN) tribunals have been built upon a tenuous relation between the perpetrator and the crime.In their early days, the tribunals often dealt with low-to mid-range perpetrators.Questions about tenuous relationships became more prominent when the tribunals started dealing with leadership cases and faced the challenge of linking those persons to the scene of the crimes (Zahar 2014, 242-44).As discussed in the third section of this article, the use of joint criminal enterprise III was particularly controversial in this regard.
Modes of Liability and Mass Atrocities 519 and was merely "foreseeable" by the accused. 17The ICTY and ICTR's interpretation and application of the "command responsibility" principle-namely, holding superiors criminally responsible for failing to address the crimes of their subordinates-has triggered similar concerns for expanding the scope of personal culpability (Damaška 2008, 350).Specifically problematic appeared to be the tribunals' finding that command responsibility did not require a causal connection between the commander's inaction and the crimes (Robinson 2020, 146).One authoritative commentator called command responsibility "the silver bullet of the prosecution" (Schabas 2007, 222).
This broad interpretation of the culpability principle at Nuremberg and the UN tribunals sometimes came "dangerously close" to abrogating the culpability principle by obfuscating the accused's personal contribution toward the crime (May 2008, 266).This approach was premised on two sets of background understandings of the scope and function of legal norms dominating the ICL field at the time: the human rights/ humanitarian logic of taking into consideration the interests of victims and the society at trial and the common law practice of relying on broad theories of group criminality.The justification of theories such as JCE revealed a humanitarian concern to serve justice to the victims and prevent the perpetrators of mass atrocities from escaping punishment by exploiting the limits of the culpability principle.One authoritative judge at the UN tribunals argued that the creative interpretation of the concept of "culpability" found in JCE was "necessary" for the fulfilment of the ICTY's "mission to administer international criminal justice" (Shahabuddeen 2010, 197).
The ICTY's first president Antonio Cassese and colleagues (2009,294) argued that JCE served to prevent individual perpetrators from escaping criminal accountability by hiding behind "the fog of collective criminality."The approach of Cassese, a highly influential figure in ICL, has been described as one "that sought to move the law forward and render it more humane," possibly reflective of his earlier interest in philosophy and the humanities (Cryer 2012(Cryer , 1048)).JCE was useful for holding criminally responsible those persons who were not directly involved in the crimes, especially leadership figures.From this humanitarian perspective, the culpability principle cannot be interpreted the same way in ICL as in domestic systems because many persons would escape punishment due to the difficulties of delineating each individual's contribution to the mass atrocities.Echoing the understandings shared within the human rights and IHL communities of practice, the ICTY judges considered that the preservation of justice and fairness toward the accused had to be "balance[d]" against "the preservation of world order." 18n additional factor that influenced the jurisprudence of international tribunals was the spread of common law practices.The United States exerted significant influence over the process of developing the legal theories that were used at Nuremberg to link the accused Nazis to the crimes.The original idea that the prosecution of Nazi crimes required special theories of liability, such as conspiracy and membership in criminal organizations, to reflect the collective nature of those crimes was that of US Colonel Murray C. Bernays (van Sliedregt 2012, 22-23), described as "the thoughtful and bookish chief" of the Special Projects Office in the War Department (Borgwardt 2007, 221).The US Secretary of War Henry Stimson reportedly endorsed the idea because he had personally relied on the concept of conspiracy when practicing as a young lawyer (222).Another American-Robert Jackson, the head of the US delegation to the London conference and future chief US prosecutor at Nuremberg-reportedly became the "main force" behind the inclusion of the conspiracy charge in the Nuremberg Charter, 19 albeit in a somewhat different version from Bernays's original idea (Yanev 2015, 434).Thus, the development of the first modes of liability in ICL has been significantly influenced by the ideas shared among US officials in the postwar period.In common law, the idea of "conspiracy" seemed less controversial than in civil law.It is telling that several years after the Nuremberg judgment, the Pinkerton conspiracy became an accepted doctrine in US domestic law (429). 20 Those ideas did not remain uncontested.During the 1945 London negotiations, the French and the Soviet representatives who were trained in civil law reportedly expressed strong disapproval of the broad concept of conspiracy that diluted the link between the accused and the crime (Dubber 2006(Dubber , 1323)).But the US delegation insisted on including the conspiracy charge in the Nuremberg Charter (van Sliedregt 2012, 23).The uneasiness displayed by some members of the nascent ICL community, and especially the French judge at Nuremberg, Donnedieu de Vabres, with the concept of conspiracy has not been inconsequential.In fact, that uneasiness has often been quoted as the reason behind the limited reliance on the conspiracy charge in the final Nuremberg judgment (Gallant 2008, 117;Yanev 2015, 444).Nevertheless, the influence of common law ideas has had a long-lasting impact on the development of modes of liability in ICL.The ICTY Appeals Chamber, which established the JCE theory referred to the jurisprudence of the post-Second World War UK and US military courts and to the US Pinkerton doctrine in support of the existence of JCE. 21Like at Nuremberg, this trend has been explained in part by the fact that, in the early years of the tribunal, many ICTY judges and legal officers were trained in common law (Steer 2014, 45).
But, over time, as the ICL system developed and its community of practice began to attract practitioners from a broader variety of professional backgrounds, the assumption that conspiracy-and JCE-like modes of liability were appropriate for addressing mass atrocities became problematic.Legal scholars and practitioners with extensive professional background in civil law systems, including in Germany and the Netherlands, strongly advocated for reforming JCE or for substituting it with a different theory of criminal responsibility.Examples of those ideas include the writings of Kai Ambos 20.The two brothers Daniel and Walter Pinkerton were charged with substantive violations of the US Tax Code resulting from an alleged liquor-bootlegging conspiracy.While there was no evidence that Daniel directly participated in Walter's crimes, the court upheld his conviction.In Pinkerton, the court established that, under conspiracy law, any act in furtherance of the conspiracy is attributable to all participants.For a detailed analysis, see Noferi 2006, 94. 21. Tadić, paras. 205-13, 224.
Modes of Liability and Mass Atrocities 521 (2007, and Harmen van der Wilt (2007).Premised on a narrower interpretation of the culpability principle, the critics of JCE and specifically of its broadest articulation "JCE III," expressed concern that convicting a person for crimes that were merely "foreseeable" risked the attribution of liability without evidence of personal culpability (Ambos 2007, 174).22 From this perspective, JCE treated so many persons as potential perpetrators that the doctrine experienced difficulties explaining "why each fish caught deserves punishment for intentional wrongdoing" (Weigend 2008, 477).One ICL expert with significant experience in the Egyptian justice system, Mohamed Elewa Badar, famously called JCE III a theory developed to "just convict everyone" (Badar 2006).Overall, many critics of JCE share familiarity with civil law concepts, extensive professional background in criminal law, and concern for overstretching the culpability principle.Some criticism of JCE was also voiced inside the UN tribunals, especially from judges with a civil law background.A notable example was the Stakić case.In 2003, an ICTY chamber chaired by Judge Wolfgang Schomburg, previously serving at the German Federal Court of Justice, tried to substitute JCE with a German-influenced theory of criminal responsibility, arguing that the latter was more compliant with the principles of legality and culpability.23 It is possible that at least one more of the Stakić trial judges-namely, Judge Carmen Argibay from Argentina-was familiar with the proposed German theory of criminal responsibility, developed by Claus Roxin.That theory had been employed by Argentinean courts in trials against members of the Junta regime (Ambos 2007, 182).The attempts to substitute JCE with the German theory remained unsuccessful at the UN tribunals because JCE was "routinely applied in the Tribunal's jurisprudence" and its substitution would have generated "uncertainty, if not confusion" about the applicable law.24 JCE had already became an established practice at the UN tribunals, which most judges were reluctant to change.But, as the next section discusses, the German criminal law practices became significantly more influential at the ICC.

Drafting the Rome Statute
The unique context of the ICC's creation enabled different know-how ideas to become involved in the development of ICL, which brought up modifications in the social practices of the field.The establishment of the ICC and the drafting of its Rome Statute provided a forum for contesting the old practices, or socially meaningful patterns of legal reasoning, that had previously dominated ICL and for presenting new ideas concerning modes of liability.For the first time in ICL, a plurality of states, 22.Other international tribunals, such as the Special Tribunal for Lebanon (STL), have also found the "foreseeability' standard problematic.The STL concluded that "the better approach under international criminal law is not to allow convictions under JCE III for special intent crimes."STL, Interlocutory Decision on the Applicable Law: Terrorism, Conspiracy, Homicide, Perpetration, Cumulative Charging, Judgment no.STL-11-01/1, February 16, 2011, para. 249 legal experts from different backgrounds, and hundreds of NGOs participated in the drafting process.Furthermore, the stakes of influencing the ICC's founding document were particularly high.As the statute of "the first permanent general, future oriented international criminal court" (Kaul 2012, 5), the Rome Statute was bound to play a more profound role in the ICL field than the statutes of the temporary UN tribunals (van Schaack 2008, 170).
A major novelty of the Rome Statute was that it included a section called "General Principles of Criminal Law," which included the applicable modes of liability (Preparatory Committee 1996, 41-47).Article 25(3)(a) of the Rome Statute specified in detail a limited number of situations in which a person could be found guilty of committing a crime-namely, committing a crime personally (direct perpetration), jointly with others (co-perpetration), or by using other persons as tools to physically commit the crime (indirect perpetration). 25Furthermore, for the first time in ICL, the Rome Statute defined criminal intent or the "mental element" of personal culpability.Article 30 of the Rome Statute required that the accused "mean[t] to engage" in the crime or was "aware that it will occur in the ordinary course of events" as a result of her actions.Hence, unlike the statutes of previous tribunals, the Rome Statute introduced a system of highly detailed rules for assessing the guilt or innocence of the accused.
The comprehensive codification of criminal law principles in the Rome Statute has sometimes been perceived as an attempt by states to limit the independence of future ICC judges and avoid the legally controversial practices of previous tribunals (Hunt 2004, 61).This proposition is in line with the rationalist principal-agent model of delegation in international relations, according to which the greater the specification of the rules of operation of the international organization, the greater the control of the principal (the states) over the agent (the court) (Hawkins et al. 2006, 27).There is certainly merit to such arguments.The drafting records show that some states favored a flexible ICC statute that would have enabled the judges to elaborate the applicable modes of liability, but most states preferred that the fundamental criminal law principles were clearly articulated in the statute to avoid conferring "substantive legislative power" to the future ICC judges (Preparatory Committee 1996, 42).
But while most states may have preferred a more constrained court, it would be simplistic to attribute the Rome Statute's general part only to sovereignty interests and to ignore the variety of competing understandings among legal experts about what the future ICC statute should look like.The absence of codified general principles at previous international tribunals was influenced by the United States's important role in their establishment and the common law preference for pragmatism over the systematization of the law (Ambos 2006, 661).However, in the 1990s, many authoritative civil law scholars, including the German scholars Albin Eser and Kai Ambos, called for the categorization of the principles of criminal law within a "general part" of ICL (see Eser 1993;Ambos 1996).That proposition was also supported by the influential Siracusa draft of the ICC's statute, produced by legal experts in 1995-96 in Italy. 26While the codification of ICL principles might have reassured some states that their nationals would be protected from judicial creativity, the general part constituted for civil law practitioners a necessary component of a criminal code, especially at the international level (Saland 1999, 191;Eser 2001, 20).Instead of constraining the court and benefiting states, the codification of the principles of criminal law within the Rome Statute constituted for some scholars a safeguard against the infiltration of political interests and ensured principled judicial reasoning at the new court (Ohlin 2007, 77).Thus, the active participation of professionals with civil law and criminal law backgrounds in the Rome Statute negotiations introduced new patterns of socially meaningful rules in the ICL field-namely, the importance of codifying and categorizing legal principles.
Notably, because of the importance of expert authority within the legal field, legal scholars and practitioners became particularly influential actors during the drafting of the general principles in the Rome Statute, including the modes of liability.Because the general principles were considered the "most technically difficult part of the Statute" (Bassiouni 2005, 158) and comparatively "non-political" (Saland 1999, 193), the significance of expert knowledge increased.The drafting records reveal that the main actors who influenced the content of the Rome Statute's general part were an "informal group" of unnamed criminal law experts at the Preparatory Commission on the ICC Statute (Preparatory Committee 1996, 6) and the official Working Group on the General Principles of Law at the Rome Conference, coordinated by the Swedish representative Per Saland (1999).Furthermore, many of the general principles codified in the Rome Statute were adopted from the proposals made in the Siracusa draft that had been drafted earlier by a group of authoritative scholars (Ambos 1996, 521). 27Thus, by virtue of their expert authority, legal professionals turned out to be particularly influential actors in ICL practice during the Rome Statute's negotiations.
The codification of the general criminal law principles in the Rome Statute can thus be traced not only to state interests but also, more importantly, to the background understandings of many legal scholars at the time about the need for a more principled approach to the assessment of criminal responsibility at the future court.During that time, the so-called "liberal critique" of ICL, an academic movement defending compliance with the principles of criminal law, grew in power (Robinson 2020, 5).From a liberal perspective, the fundamental principles of criminal law are said to ensure that judgments on mass atrocities would no longer be based on the "gut-level decisions" of individual judges but, rather, on the dispassionate application of the law (Ohlin 2007, 88).Representatives of this critical liberal school in ICL include ICL academics with extensive background in comparative and domestic criminal law, such as Elies van Sliedregt, Kai Ambos, Thomas Weigend, Jens D. Ohlin, and Darryl Robinson, to name a few.
Furthermore, the very concepts of "state interests" or "state beliefs," used respectively in the rationalist and constructivist literature, become problematic when examined in practice.During the Rome Statute's drafting, there was significant disagreement 26.Draft Statute for an International Criminal Court, Siracusa/Freiburg, 1995, 55, https://www.legaltools.org/doc/39a534/pdf/. 27. Draft Statute,[55][56][57] between the delegates of countries with different domestic penal systems regarding the appropriate scope of criminal responsibility.Common law countries favored the inclusion of a provision akin to conspiracy, which they considered would have enabled the court to efficiently fulfill its purpose of ending impunity for mass crimes.But that proposition triggered stark opposition among civil law countries, which expressed worry about the compliance of such theories with the culpability principle (Saland 1999, 199).Consequently, what seemed to be "rational" or "appropriate" to states, or for that matter to legal experts, during the Rome Statute's drafting, depended on their respective background knowledge.
While the background knowledge of human rights and IHL practitioners, as well as of criminal lawyers from common law systems, long dominated the ICL community of practice, the ICC's establishment provided criminal lawyers from civil law systems with greater influence within the field, which resulted in modifications of the socially meaningful patterns of legal reasoning in ICL.In the words of one scholar, the general part of the Rome Statute and, more generally, the categorization of legal rules, "emancipated" ICL from the "rudimentary nature of international law" and turned it into a mature system of criminal justice (Militello 2007, 943).From this perspective, the drafting of the Rome Statute constituted a unique opportunity to move past the considerations of "developing" ICL and "ending impunity" for mass crimes and to prioritize instead procedural fairness and the defendant's rights (Amann 2003, 180-81).As the next section discusses, most ICC judges shared the understandings of their colleagues who drafted the Rome Statute and adopted a similarly restrained interpretation of those general principles.

Interpreting the Rome Statute
The belief in the importance of upholding a strict interpretation of the principles of criminal law and the civil law know-how became especially pronounced when the ICC began operation in the early 2000s.The fact that the general principles of law had been codified in the statute did not guarantee that ICC judges would follow such rules in practice.No matter how precise the wording of the law, the latter is always in need of judicial interpretation (Helfer and Slaughter 2005, 937; Greenawalt 2011Greenawalt , 1080)).ICC judges were left with a "[c]onsiderable discretion" in interpreting the statutory text (Schiff 2008, 45), and some commentators considered it "unrealistic" that the formal codification of criminal law principles would meaningfully restrain the ICC judges from engaging in creative interpretations of the law (Wessel 2006, 414).Therefore, some commentators thought that ICC judges would follow the practice of previous tribunals and interpret the gaps in the statutory text in a manner that would render convictions easier (Jessberger andGeneuss 2012, 1083).Yet ICC judges have not only proactively declared fidelity to the principles of criminal law but also interpreted the relevant Rome Statute provisions in a more restrained manner than seems to have been intended by the drafters (Cryer 2009, 403-4). 28  28.See ICC, Prosecutor v. Katanga, ICC-01/04-01/07-3436-tENG, March 7, 2014, Judgment pursuant to Article 74 of the Statute, paras.54-57, https://www.icc-cpi.int/CourtRecords/CR2015_04025.PDF; ICC,

Modes of Liability and Mass Atrocities 525
Some have perceived these restrained interpretations as deference to state interests (Cryer 2009, 392), but such accounts ignore the pleas of many legal experts to abandon the controversial practices, or patterns of legal reasoning, of the UN tribunals concerning the assessment of criminal responsibility (see Danner and Martinez 2005, 88;Ambos 2006, 104).The "practice" approach provides important insights into the interpretation of Article 25(3)(a) of the Rome Statute-a provision of key significance for trial outcomes-because it elucidates the transformation of the socially meaningful patterns of legal reasoning that has taken place with the establishment of the ICC.Since mass atrocities are committed by a multitude of persons, the interpretation of Article 25(3)(a) of the Rome Statute, which provides that a person could be held criminally responsible for committing "jointly with another" person an international crime, bears significant implications for the prospects of successful prosecutions at the ICC.The Rome Statute's rather "vague" definition-"jointly with another"-required further interpretation at the court (Weigend 2008, 479).At first, the legal representatives of victims at the ICC interpreted that phrase as pertaining to the notion of "joint criminal enterprise." 29Some commentators on the Rome Statute also expected the judges at the new court to be "strongly influenced" by the prolific jurisprudence of the UN tribunals on JCE (Schabas 2007, 215).But, instead, the ICC judges firmly rejected JCE and determined instead that to convict a person under Article 25(3)(a) required evidence that the defendant had exercised "control" over the commission of the crime, evidenced in the accused's power to "frustrate" its commission. 30These strict requirements were differentiated from the jurisprudence of the UN tribunals, which merely required that the defendant had made an intentional contribution toward the collective crime, regardless of the level of that contribution. 31In effect, the interpretation forwarded by the ICC judges significantly narrowed the scope of situations in which a person could be found guilty for committing an international crime.
Notably, the analysis of the official drafting history of Article 25(3)(a) of the Rome Statute nowhere suggests that states contemplated to include the requirement of proving the defendant's "control over the crime."The interpretation of Article 25(3)(a) betrayed the judges' preferences for interpreting the text rather than a "narrow, literal interpretation" of the Rome Statute (Weigend 2011, 109;emphasis added).As Leila Sadat and Jarrod Jolly (2014, 775) point out, Article 25(3) is a "a consensus provision," the product of the complex negotiations among experts with different legal backgrounds and, consequently, "lacks a strong and logically cohesive theoretical underpinning."ICC judges themselves acknowledged that their interpretations were not "the only" one compatible with the statutory wording. 32Rather, Article 25(3) has been described as a "Rorschach blot" because experts read into the provision their own professional experience and background ideas about criminal liability (774).While ICC judges could have used their constrained independence in interpreting the Rome Statute to expand the scope of situations in which a defendant could be found guilty-thus, following the practices of previous tribunals-they chose to do the opposite.
The judges' decision should be understood in the context of the transformation of the socially meaningful patterns of legal reasoning taking place within the ICL community of practice, facilitated by the increasing influence of criminal law expertise and, especially, that of civil law countries like Germany (Minkova 2022).The substitution of JCE for the "control-over-the-crime" theory reflected the backlash against the victimoriented humanitarian approach of previous tribunals.Authoritative criminal law experts with civil law background had expressed the opinion that JCE had no place at the permanent ICC (see Ambos 2007, 172-73).The ICC judges' adoption of the "control-over-the-crime" theory did not gain unanimous support in the ICL field, but it was welcomed by those criminal law academics who had for long expressed concern with the UN tribunals' controversial jurisprudence (Fletcher 2011, 190).Already before the ICC adopted the "control-over-the-crime" theory in 2007, a few German scholars, well-known in the ICL field, including Kai Ambos (2006, 664), Gerhard Werle (2005, 124n196), and Albin Eser (2002,796), had pointed out the merits of relying on that theory in relation to Article 25(3)(a).The adoption of the "control" theory instead of JCE may also have something to do with the presence of civil lawtrained judges at the ICC's pre-trial chambers in Lubanga and Katanga and Ngudjolo, including Judge Sylvia Steiner from Brazil, Judge Anita Ušacka from Latvia, and Judge Claude Jorda from France (Steer 2014, 45-46).The "control" theory (Tatherrschaft) was developed by the German legal scholar Claus Roxin ([1963] 2011) in 1963.Even though the ICC judges argued that theory was applied by "numerous legal systems," 33 it has generally been recognized in Germany and some Latin American countries (Ambos 2007, 182;van Sliedregt 2012, 107).
In fact, one ICC judge with common law professional background-namely, Judge Adrian Fulford from the United Kingdom-objected to the adoption of the "controlover-the-crime" theory, which he saw as excessively demanding and placing "an unnecessary and unfair burden on the prosecution." 34Whether that opinion was informed by his common law background, a humanitarian concern for delivering substantive justice for mass atrocities, or both, Judge Fulford preferred a more pragmatic and less strict interpretation of the Rome Statute.However, his position remained the minority one at the ICC, which was dominated by the standards of acceptable criminal law practice found in civil law countries.As the next section discusses, the restrained approach to criminal responsibility, informed by the dominance of criminal law expertise at the ICC, has significantly challenged the prosecutor in establishing a link between the accused and the crimes, which has contributed to the acquittals and dismissals of charges in several cases.33.Lubanga 2007, para. 330. 34. Lubanga 2012, para. 3 (emphasis added).

Insights into the ICC's Trial Record
For nearly two decades of operation, the ICC has convicted only five persons for core international crimes.Four persons have been acquitted, 35 the proceedings against two persons have been terminated, 36 and all charges against four others have been dropped before trial. 37At first, it appears that the scarcity of successful ICC prosecutions may be the result of the challenges that the international court faces in the realm of interstate politics.Indeed, states can refuse to surrender a suspect to the court, as it happened in the case against the former Sudanese president Omar Al-Bashir.The systemic nature of mass atrocities and the time lapse between the crimes and the investigation further obstructs the collection of incriminating evidence.But this state-centered account of how international law operates presents only part of the picture.As this section demonstrates, the analyses of the practices of ICC judges, and specifically the agreed-upon strict standards of performance regarding the assessment of criminal responsibility at the court, provide new insights into the ICC's trial record.The importance of background understandings about the appropriate application of the modes of liability at the court is particularly evident in the cases against accused without any apparent political backing by states.
The ICC judges' interpretation of Article 25(3) concerning "joint" perpetration of a crime by a multitude of individuals in accordance with the demanding German "control-over-the-crime" theory proved a significant obstacle to establishing a link between the accused and the crimes.The "control" theory required that the crime was committed pursuant to a "common plan" among the perpetrators, which needed to contain an "element of criminality." 38Influential members of the ICL community of practice with extensive professional background, such as Judge Christine van den Wyngaert, cautioned that the judges should delineate a clear criminal element of the alleged "common plan."Otherwise, the argument goes, the link between the accused, who had participated in what could otherwise be a largely political or military plan, and the specific crimes that resulted from the implementation of that plan would be obfuscated. 39 This would infringe upon the principle of personal culpability in a similar manner as the controversial JCE III (van Sliedregt, Ohlin, and Weigend 2013, 735-39).But the ICC judges have lived up to the standards of competent performance agreed upon by criminal law experts, even if that meant acquittals of the accused, as was the case in 2019 during the trial against the former Ivoirian president Laurent Gbagbo and his aide Charles Blé Goudé.The ICC judges were critical of the fact that "[u]pon the pretext" that the common plan of the perpetrators must not be exclusively criminal, the prosecutor had presented a lot of evidence regarding the non-criminal aspects of Gbagbo's alleged plan to stay in power in Côte d'Ivoire after losing the 2010 elections, but failed to provide sufficient evidence regarding the criminal elements of that plan. 40 Consequently, the judges acquitted both accused by majority.
Another difficulty for the ICC prosecutor has been the judges' preference for inferring "control" over the crimes from the accused's direct involvement in the crimes rather than from their official position within the criminal group.Particularly illustrative in this regard are the 2010-11 proceedings in Abu Garda and Banda and Jerbo, concerning the 2007 attack on the Hashkanita compound in Darfur, Sudan. 41The prosecution argued that, as the alleged commanders of the attacking troops, Bahar Idriss Abu Garda, Abdallah Banda Abkaer Nourain (Banda), and Saleh Mohammed Jerbo Jamus (Jerbo) had jointly perpetrated the attack on Hashkanita. 42Regarding Abu Garda's control over the crimes, the prosecutor drew on evidence of the suspect's alleged participation in the meetings when the attack was planned.But the judges, concerned by the lack of "reliable" evidence of Abu Garda's personal participation in the attack, dismissed all charges against him. 43Similarly, in Ngudjolo, the judges acquitted the accused due to a lack of evidence that he had controlled the militant group, which had attacked civilians. 44By contrast, several witnesses confirmed that Banda and Jerbo had personally led their forces in the Hashkanita attack and had taken part in the looting of the compound afterwards. 45Since Banda and Jerbo's personal involvement in the crimes was established, the judges confirmed the charges of co-perpetration against these suspects. 46 In 2019, similar dynamics occurred in Yekatom and Ngaïssona.Yekatom was a commander of an "Anti-Balaka" armed group, allegedly involved in attacks against the Muslim population in the Central African Republic.Patrice-Edouard Ngaïssona was a senior figure with political and diplomatic functions who operated at the national level of the Anti-Balaka structure. 47The judges confirmed the charges of joint perpetration of crimes under Article 25(3)(a) against Alfred Yekatom but were not convinced that Ngaïssona had enjoyed "control" over the anti-Balaka troops, given their "high degree of autonomy" in military operations. 48Thus, ICC judges have displayed preferences for establishing the accused's personal involvement in the crimes.
The background understandings that ICL requires strict interpretation of criminal law principles by view of protecting the defendant's presumption of innocence helps explain how most ICC judges have interpreted and applied the Rome Statute's provisions concerning criminal responsibility.But the practice perspective also elucidates where the dissenting opinions to some of those judgments came from.The humanitarian view of legal practice was still upheld by some within the ICL field.For example, Judge Olga Venecia Herrera-Carbuccia dissented from the decision of her colleagues to acquit Laurent Gbagbo and Charles Blé Goudé, arguing that the ICC should not "allow" a president "to target citizens" and go unpunished. 49From Herrera-Carbuccia's perspective, the ICC's mandate was not limited to the assessment of individual criminal responsibility but included broader goals such as deterrence and social restoration.NGOs have also expressed regret that the victims of mass atrocities would not see justice being served when a case has ended up in acquittal (Amnesty International 2019).But while the victim-oriented humanitarian vision of ICL was particularly prominent at Nuremberg and the UN tribunals, the establishment of the ICC has seen a significant shift toward defendant-oriented criminal justice, which no longer recognizes the previous practices for delivering easier convictions as acceptable (Minkova 2022).
Thus, the practice-based analysis of ICC's decision making provides new insights into trial outcomes that have generally remained overlooked from a state-centered international relations perspective.International relations studies of the ICC generally note that, due to its dependency on state cooperation for executing arrests and collecting evidence, the ICC needs to "accommodate" states' interests and abstain from prosecuting figures with powerful political backing (Bosco 2014, 20).Thus, the scarcity of successful international prosecutions is attributed to the embeddedness of international law within state power politics.This provides an important, but limited, insight into the ICC's operation.Notably, in all the aforementioned cases, which have ended in acquittal or dismissal of (some) charges, the accused constituted either deposed officials or rebel commanders without significant government backing.Banda and Jerbo were members of insurgent groups who fought against the Sudanese government forces in Darfur.Yekatom and Ngaïssona were commanders of anti-Balaka groups in the Central African Republic, reportedly loyal to the deposed president François Bozizé. 50Regarding Gbagbo, while it falls beyond the scope of this article to disentangle the variety of state interests surrounding the trial, it is worth noting that the Ivoirian government opposed Gbagbo's release, arguing that his return home would destabilize the country (Wakabi 2020).The state-centered international relations scholarship helps explain why these persons may have ended up at the ICC-namely, because no government has opposed their transfer to the court but falls short of enquiring into what happens with their trials afterwards.By contrast, because the practice perspective elucidates how the ICL field operates from within, it also provides novel insights into the assessment of criminal responsibility, regardless of the political background of the accused.
Finally, it is important to clarify what this analysis is not suggesting.First, the scope of this article is limited to the construction of a specific set of ICL rules, premised on criminal law principles-the modes of liability.Because ICL is a compilation of different legal systems, it might be the case that the ICC judges have employed different interpretative approaches concerning other legal rules that do not concern the assessment of criminal responsibility, such as the court's jurisdictional reach or the scope of the Rome Statute's core crimes.Second, to suggest that the strict criminal law standards of the assessment of criminal responsibility have dominated ICC jurisprudence is not to say that the judges aim to deliver acquittals or that no cases can be successful at the court.It is simply an observation that the vision of ICL that came to dominate the ICC has prioritized the narrow technical application of legal rules, regardless of the trial outcome.So far, this priority has resulted in several high-profile acquittals and dismissals of charges at the court, but this need not be the case in the future if the prosecutor adapts to the heightened expectations of the judges regarding the quality and logical consistency of the evidence.Indeed, in cases concerning mid-level insurgent commanders who have been directly involved in hostilities and personally engaged in criminal conducts, such as Dominic Ongwen 51 and Bosco Ntaganda, 52 the prosecutor has obtained convictions.
Finally, the background knowledge of the participants in the ICL field need not be the only determinant of trial outcomes.Other factors, such as the limited opportunities for collecting evidence of mass atrocities could significantly challenge the prosecutor.Scholars have observed the pervading problems with the early ICC investigations, such as the use of intermediaries to collect evidence, the lack of local knowledge, and the short field trips of the prosecutor to the investigation sites (De Vos 2013).Notably, different factors such as the problems of collecting evidence, the interests of governments, and the politics of the legal field can influence trial outcomes in conjunction.The availability of government assistance is crucial for the prosecutor's ability to collect evidence.Furthermore, depending on the standards of competence recognized among the judges, they may be willing to tolerate to a varying degree the prosecutor's investigation hurdles.Nancy Combs (2010) observes that the UN tribunals have often lowered the burden of proof on the prosecutor by admitting evidence that would otherwise not meet the standards of domestic proceedings.By contrast, in cases like Bemba or Gbagbo and Blé Goudé, the ICC judges have demonstrated "zeal for impeccable standards" and a "hypersceptical" attitude toward incriminating evidence, which has precluded the prosecutor from convincing them of the defendant's guilt (Robinson 2019).Thus, while international prosecutors generally face the challenges of obtaining state cooperation and collecting evidence for mass atrocities, the strict interpretation of the applicable law by the judges may amplify these hurdles.

Modes of Liability and Mass Atrocities 531 CONCLUSION
The inquiry into the background assumptions of actors that are constantly (re)produced in everyday practice provides important insights into the construction and application of the modes of liability in ICL.Specifically, this article has proposed that the restrained approach to criminal responsibility at the ICC resonates with the socially meaningful standards of sound legal reasoning recognized by many members of the ICL community of practice.The construction of the ICC's modes of liability reflects a mixture of ideas: the desire to present the ICC as a stand-alone international court and a trendsetter in international justice, the preference of many legal experts inside and outside the court for separating procedural justice from substantive justice and for prioritizing the latter during trials, and the conviction of some ICC judges and legal scholars that greater categorization of the modes of liability will improve the quality of the legal process.
The implication of this observation is that the way in which the modes of liability are currently construed at the ICC might still change as the normative context transforms.There are already indications that such transformations might be taking place.While many members of the ICL epistemic community have become increasingly concerned that the practices of the UN tribunals have come dangerously close to the attribution of guilt by association, the restrained approach of the ICC judges has begun to trigger the opposite reaction.NGOs have criticized the ICC judges for applying "overly strictly" the culpability principle (Women's Initiatives for Gender Justice 2018, 147).Legal experts have cautioned that the ICC's interpretation of the modes of liability has had a particularly obstructive impact on the prosecution of gender-based crimes (SáCouto, Sadat, and Sellers 2020).Even scholars who have displayed a strong commitment to the protection of the defendant's rights in ICL have expressed concern that the ICC has "overcorrec[ted]" the practices of the UN tribunals (see Robinson 2020, 5).Some ICC judges, who have shared a humanitarian concern for protecting civilians during conflict, also have challenged the approach to criminal responsibility taken by their colleagues, as illustrated by the dissenting opinion of Judge Herrera-Carbuccia in Gbagbo and Blé Goudé.Some legal scholars have more generally cautioned against the search for "uniformity in all aspects of doctrine and practice" in ICL and argued instead for taking more seriously the domestic legal rules of the countries that would normally exercise jurisdiction over each particular case (Greenawalt 2011(Greenawalt , 1067-68)-68).
More than seven decades after the end of the Second World War, the international criminal justice community has still not arrived at a solid consensus on the scope and requirements of individual criminal responsibility for mass atrocities.Despite (or maybe because of) the detailed codification of modes of liability in the Rome Statute, the ICC has become the main institutional forum for promoting, contesting, and justifying ideas about the scope and requirements of criminal responsibility for international crimes.
19. Charter of the International Military Tribunal, reprinted in Agreement by the Government of the United Kingdom of Great Britain and Nortern Ireland, the Government of the United States of America, the Provisional Government of the French Republic and the Government of the Union of Soviet Socialist Republics for the Prosecution and Punishment of the Major War Criminals of the European Axis, 8 August 1945, https://www.un.org/en/genocideprevention/documents/atrocity-crimes/Doc.2_Charter%20of%20IMT%201945.pdf.
9. Including Journal of International Criminal Justice, European Journal of International Law, Leiden Journal of International Law, and International Criminal Law Review.