International crimes through the lens of global constitutionalism

Abstract Scholars of global constitutionalism have recently come to examine international criminal law (ICL) and its associated institutions, in particular the International Criminal Court (the ICC). This article prolongs these efforts by pointing to and remedying two deficits of that project with particular emphasis on the Rome Statute crimes. First, how does one account for the role of the international trial in global constitutionalist terms? Second, can global constitutionalism insightfully explain the content and scope of these crimes – that is, both their substantive definition and the predominant modes of liability developed by the ICC? This article answers both questions affirmatively and offers an account of their nexus. It first shows that the Rome Statute crimes are often perpetrated through a hierarchically organized apparatus of control, and interprets their global constitutional significance via the principle of constituent power. It then makes use of Antony Duff’s relational account of criminal liability to offer an account of the international trial. In the international context, one can conceive of the trial as allowing state or state-like authorities to call each other to account, which renders justice to the core function of enabling and limiting political authority on which global constitutionalism centres.


I. Introduction
Scholars of global constitutionalism have recently come to examine international criminal law (ICL) and its associated institutions, in particular the International Criminal Court (ICC). In conformity with the dual descriptive and normative agenda of global constitutionalism as a scholarly field, the emerging literature has identified a number of global constitutional aspects in the ICL system and provided a normative justification for them. For example, in the context of the ICC, it has been argued that the relationship between the Assembly of State Parties and the Office of the Prosecutor exemplifies the separation of powers that should characterize a global constitutional institution. 1 Similarly, it has been suggested that international criminal tribunals play a global constitutional role when they prosecute state or state-like leaders for having grossly abused their authoritywhat has been termed 'constitutional impact'. 2 One may readily wonder about the purpose of viewing ICL through the lens global constitutionalism. Why should one care? International crimes enshrined in the Rome Statute, for example, display a strong verticality that strongly appeals to the interpretive and normative agenda of global constitutionalism. This verticality is most ostensibly reflected in the fact that a supranational judicial organ such as the ICC may claim jurisdiction over certain crimes when the state in which the crimes were committed is 'unwilling or unable' to do so (Article 17 of the Rome Statute). As we shall see, global constitutionalism centres on the guiding idea that international law both enables and limits domestic political authority. This ICC exemplifies this dual function particularly well. This calls for further research regarding how far global constitutionalism can go in accounting for the constitutive and distinctive components of ICL, both substantive and procedural.
This project does, however, face some hurdles. While accurately highlighting central aspects of ICL, the nascent literature tends to overlook salient differences between global constitutionalism and ICL that cannot be neglected for the comfort of theorizing. Indeed, ICL poses distinctive challenges to global constitutionalism as an interpretive and normative framework. Institutionally, international crimes imply a distinctive process the trialthat is largely unknown to constitutional processes. It is nonetheless via the trial and its assigned functions that international criminal tribunals assert their distinctive authority. Authority is exercised and deployed in a very particular way at this stage, with specific roles and statuses (e.g. defendant, victim, prosecutor, witness, etc.) assigned and specific processes (e.g. interrogating, hearing, evidencing, convicting, sentencing, imprisoning, etc.) taking place. How could global constitutionalism account for the particular function(s) and the legitimacy of the international trial?
The verticality of ICL is further reflected in the nature and scope of these crimes: international crimes are often committed by or with the complicity of state or state-like institutions and their unique resources. One may further contend that international crimes quintessentially violate basic human rights whose protection constitutes an established global constitutional principle. However, as any criminal lawyer will know, the principle of culpability implies that criminal liability is always individual. This applies even when the crime requires 'a state or an organizational policy', as in the case of crimes against humanity (Article 7 of the Rome Statute). This raises a second challenge: How does a global constitutional account square the irremediably individualistic dimension of criminal liability?
This article pursues the objective of remedying these two deficits: how to apply global constitutionalism to the distinctively criminal dimension of individual liability; and how to conceive of the role and legitimacy of the (international) trial in global constitutional terms. In so doing, it aims to strengthen the applicability of global constitutionalism to the field of ICL. What preserves global constitutionalism as an attractive approach is the pool of abstract and normative principles that it offers, which can be calibrated and tailored to the various forms and shapes that 'rule-based authority' 3 may take in the international realm. I understand this approach as an exercise of 'constitutional imagination' that global constitutionalism generally encourages. Rather than tying constitutional imagination to established narratives, global constitutionalists 'argue (in a Kantian vein) that imagination should broaden and deepen the context of judgment and, in so doing, foster self-reflective attitudes'. 4 Methodologically, therefore, more efforts should be put on reconstructing the distinctively criminal aspects of ICL and reverting back to normative evaluation from there. Accordingly, I suggest first zooming in on what I call the collective aspect (A1) of international crimes. This aspect refers to the fact that the perpetrator(s) and the planner(s) of these crimes are often multiple and physically distanced from each other. More precisely, I explore the modes of criminal liability that have been created to accommodate for this aspect in the jurisprudence of the ICC in recent yearsin particular, the 'control theory' of Claus Roxin and the notion of 'Organisationsherrschaft.' 5 I show that under the Rome Statute, crimes against humanity are often perpetrated through a hierarchically organized apparatus of control. This vertical trait is attractive to global constitutionalismbut how should we evaluate it normatively?
Moving from the descriptive to the normative, I argue that the principle of constituent power can best delineate the global constitutional significance of international crimes. In a nutshell, constituent power insists that the legitimate authority to make and unmake law ultimately lies with the people over which authority is exercised. Based on constituent power, I suggest reinterpreting these particular modes of liability as concerned not only with the gravity of the crimes committedthe harm inflicted on the victimsbut also with the political responsibility of having large-scale, organizational control. I also point out that the limits of this claim as accessories to the crime (based on Article 25(3) of the Rome Statute), for instance, do not exercise that level of control. I further elaborate on this point through recent research on the criminological approach to the control theory. 6 Having explained how global constitutionalism can account for the issue of the special and limited liability of international crimes, one can logically move to the second aspect (A2): the role and legitimacy of the trialwhat makes the ICC legitimate in prosecuting these particular crimes from a global constitutional standpoint. On this point, I make use of Antony Duff's relational account of the trial. 7 I aim to show that this account is particularly well suited to providing the framework of a global constitutional account of the trial based on A1. More precisely, Duff precisely focuses on the relational question of who has the authority to call wrongdoers to account. For Duff, there must be a normatively defined communityof which the prosecution and the accused wrongdoer(s) can be shown to be membersthat justifies criminal courts exercising authority. This is where I suggest linking A1 to A2: in the international context, one can conceive of the trial as allowing state or state-like authorities to call individuals exercising large-scale control to account for grossly attacking their subjects. This renders justice to the core and guiding function of enabling and limiting political authority on which global constitutionalism centres.
II. ICL and global constitutionalism: Mapping the field In a recent anthology of ICL, Sarah Nouwen aptly suggests that the theory of international criminal law often replicates categories of the domestic level (in particular, retribution, deterrence, incapacitation or rehabilitation) and the distinctive problems that come with their application at the international level. 8 Nouwen labels these theories 'foundational', and distinguishes them from 'external' theories originating in other social-scientific disciplines, such as political science, sociology or economics, which address 'the effects of international criminal law rather than the foundational theories'. 9 Less explored is the attempt to conceive of ICL and its associated institutions through the conceptual and normative lens of another, yet still legal, discipline. This article argues that some important and distinctive aspects of ICL have not been scrutinized sufficiently by global constitutionalism. As a matter of legal facts, the article concentrates primarily on the crimes enshrined in the Rome Statute and the associated practice of the ICC. This means that the article does not specifically examine other institutions of ICL and their practice, such as hybrid tribunals, commissions of inquiry or domestic courts implementing the Rome Statute.
The first and preliminary step of this article is to start with a reminder of the central tenets of global constitutionalism as a broad scholarly agenda. This field of research is vast, and I certainly cannot do justice to its breadth within the remit of this article. Rather, I want to focus on some central tenets, both methodological and substantive, that are particularly relevant to my subsequent investigation of the international criminal realm.
Global constitutionalism: A brief reminder Global constitutionalism as a scholarly agenda refers to the rise of constitutional principles and values (human rights, democracy, the rule of law, constituent power, separation of power, proportionality, subsidiarity and so on) at the regional, international and global levels. Methodologically, global constitutionalism's distinctive ambition is to both describe and to evaluate these rapidly changing normative systems. The first, descriptive task amounts to reconstructing an existing institutional context where rule-based authority is exercised by or over the makers and subjects of international law, namely, states, individuals, IOs, international courts and tribunals, NGOs, or corporations. In his recent attempt to clarify 'the object of interest' of global constitutionalism, Mac Amhlaigh explains that 'the context within which constitutionalism is "apt" is the existence, in fact, of a pattern of rule-based obedience to an authority, an empirically verifiable "habit of obedience"'. 10 The rule-based authority of the ICC governed by the Rome Statute and its legal framework in which it operates (both procedural and substantive) is the institutional and descriptive perimeter of this article.
This descriptive task of reconstructing the actual exercise of rule-based authority in turn shapes the task of attempting to normatively justify it. There is no reason to assume that such reconstruction may be achieved with the help of one all-encompassing normative political theory. Here again, I follow Mac Amhlaigh's suggestion that the normative basis of global constitutionalism is a blend of republicanism and liberalism. For example, the core republican idea of self-legislation and its concern for non-domination informs the global constitutional tenets of constituent power and democracy. This explains that, 'in institutional terms, the accent in republican theory is on deliberation, contestation and participation, which makes it the natural foundational theory for political forms of constitutionalism'. 11 Liberalism, in turn, places emphasis on a catalogue of values and rights geared towards the self-development of individuals -'its belief in the ability to legally isolate certain values as fundamental to individual flourishing by reference either to metaphysical ideas of natural rights'. 12 Liberalism hence better informs the tenets of human rights and their possible restrictionsproportionality, for example.
Still, whether and how these principles apply at the supranational level remains an interpretive enterprise open to contestation and revision. In that regard, I follow Aoife O'Donoghue in characterizing global constitutionalism as an 'aspirational process … where some particular norms are present, others nascent and others non-existent, while still others merely have the potential to emerge as the system matures'. 13 Whether the international level replicates the domestic level in substance and form, or whether they find a distinct institutional expression, cannot be established a priori. O'Donogue takes the example of the separation of powers: 'certainly, the executive, legislative and judiciary separation of powers model does not need replication within global constitutionalization. Instead, funneling power into different avenues to prevent the over-grasping nature of power's character is of central importance and underlies the division required.' 14 In the next section, I illustrate this protean character of the separation of powers in the ICC context with the intra-institutional interaction between the Prosecutor and the Court.
Finally, it worth noting that despite these aspirational and multifaceted features, the overarching function of (global) constitutional principles remains the same, namely a dual enabling and constraining function, which applies to the various levels of authority implicated. As Anne Peters puts it, 'the idea is not to create a global, centralized government, but to constitutionalize global, polyarchic, and multilevel governance.' 15 The complementarity regime of the ICC is a good example; the ICC may trigger jurisdiction only when states are 'unable or unwilling' to address a situation (Article 17 of the Rome Statute). The subsidiarity regime of the European Court of Human Rights (ECtHR) is another example: state parties to the ECHR, for example, hold primary responsibility to interpret and enforce the ECHR and rights-holders must have exhausted 10 Mac Amhlaigh (n 3), 189. 11 Ibid., 191. 12  domestic remedies in order to secure admissibility at the ECtHR (Article 34 of the ECHR). These regimes both empower and limit the authority of states and supra-state institutions.
What has global constitutionalism so far achieved, and why should we even look at it (further)? Having surveyed the basic principles and agenda of global constitutionalism generally, one can move to the specific nexus between global constitutionalism and ICL and critically review how it has been addressed in the nascent global constitutional literature so far. This is done with a view to identifying deficits, which the article will then move on to address. I make two claims in this section. First, I take the example of the principle of the separation of powers to show that while global constitutionalism has been aptly applied to the interplay between the organs of ICC, it has not yet paid sufficient attention to the very function of the international trial. Second, I show that the specific kinds of conducts that fall within international criminal liability have not been considered carefully. Scholars of global constitutionalism have identified the protection of human rights as the broader global constitutional function of ICL, but failed to explain how this function applies to individuals, which are the only entities liable for committing these crimes.

Example 1: The separation of powers
Let us now zoom in on one global constitutional principle identified in the preceding sectionthe separation of powersand see how it plays out in the ICL system and at the ICC in particular. While this principle classically refers to the tripartite division between the legislative, the executive and the judiciary within the state, its core normative function is to protect subjects from the abusive, and in extreme cases tyrannical, exercise of authorityhistorically, from the potential tyranny of the absolute monarchalthough its operation may also serve accountability functions (promote transparency, publicity and effectiveness) or simply the rule of law. As Carolan explains, 'these process values promote institutional accountability, which in turn fosters the substantive principle of non-arbitrariness, which is at the core of legitimate governance'. 16 Moving from the normative to the descriptive, how is the separation of powers principle reflected in the international criminal procedure at the ICC? Anthony Lang and Andrea Birdsall take the example of the first ever completed trial at the ICC, the one of Thomas Lubanga Dyilo, 17 and in particular the interaction between two constitutive organs of the ICC, the Office of Prosecutor (OTP) and the Pre-Trial Chamber of the Court (the Chamber). While the former receives information pertaining to the potential commission of international crimes (including referrals from the UN Security Council), the opening of investigations is premised on the fulfilment of admissibility criteria, which is examined by the Chamber. In the Lubanga case, the Chamber was notably concerned that the OTP did not abide by a number of evidentiary rules, such as the disclosure of documents that may have a 'disculpatory effect'. 18 In doing so, the OTP may then interfere with the defendant's right to a fair trial, which reflects another global constitutional principle (the protection of human rights). Lang and Birdsall conclude that the Chamber both enables and limits power by making the OTP's investigative powers contingent upon certain conditions, thereby illustrating the dual function of global constitutionalism highlighted earlier 'by both limiting the power of one particular actor within the system while at the same time enabling power, that power is channeled towards productive and useful ends'. 19 While I fully agree that global constitutionalism and the separation of powers help to interpret and justify this intra-institutional process, it leaves out the core and abstract function of the ICC calling to accountand potentially punishingwrongdoers through the particular device of the trial. That is, the separation of powers principle captures the interplay between different constitutive organs of the ICC, but only when a particular situation has been opened and investigated. Yet, clearly the ICC operates other functions and hence exercises authoritybeyond that particular stage. I suggest that one of these distinctively criminal stages is the international trial. Authority is exercised and deployed in a very particular way at this stage: specific roles and statuses (e.g. defendant, victim, prosecutor, witness) are assigned and specific processes (e.g. interrogating, hearing, evidencing, convicting, sentencing, imprisoning) take place. Can global constitutionalism account for these functions, and if yes, which principle(s) could help interpret and normatively account for them? Global constitutionalism needs to tackle this issue if it aims to apply to ICL specifically. The ICC does not prosecute any alleged perpetrator over any crime, however. One therefore needs to zoom in on the substantive provisionsthe content and scope of criminal liabilityon which it concentrates.
Example 2: the protection of human rights Another example of the ongoing application of global constitutionalism in the ICL context is the protection of human rights. Clearly, this constitutional function is most clearly instantiated by human rights courts and/or by domestic courts and authorities applying human rights norms. As Stephen Gardbaum suggests, 'whatever the general degree of analogy or dis-analogy between international and constitutional law, domestic bills of rights and international human rights law perform the same basic function of stating limits on what governments may do to people within their jurisdictions'. 20 ICL tribunals such as the ICC perform this function in their own way. The basic idea here is that, as far the Rome Statute is concerned, the ICC may interfere with the sovereignty of states when these states are 'unable or unwilling' to remedy gross violations of these rights within their jurisdiction.
I will here focus on another relevant contribution, that of Elderkin. 21 This contribution helpfully points to the 'constitutional impact' of the ICC that may outweigh that of human rights systems. His approach focuses on the on the key proposition that the ICC's prosecutorial practice centres on persons in positions of authoritywhich proved decisive to the very possibility of committing crimes of that scale. This personal jurisdiction renders the ICC prima facie global constitutional. It is global constitutional in two ways. The first is that it applies to a function prototypically reserved to constitutional and human rights lawthat is, 'constraining officials from certain abuses of their constitutional powers; and offering a means for officials to be removed from office where they have committed such abuses'. 22 Elderkin's distinctive contribution is to examine how the global constitutional function of ICL differs from human rights law and constitutional law: whereas constitutional law and international human rights law exhibit a degree of subject matter overlap insofar as they both concern rights provisions, constitutional law and ICL exhibit a kind of jurisdictional overlap in the sense that they both apply limits to individuals in positions of power who have the capacity to cause massive harm if they should abuse the trust and authority vested in them. 23 It should be noted that the 'jurisdictional overlap' that Elderkin identifies as an 'abuse of trust' 24 has been carefully conceptualized by the theorists of ICL, in particular David Luban, who famously argued that crimes against humanity constitute a 'perversion of politics': 'for a state to attack individuals and their groups solely because the groups exist and the individuals belong to them transforms politics from the art of managing our unsociable sociability into a lethal threat'. 25 The second important way is comparative: the constitutional intrusion of ICL may, in principle, be greater than the intrusion of other global constitutional institutions such as human rights courts. This results from two more specific principles: on the one hand, Rome Statute crimes may benefit from universal jurisdiction, which detaches the jurisdiction of courts from the location of the crimes. On the other hand, with respect to Rome Statute crimes, the ICC may claim jurisdiction only if domestic authorities are 'unwilling or unable' to actwhich could be more effective than human rights law when the rule of law has vanished, for example. Elderkin rightly points out that 'in these circumstances, IHRL may be a source of moral comfort, confirming that the minority's rights have been infringed, but it does not provide any mechanism by which to sanction the president. ICL, however, does offer the possibility to act'. 26 The point is well taken.
While I fully endorse Elderkin's key points, I contend that his approachvery much like Lang and Birdsall'sis limited in accounting for the specifically criminal nature of ICL. This is because Elderkin's contribution conducts a comparative analysis of the ICC performance vis-à-vis human rights courts, which are not criminal. Yet, as I shall explain in more detail in the next section, the core problem with the ideas of 'constitutional impact' and 'jurisdictional overlap' is that one may easily agree that Rome Statute crimes pertain to offences committed by individuals in positions of power, and that this trait denotes a verticality that remains of prime interest to global constitutionalism aiming to enable and limit political authority. However, only individuals are liable to international crimes under ICL. This raises a significant interpretive and normative challenge for global constitutionalism: Can its principles square the irreducibly individualistic category of criminal liability? 22 Ibid., 232. 23 Ibid., 236. 24  This concludes the first, critical part of the article, which has identified two major deficits with respect to the nexus between global constitutionalism and international crimes (with emphasis on the Rome Statute). First, while I agree with Lang and Birdsall (on the separation of powers), it remains to be explained why, unlike other global constitutional institutions, the ICC is a distinctively criminal court for which the (international) trial remains the central institutional function. Second, while I agree with Elderkin (on the broad constitutional trait in that ICL focuses on wrongdoers in positions of authority), I submit that this point is not sufficient to explain how this trait is reflected in specifically criminal practices, which only concern individuals. As we shall see next, this problem of over-determination looms large: the executor and the planner of these crimes are often largely physically distant from each other.

III. Global constitutionalism and international crimes
In this section, I aim to remedy the two deficits identified above. First, I focus on the collective aspect of international crimes (A1) through the distinctive modes of liability favoured by the ICC and interpret them in light of the principle of constituent power. Second, I turn to the justification of the international trial (A2), which I achieve by introducing and applying Antony Duff's account. I show that the ICC embodies a community of state or state-like authorities that can call each other to account via the device of the trial. I then offer an account of the nexus in light of the guiding idea that ICL qua international law enables and limits political authority: state or state-like authorities have set up the ICC to call each other to account for the crimes committed upon individuals when these authorities are unable or unwilling to do so themselves.
The collective aspect (A1) International criminal tribunals are rarely concerned with the criminal liability of the 'rank-and-file perpetrator of the crime'; 27 rather, they focus on those 'most responsible', as the Preamble to Rome Statute indicates, who often are persons in positions of authority who are quite far removed from the commission of the reprehensible acts, but whose role(s) in planning and controlling the chain of actions and decisions leading the attack(s) remains decisive. More generally, it is a widely established proposition that international crimes have a significant collective aspect (A1). This has been asserted by both doctrinal and criminological scholars. The literature has even created the category of 'international criminality' for designing corresponding modes of liability: 'the gist of the argument is that the extraordinary nature of international crimes calls for a departure from "ordinary", domestic concepts of liability'. 28 Criminologists use the notion of 'structural violence' to capture this aspect. In the words of Smeulers: the outstanding legal characteristic is that they can only be qualified as international crimes if they are committed as part of a more structural form of violence. Most importantly, this implies that collective entities are involved in the commission of 27 Jain (n 5). 28  international crimesso that 'the sources' of the behavior is not the individual, but the collective. 29 Some of the Rome Statute crimes have this collective aspect enshrined -for instance in the requirement of 'a state or organizational policy' in the specification clauses of crimes against humanity (Article 7 of the Rome Statute). 30 The OTP has consistently taken this criterion as demarcating which course of conduct can ultimately qualify as a crime against humanity. 31 Other core crimes such as genocide and war crimes have been placed in this collective dimension, induced by the requirements that the crime is committed 'with the purpose to destroy a group as a whole or in part' (genocide, Article 6) or that armed conflicts occur (war crimes, Article 8). However, as I explain below, the collective aspect cuts across these categories through the modes of liability that have been designed specifically for ICL. I purport to show that global constitutionalism has the internal resources to interpret and normatively account for this collective aspect altogether (A1).
Intuitively, one may be sceptical of the very idea of criminalizing an organization as a whole. Indeed, (international) criminal law remains categorically different from global constitutional processes through the principle of individual culpability (Article 25 of the Rome Statute). It goes together with the nulla poena sine culpa principle, according to which no one should be held responsible (and hence have their freedoms severely limited) without having committed, contributed to or furthered a criminal act. It follows that in principle any individual, but only individuals, can commit, contribute to or further an international crime (Article 27 of the Rome Statute). In the words of Van Sliedregt, 'international criminal law subscribes to the liberal justice model, requiring proof of personal culpability for a finding of guilt and the imposition of punishment'. 32 Yet the principle of culpability has not prevented the discipline and practice of ICL from accommodating the irremediably collective dimension of international crimes in its conceptual architecture. While the ICC Statute does not comprise any specific reference explicitly tailored to a criminal mastermind, one may look at the development of particular modes of criminal responsibility for this purpose, which answer the central question of who is to blame. As Jens Ohlin explains, 'if we insist on individual blameworthiness to the point of ignoring the reality of collective conduct, then we have enforced a fallacious fidelity to the principle of culpability that is blind to the reality of human collaboration'. 33 The ICC therefore has to accommodate an organizational context under individualistic categories of criminal liability.
Collective modes of liability are widely discussed among scholars of ICL, and this incursion certainly cannot render justice to their fascinating complexity both in theory and practice. In nutshell, approaches have oscillated between two models. On the one hand, a common law approach is based on the notion of Joint Criminal Enterprise (JCE) focused on a collective design and execution of the crime. As Elies Van Sliedregt puts it, 29 Smeulers (n 6). 'central to common purpose liability is the common plan or purpose, which compensates for the lack of physical involvement in the crime and enables imputation of the crime at the same level as the physical perpetrator'. 34 On the other hand, a civil law approach informed by Claus Roxin's control theory of perpetration emphasizes a single but indirect perpetrator controlling and instrumentalizing an entire organization (human and material resources) to perpetrate the crime. 35 Criminologists, for instance, argue in favour of this approach: 'the strength of the control theory is that it adequately reflects the social reality of state-sponsored international crimes'. 36 Each approach has its shortcomings. The JCE does not differentiate between principals and accompliceshence blurring the hierarchical structure and the distance between the planner and the executor of the crime. The control theory may obscure the fact that several organizations are intertwined with a 'long and diffuse chain of command'. 37 This highlights the conceptual limits of the global constitutional approach as a result: not all modes of responsibility reflect the threshold of collective and institutional control and not every crime lends itself to it as a result. For example, an accomplice to a war crime (Article 8 of the Rome Statute) will not have needed to benefit from institutional control to commit the reprehensible act(s).
Again, while the relevant articles of the Rome Statutein particular, Article 25(3)do not clearly determine the applicable model of criminal liability, the ICC has clearly come to privilege Roxin's model of Organisationsherrschaft as a version of indirect perpetration in its practice. What is again central to this model is the individualthe so-called Hintermannsitting at the top of a hierarchical apparatus of power that is instrumentalized to enable and commit the crime. The defendant is convicted as principal perpetrator without having committed the actus reus of the crime. The Kenya case of The Prosecutor v. Muthaura, Kenyatta & Ali particularly reflects that approach. Here the OTP precisely had to demonstrate that the defendant and his co-perpetrators controlled and instrumentalized a hierarchical organization. The Decision of Confirmation on the Charges 38 states that the Prosecutor submits that Mr. Muthaura and Mr. Kenyatta activated and utilized pre-existing structures, such as the Mungiki to perpetrate the widespread and systematic attacks and the Kenya Police to ensure that the Mungiki operations were not interfered with (para. 103).
The Mungiki refers to an ethnic organization that was effectively performing public functions at the time of the alleged atrocities in the Kenya post-electoral context: The Prosecutor avers that '[u]p until the time of the [post-election violence], the Mungiki controlled the public transport system, provided power through illegal electricity connection, demanded a fee for accessing public toilets and sold water to residents in the poorest parts of Central Province and Nairobi. It also provided 34 van Sliedregt (n 28), 1175. 35 Roxin (n 5). 36 Smeulers (n 6). 37 Ibid. protection services to businesses and was enlisted by politicians to intimidate opponents' (para 103). Now, these modes of criminal liability specific to ICL are not only central to accounting for the collective nature of international crimes such as crimes against humanity (Article 7) and the basic purpose of convicting perpetrators within the realm of criminal law. They also are central to better identifying and imagining their global constitutional relevance.
More precisely, our incursion above lays the ground for refining Elderkin's notion of 'jurisdictional overlap'. It appears that one does not need to be a state or even a state-like leader to fall within the jurisdiction of the ICCthe precise formal or informal status of the Hintermann does not matter much. Rather, it is the act of (mis-)using or usurping a hierarchical organization to commit the crimes that is distinctive of international criminality. The example of the Mungiki organization is quite telling; it is not a state organ created and maintained to commit the alleged crimes. Rather, it is their instrumentalization by individuals at the top of the chain that enabled these atrocities. This instrumentalization illustrates the criminological perspective mentioned earlier very well, namely that 'people can control others via organizations and that they can influence and control the will of the people functioning within them'. 39 How can global constitutionalism interpret and normatively account for the collective aspect of international crimes? I suggest returning to the catalogue of normative principles of global constitutionalism surveyed in the first part of the article and to introduce another principle: constituent power. This is where our 'constitutional imagination' 40 can properly operate. Constituent power expresses the core normative idea that the legitimate authority to make and unmake laws ultimately lies with the people over whom authority is exercised. Found in various iterations in the seminal works of social contract theorists, constituent power is viewed as 'the generative principle of modern constitutional arrangements'. 41 While the term 'pouvoir constituant' is initially found in Sieyès, 42 its expression is already prominent in both Locke and Rousseau. In Rousseau, for example, constituent power takes the generative shape of the general will as the unification all individual wills and the only source of legitimate political settlement: 'that the general will, to be truly such, must be so in its object as well as its essence, that it must issue from all in order to apply to all, and that it loses its natural rectitude when it tends towards some individual and determinate object'. 43 Importantly for our purposes, the concept of constituent power has both informal and formal foundations: constituent power has an intrinsically social and political foundation that resists entrenching its authority into a formal and unamendable norm, namely the constitution. However, constituent power cannot only be the 'power of the multitude', 44 as Loughlin puts it; it must be also formalized as suchin other words, it must be constituted and constrain the power of the people. This is, in a nutshell, the 'paradox of constituent power'. What matters here is not only the basic nexus established by 39 Smeulers (n 6). constituent power between the people as a community of equals and the very possibility of legitimate political authority. It is also the implication that constituted power should be prevented from changing the basic definition of competences on which it is foundedwhat Niesen calls the 'threat of constituent usurpation'. 45 This thick nexus may be used, I believe, as a heuristic device for examining practices that are not prima facie constitutional as 'an ideal and pure type of democratic constitutional making in accordance with which we can measure and assess, that is recognize, the legitimacy of existing practices of constitutional founding'. 46 Is there any conceptual space for the principle of constituent power in reconstructing the Rome Statute crimes and their particular modes of criminal responsibility? I suggest that there is some when the ICC indicts individuals having enjoyed sufficient control to instrumentalize an entire hierarchical apparatus of power to commit their atrocities. More precisely, my claim here is that large-scale organizational control is particularly relevant to the very possibility of constituent power: if the de jure or de facto authority in placethe hierarchical apparatus of control governed by the Hintermannis used to attack a significant number of civilian subjects in a systematic and widespread manner, the necessary conditions to form and preserve constituent powerand eventually make and unmake lawsare profoundly lacking. On the one hand, this apparatus of control renders the exercise of constituent powerthe informal dimension of the conceptvirtually impossible. On the other hand, the control quintessentially marks the usurpation by constituted powerthe formal dimensionof the only normative basis on which authority can be exercised.
It is important to pause here and reflect on the implications of this argument for the broader normative significance of international crimes. A lot of ink has been spilled on the idea that international crimes are crimes that attack the 'dignity' or 'humanity' of human beingswhich Massimo Renzo views as a 'salient feature' of an international crime such as crimes against humanity: 'they constitute particularly odious offences. These crimes are so barbarous as to violate the human dignity of the victims.' 47 In the same individualistic vein, it has been argued that these crimes against deny their victims their most basic individual rights -'these rights that are necessary to all the other rights', 48 as well as their 'rights to have rights' 49 when state authorities are implicated. My emphasis on the principle of constituent power via a reconstruction of modes of liability certainly does not aim to diminish the odious and atrocious harm inflicted on individuals. Rather, my argument adds a distinctively political and collective level of wrongdoing that is fundamental on its own terms: international crimes do not just inhibit the very possibility of forming and exercising constituent power as a core condition of legitimate political authority; rather, they denote the existence of a sovereign, constituted and centralizing power that usurps the only terms that can ground a legitimate political authority. 45  This helps us to see the added value of examining these crimes through the lens of global constitutionalism: various aspects of these crimes may be apprehended and evaluated by various principlesand global constitutionalism offers a pool of resources particularly suited to interpret and evaluate the multifaceted verticality of international crimes. Now, it is equally important to point to one boundary of this argument, which also points to the boundaries of global constitutionalism. At the domestic level, Malcolm Thorburn has developed a distinctively constitutionalist account of the criminal law, which importantly overlaps with the argument presented above. The overlap extends to the key (constitutional) premise that the role of the criminal law 'is tied tightly to the very survival of the framework of rights that makes it possible for us to interact with others on terms of equal freedom', 50 rather than simply as a tool to enforce morality. This echoes the argument presented above in that the kind of morality doing the work is distinctively political. Yet, if one strictly replicates this argument at the international level, one faces the similar issue of over-determination found earlier: it is not just crimes that are part of a systematic attack on the civilian population that will be criminalized, but virtually any instance murder or rape. Here, the benefit of using constituent power emerges clearly: this principle does not apply to one individual in isolation but concerns the collective processes of forming and preserving a community of equals deciding for themselves. The scale and scope of international crimesthe fact that crimes form part of 'widespread and systematic' attack on a significant portion of the civilian population, as we have seen aboveprecisely speaks to this collective dimension with which constituent power is concerned. I shall return to Thorburn's account in the final part of the article in outlining the implications of my argument.

IV. The international criminal trial (A2)
Having offered an incursion into the modes of criminal responsibility and having interpreted their constitutional significance in terms of constituent power, I now turn to the second deficit identified earlier, namely the role and legitimacy of the international trial (A2). However, in order to calibrate our analysis, a detour through criminal law theory is necessary. Indeed, we need a conceptual framework that precisely speaks to this central and distinctive function of (international) criminal law, which criminal law theorists have long studied. In this section, I show how to fruitfully use criminal law theory in general and Antony Duff's account of criminal liability to this end, before applying the framework to A1.
I finished the first part of this article with the basic proposition that global constitutionalism is broadly concerned with international law as enabling and limiting political authority. Antony Duff's work offers a particularly smooth transition here. His account starts precisely with questioning the legitimate authority that criminal courts hold over their subjects, namely individuals. As Duff explains, 'the trial can be seen as an assertion or demonstration of the law's authority'that is, 'the authority to assert and define the polity's central values, those whose violation is to constitute a public wrong. It does not claim to create those values, or to create the wrongs that it punishes.' 51 The starting point of Duff's reconstruction is indeed that criminal wrongs are public wrongs (unlike the wrongs of private law) and that only these public wrongs justify the state triggering its authority in these matters. But what makes these wrongs public? Duff posits that criminal courts first ought to establishin normative termsthe relation in virtue of which they can call wrongdoers to account, and therefore render their authority public. This community supposes that both the authority (the criminal court) and its subjects (the defendant called to account) are members. Only then can one say that an individual can be held responsible for committing certain wrongs, and a court can recognize the wrongdoer 'as a fellow member of a normative community who is called to answer to his fellows'. 52 Duff's emphasis on the normative community as grounding the criminal law's authority then helps shed light on the particular function of the criminal trial. In his view, the trial is not only an occasion to establish the facts of the matter, acknowledge the harm done to the victim(s) and potentially convict and punish the wrongdoer; it also implies that the defendant is called to account and asked to give reasons for their actions that amount to a breach of the normative community's self-defined values: 'he comes to the trial not merely as an object of inquiry but as a subject, an agent who called to account (although of course he cannot be forced) to take part in a rational process of proof and argument'. 53 The core premise here is that every member of the normative community is criminally responsible to the extent that every member is in principle able to respond to the charge of having violated the values that constitute this community. This helps ground the very notion of criminal responsibility. As Duff puts it, 'I am responsible for X to S as Φin virtue of satisfying some normatively-laden description, which makes me responsible (both prospectively and retrospectively) for X to S.' 54 While I cannot reconstruct Duff's account fully here, I shall retain and develop two essential points in transitioning to the context of Rome Statute crimes. First, Duff's account specifically focuses on the trial's function(s) as the main explanandum of criminal law theory. As such, it offers a framework that I hope to prove useful in remedying the deficit of the current literature on ICL and global constitutionalism. It is worth insisting that Duff's justification of the trial is non-instrumental: it focuses on the standing of and the communication between the prosecutor and the court on the one hand and the defendant (and their counsel) on the other. The exercise of (international) political authority is here particular to (international) criminal law: the ruler (the prosecutor, the court) calls the ruled (the defendant and her counsel) to account for their wrongful actions. By calling them to account, the defendant is treated as a responsible member of the normative community of which she is a part: 'the trial must initially, if it is to respect the presumption of innocence, address the defendant as a citizen who is presumptively innocent of wrongdoing, although he has been accused of it'. 55 Then the defendant is called to account on the basis of charges brought by the prosecution. Finally, the stage of the criminal conviction (if any) becomes essentially communicative. It 'communicates to the defendant the judgment that he committed the wrong described in the charge. As a communicative act, it is intended to elicit an appropriate responseof understanding and (ideally) remorseful acceptance. Second, the structure of Duff's relational account remains independent from the precise nature and scope of the normative community in virtue of which the criminal lawand any criminal courtclaims to derive its legitimate authority. To this extent, it is a functional account of the criminal process and the relevant function-holder is defined based on independent variables. In the domestic context, this normative community consists of citizenship. In a recent contribution, Duff also contemplates extending his account to the European level: 'it can portray systems of transnational criminal law, such as EU criminal law, as the criminal law of a (nascent or developing) supranational political community'. 57 Now, if we apply Duff's initial question of legitimate authority to the ICC, which community does this international court embody? Duff has himself offered a response, which is to consider the relevant normativity community as the moral community of humanity: 'what gives it the right to intervene on behalf of members of more local polities whose national courts have let them down is our shared humanity; but that is not far from saying that the perpetrators should have to answer not merely to their polity, but to humanity'. 58 Yet, as I argue further later, this moral community only derives from the harm inflicted upon the victims. Here again, I suggest that from a global constitutional standpoint, the harm inflicted on individuals is a necessary condition but not a sufficient one when one specifically examines the collective aspect identified earlier (A1).

Building the nexus between A1 and A2
Having explained the architecture of Duff's account, I now return to the ICC context with a view to applying A2 to A1. There is, though, a preliminary methodological similarity between Duff's relational account and global constitutionalism that is worth mentioning. Duff's initial motivation is to question the legitimate authority of the (international) criminal law and criminal courts, as we have seen. In that sense, the relational approach is close to the global constitutional enterprise of identifying (descriptively) and evaluating (normatively) the exercise of global political authority: 'a normative account needs to be grounded in our existing criminal processes, as a rational reconstruction of the ends that such processes could be taken to serve'. 59 This methodological stance is not found in all approaches to the criminal law: a retributivist or a consequentialist about the criminal law will view the trial as essentially truth-seekingas a means to some further end, in particular punishment, without reflecting on the normative significance of the trial beyond this instrumental function. As such, it cannot explain the process of deliberation and communicationcalling the wrongdoer to accountthat is distinctive of the (international) criminal trial. Duff's relational account is here highly relevant to a global constitutional perspective to the extent that the authority of an international or supranational institution is not giveneven when it concerns acts that 'shock the conscience of humanity' as the Preamble to the Rome Statute stipulates. The normative community is a distinctively political community citizenship in the domestic contextand, as I will argue, its collective aspect (A1) can be seen as embodying a global constitutional community in the international context. The dual objective of identifying and evaluating global criminal authority also contrasts with revisionist accounts of ICL. For instance, Massimo Renzo (following Duff's own intimations) has argued that international crimes such as crimes against humanity 'deny their victims the status of being human'. 60 Renzo focuses on the particular normative value denied by these violations, namely 'humanity'. Since this status is inherent to the human condition, any act (such as murder, rape, or deportation) becomes an international crime independently of the specifics the victim or the offender. The normative community of the Rome Statute, on that view, is simply the community of humanity, and wrongdoers are called to account for violating the community's basic value: 'crimes against humanity are those that properly concern the whole of humanity, where this means that they are wrongs for which we are responsible (i.e., accountable) to the whole of humanity'. 61 Notwithstanding the huge extension of ICL's scope, this moral account does not necessarily contradict the political account privileged in this articleas already explained, they can coexistwhile the 'rank-and-file perpetrator' committing atrocities may be called to account by the human community as a whole, the Hintermann may be called to account in virtue of the organizing and institutional role they hold, and that is where global constitutionalism works best. This is where A1 comes to play a crucial specification role: I showed in the first part of this article that the collective dimension of Rome Statute crimes and the modes of liability privileged by the ICC concern the grossest atrocities and abuses committed by individuals having instrumentalized an hierarchical apparatus. This is also supported by a more criminological approach to these crimes: Arlette Smeulers has explained how the mastermind creates a 'context' that 'instigates and induces others to commit such crimes'. 62 The implication here is that under the Duffian framework, the accuser (e.g. the ICC) and the accused are assigned the same normative status, but one that is informed by the role and status they share. As far as this collective dimension is concerned, I argue in that sense that the authority of the ICC embodies a community of responsible states and state-like authorities calling each other to account. It is crucial to specify the steps that led to this conclusion: first, community membership is not formally defined by membership in the international community of states (and/or having signed and ratified the Rome Statute). Rather, it is the very fact of committing these offences through a hierarchical organization that confers membership and hence assigns criminal liability. It is, in other words, a properly normative community of a global constitutional kind. This could explain the particular jurisdictional regime of the Rome Statute, which allows virtually any state to potentially prosecute international crimes based on the principle of universal jurisdiction. This clearly remains a normative account since as a matter of practice many states do not provide for universal jurisdiction in their domestic systems.
Second, the deliberative and communicative process of the trial will be informed by global constitutional principles. For Duff, the defendant is not called to account solely as an individual or citizen, but also as an army general, a warlord, a president or a minister. The prosecutor shall treat them in accordance with that institutional and authoritative status: 'it requires that we respect the other as a participant in this process, so as someone to whom we must be willing to listen as well as talk'. 63 Surely, calling to account is not the same as establishing criminal conviction. The prosecution will have to demonstrate the 60 Renzo (n 47), 448. 61 Ibid., 460. 62 Smeulers (n 6). 63 Duff (n 56), 90. defendant's guilt by establishing their role in the commission of these horrendous crimes. A number of distinctively criminal categories matter here, in particular the mens rea requirement: the knowledge that performing certain actions will lead (or contribute) to committing (international) criminal offences. This is particularly important to international crimes as the planner of the crimes and the executors are often physically distant from one another, as we have seen. Showing that the planner enjoyed control over the actions of the executor constitutes the subjective element of the crime as opposed to the objective element constituted by the criminal acts themselves. As Jain further explains, 'the perpetrator is part of and acts within a social structure that influences his conduct, and he acts with the consciousness that he is part of a common project'. 64 Let me pause here and replace these findings in the overall structure of the article. I showed how the global constitutional subject-matter of the ICCcaptured descriptivelycan operate within an account of the international criminal trial borrowed from Duff. That is, for Duff there must be a community of which the accuser and the accused can be both members of, so that they can call each other to account via the trial. A1 enters the stage at this point: my interpretation based on constituent power suggests that it is states and state-like entities that can be conferred such membership. Note that how this argument speaks to the core idea that international law both enables and limits the authority of these entities: they can both call wrongdoers to account and themselves be called to account. V. A further challenge for global constitutionalism: The coercive dimension of ICL Having explained how A1 and A2 can form part of a coherent nexus, in this final section I aim to raise and discuss a further challenge for a global constitutional account of ICL. For reasons of space and concision, I shall only lay down the challenge, offer some preliminary ideas about how to tackle it. We have seen above how the international trial can embody a community of states and state-like authorities that call each other to account. Now, as critics of Duff have pointed out, this account has one important limitation: it cannot fully explain the coercive dimension of (international) criminal law. Indeed, the deliberative and communicative process of the trial may very well occur in a number of communities, not all public, and not all having the coercive powers of the criminal justice system. According to Thorburn, 'because Duff's account models criminal justice on the practice of a community calling its members to account for wrongdoing, it is unable to account for the criminal law's fundamentally coercive nature'. 65 While on Duff's account the criminal law depends on the definition of a normative community of which the suspected offender is a member, not every community can exercise coercion the way the criminal law entitles criminal courts to. A family, a sports club or a political party may determine what counts as a wrong according to these communities' values and rules. However, as Thorburn explains, 'they are not entitled to take away our vested entitlements or to deprive us of liberty as the state may do in response to crime'. 66 Indeed, while Duff requires paying the utmost attention to the trial, his account leaves the coercive functions of arresting, interrogating, judging, imprisoning and punishing somewhat unaccounted for. This equally is a challenge to a global constitutional approach to the Rome Statute crimes. Can our global constitutional imagination make sense of this coercive dimension?
In what follows, I aim to offer one avenue for gaining further clarity on how to address that question. This avenue is more strategic than conclusive. In examining A1, we have identified the extent to which international crimes and human rights violations are analogous. Based on A1, one can interpret international crimes and human rights violations as analogous only to the extent that are committed by public authorities or by individuals in position of public authority. But surely they differ in the kind of institutional response and sanction imposed by domestic, international or supranational courts and tribunals for their violations. Most human rights treaties specify the grounds on which it may be permissible for state authorities to interfere with these rights. For example, the right to privacy in human rights convention (e.g. Article 8 of the European Convention on Human Rights) may be curtailed in order to secure 'national security', or the right to freedom of expression (Article 10) may be restricted to preserve 'public morals'and therefore protect other basic rights of individuals. All legitimate interferences aim to protect a greater public good that can also be cast in terms of rights. 67 The dis-analogy between human rights law and international crimes may arise precisely at this point: while A1 suggests that certain perpetrators of international crimes are exercising public authority by instrumentalizing an organizational framework, there is far less textual space in the Rome Statute to argue that these acts can, in principle, be instrumental to pursuing a 'legitimate aim' with a view to protect other basic rights of their subjects. Article 33(2) of the Rome Statute, for instance, establishes that 'the law assumes that orders to commit genocide or crimes against humanity are manifestly unlawful'. The closest analogy to the restriction clauses found in human rights law would be defences for excluding criminal liability. Indeed, some international crimes such as war crimes (Article 8) have an in-built proportionality component in the definition of the office. However, the general grounds for triggering defences are narrower in the Rome Statute (e.g. destruction of capacity, imminent self-defence, etc.), which adds to the general fact that prosecutors of international crimes concentrate on individuals with the least chance of benefiting from defences. Proportionality therefore finds a more limited place in defences with respect to self-defence and the defence of others. 68 The hypothesis here is that it is the intentionality of the perpetrator that is distinctive of international crimes warranting the imposition of punishment. On this account, the global constitutional gravity of these crimes also lies in the fact that international criminal offenders commit crimes through an institutional channel that is intentionally instrumentalized (A1). The advantage of this account is that it fits the core idea of 'role inversion' that has been advanced in the literature. Two authors have brilliantly explained how international crimes pervert political authority. In the specific case of crimes against humanity, David Luban seminally diagnosed them as 'politics gone cancerous', 69 while Richard Vernon argued that the triad of 'administrative capacity, local authority and 67 The proportionality test is the device used in constitutional and human rights adjudication to establish this necessity in given circumstances. In conducting the test, courts distinguish between the legitimate aim underlying the interference with a right from the measures employed to secure that aim (or proportionality stricto sensu).
68 Cf. Articles 31-33 of the Rome Statute of the ICC include mental incapacity, intoxication, self-defence or the defence of others, and duress as the grounds for potentially excluding criminal responsibility. 69 Luban (n 25), 90.
territoriality' distinguishes the travesty of states when they attack their subjects: 'when, therefore, they play an essential role in an attack on a group of a state's subjects, that group is absolutely worse off than it could be in the worst-case scenario of statelessness'. 70 As we have specified with A1, however, it is not so much the nature of the criminal agent that ultimately matters but rather the function of instrumentalizing an institutional apparatus that is constitutionally relevant. In addition, the constitutional wrongdoing is not limited to attacking subjects when they should be protected but extends to the conditions for the formation and preservation of their constituent power. The premises and contours of this approach need a lot of further scrutiny and refinement. Most importantly, if the intentionality of the wrongdoer is central, it remains to see for which reprehensive conduct(s) this wrongdoer should be coercedand possibly punished. This raises significant issues of evidence: how does one establish the global constitutional wrongdoing of annihilating constituent power? And how does one conceive of punishment on these terms? A global constitutional account would certainly pay the utmost attention to the rule of law principles of reasonable suspicion, burden of proof or proof beyond reasonable doubt, for example. In other words, it is one thing to develop the global constitutional significance of international crimes based on established modes of liability, but quite another to include constitutional wrongdoing as a distinct criminal category with clear and practicable evidentiary requirements. It could therefore be that global constitutionalism's lens can be used to interpret the global constitutional significance of international crimes but may remain limited in informing and potentially redefining categories of criminal liability. This hurdle seems to concern the concrete applicability of global constitutionalism to ICL, but certainly not its imaginative resources.

VI. Conclusion
This article has been an exercise in global constitutional imagination. I have used global constitutionalism to reconstruct some aspects that are distinctive of the international criminal justice realm. What makes global constitutionalism attractive is that it uses a normatively rich and conceptually flexible set of normative principles to reconstruct and evaluate legal and judicial normativity. What makes ICL attractive to global constitutionalism is the verticality displayed by its institutional regime and its substantive provisions. However, how this verticality is instantiated in ICL and how global constitutionalism can conceptually and normatively exist have not been sufficiently explored.
The first step was to critically appraise recent contributions with respect to two aspects of this verticality. I pointed to two important deficits pertaining to the core function of the trial and the irremediably individualistic nature of (international) criminal liability. I then used the global constitutional principle of constituent power to interpret the global constitutional significance of international crimes via a reconstruction of the predominant modes of liability that capture the collective and organizational aspect of these crimes.
Moving to the conceptual and evaluative assessment of the trial, I argued that Antony Duff's relational account can offer a framework to fit the global constitutional subject matter previously circumscribed (A1). I inferred that as far A1 is concerned, the ICC calls public authorities to account for failing to stand by the principle of constituent power that is necessary to form and preserve a legitimate political community (A2).
Finally, with a view to encouraging further research on this nexus, I raised the challenge of accounting the coercive nature of ICL. I suggested that the intentionality of the perpetrator might be a fruitful avenue to justify coercion and ultimately punishment, but significant hurdles remainparticularly in relation to evidence.