1. Introduction
1.1. The story so far
In this article, I argue that two prominent frameworks for evaluating and developing international criminal law (ICL) can be reconciled into a new framework that absorbs the best insights of its predecessors. The simplest way to introduce this framework is to outline three major ways of thinking about ICL so far.
The first approach was particularly dominant during the rapid construction of the field of ICL. The reasoning employed in this pioneering period was primarily formalist, i.e. deriving rules from authorities and precedents, combined with teleological arguments that generally focused on expanding the reach of ICL.Footnote 1 It is not a criticism when I say that reasoning in this phase often had a rushed, incomplete, or instrumentalist character. Jurists were rightly preoccupied with the urgent task of constructing an entire legal system. In 15 years, thousands of hands hastily wove together, from sparse and inconsistent fragments, the elaborate tapestry of definitions and rules of ICL that we recognize today. It was a fast-paced transnational conversation with judicial, scholarly, and legislative inputs. There was not time to ruminate upon every subtle question.
The second step in the story is the liberal critique of ICL. As the corpus of ICL took shape, scholars began to point out that, although contemporary ICL proclaims its exemplary compliance with fundamental liberal principles, it often seems to contravene these principles, at times rather dramatically. The observation was not necessarily that the ‘emperor has no clothes’, but that there were at least some serious ‘wardrobe malfunctions’. Under a liberal approach, formalist (source-based) and teleological reasoning is not sufficient; one must also consider fundamental principles such as the principle of legality and personal culpability. Among the first pioneers in this respect were George Fletcher, Jens Ohlin, Allison Danner, Jenny Martinez, Kai Ambos, and Mirjan Damaska.Footnote 2 Concerns initially tended to focus on ‘joint criminal enterprise’, but critical attention quickly spread to other doctrines, such as the tribunals’ approach to command responsibility and duress. In very recent years, liberal principles have garnered growing scholarly and judicial attention in ICL scholarship.
The third step in the story is the critique of the liberal critique. Scholars such as Mark Drumbl, Mark Osiel, and others have pointed out that the assumptions and principles of ordinary criminal law may not be applicable or appropriate in the context of international crimes and thus should not be extended automatically to the international plane.Footnote 3 For example, ICL crimes involve extraordinary collective dimensions and extensive communal engagement, in which participation is not so self-evidently ‘deviant’, frustrating classic assumptions about ‘moral choice’ and individual agency.Footnote 4 Western principles should not be imposed on others, particularly principles which assume the individual as the central unit of action and attempt to shoehorn collective activities into individualist paradigms.Footnote 5 Thus, it is argued that objections to departures from orthodox principles are ‘exaggeratedly heated’, that departures from principles of individual responsibility may be necessary to deal with collective violence, and that it is possible that the principle of culpability may need to be modified or abandoned.Footnote 6
1.2. The argument
My objective in this article is to show the possibility of a fourth step. Although the liberal critique and the critique of the liberal critique appear to be in opposition, I aim to show that they can be reconciled into a new account that overcomes the most plausible objections to each of the two prior approaches. One can coherently agree with both strands of thought, provided that some important clarifications and refinements are made to each.
The result is a more careful liberal account. I embrace the critique that we cannot simply project familiar national principles onto ICL. We must inspect and re-articulate those principles to take into account the special contexts encountered by ICL, which include massively collective action, state criminality, and non-legislative forms of law creation. I attach an important caveat, however: that the special contexts do not mean that we are free to discard our underlying deontic commitment to our fellow human beings. Thus, my account remains a liberal account, in that it still respects principled constraints rooted in respect for the moral agency of individuals. It differs from the straightforward approach of replicating familiar principles, in that it explores how our deontic commitment might manifest differently in different circumstances.
My hope and aspiration is that most scholars adopting a ‘liberal’ approach to ICL would, once presented with the issue and the arguments, favour the more subtle variation I outline here. Their aim would not necessarily be to replicate the articulation of principles but rather to respect the underlying deontic commitment. Similarly, it is my hope that many of those emphasizing the distinctiveness of ICL and the need to avoid unreflecting mimicry of domestic law would agree that the refashioned rules must still comport with a credible account of just treatment of individuals.
In the present space, I can only aim to sketch out the outlines of the framework. I hope to do so sufficiently to convince you of its feasibility and comparative appeal. I will suggest that this account may generate some uniquely helpful questions about current problems and controversies in ICL, including duress, superior orders, and non-retroactivity. For example, rather than reproducing familiar understandings of the legality principle, or rejecting it as inapplicable, we can use a system of diffuse forms of law creation to explore its underpinnings and ramifications in this novel context.
Finally, as an interesting by-product, the account may also generate insights for mainstream criminal-law theory, because it leads us to a liberal theory of abnormal criminal law. For example, most liberal accounts focus on the familiar dyad of the individual and the state. However, our aspiration of examining desert in the abnormal contexts faced by ICL – where law may be applied not by a state but by an international tribunal – leads us to into some new and largely unexplored territory. We lose the familiar backdrop of most criminal-law thinking: i.e. a generally orderly society, in which a single overarching state is the law-giver, law-adjudicator, and law-enforcer. We lose many of our familiar tools of thought based on the relationship between state and citizen. These abnormal features compel us to explore a more general account of criminal justice under very different conditions. The cosmopolitan features of the account include not only its resistance to parochial replication but also its exploration of criminal law under governance structures other than the state.
Examining criminal law in ‘abnormal’ situations can teach us about our theory of the ‘normal’ situations. ICL – criminal law in extremis – presents a new set of problems, which may expose unnoticed parameters and subtleties in concepts such as moral choice, fair warning, or authority. Such inquiry can demonstrate that familiar and seemingly fundamental principles are contextually contingent manifestations of more complex commitments.
This article builds upon a previous work, ‘The Identity Crisis of International Criminal Law’.Footnote 7 My focus in that work was unusual: it was not about liberal principles per se, but rather about techniques of reasoning employed in ICL. I explored certain recurring techniques of reasoning and structures of argument, regarded as appropriate in a liberal legal system, and showed how they engendered internal contradictions with the system's stated principles.Footnote 8 Because my interest lay in reasoning and internal contradictions, I looked at compliance with fundamental principles as articulated by the system itself.Footnote 9 I emphasized that I was not purporting to evaluate those contradictions, as that would require a conclusion on what principles the system ought to follow.Footnote 10 I noted that an evaluation of how best to resolve those internal contradictions would require inter alia a philosophical analysis of the merits of particular principles and doctrines and would be part of a much larger research project.Footnote 11 The present article advances the next step in that project.
The prescriptive claim in ‘Identity Crisis’ was quite modest: that ICL should avoid self-contradiction, i.e. promising compliance with certain principles and then contravening them.Footnote 12 I noted that, from the point of view of coherence (non-contradiction), there are multiple ways to resolve these internal conflicts, such as reforming the doctrine, refining the principle, or rejecting the principle.Footnote 13 In this article, I now advance a stronger claim: that such contradictions should be resolved in favour of fundamental principles. However, as foreshadowed in ‘Identity Crisis’, this must be accompanied by a careful methodology to inspect and articulate the fundamental principles properly applicable in ICL.Footnote 14 I suggested that there was scope to re-examine the specific articulations of fundamental principles as found in national law while retaining fidelity to the fundamental liberal commitment of treating persons as ends and not solely as means.Footnote 15 This article develops and explains those lines of thought.
1.3. Outline and terminology
In section 2, I show why we should care about liberal principles. In section 3, I review and agree with critiques of liberal accounts that simply replicate articulations from national systems (what will be called here a ‘parochial’ approach). I accordingly introduce a ‘cosmopolitan’ liberal approach. In section 4, I outline various objections to liberal accounts and argue that a careful liberal account can indeed contend with pluralism, collectivity, and community. In section 5, I clarify that the suggested framework does not purport to generate incontrovertible answers; it merely provides a framework for asking certain questions. Section 6 briefly outlines some potential implications of this account for specific controversies in ICL.
In this article I use the following terms in the following ways. A ‘doctrine’ is a rule, posited in the legal system, stating for example the elements of crimes against humanity or the requirements of command responsibility. A ‘fundamental principle’ is a principle of a liberal system, such as the principle of culpability or the principle of legality. An ‘articulation of a fundamental principle’ is a certain understanding of the concrete features of a fundamental principle, for example the proposition that the principle of legality requires prior legislation published in the official gazette. Finally, the ‘underlying deontic commitment’ refers to the basic commitment from which these fundamental principles are derived. For now, we can simply say that it is the commitment to treat persons as moral agents, possessed of dignity and capable of directing their behaviour by reason.
2. Why a ‘liberal’ account?
In ICL discourse, fundamental principles such as the principles of legality and culpability have often been treated as mere doctrinal rules – rules with the unfortunate effect of hindering the pursuit of justice – and thus various arguments have been employed to avoid, minimize, or circumvent the perceived ‘inconvenient obstacles’.Footnote 16 If one sees fundamental principles as mere stipulations of positive law, it is entirely understandable to use the clever doctrinal techniques we use to avoid or minimize any problematic rule.Footnote 17 Moreover, if they are merely rules from national systems, then one may ask whether such rules are legally applicable in ICL at all.Footnote 18 Accordingly, against this jurisprudential backdrop, it is pertinent to very briefly outline why such principles are of more than formalistic importance.
The basic idea is that a ‘liberal’ system of justice is constrained by respect for the autonomy, dignity, or agency of the human beings subject to the system. Even if one sees a utilitarian purpose for the criminal-law system as a whole (to protect society), the question of ‘what justifies us in applying the system of punishment to a particular individual’ cannot be decided solely on utilitarian grounds.Footnote 19 Consequentialist arguments fail to capture our abhorrence of imposing punishments that are not deserved: such punishments are simply unjust.Footnote 20 Accordingly, while utilitarian arguments are relevant and important, we also require a deontological justification, showing that the punishment is ‘deserved’. A system that neglects the constraint of desert is arguably not a system of ‘justice’,Footnote 21 and is arguably not a system of ‘criminal law’ but rather an exercise of ‘police’ power.Footnote 22
The conviction that persons are ‘subjects’ and not ‘objects’Footnote 23 is the core commitment of a liberal criminal justice system. This basic commitment translates into certain fundamental principles that constrain a liberal system of justice. One such principle is the principle of personal culpability, which means that persons can only be punished for crimes for which they bear some personal responsibility, by virtue of their acts and omissions.Footnote 24 ICL jurisprudence declares that the principle of personal culpability is ‘the foundation of criminal responsibility’.Footnote 25 A second principle is the principle of legality (nullum crimen sine lege), which stipulates that persons can only be punished for conduct that was criminal at the time of commission (i.e. no retroactive criminalization). This principle allows individual actors to have fair notice of criminal prohibitions so that they can order their affairs accordingly, and restricts the scope for arbitrary exercise of coercive power by officials.Footnote 26 ICL recognizes the principle of legality as a ‘solid pillar’ without which ‘no criminalization process can be accomplished and recognized’.Footnote 27
I want to emphasize very clearly that I am using the term ‘liberal’ in this specific and minimalist sense. The term ‘liberal’ means many different things to different peopleFootnote 28 and has very different meanings within different debates. In criminal-law theory, it commonly conveys that the system is constrained by respect for the autonomy, dignity, or agency of the individual. That is the minimalist sense in which I use the term; that a system of justice entails some principled (non-consequentialist) constraints on the pursuit of societal protection.
As I will explain in sections 4 and 5, this minimal condition is compatible with more than one political philosophy or moral philosophy (whether or not they are labelled as ‘liberal’ in the many other senses of the word). It is also compatible with more than one concept of the individual or cultural tradition. I assume for now that a liberal system of ICL will include both the culpability principle and the legality principle in at least some form. The precise contours and ramifications of those principles remain to be determined. The point of the proposed account is to provide a framework for inquiry about those contours.
In sum, fundamental principles may at times seem to inhibit the pursuit of maximal societal protection. However, the alternative – to create a punitive system for the ‘administrative elimination of wrongdoers’Footnote 29 in the name of advancing human rights – seems philosophically incoherent.Footnote 30 If ICL wishes to instil the value that human beings must be treated as moral agents possessed of dignity, it must in turn treat persons as moral agents possessed of dignity. To treat persons as objects in order to send a message that persons may not be used as objects is to embark on a project riven with self-contradiction.
3. Why a ‘cosmopolitan’ liberal account?
3.1. The critique of replication, and a caveat thereto
In the foregoing sections, I sought to show why we should value liberal principles of criminal justice. However, the liberal approach to international criminal law – or at least the way the liberal approach is commonly employed – has itself been subject to excellent criticisms. In particularly important contributions, Mark Drumbl, Mark Osiel, and others have convincingly argued against the automatic replication of the assumptions, methods, and principles of national doctrinal frameworks in ICL.Footnote 31 Drumbl argues, for example, that the paradigm of individual culpability may be suitable for deviant isolated crimes but not for mass crimes, which involve organic group dimensions.Footnote 32 Many scholars rightly emphasize that, whereas ordinary crime involves ‘deviance’ from societal expectations, ICL faces situations of ‘inverted morality’Footnote 33 in which there is strong social pressure to participate in crimes.Footnote 34 In ICL contexts, it is often abstention from crime that would be deviant. Scholars also warn against extending ‘western doctrines onto the transnational plane without considering the implications for societies not sharing similar assumptions’.Footnote 35 For these and other reasons (see section 4), it is argued that principles such as culpability may have to be adapted, modified, or even abandoned.Footnote 36
I agree that ICL need not replicate particular doctrines, or even particular articulations of fundamental principles, merely because they appear in national systems.Footnote 37 We may and must critically inspect articulations of fundamental principles to assess their relevance and soundness in new contexts. However, I would add a crucial caveat to these important observations. This latitude for reinspection does not entail that we are free to abandon the underlying deontic commitment to treat humans justly as moral agents. Thus, we must still grapple with the question of desert. Common formulations of principles may be re-evaluated and rearticulated, but the articulations require a plausible deontological justification.
Insofar as scholars such as Drumbl and Osiel are simply calling for thoughtful inspection of liberal principles,Footnote 38 the position I outline is compatible with theirs. There are only a few passages in Drumbl's work which seem to suggest a fundamentally different approach, in which case my caveat would be significant. For example, Drumbl observes that international tribunals’ ‘recourse to generous – and at times somewhat vicarious – liability theories becomes eminently understandable’ in light of the collective and organic sources of violence.Footnote 39 I agree that the pressure to expand liability doctrines is understandable in a psychological sense. My caveat is that, if vicarious liability refers to liability without culpability, it does not seem justifiable in a system of criminal law. Responsibility short of personal culpability should be addressed through other mechanisms. Some passages of Drumbl's work suggest that he would likely agree with this caveat.Footnote 40
I believe my proposed approach may also be reconcilable with that of Mark Osiel. (Some readers of Osiel's thoughtful work, Making Sense of Mass Atrocity, may find this optimism surprising, because in that work he presents my approach as being in opposition to his own.Footnote 41 The approach he ascribes to me would fall within what I describe as a ‘parochial’ liberal account. The present article introduces and explains my approach and, I hope, illuminates the brief clarifications and foreshadowing that I tried to provide in previous works.Footnote 42)
It is difficult to be certain whether my caveat is compatible with Osiel's position, as he refers to culpability in ways that differ subtly. At times he emphasizes the need for compliance with principles,Footnote 43 while arguing that there is scope to adapt liberal principles,Footnote 44 which is particularly akin to the approach I outline here. At times he contemplates some degree of non-compliance, suggesting that ICL should ‘ideally’ comply, Footnote 45 that it should not ‘unduly’ depart,Footnote 46 and that incompatibility should be kept to a ‘morally acceptable minimum’.Footnote 47 These suggestions could also be compatible with the approach I outline here, if a higher-order theory can be developed that justifies departures from deontic principles (see section 5, below). However, at times he seems still more sceptical, lamenting the prevalence of deontological thinking in criminal theory and the ‘reverential status accorded to the culpability principle in current criminal theory’.Footnote 48 My approach is compatible even with these more sceptical statements about culpability, if they are taken to refer to historically contingent articulations of the principle appearing in national laws, or to the ‘punctilious’ manner in which it is sometimes applied.Footnote 49 Our approaches would diverge, however, if he is referring to the underlying deontic commitment. I would argue that careful respect for the culpability and legality principles, once they are properly delineated, is entirely warranted in criminal law. Thus I could agree with modification, but not abandonment, of the culpability principle. For reasons outlined in section 2, if we punish without culpability, we are arguably no longer engaged in criminal law but rather – to use Markus Dubber's term – an ‘ethical–administrative enterprise’.Footnote 50
Osiel raises an interesting and valuable point when he argues that public policy decisions cannot be based on ‘philosophical “principle” or metaphysics’, because ‘normative questions are . . . at stake here, not metaphysical ones’.Footnote 51 I can largelyFootnote 52 agree with the proposed prioritization of the ‘normative’ over the ‘metaphysical’, insofar as ‘metaphysical’ sounds like abstract speculation about non-verifiable matters and ‘normative’ sounds like an evaluation of practical questions. However, that still leaves questions: are we speaking of a normativity of the good or of the right? Are we simply maximizing general public welfare (the good) or are we also respecting the autonomy, rights, and agency of others (the right)? I suggest that the question of culpability is a normative question, an urgent and important one which delineates the limits of a system of justice.Footnote 53
3.2. ‘Parochial’ and ‘cosmopolitan’ liberal approaches
For ease of reference, I will use the label ‘parochial’ to describe a liberal approach that simply replicates familiar principles from one's legal system, or even from several legal systems.Footnote 54 In doing so, I am emphatically not suggesting that any scholar is in fact committed to a ‘parochial’ liberal approach. I am simply using it as a label for the problematic approach described in critiques, in order to distinguish such an approach from the more careful version I suggest here. The two approaches (replicating national articulations or digging down to the underlying commitment) are not normally differentiated, and so it is commonly sufficient in ICL scholarship to simply refer to national principles. As mentioned in the introduction, my hope is that, once presented with the distinction, most scholars adopting a ‘liberal’ approach to ICL would favour the more subtle variation I outline here.
I tentatively adopt the adjective ‘cosmopolitan’ as a label to distinguish the approach that I advocate here. The term offers an obvious contrast to the ‘parochial’ approach of reproducing one's system. Further, the account resonates with cosmopolitanism as developed in literature on international relations and international legal theory,Footnote 55 and referenced in ICL literature.Footnote 56 While the term is used with slightly differing connotations and permutations, there are three recurring features. First, cosmopolitanism searches for commonalities between cultures but it also recognizes and respects differences, thus embracing pluralism and the building of a modus vivendi.Footnote 57 Second, our regard for others does not stop at the boundaries of a given state; we are rightly concerned with the global community of humans.Footnote 58 Third, the emphasis is placed on humans rather than states; states are merely of derivative significance as one instrument to advance the ends of human beings, and they may be supplemented by other governance structures as needed.Footnote 59 The first point captures the anti-parochial aspiration of the framework. The second and third points are particularly pertinent, as ICL entails the ‘unmediated’ application of law to individuals by international governance mechanisms.Footnote 60 It applies in many states and thus cannot rely on standard thinking about criminal law, which largely treats each state as an isolated compartment and analyses problems within that compartment. Thus, I believe that cosmopolitanism resonates with the aspirations of ICL: a concern for human beings that extends beyond borders, a willingness to embrace alternative governance structures to supplement state structures,Footnote 61 and inclusiveness of the concerns of the ‘international community as a whole’.Footnote 62
A feature of cosmopolitanism that is particularly relevant for this project is its contrast with a state-centric approach, which is a very challenging idea in a criminal-law context. Cosmopolitanism offers an understanding of individuals not only as citizens of a given state, but as members of overlapping networks. States are an important and prominent centre of authority, but they are not the only possible centre of authority. One can envisage a ‘neo-medieval’ landscape with a more diverse array of governance structures.Footnote 63 By contrast, criminal-law theory traditionally assumes the modern state as its centrepiece, which is entirely understandable given our historical experience of criminal law. Apart from peripheral cases of overlapping jurisdiction, most thinking about criminal law regards each country as a more-or-less closed system, with the players being the single state and the individual inhabitants. In this picture, one can readily rely on concepts such as citizenship or community to help explain aspects of criminal law. ICL, especially when applied by international tribunals, provides examples that do not fit this familiar picture. It can help us see that the familiar picture is only one instance of a more general case. Although many regard the state as a strictly essential requirement for criminal law, we might find on inspection that what is really required is not the entire package of the modern Westphalian state, but rather certain features of the state, and that those features could be allocated differently or vested in other institutions.
I take the term ‘cosmopolitan liberal’ (which has been used in different literatures with slightly different connotationsFootnote 64) to combine two humanistic ideas. One is a liberal respect for the individual as an agent with autonomy and dignity, and the other is a cosmopolitan regard for all human beings globally as ends of law (which implies a somewhat reduced emphasis on states and their borders).Footnote 65 The approach is ‘liberal’ in its concern with justifiable public coercion, and ‘cosmopolitan’ in that it is willing to contemplate such questions not only in the familiar compartment of a state (a legal entity linked to a defined territory, population, and government) but under other possible configurations.
There are many other ways in which the special contexts raised by ICL may raise new challenges, because they depart from the familiar context in which we normally think about criminal law. Mass co-ordination may increase the harm inflicted while also possibly diminishing the agency of individuals. Causally over-determined crimes raise questions about causation and blame.Footnote 66 Crimes of obedience raise questions about deviance and wrongdoing. Criminal governments overturn the normal role of state as law-provider. The alternative means of law creation used in ICL call for reflection on the meaning of fair warning and the nature of the legality principle. The examination of the special problems faced by ICL can lead us to learn more about the principles as used in everyday experience.
4. Embracing pluralism, collectivity, and community
This section explores some specific objections to liberal approaches. My suggestion is that a more thoughtful liberal account can embrace and absorb these critiques and be strengthened by them. The most common objections to liberal accounts include: that they insist on a legalistic criminal-law response and exclude other alternatives, that they are fixated on the individual and cannot process collective dimensions, that they conceive of persons as socially unencumbered individuals and fail to account for communitarian values and duties and social meaning, and that they impose Western constructs.
4.1. Welcoming pluralism of responses
Valuable criticisms have been made of the tendency of the ‘liberal legalist’ modelFootnote 67 to treat criminal law as the ‘preferred ideal type’ and ‘first best practice’ for responding to mass atrocity, squeezing out consideration of other mechanisms.Footnote 68 Many scholars have emphasized that other mechanisms, such as truth commissions, civil liability (including of states or corporations), legislative and institutional reform, and lustration and commemoration, may be better suited to the collective nature of mass atrocity. Such mechanisms can better address root causes and structural features, or address the contribution of bystanders and other enablers who did not make a criminal contribution and yet share some responsibility.Footnote 69
It is indeed important to recall the limits of the criminal sanction as a response and the functions that it does and does not attempt to perform. A cosmopolitan liberal account is perfectly compatible with a pluralistic approach and the use of other mechanisms, operating alongside or perhaps as an alternative to criminal law.Footnote 70 It is compatible because it is only a theory about the constraints of criminal law. A cosmopolitan liberal account simply requires that, once we adopt a criminal-law response, and thus choose to stigmatize and punish individuals, we do so in a manner respectful of the moral agency of human beings.
4.2. Grappling with collectivity
As was discussed in section 3, many scholars have emphasized that the collective dimensions of mass atrocity and the attendant social pressures create severe challenges for orthodox ideas about crime and ‘conformity’, ‘deviance’, ‘agency’ and ‘moral choice’.Footnote 71 As noted above, I agree that these features call for a fresh inquiry. Nonetheless, once one chooses to employ criminal law and thereby to blame, punish, and stigmatize individuals for crimes, one has no choice but to grapple with individual agency, choice, and desert.
Liberal criminal theory is sometimes presented as regarding the individual as the central ‘unit of action’.Footnote 72 However, I think this slightly misses the reason for the focus on the individual; it is not because of any prior commitment to a blinkered form of methodological individualism. Instead, I think that the focus on the individual arises because, once criminal law is employed, the individual is the unit of punishment. Once we decide to punish and stigmatize individuals for crimes, we are obliged to determine what we are punishing them for. This inevitably brings the question of individual agency – to identify the actions and contributions for which that individual is to be held responsible.
A liberal account does not require that we parcel out each contribution so that each harm is attributed to one and only one individual.Footnote 73 A liberal account can recognize that when individuals pool their efforts together, they can create consequences of a scale and nature that they could not achieve on their own, and that they can share in various forms of responsibility for their collective doings.Footnote 74 On a careful liberal account of mass crimes, we would contend with the intriguing challenge that collective action can both expand agency, by allowing attainment of aims that could not be attained alone, and also diminish agency and moral choice in situations of social pressure, propaganda, or demands of authority.Footnote 75
We should also guard against overstating the limitations of criminal law. It would be a mistake to suggest that criminal law or liberal criminal-law theory is so fixated on individuals that it is completely unequipped to grapple with collective action.Footnote 76 Collective action should not be seen as a form of kryptonite before which criminal law finds itself powerless. Individuals have been working together to commit crime since ‘crime’ was first conceived. Criminal-law doctrine and theory draw on centuries of thought and experience concerning individuals pooling their efforts to produce crimes. This has generated tools such as joint commission, commission through an organization, complicity, and the distinction between principals and accessories. In the context of macro-criminality, much interesting thought has been given, for example, to how to address the Hintermann – the ‘man in the background’ – who is not present at the crime scene but who masterminds the crime.Footnote 77 Further reflection is certainly needed to better gauge individual agency in mass crimes, but it is premature to say that criminal law is unable to do so, or that it suffers from some defect that allows it to only see the world as isolated individuals acting on their own and prevents it from discerning co-ordination among individuals.
Moreover, problems of agency are not unique to ICL. National systems confront puzzles of foetal alcohol syndrome and operate in communities in a state of anomie where criminality is normalized. The agency issues faced by ICL may be different in scope or nature, but they are not exclusive.
4.3. Acknowledging community and social construction
A related critique is that liberal accounts are so individualistic and abstract that they miss out on communitarian values and duties, or on social meaning.Footnote 78 The concern is that liberal theory misconceives of the individual as completely separate from society, and must disaggregate events into ‘socially unencumbered individuals independently interacting’, producing distorted understandings.Footnote 79 Certainly, some forms of political liberalism, particularly classical contractarian theories, may be vulnerable to such critique.
However, we can advance a liberal criminal-law theory without necessarily subscribing to an empirically unsound world view in which we were all atomistic, self-created individuals who entered into a social contract to advance our personal aims. As Alan Brudner notes, ‘[c]ontrary to a common belief, a liberal theory of penal justice is not necessarily one that conceives the individual as an abstract subject or person uprooted from its social and ethical environment.’Footnote 80 A liberal theory can acknowledge that we are social and political animals, that we were born in society, and that our identities and our realities are richly socially constructed.
A liberal criminal-law theory does not have to ignore the importance of society. It simply requires that we justify our actions against the individual on behalf of society. We may indeed be social animals, but we are not drones in a hive society, to be used without concern as instruments for the collective good. A liberal account recognizes that there is some attribute of sentient persons, whether it be labelled autonomy or dignity or capacity for reason, that requires us to pause and justify our punishment and treatment of them.
Indeed, in a careful liberal theory, social context and social roles may play a powerful role.Footnote 81 The acknowledgement of community and of social roles may bear fruit (as we will see in section 6 in the discussion on duress) by bringing into question some of the ready conclusions of more atomistic liberal theories.
4.4. A global, rather than ‘Western’, account
Finally, I come to the most difficult challenge for a cosmopolitan liberal account. A frequently advanced objection to liberal principles is that they are a ‘Western’ construct.Footnote 82 Such warnings rightly alert us to the historical and cultural contingency and potential non-universality of familiar articulations of principles, as well as the inappropriateness or even neocolonialism of extending such principles in other contexts. Cosmopolitanism attempts to avoid such criticisms through its readiness to absorb, benefit from, and accommodate perspectives of different regions and traditions.Footnote 83 Hence a cosmopolitan liberal account looks to diverse legal systems and traditions for its inspiration, to identify widely shared principles rather than a regional conception.Footnote 84
This aspiration opens up a potentially powerful objection to the cosmopolitan liberal account. One could argue that liberal theory is so thoroughly and irreducibly ‘Western’ that a ‘cosmopolitan liberal’ account is inherently contradictory. In the present space, I aim merely to sketch out what I believe would be the two main lines of response.
The first line of response is empirical. Without denying the affiliation between ‘liberalism’ and Western thought, one could question whether basic principles such as fair warning or personal culpability are truly only values of the West. For example, in the negotiation of the Rome Statute, delegates from all regions and legal traditions exhibited a shared commitment to the fundamental principles of legality and personal culpability.Footnote 85 One could respond that the delegates were subject to Western pressure or reflected a Westernized elite. However, a survey of domestic systems would indicate that the principles seem to attract global recognition and support.
One could nonetheless respond that, even if there is shared support across legal traditions and regions today, liberal principles of criminal justice are still Western in origin and were imposed and exported during waves of colonization. Yet even that proposition is murkier than is widely assumed. To give just one example, in China, as early as the eleventh to eighth centuries bc, royal instructions required local rulers to ‘set forth’ and make accessible the laws on offences and punishments and ‘to ensure that officials apply the existing law and not on their own initiative introduce innovations’.Footnote 86 There followed a considerable legacy of codification and publication,Footnote 87 including placing descriptions of penal laws outside the palace for the information of the public.Footnote 88 Ancient laws reflected not only the principle of legality but also the principle of culpability, including distinguishing intentional from accidental acts and mitigating punishment for the young.Footnote 89 Such developments, millennia before Europe saw its ‘Enlightenment’, suggest that these principles of fairness should not necessarily be credited and ascribed to a single region; they may have much deeper roots and broader appeal.Footnote 90
Furthermore, empirical evidence also suggests that common intuitions of justice reflect the principle of culpability, and include quite subtle distinctions tracking criminal law and (deontological) moral philosophy.Footnote 91 Cross-cultural studies show a remarkable confluence of intuitions in subjects from the USA, China, Puerto Rico, India, Indonesia, Iran, Italy, and Yugoslavia.Footnote 92 Similarly, anthropological work suggests that very basic concepts of responsibility are quite widely shared.Footnote 93 Such findings provide further reason to at least hesitate about the assumption that fundamental principles are ‘Western’.
The second line of response is conceptual. Rather than investigating the empirical origins of liberal principles, one would shift to merits, and ask whether there is any attractive alternative. It is not clear that we have any plausibly attractive model, even in theory, for a criminal law that rejects the culpability and legality principles. Presumably one could point to authoritarian or totalitarian systems as examples of a different model, but it would be hard to make the normative case that ICL should emulate them. Could we advocate a criminal law that, for example, rejects culpability and hence convicts persons for crimes in which they were not involved? One could also ask about coherence: would such doctrines be coherent with the values of ICL, or would it culminate in internal contradictions?
One could object that the foregoing appeal not to reject the constraint of culpability sounds plausible only because it is expressed at a high level of generality, and that legal traditions diverge when they articulate the concept in more detail. However, my point is precisely to distinguish between these levels of generality. Here, I am addressing objections to the constraint of culpability in general. Once we agree that culpability matters, we then turn to the more precise task of articulating the content of the constraint. At this more granular level, we consider the divergent articulations, as they show us solutions worked out through experience, and awaken us to different traditions of thought. I should emphasize again that we do so for purposes of inspiration rather than replication. It is possible that ICL may adopt deontologically justifiable articulations of principles that are different from any existing system.
One could also rightly warn that the language of cosmopolitanism has often been used, both advertently and inadvertently, as a mask for hegemony.Footnote 94 The point is sobering, but it does not seem to be a reason to decline to even attempt a genuine cosmopolitan conversation. It serves as a potential critique of a particular application of the methodology, where it is shown that a plausible perspective has been neglected.
5. Clarifications of the cosmopolitan liberal account
The objective of this article is to show that one can coherently agree with the best insights of those advocating a liberal approach and those questioning the applicability of familiar principles. My basic point is that a synthesis is possible: we can acknowledge the distinct contexts of ICL and yet still require fidelity to an underlying deontic commitment. It is not possible to explore all underpinnings and implications of this approach here, but I will elaborate on such questions in a later treatment. For now I aim only to sketch the outlines of the approach, so that you can see how it relates to other approaches. In this section I briefly address the most likely questions, and in the next section I will outline some of the questions that the approach would pose for current ICL problems.
5.1. A minimalist framework for questioning
First, I emphatically do not suggest that the cosmopolitan liberal account, or deontological inquiry in general, can pretend to produce ‘answers’ with anything remotely akin to mathematic certainty or precision. It is merely a framework requiring us to pose certain questions. It is a discipline that requires us to go beyond revulsion and the wish that someone be punished, and to consider the constraints of justice. It requires that we not only look at questions of efficacy, but that we also confront questions about fairness.
Views as to what makes a person culpable in relation to a particular act may plausibly differ. For example, liberal criminal theorists are divided as to whether criminal negligence is sufficient to ground culpability for punishment.Footnote 95 Liberal criminal theorists have divergent views on what form of causal contribution is required for culpability, or whether causal contribution is always required.Footnote 96 The point is simply that we grapple with these questions. It is the frame of a conversation. A liberal account requires us to provide a coherent and plausible claim that we have upheld our commitment to respect the agency of the persons subject to our system.
Second, the proposed framework seeks to make minimal assumptions about underlying political philosophies and about theories as to why persons deserve respect, and thereby to be compatible with multiple viewpoints. We need not resolve all such riddles, because points of ‘overlapping consensus’Footnote 97 generally provide a sufficient basis to work on problems of the middle range, such as examining criminal-law doctrines.Footnote 98 For example, when scholars speak of the attribute of persons that warrants respectful treatment, they refer variously to attributes such as agency, autonomy, dignity, the capacity for reason-directed behaviour, personhood, worth, and so on. It is conceivable that some criminal-law problem may arise that requires us to specify the relevant attribute with more precision, or to distinguish between different political philosophies, in order to generate an answer. Until then, for most questions, liberal criminal-law theory can provide an adequate tool for analysis without taking sides in such controversies, and by adhering to what is ‘uncontroversially generic to the various schools of liberalism’:Footnote 99 that human beings have intrinsic worth and thus that we cannot use them as mere objects for a lesson to deter others or to convey socially valuable messages. Indeed, given the minimal sense in which I use the term here, a ‘liberal’ criminal-law theory may be compatible with political philosophies such as communitarianism, provided they also value autonomy, and hence embrace and explore constraints such as culpability.Footnote 100
Third, the approach I outline seems to be compatible with ‘absolute’ or ‘moderate’ deontological approaches. The most straightforward deontological position is the ‘absolute,’ position that fundamental principles must never be breached; society may never treat persons unjustly in its pursuit of societal aims.Footnote 101 However, some liberal theorists seem tentatively open to the possibility that departures from deontic principles might be justifiable in exceptional circumstances.Footnote 102 The latter view could be supported by ‘moderate’ (or ‘threshold’) deontological approaches, in which overwhelming circumstances (such as avoiding catastrophe) might override our commitment to the individual.Footnote 103 The ‘moderate’ view has intuitive appeal but is more difficult to develop, as one would have to explain why thresholds exist, what they are, and how and why deontological and utilitarian arguments are mutually delineated and constrained.Footnote 104 My point for now is simply that a cosmopolitan liberal account does not seem to be incompatible with either the absolute or moderate view. The objective of the account is that one requires a plausible justification for ICL doctrines, respectful of the agency and personhood of the persons under the system. There is no reason at present to rule out that a plausible justification might be based in an absolute or moderate account.
5.2. Liberal criminal theory without ‘the state’?
I suggested above (section 3) that in ICL, the traditionally central character in liberal criminal-law theory, the state, ceases to play that central role. This may sound puzzling. After all, states seem pretty ubiquitous in ICL: they are victims and villains, they order crimes and try to halt crimes, they compete for jurisdiction and bestow jurisdiction, they carry out arrests, and they shape ICL doctrines as well as the priorities and policies of international justice efforts. What I am saying is that ICL confronts us with criminal law without ‘The State’, that familiar conceptual framework for criminal law thinking: a single, monolithic, omnipresent entity, claiming a monopoly of force in a given territory, exercising supreme authority, and acting as law-maker, law-interpreter, law-enforcer, impartial dispute adjudicator, keeper of the peace, custodian of public right, embodiment of the community, and beneficiary of duties of allegiance.
In the normal case, criminal-law theory can assume that all of these roles and attributes are merged in one posited entity, which is also the entity creating and enforcing criminal law in the case under question. This package, this unity, may not be present in ICL contexts. These functions may be disaggregated over different entities, or they may be duplicated in more than one entity purporting to exercise them in different ways.
When we lose the single (comparatively) tidy package, we lose some of the devices and solutions that have been used in discussing criminal law. The idea of ‘citizenship’ may not be an appropriate explanatory tool if bonds of citizenship are not present between accused, victims, other states asserting authority, and international tribunals.Footnote 105 The tool of ‘community’ is considerably complicated in ICL. The concept of ‘state authority’ is less straightforward where the entity applying ICL is not the state exercising the authority in question; one must ask to what extent ICL should accommodate state authority and why. When these roles and attributes are disaggregated instead of bundled, we have both the burden and the opportunity of isolating the significance of those different roles and attributes for criminal law.
At a quite literal level, ICL is routinely carried out not by states but by international tribunals, providing a direct, unmediated application of criminal law to persons. One could respond that this is still in some sense criminal law by the ‘state’, because an international tribunal acts like a state insofar as it applies criminal law.Footnote 106 A long tradition of thought requires a ‘state’ for criminal law.Footnote 107 But perhaps what is really signified is something broader, such as ‘governance’ or ‘public authority’, and the institution of the Westphalian state is actually just the most familiar species of a broader genus. ICL provides the opportunity to explore criminal law under alternative forms of governance, and, in so doing, to learn more about the more truly general case of criminal law.
It would be sound to say that thinking about ‘states’ is still relevant, in that we may certainly turn to rich and well-developed thinking in the context of states as a ‘reservoir’ of ideas for governance under international institutions.Footnote 108 But my point of caution is that we draw from that reservoir of ideas with care, so that our net does not include the accumulated detritus that is particular to states but not to other mechanisms of governance. Some features of a state may be needed for criminal law, others may not, and others may require modification of our thinking. We shouldn't try to equate tribunals with states and chastise them for not having the features of a state. For example, ICL does not feature a legislature per se. We must distinguish the rules that have grown around particular contingent features from what is essential in order to respect human agency.
6. Implications
The foregoing sections provided a quick and abstract sketch of the proposed account. I will provide three examples to foreshadow how the account may provide a new way of looking at concrete problems in ICL. I am not attempting to stake out a conclusion on any of these issues. I am merely demonstrating how the account can pose new questions about familiar problems.
6.1. Legality and lex scripta
Scholars such as George Fletcher have argued that, in order to comply with liberal principles, ICL must comply with the principle nullum crimen sine lege scripta: no crime without written law. Accordingly, he argues that the reliance on customary international law as a source of prohibitions is an error produced by habits of international lawyers and reflects a failure to understand the full implications of legality in criminal cases.Footnote 109 I certainly agree that some problems in ICL flow from habits and thought patterns of international law.Footnote 110 This may, however, be an instance where the seeming departure is one that, on further inspection, may prove to be justifiable.
The requirement of lex scripta (written law) has emerged as a very plausible articulation of a fundamental principle in the context of the modern state, which features a separation of powers (and thus a legislature) and heterogeneous societies subject to voluminous regulation. In almost all of contemporary human experience with the criminal sanction, it seems quite plausible to say that lex scripta is a precondition for just punishment. However, it is studying the exception that may teach us the most about the rule. Lex scripta may simply be a technique to satisfy a more elementary requirement. If so, it may emerge as a requirement of justice only under certain conditions.
I think we could readily come up with examples in which just treatment would not require written law. We could imagine, for example, a small society trapped on a single island developing a system to enforce a few basic prohibitions. Our concept of fair warning would likely not require written prohibitions in that situation, if prohibitions were otherwise known or ascertainable. Note that I am not suggesting that ICL is analogous to the island situation. I am using the island situation to demonstrate that the lex scripta requirement admits of exceptions and must therefore be contextually contingent. It is not an elementary requirement but rather a technique. ICL, a legal system without a legislature, which of necessity largely relied on customary law for its basic rules, and which has recently and rapidly transited from an embryonic to a relatively mature system, provides a wonderful setting to try to explore the parameters of the lex scripta requirement. Under what circumstances does our concept of justice require written law and why? If writing is required, does it have to be in one place (a code), or can the writings be scattered in multiple places, as it is in common-law jurisprudence or most of customary international law?
ICL provides further opportunities to explore legality. ICL frequently confronts massive evils in situations where positive law is lacking or laconic. The stakes in cases of such lacunae are often higher than in ordinary criminal law, because the crimes are generally far more horrific and positing new law (for example by multilateral treaty) is more difficult. ICL jurists have developed various strategies of argument to justify punishment in such cases. I expect that some such punishment can indeed be justified. However, each of the major argumentative strategies employed to date has shortcomings.
For example, it is often argued that the prohibition on retroactive law was not applicable in ICL, at least in its early stages.Footnote 111 The argument may indeed be right as a matter of positive law, but it doesn't help us at all with our question whether retroactivity is normatively acceptable, i.e. whether it is ‘just’. Another popular argument is that the ‘formal justice’ enshrined in the principle of legality must give way to the ‘substantive justice’ of not letting persons escape punishment for heinous deeds.Footnote 112 The argument is appealing but incomplete, because it has a purely formal structure: it does not specify any content for its exception. How do we know that the prohibition we wish to impose falls within this concept of ‘substantive justice’? There must be some ascertainable limits to the set of norms that purportedly displace the requirement of ‘formal justice’. So this form of argument does not get us far either; we are thrown back into the search for some method or source to delineate the punishable prohibitions.
As may be seen, the account I propose would not transplant the familiar requirement of written legislation (or its international analogue, a treaty). Nor would it indulge the arguments that would circumvent legality and fair warning altogether.Footnote 113 It uses these hard cases to try to isolate what form of prohibition or warning is truly needed before we can prosecute a person.
6.2. Duress and social role
A cosmopolitan liberal account can also generate new questions in assessing the defence of duress in the peculiar contexts raised by ICL. For example, in the Erdemović case, the accused had enlisted in a non-combat unit of the army.Footnote 114 One day his unit was sent to a farm, where they were informed that they were to shoot Muslim civilians. He protested the order, and was presented with a choice of either participating or joining the prisoners and being shot along with them. Faced with the unappealing alternative of sacrificing his life while saving no lives, Erdemović complied. The majority of the ICTY Appeals Chamber adopted a rule that duress is not a defence to the killing of civilians, following the lead of many common-law jurisdictions.Footnote 115
ICL scholars within the liberal tradition have widely and understandably criticized the majority decision for its insensitivity to fundamental principles and the role of moral choice in culpability, given that the only way for Erdemović to be innocent was to be dead.Footnote 116 In a previous work, I criticized the reasoning employed by the majority;Footnote 117 however, as I emphasized, this did not necessarily exhaust my analysis of the outcome.Footnote 118 While I agree with the liberal critiques, there is at least room for further analysis if we take into account the social dimension of human experience.
In theories that conceive of atomistic individuals entering into a notional social contract to better advance their personal aims, the freedom to preserve one's own life is the ultimate domain reservé. Footnote 119 In such a model, a law that requires one to die in order to avoid censure is close to nonsensical.Footnote 120 As Paul Kahn has noted, traditional contractarian liberal theories have trouble grappling with sacrifice (both the individual's willingness to sacrifice and the state or community's claim to expect sacrifice).Footnote 121 If we acknowledge the richly social world of human beings, as discussed in section 4, our analysis might change. Social roles can change the expectations placed upon us. A person assuming the role of ‘soldier’ is expected to carry out dangerous and life-imperilling acts, including, for example, charging a machine gun nest if ordered to do so. Experience suggests that many humans can conceive of fates worse than death (hence the phrase ‘death before dishonour’ or indeed, ‘a fate worse than death’). Thus, punishment for a refusal to fulfil an almost certainly lethal duty is not necessarily an absurdity. Criminal laws may punish soldiers for desertion or insubordination or cowardice in the face of the enemy.
Perhaps duress is based on the expectations of firmness that we can fairly expect from members of society.Footnote 122 Normally criminal law would not and could not demand heroism. It is possible that we can justly impose higher expectations on persons assuming the role of soldier. Just as we hold soldiers liable for desertion in the face of the enemy, perhaps we could hold them to a similar standard concerning their duty not to fire on civilians. Although the reasoning of the majority decision in Erdemović may be faulted for inadequate engagement with deontological considerations, it is at least conceivable that the conclusion reached might be justified, insofar as it was restricted to soldiers,Footnote 123 using an account that considers these social dimensions of human experience. To be clear, I am not here advancing a conclusion one way or another. I am merely highlighting the type of question that a more careful liberal account can raise.Footnote 124
6.3. Superior orders and authority
Finally, the controvertedFootnote 125 ICL defence of ‘superior orders’ is another example that may pose puzzles for criminal-law theory. It does not immediately fit into the category of justification or excuse; it does not necessarily preclude wrongfulness, nor does it necessarily preclude individual blameworthiness. It does not fit with common rationales such as that the actor averted a greater harm or that the actor had diminished capacity.Footnote 126 There are clear affinities between ICL's partial accommodation of superior orders and the concept of ‘state authority’. Malcolm Thorburn has advanced an intriguing explanation in the context of justifications that highlights the authorization by the state, which looks into the role of the agent, the agent's reasons for acting and the relevant scope of discretion.Footnote 127 The concept of ‘state authority’ is very likely part of the apparatus needed to explain the superior orders doctrine.Footnote 128
Nonetheless, ICL contexts also complicate the relationship between state authority and criminal law. In a ‘normal’ context, a single state is both the applier of criminal law and the authorizer of the act. It therefore seems more obvious why the state would grant deference to its own authorized acts in its criminal law. In ICL contexts, there may not be a unity of identity between the applier of criminal law and the authorizer of the act. Furthermore, there may be multiple authority structures asserting authority in conflicting ways: the state of nationality of an official may be authorizing and ordering an act, the territorial state may be forbidding the act, and an international tribunal with jurisdiction may also proscribe conduct while accommodating assertions of authority within circumscribed boundaries. In these messier contexts, we are compelled to ask additional questions: why should criminal law accommodate authority? What is the proper scope for that accommodation?
7. Conclusion
In this article, I have outlined a new framework for evaluating and developing ICL. Reasoning in ICL jurisprudence and literature has often adopted a formalistic approach (reasoning from sources) coupled with fairly simple teleological arguments (aimed at maximizing efficacy), without sufficiently engaging with fundamental principles as constraints of a system of justice. This tendency gave raise to the liberal critique of ICL, which in turn gave rise to a critique of the liberal approach.
My argument is that it is possible to agree both with the liberal approach and with the critiques of the liberal approach, provided that some caveats are made to each. ICL should not uncritically replicate principles from national systems, nor should it uncritically abandon them. To constitute a system of justice, it must grapple with legality and culpability. We can draw inspiration from national approaches as we discuss possible adaptations to uphold the deontic commitment in the special contexts faced by ICL.
I use the term ‘cosmopolitan’ liberal account. Cosmopolitanism and liberalism are both humanistic ideas rooted in respect for individuals; the ‘liberal’ aspect emphasizes respect for the moral agency and worth of the individual, and the ‘cosmopolitan’ aspect emphasizes that this respect transcends state boundaries. The idea of the state is merely one useful device for human governance. Hence the approach re-examines thinking about criminal law under the Westphalian state and considers criminal law under other authority structures. It aims not to impose ‘Western’ values but to draw inspiration from the world's legal systems. However, rather than simply mimicking the approaches of national legal systems, it engages seriously with the contexts faced by ICL – including collective action, inverted morality, and causally overdetermined crimes – in order to identify individual desert in such contexts. The approach points to some new questions about ICL controversies, and may help build a system that is both effective and just.
In addition to shedding light on the interpretation and evaluation of ICL, this line of thought also has exciting implications for general criminal-law theory. The study of abnormal situations can help us discern conditions and parameters embedded in what we thought, based on our everyday experience, to be elementary principles. Doing so helps us to develop a theory that is truly more ‘general’. Standard criminal-law theory is generally predicated on a ‘normal’ situation, in which a functioning state adopts laws (through a legislature) and applies those laws to persons within its jurisdiction, and in which there is relative stability, so that criminal conduct is generally deviant from societal norms and the state itself is not criminal. ICL – law in extremis – presents new and extreme problems that may lead us to discern subtleties in notions such as fair warning, choice, culpability, causation, or authority that we had not previously considered. ICL problems may help us to discover that formulations of principles that seem foundational are contextually contingent manifestations of a deeper deontic commitment.
