Beyond cosmopolitanism: towards a non-ideal account of transnational justice

Ethics & Global Politics

Beyond cosmopolitanism: towards a non-ideal account of transnational justice

Christine  Chwaszcza*

Chair of Social and Political Philosophy,European University Institute,Florence,Italy


Cosmopolitanism in normative theory of transnational justice is often characterized by the thesis that the moral and legal status of states must be entirely derived from the moral status of the individuals who constitute them. Although the thesis itself is rather indeterminate in substantive and analytical content, it is generally understood as the claim that states should not be granted the status of moral and legal agents sui generis. This article argues that such a view is analytically and methodologically misleading, and that any fruitful approach towards a liberal theory of transnational justice must face the challenge of coming up with a more complex concept of statehood, and acknowledge that in international relations and international law states are collective moral agents in their own right that can be addressees of genuinely collective forms of responsibility. The argument starts with a critical examinations of two common interpretations of the cosmopolitan thesis, a reductivist reading, which suggests that we can reduce the moral and legal status of states to the rights and duties of individuals (section I), and a methodological reading, which suggests that the moral status of individuals must based on the acknowledgment of ‘universal’ individual rights (section II). For different reasons, both readings are argued to fail. Section III then presents an outline of how to conceive of states as agents that possess moral and legal status sui generis and are addressees of collective responsibility.

Keywords: cosmopolitanism; statism; ethical individualism; methodological individualism; collective agents; collective responsibility

*Correspondence to: Christine Chwaszcza, Department of Social and Political Science, European University Institute,Badia Fiesolana, San Domenico di Fiesole (Firenze),Italy,

Citation: Ethics & Global Politics, Vol. 1, No. 3, 2008, pp 115–138. DOI: 10.3402/egp.v1i1.1859
© 2008 . CChwaszcza. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (, permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Charles Beitz has characterized cosmopolitanism as the thesis that the moral and legal status of states must be entirely derived from the moral status of the individuals who constitute them.1 From the cosmopolitan moral point of view, socio-political institutions are mere instruments for the realization of justice, but they do not constitute ‘entities’ with a moral standing of their own.2 Accordingly, the assignment of normative competences and responsibilities to states in international law and international relations must derive entirely from the rights and obligations that individual persons owe one another. To the extent that political associations—communities, nations, or societies—are considered to differ from states, this reducibility thesis applies equally to these collectives.

Although cosmopolitans rarely argue for the reducibility thesis or explain its meaning more precisely, they seem to take for granted that it derives from a liberal commitment to ‘ethical individualism’ or articulates an unproblematic expression of it. But no such claim is immediately obvious, and in fact the substantive point of the cosmopolitan thesis is not unambiguous. Sometimes the cosmopolitan thesis is interpreted in a reductivist sense that is best understood as an expression of methodological individualism. Other interpretations take it to be equivalent either to the idea that certain specific individual rights have universal scope in the sense of addressing ‘mankind,’ or to a general metaethical commitment to a rights-based concept of justice. I will consider both trends of interpretation, because they are in a non-trivial sense interconnected, and will argue that the cosmopolitan thesis is misleading in both of the suggested readings, either for analytical reasons or for methodological reasons.3

Any liberal approach towards problems of transnational justice must assign a moral and legal status to states—and equally to political associations—that cannot be reduced to the rights and duties of individual persons. Some readers might think that such a stance requires one to give up ethical individualism, but it does not. As I understand it, ethical individualism is a metaethical principle that guides normative justification by insisting that legal institutions and socio-political practices are legitimate only to the extent to which they can be considered to be justifiable from the perspective of the individuals who constitute them or are affected by them or both. The reductivist interpretation of the cosmopolitan thesis, by contrast, expresses an ontological claim because it insists that all normative attributes of states ought to be reducible to the normative attributes of individual persons. Such a claim does not follow from a commitment to ethical individualism—at least not without further argument—because it can easily be argued that rational individuals would consent to investing institutions with legal competences and moral responsibility that go far beyond those that can be assigned to individual persons. As a matter of fact, I think few cosmopolitans would deny that.

The problem of the cosmopolitan thesis does not concern value-claims, but rather the appropriate structure of ethical reflection and the question of who are the addressees of requirements of transnational justice.4 The most appropriate answer to that latter question is certainly that primary addressees are states and political associations considered as institutional and collective agents, which is to say that neither individual persons nor a global sovereign are (primary) addressees of transnational justice. If that is the case, however, no plausible account of transnational justice can avoid coming up with an account of how we can conceive of the moral agency of states, and of how moral responsibility can be assigned to institutional agents, such as states and political associations.

Unfortunately, the role of states as moral agents has hardly been explored so far.5 The present article proposes an account of how that exploration can be done, mainly by tackling some assumed theoretical obstacles that are usually invoked against it. By far the main theoretical obstacle derives from debates about agency. Agency is often defined in terms of specific mental or psychological capacities of agents. That suggests that conceiving of states as agents requires us to conceive of them either as forms of super-organisms6 or as entities that possess a ‘common mind.’7 As my analysis will show, neither assumption is required (or even sensible) for considerations concerning the moral and legal status of individual persons as well as institutional actors. Since the focus of inquiry concerns problems of the analysis of moral agency and the assignment of moral responsibility to institutions such as states and political associations, I will suppose throughout that liberal values can be justified and that they apply to problems of transnational justice.

My argument will proceed in three steps. I will first take the cosmopolitan thesis at face value, that is, as arguing that all moral properties of states can be reduced to moral properties of individuals, and I will show why such a reductivist thesis is analytically problematic for reasons connected to the philosophy of meaning and action. Second, I will turn to a more common reading of the cosmopolitan thesis, one that understands it as a methodological commitment to the traditional liberal structure of rights-based argument, and I will argue that ethical reflection on transnational justice requires a quite different ethical framework that must be based on a theory of the moral agency of states. The final section will then take a new start, by showing how we can conceive of states and political associations as collective agents and subjects of collective responsibility in a way that is fully compatible with ethical individualism, and that can serve as the basis for an agency-oriented approach to transnational justice.


If the cosmopolitan thesis is supposed to express a morally substantive claim, we must assume that we can somehow identify the moral and legal status of individuals independently of the specific institutional structure of the socio-political environment. I will call this the assumption of context independency, and I will argue that it is highly problematic. What is meant by context independency is best explained by giving some examples of what I mean by context. Consider the following propositions that represent ordinary ways of talking about states:

1. ‘The USA signed a trade agreement with Mexico.’

2. ‘Americans benefited more from the agreement than Mexicans.’

3. ‘In American democracy, the Supreme Court acts partly as a political legislator.’

4. ‘The people of the USA elected a new president.’

5. ‘Hobbes argued that all sovereign political power derives from the consent of those subjected to it.’

Each proposition obviously is about states, but each addresses a different socio-political phenomenon: (1) refers to the USA and Mexico as legal persons in international law; (2) to the population of the USA and Mexico as a collective of natural persons; (3) is about the institutional order of the USA; (4) addresses a particular part of the overall population in their role as citizens of the USA; (5), finally, represents an abstract view of the structure of the socio-political order.

The five different kinds of reference do not offer an exhaustive classification of institutional contexts, but they should suffice to illustrate an important truth about political and social ontology: socio-political institutions are to a significant extent constituted by social conventions and legal practices. They, therefore, cannot be said to ‘identical with’ or ‘nothing but’ the individual persons who constitute them, because the members of a group of individual persons can be related to one another in a variety of different socio-political and legal ways. We certainly do not refer to a different group of natural persons when we ask of the first statement, ‘Who are the individuals who constitute the United States?’ or of the second, ‘Who are the individuals who constitute the “Americans”?’ But it would be wrong to identify the status of states as legal persons in international law with collectives of natural persons in (2) or even with the rights and duties of citizens as in (4).

Why is socio-political ontology relevant to the cosmopolitan thesis? The answer is that differences in institutional perspective play a constitutive role in the specification of the actions, rights, and duties of the individual persons who constitute the socio-political collective. In other words, the substantive content of the rights and duties of individuals cannot be specified independently of the socio-political context in which those individuals act.8 The point is most obvious with respect to civil and political rights and duties: citizens do not have a right to cast their vote because they are individuals, but qua being members of a particular political association. The case might be less obvious with other rights, but civil and political rights do not constitute an exceptional case or class of normative rights in that respect. Basic human rights, such as the right of detainees to be brought before a judge within 24 hours, are similarly context dependent. Almost all rights and duties that are traditionally assigned to individual persons—be they natural persons, citizens, or office holders—receive their ethical and moral or legal significance against the background of institutions within which individuals act and interact.9 Since individual agency in socio-political contexts is overwhelmingly ‘conventional,’ in the sense of being founded in legal and socio-political conventions and general practices, the competences, rights, and duties of individual agents cannot be specified without reference to those conventions and practices that constitute the socio-political environment. For that reason, the identification of individual rights and duties cannot be detached from the socio-political background institutions that constitute the context of analysis.

This problem has been discussed at length in debates about ontological and semantic individualism, and I would like to turn briefly to that discussion, because it is instructive for the analysis of the concept of individual rights and duties. The general problem of context dependency is related to the identification and individuation of actions. Most actions—especially social actions—receive their significance against the background of certain institutional structures. To understand, for example, what goes on when Mr. Smith cashes a check at the bank, the relevant pieces of behavior cannot be reduced to an observation of Mr. Smith's entering a building and exchanging words and paper slips with another individual who sits behind a counter. No meaningful individuation of Mr. Smith's action and no meaningful description of his behavior is possible outside of or beyond the background of the specific institutional contexts of monetary currency, banking systems, and so on. Defenders of methodological individualism would insist that we can in principle redescribe the relevant financial and economic institutions by explaining them as results or effects of the actions and dispositions of the individual persons who are involved in bringing those complex structures about. Notoriously, however, it is far from obvious how we can do that without referring to those very institutions, because the relevant actions and dispositions can be said to be meaningful only against the background of those institutions. Social ontology, it seems, is tightly connected with the philosophy of meaning.

We need, therefore, to be careful with regard to the conclusions or inferences to be drawn from insights that are as general and abstract as that of context dependency. For it is indeed a quite ubiquitous phenomenon in the analysis of social action. As Anscombe reminds us in her article ‘Brute Facts,’ the analysis of social behavior must—almost always—assume the existence of some background institutions of some kind. In her example, ‘I owe the grocer one shilling because he supplied me with potatoes,’ context dependency is not only crucial for ascribing a (moral) duty to Anscombe to pay her grocer one shilling. Also the description of the grocer's behavior as ‘supplying her with potatoes’ is context dependent too, because it (at least implicitly) assumes the existence of a system of economic exchange and services: the grocer did not just dump dirt in her kitchen or leave a present for Anscombe.10 An individuation and description of social action will almost unavoidably be embedded in cultural practices and conventions of various kinds: moral, economic, cultural, legal, and so on. Yet, neither our understanding of a person's social behavior nor our explanation of it will necessarily become more precise or accurate if we explicitly describe and rearticulate the full institutional background. We might gain deeper insight in some cases by doing so—for example, if we study a foreign culture or a very complex socio-political institution such as economics. But in other cases, especially in cases of familiar behavior, the task might turn out to go beyond what we in fact do understand, because we often understand contextual actions, such as ‘cashing a check,’ much better than their complex socio-political institutional background, in this case, banking systems and monetary currency.

The relevant insight to be gained from the ubiquity of context dependency, as I see it, is that we lack a clear contrast between the supposed individualistic and holistic aspects of social agency. To be cautious, that conclusion might not hold for all instantiations of human behavior, but nevertheless for a rather huge sphere of social agency it must be said that we do not know how to give a meaningful description of people's behavior that is independent of the institutional context within which it is exhibited. That is especially true for actions that instantiate general practices. Although personally, I do not think that there is a significant difference between moral and non-amoral practices, the point has been primarily discussed for moral practices, such as making a promise and other forms of putting oneself under an obligation. Criticizing an argument of Searle's, that the fact of making a promise allows us to infer that promisor ought to do what he promised, and thereby to derive an ‘ought’ from an ‘is,’ John Mackie has suggested that we can take two different stances towards the act and what he calls the ‘institution’ of promising, or what I would prefer to call a ‘general practice’: One stance from ‘inside the institution,’ from which making a promise indeed generates an obligation to do as promised for those who participate in that institution, and one stance from ‘outside the institution,’ from which the promissory statement ‘I promise’ is a mere fact with no normative consequences.11

But what kind of fact could the utterance ‘I promise’ or ‘I give you my word’ be if considered from the outside perspective? What could the phrase mean? I am inclined to say that from the outside perspective it must be meaningless. It would be no more meaningful than exchanging slips of paper outside monetary and financial institutions. Referring to a parallel between moral institutions and games such as chess that Mackie himself draws upon, I have no more clue what ‘I promise’ might mean outside the institutions of promising than I have about the non-metaphorical meaning of uttering ‘checkmate’ outside the game of chess.12 To say that certain speech acts are meaningful only to the extent that certain institutions exist, therefore, must not be understood as indicating that we can separate the institution of promising from the relevant speech act or can say that one aspect is logical or temporarily prior to the other, but rather merely as indicating that uttering the relevant phrase (performing a speech act) is a particular instantiation of a general social or moral practice. That, of course, is to say that it is not a ‘brute fact.’ Returning to Anscombe's point, there are rather few ‘brute’ facts in social and moral agency, or at least very few facts that are of philosophical—or even cultural—significance.

The crucial insight to be drawn from the discussion of context dependency is this: context–dependent description of actions cannot be meaningfully separated from the background institutions in which they occur. That, however, implies that we do not adequately conceive of actions if we abstract from their institutional context. I will, therefore, call them ‘conventional’ in order to contrast them with ‘brute facts.’

It may well be the case that context dependency is not relevant for all forms of human behavior—scratching one's back (in the literal sense) or turning the light on are certainly context independent—but most instantiations of social agency receive their specific meaning only against some form of background institution. Furthermore, the substantive contents of moral and legal rights and duties are very unlikely candidates for context-independent descriptions. For all context-dependent action descriptions, however, we lack a clear idea of how we can make sense of what a ‘reduction’ of the description or explanation might look like, especially when the relevant background institutions play a constitutive role for personal actions and interpersonal interaction. Accordingly, the theoretical insight that I think has to be drawn from the ubiquity of context dependency is that, given the semantic texture of social agency, we lack a well-defined concept of reduction that would allow us to separate a substantive description of agency from the socio-political background institutions that constitute the context in which agents act.

This brings us to the initial problem with the cosmopolitan thesis. As I mentioned at the beginning of the section, if the cosmopolitan thesis is to make a substantive point, it must be possible to give an informative account of the rights and duties of individuals that is independent of the institutional background of the state. The cosmopolitan thesis therefore faces a double challenge. First, the substantive content of many individual rights and duties makes sense (only) against the background institutions of the domestic perspective of the state, although these background institutions are rarely explicitly addressed, but rather constitute an implicit framework. That is not to say, of course, that individuals do not have any moral and legal standing in international law and international relations at all, but is instead to say that that standing is most likely different from their domestic rights and duties. Second, cosmopolitan references to ‘the state’ can refer to specific institutional contexts—as they are expressed, for example, in propositions (3) and (5) above—that are quite different from the institutional structure that prevails in international relations as in (1).

Postponing the second challenge to the next section, I would like to conclude with regard to the first that the specification of individual rights and duties in the context of transnational justice still has to be determined.13 To avoid a likely misunderstanding: None of what has been said so far excludes the possibility of ethical norms of or ethical behavior in international relations. The point is rather to show why ethical analysis cannot proceed as if the fact that the international order is constituted by states makes no normative difference for ethical analysis and ethical argument. That notwithstanding, there are certainly some familiar background institutions that we can take for granted not only in domestic but also in international law. We assume, for example, that individual property rights are in some sense global and that travelers do not give up their property rights to their car or their wallet whenever they cross a border. Similarly, we certainly hold that persons ought not to be killed or maimed for arbitrary reasons or enslaved, independently of whether they are citizens or foreigners. In addition, principles such as pacta sunt servanda, or that promises ought to be kept, do not require a specific institutional background beyond the existence of some form of functioning legal and moral order. Other rights and duties of individuals, however, are more closely tied to the framework of state institutions and political associations. Those rights and duties cannot be taken for granted, because international relations and international law do not constitute a global state. That raises the problem that we lack a clear understanding of what the rights and duties of individuals are (or ought to be) outside the socio-political background institution of the liberal paradigm of the (national) legal state. Accordingly, it is doubtful that the cosmopolitan thesis provides us with a substantively informative guideline for moral reflection on problems of transnational justice.


I will now turn to the second challenge. The cosmopolitan thesis can also be understood in a different sense: as a general commitment to a rights-based theory. But even then, and even if we assume that the problem of determining the substance of individual rights outside state institutions has been solved, the cosmopolitan thesis still remains problematic for methodological reasons.

As John Rawls has correctly pointed out, liberal theory considers justice to be the first virtue of institutions.14 The institutional background of liberal political theory can be specified more precisely as the institutional framework of the liberal model of the (democratic) legal state, as expressed, for instance, in propositions (3) and (5) at the beginning of section I. Indeed, the cosmopolitan concept of ‘the state’ or references to ‘states’ are best understood as referring to exactly that model. The liberal model of the legal state, however, is essentially an abstract ideal type construed for a specific purpose, that is, the justification of legitimate constraints of political government. It addresses problems concerning the normative structure of constitutional, public, and private law15 and is tightly connected to the development of what is sometimes called the ‘rule of law’ (Rechtsstaatlichkeit). But it leaves other questions out, most notably the question, ‘Who is the addressee of requirements of justice?’

Such a question does not arise within the liberal model of the legal state, for the trivial reason that the model itself is meant to answer it. The liberal model of the legal state is an institutional organization of agencies and offices that are invested with all the relevant competence and responsibility to facilitate justice, peace, welfare, and social and political life. For that reason, liberal normative theory primarily aims at determining the normative limits of the task and scope of those agencies and of devising reliable and suitable measures of control or checks and balances between them. It is indeed a characteristic feature of the liberal model of the legal state that it invests institutional offices and agencies with rights and duties that go far beyond what individual persons can be expected to perform, and grants to institutions competences and powers that individuals could not be entrusted with. We do not expect ordinary citizens to risk their lives, for example, in order to prevent other persons from committing a crime. Police officers, by contrast, are expected to do so, because it is ‘their job.’ Although ordinary citizens are still obliged to support the police by reporting to them, testifying, and supporting their investigations, they are not obliged to face serious risks to their lives and health, nor is it considered to be their genuine task or responsibility to prevent crimes.16 It suffices that they do not commit them. The idea, to put it in a nutshell, that institutions are not only more efficient but can also be organized in a way that makes them much more reliable than individual persons articulates the very legacy of liberal theory. It marks the shift from political philosophy as a branch of virtue ethics to the liberal institutional paradigm. The liberal model of the legal state articulates in a crucial sense an ‘impersonal’ structure, and more importantly a structural arrangement of powers and competences that is conceived to be more reliable and trustworthy than natural persons. The question of ‘who is the addressee?’ of liberal requirements of political justice, therefore, does not arise.

The institutional structure of international law and international relations, however, does not resemble the liberal model of the legal state, and that again brings the question back to the foreground, ‘Who is the addressee of requirements of transnational justice?’ The only practically convincing answer to that question is that primary addressees are ‘states,’ but this time the reference is to states conceived of as legal persons in international law and international relations, as in proposition (1) at the beginning of section I, which is quite different—not least in normative respects—from the liberal model of the legal state in propositions (3) and (5). Given the considerations concerning context dependency in section I, how states are conceived of makes a difference with regard to what they are and what they can be asked to do.

The problem of the cosmopolitan thesis, understood as a methodological commitment to the liberal tradition of rights-based normative theory, can now be restated: it implicitly takes for granted (or pretends) that the liberal model of the constitutional legal state articulates an apt framework for the discussion of transnational justice. Implicitly, the thesis assumes that the specification of transnational rights and duties can be based on the assumption that international relations can be analyzed as if they resemble the liberal model of the legal state. Restated, the substantive point of the cosmopolitan thesis is a methodological one that concerns the structure of normative argument and a metaethical framework for normative reflection. The tendency implicitly to refer to the model of the (constitutional) legal state will be called ‘methodological statism.’

Since I will argue that the institutional structure of international law and international relations is not just morally imperfect or less ideal when compared with the ideal type of the legal state, but is instead genuinely different, and that this raises entirely different questions from those that traditional liberal theory addresses, I would like to emphasize briefly some of the most important structural differences. Notoriously, the general order of international relations and international law is decentralized and lacks reliable mechanisms of norm enforcement and other structures of the institutionalized division of labor. As a matter of fact, nation states are the authors, addressees, and guardians of international law at the same time. That has the effect of making international law more often motivated by political than by normative reasons. International law and the organization of international relations also lacks an institutionalized division of moral labor and institutionalized mechanisms of control, redress, and checks and balances. As a consequence, not only norm-conforming behavior but also the willingness and capacity for norm enforcement lie entirely on the shoulders of states, which can turn out to be a rather costly enterprise in financial as well as moral terms for those who engage in it. Even if we can determine the substantive content of individual rights in the context of international relations, it will not follow that states have an unqualified obligation or responsibility to guarantee them. Such an assumption can be made within the institutional context of the liberal model of the legal state because its organization guarantees the existence of reliable and centralized mechanisms of norm enforcement. But in a decentralized system of norm enforcement, agents can at best be obliged not to violate rights themselves. It does not follow automatically that they can also be required to prevent or stop others from violating the rights of third parties—at least not under all conditions.

Although the existence of international organizations such as the United Nations (UN) and regulatory regimes such as the World Trade Organization (WTO) would seem to prove that there exists some degree of institutionalization, Chris Brown rightly points out that the international system does not form a ‘society’—not even in a metaphorical sense.17 Relevant international organizations such as the UN and institutions such as the UN's Security Council are aggregations of member states rather than integrated systems. Far from exerting influence or power over their constitutive members, the constitutive member states of these institutions exert power through those organizations—as much power as they want and according to the terms they think best, which can change from one decision, crisis, or situation to the next.18 Obviously, that does not imply that any single member state can dominate international organizations, although some states can effectively block their functioning, but it is meant to imply that the decisions and measures undertaken by international organizations do not represent a common good or shared purpose, but instead reflect the national interests of their member states and the power relations that hold between them. Organizations and regimes, such as the European Union or the WTO, which have developed reliable mechanism of norm enforcement, are strictly limited with respect to their agenda and focus primarily on regulation of economic policies.19

In addition, in contrast to domestic legal systems, international law lacks an agreed-upon body of authorized legislators, a developed system of jurisdiction, procedures for redress, and a system for the control of powers. Besides treaties, the major source of norm generation is still customary international law, that is to say, de facto state practice. As a consequence, not all norms of international law should be accepted uncritically as morally desirable, and not all breaches of international law are morally repulsive, either. Although it seems true that international courts of justice are slowly but increasingly becoming an additional source of international law, especially under the umbrella international human rights regime, that development itself remains contested, cannot be enforced in practice, and occurs so rarely that it cannot be considered a reliable factor.

To summarize, the institutional environment of international relations and international law is not just less ideal than the liberal model of the legal state, it poses a completely different challenge, because it lacks most—if not all—of the constitutive mechanisms that allow us to shift the burden of morals from the shoulders of agents to the mechanisms of an institutional structure. Given the structural differences, ethical analysis requires not only an explicit specification of the addressees of requirements of transnational justice but also, more precisely, a specification of what can be required from whom and under the prevailing conditions. Those conditions are, among other things, characterized by the absence of centrally organized mechanisms of norm enforce and institutional division of labor. The cosmopolitan reliance on methodological statism is therefore unwarranted.

Such a reliance is implicitly deceptive, because it suggests that ethical reflection of transnational justice can be pursued in analogy to domestic justice, and this assumption is methodologically misleading in a crucial respect. Whereas the liberal model of the legal state can employ the ideal assumption that all ethically relevant agents are willing to obey moral and legal norms—for whatever reasons—such an idealization would misrepresent a decisive structural feature of the institutional order of international relations. The possibility of full compliance, however, can be argued to articulate a crucial assumption in support of the principle that all ethically relevant agents ought to be treated equally, because it allows us too disregard the problems and challenges that arise from deviant behavior.20 If norm-conforming behavior can be generally expected (or ensured), the challenge of normative arguments can focus on the justification of (the compatibility of) general norms for the regulation of conduct and equal rights of moral and legal subjects. If, however, norm-conforming behavior cannot be assumed to express a structurally appropriate idealization, ethical reasoning becomes much more complex. It must, first, distinguish between compliant and non-compliant agents; second, differentiate between principles of conduct towards compliant and non-compliant agents; and, third, must be crucially concerned with the specification of moral duties in an imperfectly moral environment. Such considerations require quite different forms of deliberation and normative argument than the justification of general norms.

Given the structure of international relations, no theory of transnational justice can avoid considering states as the primary addressees of requirements of transnational justice, because we certainly cannot require from individuals qua natural persons that they counteract injustice committed by foreign governments or states. Addressing states, however, requires an account of statehood, or, rather, of the moral agency of states, that goes beyond the liberal model of the legal state and considers states as institutional agents in international law and international relations. To be sure, such an account cannot abstract entirely from either internal conditions of legitimacy for states considered as organizations for political association, nor can it entirely abstract from the role of states as legal persons in international law—neither of which, obviously, can be reduced to the other nor to the rights and duties of individual persons—, and it cannot even rely merely on ‘abstract’ or ‘ideal type’-like models. Rather, it must show how the various faces—abstract and non-abstract—of statehood can be combined—and, if the theory is supposed to be liberal theory, combined in a way that is compatible with the principle of ethical individualism.

In the remainder of the article, I propose the first step in conducting such an analysis, beginning by eliminating some of the theoretical obstacles that have been invoked against conceiving of states—and other collectives or institutions—as agents sui generis.


One set of frequent objections to conceiving of institutions or collectives of persons as agents with a moral and legal status sui generis derives from a specific view concerning the nature of agency, personhood, and responsibility. I will call these objections ‘metaphysical,’ although they usually refer to natural persons, because they are often connected with views of human nature that resemble Immanuel Kant's concept of persons in The Metaphysics of Morals.21 The metaphysical view holds that only beings that fulfill certain mental or psychological requirements that are held to be necessary for moral reflection and moral self-correction can be conceived of as moral persons. Moral persons according to the metaphysical view are agents not only in the sense that they can be the subject of certain acts, and consequently can act correctly or incorrectly, but are subjects whose actions are understood to manifest the moral quality of the person. The systematic problem that motivates the metaphysical view concerns the foundations of judgments concerning virtues and vices—that is to say, the moral character or moral merit of persons. Accordingly, the concept of moral responsibility that is most commonly used by theorists, who take the metaphysical view as their starting point, concerns the moral autonomy of persons, i.e. the capacity to choose one's actions on the basis of moral reflection. Discussions of moral responsibility in this sense are tightly connected to early modern debates on free will and related speculations about the structure of the mind or ‘the will’ that accompany that debate.

The concepts of agency, personhood, and responsibility, however, can also be analyzed from a non-metaphysical perspective, which I will for convenience call the semantic view. The latter type of analysis is motivated by problems related to the individuation of actions—that is to say by problems of identifying and individuating pieces of behavior as actions of such-and-such a kind.22 By identifying a piece of behavior as an action, we bring it under a description that satisfies common standards of intentionality. As Anscombe has convincingly argued, by identifying behavior as ‘intentional’ we do not ‘refer’ to any form of mental or psychological state or states that accompany or cause the action, or any other kind of ‘psychological fact.’ Rather, by identifying a piece of behavior as an ‘intentional action,’ we recognize it as intelligent behavior that can be brought under a specific form of a meaningful description of what is going on. What is commonly called the ‘intention’ of the action, or the ‘intention’ of the agent in doing so-and-so, is not a psychological property or any form of observable fact, but a specific structure of the description of the agent's behavior. Whereas, Anscombe famously argued that the paradigm form of such descriptions is Aristotle's model of the practical syllogism, Michael Bratman more recently argued for a broader and less formal account of intentions that he calls the planning theory of intention.23

The view that the concept of intention—or intentionality—does not ‘refer’ to any mental or psychological states, of course, does not settle any of the questions that bother the metaphysical debates, but it prevents us from falling into two types of errors. It prevents us, firstly, from defining too hastily pseudo-empirical standards of mental activity that a subject must have in order to count as an agent, e.g. having the capacity to form desires and beliefs or to reason; and secondly, it highlights the highly conventional character that guides the individuation of actions, and that has already been pointed out in section I. The concept of agency according to the semantic view concerns the meaning of certain forms of behavior and utterances, which, for reasons emphasized in section I, is highly relevant for the ascription of liberties, rights and duties, or competences and powers to agents. Agency in the second view is not—or at least not primarily—defined by mental or psychological properties of actors, but by the conventions and general practices that define what counts as an action of such-and-such a kind. The related concept of responsibility, accordingly, concerns first and foremost standards of the conduct that can be applied not only to natural, but also legal persons, including the assignment of rights, liberties and duties, or competences and powers.

Obviously, the metaphysical and the semantic view tackle quite different types of problems that should not be confused, and that are not necessarily interrelated. Most importantly, the semantic concept of agency responds to the fact that we are able to qualify actions as such-and-such from a third person perspective, even though we usually do not have access to the mind or psychological states of other persons. That fact as such already reveals that our concept of agency is not as closely tied to the assumption that agents possess specific mental or psychological capacities, as theorists are often inclined to think.24 As the general practice of law courts reveals, the ascriptions of actions and intentions to other persons from the third person perspective relies on standardized expectations concerning human behavior, not on a verification of the fact that the agent was in a specific state of mind at the time she acted. Nevertheless, the practice is quite reliable. Notwithstanding inevitable errors, or misjudgements, or cases of complete ignorance, judgments of whether a person acted ‘mens rea,’ ‘on purpose,’ ‘voluntary,’ ‘negligiently,’ ‘inadvertently,’ ‘in ignorance,’ and so on, are highly relevant for assessments of the moral as well as the legal quality of actions—which is something different from assessing the moral merit of persons. The semantic analysis of agency, to summarize, is fully compatible with the acknowledgment that non-natural persons—or even ‘mindless’ beings25 —can be considered as agents, even though they are not moral persons in the metaphysical sense.

A common source of confusion in debates on collective agency and collective responsibility seems to derive from the fact that natural persons are moral agents in both respects. They are agents to whom we assign rights and duties, and they have the status of moral persons, whose behavior and conduct is considered the basis for judgments of moral merit, virtue and vice. Even with regard to natural persons, however, the two aspects do not always coincide. A natural person who inadvertently causes a car accident in the sequence of which another person dies, can—under certain circumstances—be said to have ‘killed’ that other person negligently (despite the fact that she did not intend the death of the person). Nevertheless, we would not therefore consider her necessarily as a wicked or mean character, even though in principle we consider her as a moral agent in the metaphysical sense. As the example shows, from the semantic point of view we hold persons responsible even for actions that they ‘did not intend’ and the consequences following from such actions.

Whereas, assessments of personal merit or character presuppose that agents have a certain rational nature or specific psychological properties that we commonly ascribe to natural persons only, the assignment of moral and legal status to persons imposes no such requirements: newborns and brain-damaged persons, for example, can be moral agents in the semantic sense and legal persons, as can companies and institutions, even though we usually find judgments regarding their character or personal merit to be misplaced.

In contrast to judgments concerning the moral character of a person and his or her personal merit, the specification of an agent's moral and legal status involves what he or she is required, permitted, or forbidden to do. As has been argued in section I, these spheres of activity are to a large extent embedded in social practices and determined by social practices and legal conventions. They are not directly related to psychological properties of natural persons and need not even presuppose their existence. Companies, for example, are common subjects of economic law, though they certainly do not qualify as bearers of psychological properties, nor are they natural entities of any kind.

Conceiving of states as institutional or collective agents is philosophically entirely unproblematic if we adopt the semantic analysis of moral agency. The fact that collectives and institutions are not natural entities and cannot act or reach decisions in a literal sense certainly excludes judgments concerning their character.26 But it does not exclude the possibility that they can be represented by natural persons or that their internal constitutions can include mechanisms or processes of deliberation and decision making or authorized structures of command. Their non-natural character therefore should not be invoked as an argument against conceiving of them as agents. Equally, it should not be invoked against assigning them a moral and legal status sui generis, because their competences and tasks differ significantly from those of natural persons. Finally, the fact that competences and tasks of non-natural actors are commonly determined by socio-political or legal conventions does not categorically distinguish them from natural actors, because the same holds true for many actions that individual persons perform, as has already been alluded to in section I.

The conventional character of non-natural agents, however, reveals, why their agency cannot be ‘reduced’ to acts and dispositions of the individuals who represent them. Borrowing an example of Fred Stoutland's,27 in most countries both the right and the competence to confer a PhD degree is a privilege of universities, not a competence or right of individual professors. Although the conferring is usually performed by a group of professors or a dean, we cannot reduce the conferring to the acts and dispositions of those individual persons considered as natural persons for the simple reason that their actions and speech acts receive their significance only against the fact that they are members of the faculty and authorized or invested by the university to examine and confer a degree. Similarly, members of parliament have their legislative powers only because of their office, and servants of the tax administration are allowed to check individuals’ private bank accounts only because of their job. If we were to exclude all non-natural entities or persons from the set of possible agents, the social and political universe would be a rather dismembered space.

The difficulty in analyzing states as moral agents, rather, consists (1) in not confusing agency understood as the assignment of competences, tasks, and responsibility with the rational or psychological requirements that must be met when we judge the character or personal merit of natural persons; (2) in not confusing the standards upon which we base our judgment about agency performance of natural persons with our assessments of the performance of non-natural agents; and finally (3) in not assuming that the analysis of institutional, corporate, or collective agents can always be meaningfully reduced or de-composed into the actions and decisions of the persons who represent them.28

With regard to the assignment of responsibility to non-natural agents, two major mistakes are particularly to be avoided: treating institutional agents and questions of collective responsibility as analogous to the agency of natural persons and their responsibility, and falling into the trap of conceiving of collective agency and collective responsibility in terms of an aggregation of individual actions and the aggregated responsibility of individual persons.29 Even beyond the distinction of personal merit on the one hand and conventional ascriptions of agency and responsibility on the other, there is no reason to assume that we ought to apply the same standards to natural and to institutional agents. Institutional agency in that sense can be understood as being determined, or at least circumscribed, by the competences and tasks that an institution is supposed to serve or perform according to legal conventions and socio-political practices that are constitutive of these agents’ establishment and their internal functioning. In the case of states as addressees of transnational justice, we must consider both their internal constitution, in accordance with liberal theory, and the competences and task of states as legal persons in international law and international relations.

That brings me to a different obstacle that has been invoked against conceiving—especially—of states and political associations as agents with a moral and legal status sui generis. I will call it the metaethical obstacle, because that is how it as been perceived—although, I think, mistakenly so. In a sense, the cosmopolitan thesis responds to a claim of Michael Walzer's that political associations have an irreducible moral quality that ought to exempt them from humanitarian intervention, regardless of conditions of internal legitimacy. In opposition to Walzer's thesis, cosmopolitans have denied not only that the moral and legal status of states (political associations) ought not to be considered entirely independently of conditions of internal legitimacy—which certainly can be defended as a requirement of ethical individualism—but that it ought to be entirely reducible to the rights and individuals who constitute them—which, as argued above, is a highly problematic thesis for analytical reasons.30

The controversy reveals quite lucidly the fruitlessness of contrasting individualism and holism in social theory and rather misrepresents the systematic problem that Walzer (not very successfully) tried to address. It is certainly right that the fact that the organization of a particular state violates the rights of its citizens does not generate a right—or even a duty—for other states to protect those whose rights are violated without further argument. The reason, however, is not that political associations are holistic entities, but the more complicated structure of ethical arguments that concern the specification of duties to counteract injustice committed by others. Although I do not think that it is difficult to provide such an argument in principle if we assume that ‘states’ have certain duties and obligations not only towards their citizens but also towards one another, and that they can be required to uphold certain standards of minimal legitimacy for subjects of international law, it is far less easy to argue for a direct link between individual rights and a transnational obligation (in the strict sense) to protect those rights in cases where that requires counteracting injustice committed by other parties than one's own. Some of the difficulties will be addressed in further detail below, but before that can be done, I first have to indicate according to what standards states can be held responsible for the transnational pursuit of justice.

For the moment, I would simply like to emphasize that no argument like the one above would be compatible with the cosmopolitan thesis either, because ascribing duties or obligations towards other states to states cannot be meaningfully reduced to individual rights and duties. Unless we accept that states in their role as legal persons in international relations can be subject to duties and obligations that could not be imposed upon individual persons, it is difficult to see how genuine forms of transnational justice can be defended at all.

It is therefore, firstly, necessary to combine internal and external perspectives on statehood. But that is easier said than done, because the two perspectives have so far been kept theoretically separate, not only for disciplinary reasons but also for pragmatic, if not moral, ones. As a matter of fact, international law no longer treats states as mere black boxes and probably did not employ such a concept of statehood even at the time when Walzer first published his ‘Just and Unjust Wars.’ Pace Walzer's original thesis, international law increasingly requires, besides the recognition of a territorially located people and efficient government, respect for at least minimal standards of domestic human rights as conditions of legitimate statehood. Nevertheless, even ‘failed states’ such as Somalia cannot easily be claimed to lack any form of statehood at all. Not only is the state of Somalia, despite its internal constitution, subject to quite a few international treaties and contracts that have not been suspended, but denying Somalia external recognition of its statehood would most likely generate a complete legal vacuum that annihilates the very object of any form of international intervention (which, of course, must not always be military but can also be diplomatic, political, economic, and so forth), because non-recognition would turn Somalia into a legal non-entity.

The two extreme positions—the reduction of external aspects of statehood to internal conditions of legitimacy, and the complete detachment of external conditions of statehood from a state's internal constitution—are equally implausible. What is required, rather, is the interconnection of the two perspectives. In some cases, that might require no more than resisting the temptation to identify the state as a form of organization of political institutions with a specific government or representatives—which should not be an insurmountable problem, as the following paragraph will indicate. But certainly in some cases (and maybe Somalia is one) we might be left with the problem that we do not know how to specify whether a state should—or ought to—still be considered a state or not (and whether it is still one state or more than one). In such a case, we may have to take whatever pragmatic stance seems most apt to improve the situation for the people. Since conditions of statehood will unavoidably make normative requirements concerning legitimacy, functioning, efficiency, and so on, dysfunctional states will always pose a problem of classification.

With respect to (more or less) well-functioning institutions, however, the second challenge—resisting the identification of states with the persons who represent them—should not cause insurmountable difficulties, because the distinction between the state as an institution and its representation through specific persons is a general feature of our common understanding. Although the different branches of government, public offices, administration, police forces, and so forth are represented by natural persons or collectives of natural persons, state agency cannot be reduced to actions and dispositions of the persons who represent the state, because as institutional representatives those persons are invested with special competences and responsibilities that they would not have independently of their institutional role. The representatives’ roles in performing those competences and tasks can be more or less specified by explicit rules or administrative codes, they can also be bound to specific processes of collective or administrative decision making, or—depending on the complexity of the relevant competences and tasks—representatives can be granted a larger or narrower range of professional or political discretion. In this sense, institutional agents cannot only be said to execute tasks or pursue goals but also to make decisions as institutional agents that can be more or less efficient, legitimate, rational, norm conforming, considerate, or responsible according to their internal constitution.

Separately from its institutional structure, we can also assess the performance of those natural persons who represent the offices and judge whether they were up to their tasks or inept, honest or corrupt, considerate or sloppy, smart or dumb, responsible or irresponsible, and so on. Obviously, the two assessments are neither entirely independent nor the same, because assessments of institutional agency and decision making rely on standards that are different from those for individual persons. A considerate person, for instance, might be outvoted in a collective decision but nonetheless bound by it. It is not always individual participants in decisions who are to blame for bad results, because rules of decision making cannot be adjusted ad hoc to particular situations, and, generally, processes of decision making can be simply inefficient for structural reasons. Even within the limits of the possible, faults cannot always be assigned to incumbent persons, and the best institutional design can be worthless if the incumbent personnel is not up to the task. State agency is in this sense not entirely independent from the performance of the natural persons who represent the institution, but neither should it be constructed as an ethics of ‘statesmanship,’ as indicated in Rawls or Buchanan and Keohane.31

A conception of states as agents as outlined above is fully compatible with the liberal idea that the primary normative purpose of socio-political institutions is the facilitation of peaceful and cooperative forms of social life, including the promotion of justice and the design and execution of policies. Normative requirements and constraints determine the scope of competences and tasks and exclude certain means and measures for achieving them. But they also allow us to integrate those requirements of state agency that result from the role of states as legal persons in international law and international relations with the internal perspective. Such an integration is not necessarily prevented by the fact the moral and legal status of states in international law and international relations cannot be reduced to the rights and duties of individual persons because states must perform tasks that go beyond the competence and authority of individual persons and even generate duties (and maybe also rights) for citizens. Again, as argued above, from the internal perspective, institutional competences and responsibilities already go beyond the rights and duties of individuals. In addition, and again as considered from the internal perspective, institutions generate rights and duties for citizens that we do not assign to them as natural persons but as citizens only, for example through political legislation, international treaties, and general practices of state conduct. To remind the reader of Anscombe's point, there are very few ‘brute facts’ in social agency.

In combining internal and external perspectives of statehood, and in upholding the distinction between judgments concerning institutions and judgments concerning the persons who represent them, we must, thridly, also distinguish between ideal and de facto conditions of statehood. Although the task sounds rather complicated, it is far from impossible given that the overall number of states is about 200 (depending on how one counts some contested cases). It requires, however, that the normative ideal be ‘checked’ from a more realistic perspective. Beyond being fully compatible with the principle of ethical individualism, such an account of the agency of statehood would even allow us to distinguish different degrees of legitimacy of and competence in state agency.

The true theoretical challenge of an account of states as primary addressees of requirements of transnational justice, therefore, derives neither from metaphysical nor from metaethical concerns, but rather from the fact that the moral responsibility—or duty—of states with regard to transnational justice must reflect the specific structure of international law and international relations, as discussed in section II above. When it comes to the specification of moral requirements of genuinely transnational justice, we cannot rely on the assumption of methodological statism and argue as if international law and international relations would resemble the liberal model of the legal state.

Against the background of the institutional structure in international relations, the difficulty concerns especially the specification of moral responsibility—or duty—towards morally deviant agents in a morally imperfect environment. If internal and external conditions of statehood are interconnected, internally illegitimate states cannot be treated in the same way as those that can be considered internally legitimate. Their internal moral deficiency cannot be externally ignored, at least from a moral point of view—even if it seems advisable to do so for pragmatic reasons.

The liberal commitment to ethical individualism requires that, at a minimum, violations of basic human rights, such as the prevention and punishment of genocide, make a moral difference with regard to the moral and legal status of the violator in international law. Given the structure of international law, however, that means that in practice requirements of transnational justice will demand counteraction against injustice committed by other parties than one's own. Notoriously, such requirements cannot be treated in purely deontological terms, because they almost always involve conflicts of moral rights as well as conflicts between moral norms and justified non-moral interests that can only be judged and decided on a case-by-case basis. Moral judgment in such conflicts requires an assessment not only of the competence but also of the capacity of agents, as well as an assessment of empirical circumstances and likely mid-term or short-term consequences of different ways of dealing with the conflict that resist generalized answers. To begin with, states differ not only with respect to their degree of internal legitimacy but also with respect to their capacities to intervene, their actual or historical involvement in conflicts, and so on. For that reason, it will be extremely difficult to determine general norms of what transnational justice requires and to articulate general standards of moral responsibility of states for transnational justice.

That, of course, does not exclude the possibility of normative assessment and justified moral judgment. Case-by-case judgments can be fully justifiable, non-arbitrary, and rational. But they requires a different framework of ethical analysis and deliberation than that which is suggested by methodological statism. Although philosophical tradition and legal theory provide quite a fruitful source of structures of moral deliberation and judgment that are helpful for specifying duties under conditions of injustice, they do not always coincide with standard moral intuitions.

Most obvious, an agency-oriented approach towards transnational justice that treats states as moral agents cannot treat all states equally. In addition, it will often have to be satisfied with particular and case-sensitive judgments that are better not generalized into universal norms because the justifiability of the approach depends on specific, empirically contingent aspects of the situation. In that sense, the cosmopolitan tendency to treat the problems of transnational justice in analogy with traditional accounts of liberal justice of the legal state is not only analytically implausible but methodologically misleading. To give it up, however, requires neither the renouncement of liberal values nor the principle of ethical individualism.


The cosmopolitan thesis that states should not be acknowledged to have a moral and legal standing sui generis has been criticized from two angles, first by interpreting it as the claim that all moral properties of institutions must be fully reducible to moral properties of individuals, and second, as a commitment to a rights-based method of ethical theorizing. Starting with the reductivist interpretation, section I argued that the moral and legal status of states cannot in a meaningful sense be reduced to individual rights and duties, because both the substantive content and the moral relevance of individual rights and duties cannot be determined independently from the institutional context in which persons act and interact. Addressing the methodological reading of the cosmopolitan thesis as commitment to a rights-based model of ethical theorizing, section II argued that the familiar rights-based approach of liberal political theory is tightly connected to the institutional background of the liberal model of the legal state—a condition that is not fulfilled by the structure of international relations. Consequently, crucial assumptions of the structure of ethical analysis, such as the assumptions that the principle of granting all relevant agents equal moral status are not fulfilled, because ethically relevant differences, such as willingness to comply with moral and legal norms, must be taken into account. Accordingly, ethical reflection on international law and international relations cannot follow the lead of the ‘rights-based ideal theory of justice,’ but requires a judgment- and act-oriented approach, that is an approach that can recognize states as agents and addressees of inter- and transnational requirements of justice. Section III has outlined such an approach, mainly by arguing that reasons and problems, which are commonly invoked against acknowledging agents that are not identical with natural persons, should be rejected. It has shown that focusing on states as moral agents must in no way depart from traditional liberal values or abandon ethical individualism.

States, of course, are far from being the only agents in international relations. Nevertheless, they must play a crucial role in any liberal theory of transnational justice, not only because they are for pragmatic reasons the most competent addressees of requirements of transnational justice, but, more importantly, because their internal constitution makes them the most apt and most ‘responsible’ addressees of moral responsibility, at least as long as they meet liberal standards of legitimacy.32

One need not ignore the fact that non-state agents, such as pressure groups, companies, and transnational interest groups, play a significant role in international and transnational relations, as well as in international organizations. Nevertheless, the legal and political conventions that determine their collective agency do not invest them with socio-political competences and responsibilities of the sort that are necessary and appropriate for promoting transnational justice. With regard to non-governmental organizations, it continues to remain unclear whom they represent, to whom they are responsible, and what their responsibilities are about.33 The competence and responsibility of economic agents, moreover, is so obviously directed towards the pursuit of economic interests—which can, but need not, be compatible with principles of moral and socio-political justice—that they do not qualify as agents of transnational justice from the beginning.

From a cosmopolitan perspective, one might object that the ‘normatively best’—ideal—solution would be the development of a quasi-statist global institution that resembles all the relevant features of the domestic model of the liberal legal state. Even if we acknowledge that such a vision articulates a normative ideal for the long run, the cosmopolitan approach might be conceived to contribute to such a goal by de-emphasizing the role of states in order to strengthen the role of international institutions. The soundness of the cosmopolitan objection, of course, depends on the soundness of the assumption that international organizations will develop in a way analogous to the domestic model of the legal state (and that ethical theorizing can contribute to such a development), and also that such a development will happen at least in a middle range period of time—which, of course, is empirical speculation. Personally, I doubt that the present status quo justifies an optimistic attitude in that respect. At least for the present state of affairs and the middle run, if not the long, run, I think that prudence recommends that we change our habits of ethical reasoning, rather than expecting that ethics can change the world in such a way that practice fits the preferred structure of ethical theory.


1Charles Beitz (2005) Cosmopolitanism and global justice, Journal of Ethics, 9, 11–27. See also Kok-Chor Tan (2004) Justice without borders: cosmopolitanism, nationalism, patriotism. Cambridge, Cambridge University Press. Despite the acknowledged vagueness of the term, Beitz’ account of cosmopolitanism is, of course, not shared by all who use the term or related terms. For different accounts see e.g. Monique Canto-Sperber (2006) The normative foundations of cosmopolitanism, Proceedings of the Aristotelian Society, 106, 267–283; Jeremy Waldron (2000) What is cosmopolitan? The Journal of Political Philosophy, 8, 227–243; David Held (2005) Principles of cosmopolitan order, in: Gillian Brock & Harry Brighouse (Eds), The political philosophy of cosmopolitanism. Cambridge, Cambridge University Press, 10–27. Since one article cannot do justice to all the different accounts of cosmopolitanism, I would like to emphasize that the following considerations concern the cosmopolitan thesis as stated above.

2See Robert Goodin (1988) What is so special about our fellow countrymen? Ethics, 98, 663–686.

3To avoid a possible understanding from the beginning: I am here concerned with the structure of ethical argument, not with the normative significance of state-sovereignty or political coercion, as e.g. Michael Blake (2001) Distributive justice, state coercion, and autonomy, Philosophy and Public Affairs, 30(3), 257–296; or Thomas Nagel (2005) The problem of global justice, Philosophy and Public Affairs, 33(2), 113–147, and others.

4I follow the practice to contrast transnational and international relations, the latter concerning inter-governmental relations, the former involving inter-societal relations.

5Despite considerable interest in both the analysis of agency of corporate actors and the analysis of collective responsibility of groups of natural persons, the specific problem of the moral and legal status of states conceived of as institutional actors is widely neglected; see e.g. the majority of contributions in Peter A. French & Howard Wetterstein (Eds.) (2006) Shared intentions and collective responsibility (= Midwest studies in philosophy, Vol. xxx). Oxford/Boston, MA, Blackwell. An alternative approach will be outlined in section III.

6For an attempt to reconstruct states as superorganisms see Alexander Wendt (2004) The state as a person in international relations theory, Review of International Studies, 30(2), 289–326. Although I find Wendt's argument unconvincing, I will not enter into a discussion of it, because my own analysis takes a completely different starting point from his.

7See e.g. Philip Pettit (2003) Groups with minds of their own, in: Frederick Schmidtt (Ed.), Socializing metaphysics: the nature of social reality. Lanham, MD, Rowman and Littlefield, 167–193. See also Gilbert's theory of plural subjecthood in part II of: Margaret Gilbert (1996) Living together. Rationality, sociality, and obligation. Lanham, MD, Rowman and Littlefield, 175–278. For some, as I think, valid criticisms of either approach see e.g. Seumas Miller & Pekka Makela (2005) The collectivist apprroach to collective moral responsibility, Metaphilosophy, 36(5), 634–651.

8The reference to general concepts such as ‘individual rights’ or ‘distributive justice’ without further specification of any substantive content is highly uninformative but unfortunately increasingly common in the literature on global and transnational justice. See, e.g. Arish Azibadeh (2007) Cooperation, pervasive impact, and coercion: on the scope (not site) of distributive justice, Philosophy and Public Affairs, 35, 318–358.

9A few general norms might be normatively meaningful even though they can be articulated as purely formal, such as the statement, ‘Promises ought to be kept.’ The concept of an individual right is not purely formal. It is commonly understood in standard legal theory accounts as a triadic (or quadruplic) relation of a rights-holder, a rights-content, and a rights-addressee (as well as, in some cases, a normative rationale).

10G. E. M. Anscombe (1981) Brute facts, in: Collected philosophical papers, Vol. III. Oxford, Basil Blackwell, 22–25.

11See John Mackie (1977) Ethics: inventing right and wrong. Harmondsworth, UK, Penguin, 64–73.

12The distinction between an inside in contrast to an outside perspective has been famously used by H. L. A. Hart (1961) The concept of law. Oxford, Clarendon Press, in order to distinguish different normative points of view towards legal systems considered as institutions. If we consider Roman law, for instance, as a legal system, then a judgment in conformity with the primary and secondary rules of Roman law can be said to have been valid within the legal system of the Romans, even if it appears wrong if we consider its conformity with outside, e.g. modern, moral standards or secondary rules of the legal system of the USA or Great Britain in 2008. The problem with Mackie's use of the distinction is that the outside perspective in the case of promising fails to provide a meaningful context.

13See esp. chapters 1, 5 and 7 in Christine Chwaszcza (2007) Moral responsibility and global justice. A human rights approach. Baden-Baden, Nomos for a discussion of the moral and legal status of individuals in transnational relations and international law.

14See John Rawls (1971) A theory of justice. Cambridge, MA, Harvard University Press.

15It is important to notice that such concerns do not determine or define the meaning of the concept of justice but rather select a specific area of justice. The concept of justice is much broader than political justice. As Rawls acknowledges requirements of justice also address interpersonal interaction and extra-domestic relations. But even proponents of formal definitions of ‘justice’ such as, e.g. ‘giving to each what is his own’ or ‘treating others as you would like to be treated yourself’ are perfectly aware of the fact that substantive requirements of justice must differ with regard to different areas of concern.

16By contrast, Scanlon has suggested that at least human rights generate duties for everybody to protect others from violations of their duties, regardless of contextual circumstances. See Thomas M. Scanlon (2003) Human rights as neutral concerns, in: The difficulty of tolerance. Cambridge, Cambridge University Press, 113–123. Although Scanlon's proposal seems to fit the cosmopolitan thesis, I do not think that it can be consistently defended within a rights-based approach, which Scanlon obviously wants to defend at the same time. If Scanlon's claim is understood to defend a strict or unconditional obligation to assist others with regard to the protection of their human rights, the claim falls outside the scope of deontological duties, if such an obligation conflicts with rights of equal importance on the side of those on whom such an obligation falls. In all non-trivial cases of assistance, therefore, such acts are certainly morally praiseworthy, but they fall under the class of ‘heroic’ or ‘supererogatory’ moral actions that cannot be required. If the notion of individual rights is to be taken seriously, and if it is understood as respecting the distinctiveness of persons in Rawls's sense that gives priority to the notion of ‘the right’ over the notion of ‘the good,’ then nobody can be required to sacrifice his own rights in order to protect another person's rights unless he has himself committed to do so, as, for example, professional police officers have due to their profession.

17See Chris Brown (2003) Moral agency and international society: reflections on norms, the United Nations, the Gulf War, and the Kosovo Campaign, in: Toni Erskine (Ed.), Can institutions have responsibilities? Collective and moral agency in international relations. Houndsmill, UK, Palgrave/Macmillan, 51–68.

18See Kees van Kersbergen, Robert H. Lieshout & Graham Lock (1999) Expansion and fragmentation: political change and the transformation of the Nation-State. Amsterdam, Amsterdam University Press.

19Since the European Union is sometimes invoked as a model of a supra-state, I would like to emphasize that it employs different methods of decision making for different policy-areas, and notoriously relies on methods of inter-governmental decision making when it comes to foreign policy.

20Rawls's distinction between ideal and non-ideal theory serves the same purpose. See Rawls, Theory of Justice.

21This seems to be the sense of responsibility that underlies H. D. Lewis's famous criticism of the acceptance of collective responsibility. See H. D. Lewis (1991) Collective responsibility, in: Larry May & Stacey Hoffman (Eds), Collective responsibility: five decades of debate and theoretical and applied ethics. Savage, MD, Rowman and Littlefield, 17–34.

22The systematical problem of the semantic view is: How do we select an appropriate answer to questions of the type ‘what did the person do when she …?’; e.g. how do we identify ‘a person's raised her arm’ as an act that can be described as ‘she voted for candidate A?’

23See G. E. M. Anscombe (1963) Intention. Oxford, Blackwell; Michael Bratman (1987) Intentions, plans, and practical reason. Cambridge, MA, Harvard University Press.

24As H. L. A. Hart rightly pointed out, judgements of whether a person acted ‘mens rea’, etc. are neither based on the observation of psychological facts nor definitions of psychological states in terms of necessary and sufficient conditions, but are judgments that are primarily formed according to standardized expectations about human behavior in connection with information concerning the circumstances and the history of the deed. See H. L. A. Hart (1968) The ascription of rights and responsibility, in: A. G. N. Flew (Ed.), Logic and language. First series. Oxford, Blackwell, 145–166.

25See for the latter view Daniel Dennett (1987) The intentional stance. Cambridge, MA, MIT University Press. See also Daniel Dennett (1978) Conditions of personhood, in: Brainstorms. Cambridge, MA, Bradford Books, 267–285.

26As one reviewer pointed out, ordinary persons and some theorists would insist that we can also judge the ‘character'of institutional agents. Although I would not want to defend such a position, it is also unnecessary to do so, because the conception of institutions as subjects of agency is analytically independent of whatever position one takes towards judgments of character. Personally, I consider judgments that say that a state or a corporation is ruthless, treacherous, outrageous, dull, or inert, and so on, as shorthand for saying either that the natural persons who represent the state or institutions are vicious or passive, or that the internal legal and conventional constitution of the state or corporation is dysfunctional, illegitimate, inefficient, or something similar.

27Oral conversation.

28A frequent objection to the concept of collective responsibility is the fear that acceptance of collective responsibility conflicts with the assignment of either individual responsibility or personal merit. According to the argument, acknowledgment of collective responsibility, e.g. the responsibility of Germany or ‘the Germans’ for the Holocaust or of the US Army for the My Lai massacre, undermines the possibility of assigning a special responsibility, e.g. to Hitler or Goering, or of holding individual soldiers liable for war crimes. See Lewis (1991) loc. cit.; see also Ian Narveson (2002) Collective responsibility, Journal of Ethics, 6, 179–198. The objection, I think, is mistaken, because it partly confuses responsibility and personal guilt on the one hand and ignores that the assignment of responsibility is usually directed to certain tasks or aims, whereby those tasks or aims do not have to be the same for collectives as a whole and for individual members of the collective. Whereas it is arguably difficult to assign collective guilt—or even equal collective guilt—to Germans for the Holocaust, especially for those generations who were born after 1945, it is arguably possible to assign collective responsibility to them and to expect them to continue paying reparations to the descendants of Jews or the State of Israel. Similarly, whereas the individual responsibility of a common soldier who acts (or is forced to act) on the command of his superior should be measured differently from the responsibility of the person who gives the command, we can still distinguish with regard to personal guilt between those who complied fully and heartily and those who did so reluctantly, as well as between those who could have made a difference to the course of events and of those who cannot have been expected to be able to do that. For a more detailed discussion, see Kurt Baier, Guilt and Responsibility, in: May & Hoffman (Eds), Collective responsibility, 197–218.

29The second trap is not as uncommon as it might seem prima facie. It can, for instance, be frequently found that responsibility—or even a (strict) obligation or duty—for the establishment of collective goods or the prevention of collective bads is ascribed to individuals, although it is commonly conceded that, per definitionem, no individual person can bring such results about. Individuals are therefore often taken to be obliged to ‘contribute their share’ for the establishment of those goals. But it usually remains completely obscure what kind of conduct or behavior that requires from individuals, and, consequently, what the conditions are that specify when and to what extent individuals are obliged to ‘contribute their share.’

30Michael Walzer (1977) Just and unjust wars. New York, Basic Books; Charles R. Beitz (1979) Bounded morality: justice and the state in world politics, International Organization, 33, 405–434; Gerald Doppelt (1978) Walzer's theory of morality in international relations, Philosophy and Public Affairs, 8, 3–26; David Luban (1980) Just war and human rights, Philosophy and Public Affairs, 9, 160–181; David Luban (1980) The romance of the nation state, Philosophy and Public Affairs, 9, 393–397; David Luban (2004) Preventive war, Philosophy and Public Affairs, 32, 207–248; and Richard Wasserstrom (1978) Just and unjust wars: a moral argument with historical illustrations, Harvard Law Review, 92, 535–545.

31See John Rawls (1999) The law of peoples. Cambridge, MA, Harvard University Press and James Buchanan and Robert Keohane (2004) The preventive use of force: a cosmopolitan institutional proposal, Ethics and International Affairs, 18, 1–22.

32See e.g. Daniel A. Bell & Jean-Marc Coicaud (Eds) (2007) Ethics in action: the ethical challenges of international human rights nongovernmental organizations. Cambridge, Cambridge University Press, United Nations Press.

33The conflicts e.g. between ecological movements and poverty relief organizations concerning genetically manipulated food resources illustrate vividly the narrowness of concerns as well as the fragmented policy orientations of non-governmental agents.

About The Author

Christine Chwaszcza
European University Institute

Chair of Social and Political Philosophy
Department of Social and Political Science

Article Metrics

Metrics Loading ...

Metrics powered by PLOS ALM

Related Content