Conceptualizing responsibility in world politics

Moral responsibility is a prominent concept used in political discourses and theoretical debates. Yet disagreement remains on how it could work in practice. When attempting to address global challenges such as global poverty, combating atrocities, or artificial intelligence, approaches often revert to retrospective accounts of responsibility that focus on non-compliance with regulatory frameworks. As a result, cases where prospective responsibility would be required often go unaddressed. In this article, we introduce an analytical conceptualization of responsibility that should help to guide the application of moral responsibility in such situations. In the first step, we develop a typology that distinguishes between four types of responsibility: ‘ obligatory ’ , ‘ structural ’ , ‘ prescribed ’ , and ‘ discursive ’ . Second, we identify responsibility gaps for each responsibility type. Third, we introduce different ethical principles from political theory that help to identify potential responsibility relations. We illustrate the utility of this framework with the example of climate change, where ethical principles beyond the contribution principle have already been applied. The paper facilitates new perspectives in political debates about how to allocate responsibility in light of global challenges and enhances theoretical debates in International Relations scholarship.


Introduction
Questions about responsible behaviour, moral conduct, or the normativity of policy choices play a crucial role in international politics, and yet they are mainly debated in the fields of political theory and moral philosophy. 1 Applying moral responsibility in practice seems more complicated, leading to what we call 'responsibility gaps'. Often, such gaps led to calls for more or better regulation to tackle a particular challenge. However, the application of ethical principles is dismissed on the ground that ethics is a matter of opinion and does not offer clear guidance for action. In this article, we propose to refine the notion of retrospective responsibility already existent in International Relations (IR) theory and offer a conceptualization that helps to address responsibility gaps more systematically. We do this because many global challenges, such as poverty or humanitarian atrocities, cannot be solved by identifying a 'guilty' actor. This, in turn, leads to political stalemate, as seen today in the face of climate change. If no one can be identified as a culpable actor, no one can be sanctioned according to the contribution principle. We argue that responsibility offers a conceptual entry point for addressing these challenges: it expands the set of principles for allocating responsibility from mere contribution (who caused a specific outcome?) to other ethical principles known from political theory such as the connectedness, beneficiary, or capacity principle. 2 Hence, we argue for ethical pluralization to address the global challenges of our time. 3 The way we use responsibility today, as a moral and ethical concept, was introduced in the 19 th century. 4 Traditionally an individualist concept, 5 today it is generally accepted that it also includes a collective or corporate dimension 6 and captures shared, 7 systemic, or global responsibility as well. 8 In addition to the concept's connection to different types of agency, existing definitions of responsibility often refer to dichotomous dimensions related to the quality of the action itself. 9 The distinction between prospective and retrospective responsibility, for instance, 10 emphasizes a time dimensionthat is, acting to prevent something from happening in the future vs. taking responsibility for harm done. There is some overlap here with another dichotomy: positive and negative responsibility. The difference is that positive responsibility emphasizes the proactiveness of an actor while negative responsibility also covers responsibility for inaction. 11 Positive responsibility is often also described as moral responsibility, as opposed to causal responsibility, which can only be retrospective since the harm done lies in the past. The question is whether the actor who caused it should be held to account (also called remedial responsibility vis-à-vis role responsibility). Different definitions of responsibility emphasize different aspects, depending on what the concept of responsibility is used for, such as addressing injustices, past or future challenges, taking care, and so on.
No matter which aspect is emphasized, responsibility is a concept based on the relation between a subject and an object of responsibility and a normative authority (for a detailed discussion of this relational understanding, see below). 12 Following this understanding, we contend that responsibility relations can vary on two dimensions: first, in their claritythat is, how responsibility is established, which ranges from a direct relationship between subject and object of responsibility to a diffuse 2 Barry 2003. 3 We owe this point to one of the reviewers. 4 Bayertz 1995. 5 See e.g. Loh 2017, 40. 6 Erskine 2003b. relationship in which responsibilities cannot be attributed directly to one or several specific actors. Second, in the specificity of the regulatory context. Here we distinguish between situations with a fixed regulatory framework, 13 where the question is one of applying the rules, and situations characterized by an unspecified, that is fragmented, contested, or absent regulatory framework. The first dimension determines who is responsible to whom, while the second determines on what grounds someone is responsible.
Combining both dimensionsthe clarity of relations and the specificity of the regulatory contextleads to a typology of four different types of responsibility: obligatory, prescribed, structural, and discursive responsibility. This typology helps us to analyze both regulated and unregulated global challenges. Based on the distinction between the regulatory context and the clarity of responsibility relations, we are able to identify different responsibility gaps for the respective responsibility types and suggest different ethical principles to address these gaps. Instead of dismissing such gaps as being difficult to address in terms of responsibility attribution, we argue that drawing on ethical principles that go beyond a contribution approach for answering the question of who ought to be responsible provides better ways to deliver practical answers in the face of global challenges. 14 The paper is structured as follows: in the subsequent section, we discuss how the notion of responsibility has been used in political theory and across the main theories in IR. This discussion offers a description of the conceptual family of responsibility and how it relates to the concepts of power, legitimacy, and appropriate behaviour. We demonstrate that it is especially useful to include normative questions in the analysis of global challenges. In a next step, we conceptualize responsibility in international politics, building on existing responsibility literature and combining it with political theory. Here, we develop the typology that allows us to categorize normatively problematic situations, specify responsibility gaps, and offer ethical principles to address them. Finally, we illustrate the usefulness of this typology with the example of climate change as a global challenge to responsibility to show how it has already been applied to some extent. In the conclusion, we discuss how an ethically inspired approach situated in IR theory helps to address empirical situations characterized by unclear responsibility relations, demonstrate its applicability to other policy areas, and outline its limits.
Responsibility in IR theory: between regulation and ethics IR scholarship deals with moral issues by invoking concepts such as power, legitimacy, and appropriate behaviour that are more or less linked to the concept of responsibility. In realism, for instance, the structural assumption of anarchy largely 13 Note that fixed does not mean immutable. Regulatory frameworks are always contested but robust for a specific period of time. 14 It should be noted that in this paper we offer an analytical conceptualization of responsibility. We do however not aim at providing an assessment of existing regulations and whether or not they produce morally 'right' or 'wrong' outcomes when applied. Neither do we offer practical guidance on how to overcome the implementation problem in responsibility situations. Yet, successful implementation starts with informed policy debates about how responsibility could be allocated in case of unclear responsibility relations to which this framework contributes. neglects responsible state behaviour towards third parties. 15 In liberalism, a moral dimension of responsibility is acknowledged by favouring cooperation and just behaviour over mere power politics. 16 With its focus on institutional integration along with common rules, neo-institutionalists increase the group for whom or for what responsibility should be taken based on relevant (societal) interests. A lot has been written on global regulations and standard-setting and their legitimacy. 17 However, this literature focuses almost exclusively on already regulated situations; global challenges that are often unregulated are not given special consideration. Finally, constructivists shed light on the existing normative framework's appropriateness and ask whether actors behave accordingly. 18 This has provided important insights into what is regarded as appropriate behaviour for political actors and has broadened the analytical framework to incorporate existing normative expectations into the analysis of world politics. However, responsibility is often only conceptualized as a policy norm. 19 Beyond these general accounts of how moral questions in international politics have been addressed in IR theory, there is an increasing body of literature that engages with responsibility in international politics theoretically, conceptually, and empirically, drawing in particular on political theory and philosophy. 20 We identify three different research foci in this literature: first, ethical-normative debates across policy fields in which responsibility has already been operationalized in concrete practices, such as corporate responsibility, environmental responsibility, and, most prominently, the responsibility to protect. 21 This research has revealed the possibilities and limits of moral responsibility in IR but does not offer more systematic results to be applied beyond the policy case in question.
Second, several studies contribute to the question of who might count as a responsible actor. 22 The extension of moral agency beyond the individual level to corporate actors begs for discussion. Toni Erskine, for example, derives conditions for the moral and shared responsibility of corporate actors and collectives of corporate actors. 23 Generally speaking, a corporate actor can be regarded as a moral agent if it has 'an identity that is more than the sum of the identities of its constitutive parts and, therefore, does not rely on a determinate membership; a decision-making structure; an identity over time; and a conception of itself as a 15 Herz 1961, 130f. However, Lundborg (2019 has stated that in light of the assumed structural conditions of survival, ethical behaviour of states becomes possible. Beardsworth underlines that taking responsibility can be in the 'national interest' of states (Beardsworth 2017, 104f.). 16 Moravcsik 2002. 17 For an overview, see Levi-Faur 2011;Grant and Keohane 2005;Bianculli et al. 2015;Heldt 2018. 18 Finnemore and Sikkink 1998;Chayes and Chayes 1993;Wiener 2014;Deitelhoff and Zimmermann 2019;Deitelhoff and Zimmermann 2020. 19 Park and Vetterlein 2010. unit'. 24 Drawing on Hannah Arendt's understanding of political theory, 25 others argue for incorporating the individual in IR theorizing. Jamie Gaskarth, 26 for example, employs a virtue ethics approach and assesses the individual responsibility of single politicians. This focus on the location of moral agency has broadened our understanding of who can be blamed and praised for their political actions in response to the multi-polar nature of global governance. This multiplicity has sharpened our sensitivity to different levels of analysis. 27 Third, responsibility research centres around the responsibilities of powerful actors. 28 This refers to the capacity of an actor to take on responsibility based on the 'Spiderman' argument that 'with great power comes great responsibility', 29 or, put differently, the capacity of an actor determines its responsibility to act. 30 The fact that some multinational corporations possess more capital than some states, for instance, is sometimes used to argue that they are in a better position to tackle certain problems and thus have the responsibility to do so. 31 Exploring corporate responsibilities for human rights, David Karp, 32 however, discards capacity as a principle (as well as legalism and universalism) of assigning responsibilities to corporations in favour of the publicness approach, according to which 'relevantly public but not relevantly private agents can justifiably be assigned human rights responsibility'. 33 Karp makes this argument with regards to the special obligation of protecting and fulfilling human rights. Yet in the case of more general obligations related to climate politics, global poverty, or development, the capacity principle might be more applicable.
While these three strands of literature are rich in empirical studies, theoretical considerations, and normative assessments, they do not offer a systematic approach to analysing responsibility situations. Thus, the purpose of this article is to present a conceptualization of responsibility that facilitates a systematic assessment of responsibility gaps along the two dimensions we have identified. The benefit of such a conceptualization of responsibility is twofold: first, we build on the existing responsibility literature in IR theory but go beyond an issue-or actor-specific analysis by offering a classificatory system of responsibility as an analytical tool that allows us to examine responsibility gaps more generally. The results of such systematic analyzes can, in a second step, inform policy debates about how responsibility 24 Erskine 2003a, 24.   Lebow 2003;Erskine 2008;Goodin 2013;Heupel 2013;Karp 2014Karp , 2015Beardsworth 2015Beardsworth , 2017Sienknecht 2022. 29 Heupel 2013. 30 Miller 2001, 460. A similar argument is made by English School scholars, see e.g. Bukovansky et al. 2012;Falkner and Buzan 2022. 31 Capacity does not always relate to wealth, i.e. the financial resources of an actor, but can also refer to opportunity. It did not cost Coca Cola much, for instance, to distribute condoms to the most remote areas in some African countries to help in the fight against HIV/AIDS. In that case, positive responsibility is requiredthat is, a moral compass of an actor (like Spiderman) who takes on responsibility because s/he can do so based on distributive justice. 32 Karp 2015. 33 Ibid., 116. 'Public' does not need to be the same as 'state'.
International Theory could be allocated, under what conditions, and following which ethical principles. We illustrate our conceptualization with the case of climate change where we can already observe an emerging process of ethical pluralization. We argue that a systematic analysis of that policy field based on our proposed typology will enhance ethical pluralization in policy areas that lack robust regulatory frameworks such as bioethics, artificial intelligence (AI), or global poverty.

Conceptualizing responsibility
Responsibility is the result of social relations and a dialogue between political actors intended to find policy solutions for problems. 34 Offering a typology of responsibility, we build on the triangular conceptualization of responsibility between a subject (A) and object (B) of responsibility and an authority (C). 35 This provides the normative ground for established responsibility claims or attributions and extends them into the expectational environment in which they are embedded (see Figure 1). The relationship between the subject and object of responsibility is realized through the outcome or consequences of action/inaction. The relationship between the subject of responsibility and the authority, which can enforce normative claims brought to it by the object, is established through the act of justification and sanctioning. This can be most easily illustrated when applied to a legal context: subject A has caused harm through a certain act to object B, who can claim 'justice' based on the normative authority of law (C). Depending on the specific act, law enforcement and the judicial system will process this claim, eventually discussing it in court, where A can justify his/her action and, if found guilty, will be sanctioned. This would depict an ideal-typical case of responsibility defined above as B's right to hold A accountable to a set of standards, here the law, and the agreedupon authority determining whether those standards have been met and, if not, what legal consequences result.
However, responsibility relations between A, B, and C are often not so straightforward, and claiming, attributing, or sanctioning responsibility in world politics is difficult. 36 We argue that in cases where the three poles of the triangle are not specified, the expectational environment becomes more relevant 37 as the normative context in which discussions about possible responsibility relations take place based on different ethical principles. Here, so-called 'communities of responsibility', 38 which Antje Vetterlein defines as sets of 'connected actors, individual or corporate, who recognise and value each other as a group with regard to corporate responsibility, and also interact with each other', become relevant. These communities serve as the platform where debates about allocating responsibility are taking place and where different actors voice their expectations as to who should be responsible for what, invoking ethical principles as justifications. 34 Vetterlein 2018, 562. 35 An authority can vary from God or the law to the social/normative structure that creates expectations and might be able to sanction non-compliance. Accordingly, also the authoritative agent varies. This is not to say that in legal contexts this expectational environment is irrelevant. Also, legally codified norms change as a result of debate in the expectational environment. 38 Vetterlein 2018, 558.
As the triangle in Figure 1 shows, responsibility allocations can be theorized as being inherently relational. These relations can differ on two dimensions: (1) the kind of social context or regulatory framework that determines the necessity for actors to act responsibly, and (2) the clarity of the social relations between the respective actors/poles. In the following, we introduce each dimension before explaining how they combine into a typology of responsibility.
The first dimension revolves around the question whether responsibility is already institutionalized and regulated. In the case of a fixed regulatory context, responsibility is inscribed in the regulations that govern a specific policy field, for example war crimes such as extrajudicial executions or torture that are regulated in international law. In such cases, taking responsibility depends on the subject's attributes, the resources or capabilities he/she possesses, and the power and interest to set or avoid standards and regulations. In other words, a normative setting exists, but subjects addressed by the norm decidedepending on their power status, dominant societal interests, and the guiding behavioural normswhether they act accordingly. In these regulated contexts, which are characterized by clear normative expectations, monitoring mechanisms, and sanctioning possibilities, the dominant question is whether potential subjects of responsibility act accordingly and take responsibility. However, many situations in world politics remain unspecified, if regulated at all. In unregulated social contexts, norms about appropriate behaviour might exist, but no monitoring or sanctioning mechanism institutionalizes these expectations. In these cases, some or all of the poles of the triangle remain preliminarily empty and debates over ethical principles in the expectational environment gain importance for constituting responsibility relations. Regarding the second dimension, we distinguish direct and diffuse relations. If the responsibility relations as described above, are complex and fuzzy, this diffuses the relations amongst subject, object, and authority. Concerning the subject-object relation, it might sometimes prove difficult to single out one or more actors and their share of responsibility in causing a specific outcome such as in the case of climate change, where it is difficult to identify responsible actors due to the 'problem of many hands'. 39 Further reasons for the difficulty of identifying a responsible subject might be that there is a time gap between action and harm done, and it is thus not possible to relate them to each other; and/or the actor and actions are spatially distant. The object-authority relation might be less direct when facing normative conflicts, either in cases where legal solutions lead to illegitimate outcomes or where different norms oppose each other.
Combining both dimensions generates a typology of four types of responsibility (see Figure 2). Obligatory responsibility describes the textbook case of accountability and exists when the responsibility relations are direct and clear and a regulatory framework exists. Here, problems revolve around questions of non-compliance. Prescribed responsibility describes situations of existing regulation and diffuse responsibility relations between a subject and an object. We would expect to observe struggles over competence, that is, who should be responsible and for what. We describe structural responsibility as situations in which there is no regulatory framework at hand, but responsibility relations between subject and object are actually clear and direct. In this case, the problem is the missing regulation. Finally, in a situation where there is no regulatory framework and responsibility relations between subject and object are diffuse, we speak of discursive responsibility. These are usually situations characterized by uncertainty, as the issue for which responsibility should be taken is new and/or the harm might be revealed only in the future, while a common sense of appropriate behaviour is still in the making. Identifying different responsibility gaps for each responsibility type allows us to link them to existing ethical principles that can be brought forward in debates about how to address these gaps (for a summary, see Table 1). In other words, our conceptualization can be used, first, to assess specific challenging situations with regards to the regulatory context and the clarity of the social relations between the actors involved. Based on that assessment, we are now equipped to identify the problem at hand and thus offer appropriate ethical principles. The expectational environment is crucial here as the place where normative debates take place. Arguments often fall short of a full range of ethical principles that could be applied when it comes to allocating responsibility and frequently stop with a reference to the contribution principle. When moral arguments are invoked, they are often rejected as being value-based and, thus, debatable. Yet political theory identifies other principles for allocating responsibility in such circumstances. The beneficiary principle, for instance, would attribute responsibility according to the extent to which an actor has benefited from some unjust situation. 40 The connectedness principle assigns responsibility to actors who are linked to injustice, be it geographical, community-based, or based on something else. 41 The capacity principle states that those who are capable of mitigating a harmful situation have an obligation to do so. 42 Introducing a broader range of ethical principles to the debate helps us to overcome situations where regulation is difficult or impossible to achieve but in which mere trust in voluntary action (e.g. as in the case of Corporate Social Responsibility, or CSR) leads to undesirable outcomes.
Obligatory responsibility: fixed regulation, direct actor relations The first type of responsibility focuses on relations between subject and object in which the former takes responsibility for an object based on a fixed regulatory framework. This is the typical case described above, when A takes responsibility for B and can be held accountable based on C. Obligatory responsibility describes situations in which the subject either takes responsibility or is asked to take responsibility. That is to say, even if someone denies his/her responsibility, an actor can be asked or coerced to take responsibility due to a fixed regulatory framework (e.g. a treaty or law) and a causal relationship. Think, for example, of Russia's invasion of Ukraine in February 2022. Russia can be held accountable for its actions in accordance with international law, as it clearly violates central international norms such as the territorial integrity norm. Here, obligatory responsibility stems from the attribution of causal responsibility. International sanctions against Russia attempt to address the norm violation. While we assume that responsibility allocations and relations are clearly identifiable in the case of obligatory responsibility, responsibility gaps may arise nevertheless. Especially at the international level, there is often a lack of sanctioning mechanisms to enforce the responsibility of political actors. As a result, powerful states can disregard existing norms without fear of repercussions. 43 Responsibility gaps here mainly originate from problems of compliance and/or enforcement rather than problems of attribution or identification of responsible actors.
Prescribed responsibility: fixed regulation, diffuse actor relations The second responsibility type describes situations in which a clear normative framework exists, yet responsibility relations are diffuse, making it difficult to allocate responsibility to specific actors properly. Consequently, the triangle becomes unspecific regarding the actors who ought to take responsibility but is specific about the object and the normative framework. Here, questions of autonomy and capability play a role, putting those actors in responsible positions who have the capacity to change an unwanted situation. Prescribed responsibility refers to indirect responsibility relations, embedded in a fixed regulatory context. Think, for example, of international security governance, which ascribes ultimate responsibility to the United Nations Security Council but does not automatically set an actor in a responsibility relationship to a particular object. Prescribed responsibility is mainly characterized by diffuse actor relations, which might either provoke actors to claim responsibility or to struggle over who is responsible for what. In situations in which many hands are involved, the problem of competence might arise as a responsibility gap. That is to say, the question is less about who contributed how much to a morally precarious situationwhich might not be solvable anywaybut who has the capacity to change the situation. The case of the 1998 Kosovo intervention is an example of prescribed responsibility. In light of the deadlock in the UN Security Council, NATO acted on its own and justified its intervention as an attempt to prevent genocide. 44 Structural responsibility: unspecified regulation, direct actor relations The third form, structural responsibility, characterizes situations where subject and object relations are clear but the regulatory framework is unspecified or absent. Moreover, the responsibility relation emerges from co-constitutive structures that prescribe certain responsibility relations having to play out in a particular way. Think, for example, of the economic and political relationship between the global North and the global South. Owing to centuries of colonization, exploitation, and 43 However, international jurisdiction further restricts the scope for action even by powerful states. 44 This example also shows that it matters a lot who it is that gets to prescribe responsibility as NATO's self-prescription to intervene in Kosovo when the UNSC had not sanctioned its action can be seen as problematic.
oppression, the North has a structural responsibility towards former colonies based on restorative justice, which has not (yet) been formalized in any regulatory framework. 45 The responsibility gap is grounded in the problem of non-regulation. Thus, an option to overcome this gap is to ask who benefits or benefitted the most from an unjust situation (beneficiary principle). 46 The beneficiary principle identifies the duties of actors without blaming them to have causally contributed to certain harm. However, it has been subject to discussions as the question emerges how to calculate reparations for instance for slavery, or more generally how duties arise based on past benefits when wrongdoings were not even considered unjust. 47 In this context, Christian Barry offers one commonsensical principle for filling the regulatory gap: According to the first, historical interpretation, the extent of benefit is to be judged by how much better off the agent is compared with their position before the unjust arrangement was put in place. On the second, unmoralised subjunctive interpretation, the magnitude of benefit is judged by comparing the agents' current position against what it would have been now, had some particular unjust arrangement not been instituted. Pursuant to the third, moralised subjunctive interpretation, degree of benefit is determined by comparing agents' current position against what it would now be under a fully just scheme of social arrangements. 48 Another example where structural responsibility plays a role is global poverty: factors such as economic globalization and trade policies can increase economic inequality. The subsidy policies of the European Union (EU) for dairy and agricultural products help otherwise more expensive EU products to be competitive against cheaper products from abroad. Yet this subsidization of dairy products in the EU harms local producers in African countries who cannot compete against these prices. Under the beneficiary principle, the EU would have a responsibility to change its policy in favour of a more equitable relationship to African countries.

Discursive responsibility: unspecified regulation, diffuse actor relations
Discursive responsibility relates to how responsibility claims and attributions are discursively produced and reproduced in situations of diffuse responsibility relations and a missing regulatory context. This concept of responsibility lacks clear responsibility relations between subject and object, while a fixed normative 45 Thompson (2018) speaks of 'remedial responsibility' for reparations for historical injustice, such as slavery. According to Young (2011, 180), structural injustice is often the result of 'a multitude of routine and deliberate actions within institutions', in which it is difficult to identify individual responsibility. The responsibility of actors in situations of structural injustice then emerges from being part of the system, respectively from being connected to the situation (Young 2011, 180 Barry 2003, 229. The common but differentiated responsibility (CBDR) principle is an example of the successful application of this principle as outlined in the empirical section below. For a critical discussion about the beneficiary principle, see e.g. Fanon 1967;Brooks 1999;Posner and Vermeule 2003. framework is also absent. Think about technological developments such as cloning or AI. Due to the novelty of such developments, there are usually little to no regulations. It must be negotiated discursively (see, e.g. the risk-based European AI Act). Furthermore, in case of new technologies, actor relations are diffuse. Who is responsible for the compliance of new technological developments with human rights? The user? The producer? The state? International organizations? Discursive responsibility thus describes situations in which a responsibility triangle is missing (problem of ignorance) or is in the making. How responsibility is constituted depends on the discursive settings expressed in the expectational environment. In such cases, no unambiguous causal link between an actor and his/her contribution to a harmful outcome can be established. Nevertheless, responsibility could be allocated in such cases based on the ways in which actors are linked to a harm in accordance with the connectedness principle. 49 Global value chains provide examples. At first glance, buyers of smartphones or clothes do not seem to be directly connected to unsafe or undignified working conditions in producer or supplier countries. However, the connectedness principle holds that consumers bear at least some responsibility for the working conditions of affected workers through their purchasing decisions. 50 In the following section, we will demonstrate how our conceptualization of responsibility can be applied to analytically examine situations of responsibility gaps as well as inform policy debates about allocating responsibility in the field of climate change.
Climate change: ethical pluralization in the making? 'Responsibility rests with each and every country. And we must all play our part. Because on climate, the world will succeed, or fail as one'. 51 This quote by COP26 President-Designate Alok Sharma is representative of the assumption that climate change is the greatest challenge of our time and that it can only be addressed by acting together. 52 At the same time, policy responses are often hesitant and cautious. The typology derived in this article admittedly cannot solve the implementation problem of action necessary to protect the climate, yet this conceptualization can inform the agenda-setting and policy dialogue prior to implementation. Although we observe the application of alternative ethical principles in the case of climate change already, there is still room for improvement as proven by the expected failure to meet the 1.5-degree target. Providing a conceptualization with which we can analyze responsibility gaps more systematically and identifying different ethical principles for addressing them can provide guidance to policymakers, engaged citizens, and civil society activists to help inform policy debates about allocating responsibility in the field of climate change. It could also significantly increase the pressure for action. Put differently, before policy solutions can be found, we need to allow for a broader discursive space in which a variety of policy 49 Young 2006, 2011. problem definitions can be discussed. 53 Our conceptualization of the processes taking place in this policy field can then be used in other policy areas where moral debates are not yet pluralized to allow for alternative ways to address ethically tricky problems.
Climate change politics is extremely contested and value-laden terrain. Regulating this policy field is very challenging for a number of reasons: short election cycles that are often identified as problematic for tackling such a long-term challenge; the fact that territorial or electoral boundaries do not overlap with ecological boundaries; the lack of expertise of the lay public, and the question of democratic legitimacy in taking political decisions; and finally the dominance of the nation-state-citizen relation in liberal democracies that overlooks not only future generations but also the potential rights of non-human entities or ecosystems. 54 Levin et al. (2012) call it a 'super-wicked problem' characterized by four key dimensions: 'time is running out; those who cause the problem also seek to provide a solution; the central authority needed to address it is weak or nonexistent; and, partly as a result, policy responses discount the future irrationally'. 55 In addition, we experience a 'culture of unaccountability' 56 which could be addressed through increased transparency and information sharing, more or better communication attempts that aim at explaining and justifying taken decisions or enhanced monitoring and evaluation mechanisms. More transparency however is also seen critically. 57 Thus, the problem of climate change serves as an example to illustrate the above identified four responsibility types resulting from, on the one hand, unclear social relations between the object, subject, and authority of responsibility and, on the other hand, different degrees of the existing regulatory context.

Obligatory responsibility
In the context of climate change, we can observe developments towards the institutionalization, or even juridification, of environmental politics. 58 The impulse to use the law as an instrument to address climate change is typically motivated by the belief that obligatory responsibility should apply. Indeed, the push to respect principles of intergenerational justice, 59 the attempt to frame environmental destruction (ecocide) as an international crime, 60 and the move to designate a 53 For a similar point on problem definition and how it affects the possible solutions we can find, see Cashore and Bernstein 2022. 54 Some would even argue that the problem of climate change emphasizes the limits of our liberal democracies, in particular on four dimensions, i.e. time, space, agency, and community understanding (Eckersley 2020). 55 Levin et al. 2012, 123; see also Schulev-Steindl et al. 2022, 22. 56 Najam and Halle 2010, cited in Kramarz and Park 2019, 3. 57 Gupta and Mason 2014. 58 See e.g. Kramarz et al. 2017. In this context, Verena Madner describes the law as being used as an instrument to 'bridge the gap between recommendations by climate science and (often unambitious) climate policies' (Madner 2022, 7), aiming not only at increasing enforcement but also at enhancing awareness, that is mobilizing the expectational environment. 59 UNESCO 1997;Thompson 2009;Meyer 2012. 60 Greene 2019. Proposals made in this context aim to include 'ecocide' as a crime against peace, which would lead to the possibility that the International Criminal Court could hear respective cases. green future as a human right, 61 all point in the direction of an eventual, robust, truly global regulatory framework. National and regional developments also hint at bottom-up processes that may contribute to thisfor example, the EU Emissions Trading System (EU ETS, the first major international carbon market), the designation of nature reserves, and various recent national court decisions. In Germany, for example, the Federal Constitutional Court ruled in April 2021 that the German climate bill is in part unconstitutional 62 because it only provides for emission reduction measures until 2030, putting the remainder of the climatechange burden on the shoulders of younger and future generations, whose civil liberties would be violated as a result. In 2019, the Dutch Supreme Court also confirmed a decision of the District Court of The Hague from 2015, which underlines that the Netherlands is exposing its citizens to danger and should thus reduce its greenhouse gas emissions by 25% until 2020. 63 In the 2021 Milieudefensie decision, Royal Dutch Shell was also ordered by the Hague District Court to reduce its emissions. In short, we observe obligatory responsibility regarding climate change, in which the subject of responsibility (e.g. governments or corporations) is non-compliant with its responsibility towards future generations (authoritative framework), and is thus held accountable for the protection of the climate (object of responsibility) by courts.
These cases are particularly noteworthy as they exemplify attempts to tackle the above-mentioned hurdles through obligatory responsibility, that is, by identifying a legal basis for rulings, even in cases where the harm is prospective and causal attribution uncertain (what, exactly, is Shell's contribution to climate change?). Once such a regulatory framework exists that specifies responsibility relations, if only through precedent, the main problem that remains is compliance. There can be different reasons why actors do or are able to avoid regulation. Problems can result from the fact that juridical boundaries do not overlap with ecological boundaries, which leaves us with a fragmented regulatory context and might render rules unenforceable. 64 In addition, some actors are more powerful than others in the field of environmental diplomacy, which gives them more leeway to not follow rules (as we have seen in the case of the US withdrawal from the Paris Agreement under the Trump administration) or to pay for rather than correct their misconduct (e.g. companies paying fines rather than reducing emissions).
While the climate change governance regime may be evolving towards clearer obligatory responsibility, other aspects of the climate change problem are better understood as reflecting other responsibility types warranting the application of other ethical principles that could and should serve as reference points in political debates where appropriate.

Prescribed responsibility
As already illustrated, it is typical for the field of environmental politics that regulations exist but that it proves difficult to identify who is responsible and for what harm. This is the domain of prescribed responsibility. Due to the number of actors involved it might even be impossible to hold specific actors responsible for the outcome (the 'problem of many hands' 65 ). The question of who caused a specific outcome might therefore not be the most appropriate question. Discussions about the responsibility of each individual state in the case of global warming signify how difficult to answer such a formally simple question is. Should we only consider scope 1 carbon emissions or also scope 2, and even scope 3 emissions, 66 as well as a growing desertification or declining biodiversity as factors for the calculation? Depending on the answer, responsibility relations would look different. Attribution science identifies the drivers of climate change by distinguishing between natural and anthropogenic causes, and can with high certainty assign climate change to human activities. Yet this research also acknowledges uncertainties linked to observational data and model simulations. 67 So what to do when responsibility cannot clearly be attributed to specific actors, let alone the specific actions of those actors?
Instead of asking who caused the harm, applying the contribution principle, we might be redirected to ask who has the capacity to change or mitigate a problem. This so-called 'capacity principle' 68 or 'ability to pay principle' is already applied in climate agreements, for instance in the common but differentiated responsibility (CBDR) principle. The Paris Agreement and previous climate regimes have distributed responsibility differently to individual countries based on their economic capacity.
The capacity principle is prototypically exemplified in the 'Spiderman-moment' when he concludes, 'With great power comes great responsibility'. 69 The great power literature in IR theory addresses precisely this issue and discusses the special responsibilities of states in light of global problems. 70 Great powers often get more involved in global affairs as they gain special rights for doing so. 71 Steven Bernstein, however, notes that '[n]orms or institutional arrangements that demand or recognize great power responsibilities for the global environment do not exist'. 72 This comes as a surprise, as this is not the case in other policy areas such as peace and security or economic governance. He even observes a decline in formal and informal recognition of great powers' responsibility regarding the environment. 73 While he explains this lack of acceptance of great power responsibility towards 65 Thompson 1980, 905. 66 'Scope 1 emissions are direct emissions from owned or controlled sources. Scope 2 emissions are indirect emissions from the generation of purchased energy. Scope 3 emissions are all indirect emissions (not included in scope 2) that occur in the value chain of the reporting company, including both upstream and downstream emissions' (World Resources Institute and World Business Council for Sustainable Development). 67 Steiner and Engdaw 2022. 68 Miller 2001. 69 What triggered this insight was Spiderman observing the murder of his uncle Sam without intervening to prevent it. Thus, while he was not responsible for the death of his uncle, as he did not kill him, he feels morally responsible as he had the capacity to prevent this murder from happening. the environment with 'a lack of congruence between systemic and environmental "great powers"; weak empirical links between action on the environment and the maintenance of international order; and no link to special rights', 74 the implications this absence has on global governance are crucial. Bernstein sees a shift happening from external to more internal responsibility, 75 that is, states are seen to be responsible for their own environmental outcomes. 76 Combined with a shift from legally binding rules based on state responsibility to soft law then leads to a 'more horizontal assignment of responsibility' 77 towards country ownership as well as away from states towards corporations or partnerships with non-state actors. This can be observed in global instruments to tackle environmental challenges. Bernstein outlines how the 2015 sustainable development goals' focus on holding countries accountable with country-level reporting shifts attention away from external responsibilities and moral commitments. The same applies to the Nationally Determined Contributions NDCs of the Paris Agreement. Thus, we somehow observe an individualization of environmental responsibilities (towards individual states) as well as a focus on rights that crowds out the actual spirit of the moral responsibility of great powers à la Spiderman. In this context, referring to the capacity principle in its moral meaning in policy debates would help to identify responsible actors in situations of prescribed responsibility and would allow us to rethink global environmental governance.

Structural responsibility
Situations with direct actor relations but without a regulatory framework characterize structural responsibility. The case of the exploitation of the world's fossil resources exemplifies structural responsibility. In these cases, the relationships between the actors are clear (it is obvious that in particular major fossil fuel companies are responsible for exploiting most of the earth's resources) but the regulatory framework that holds these actors accountable is lacking. In such situations, the beneficiary principle would help to identify actors who have benefited most from an unregulated emissions practice and attribute responsibility on this basis. According to the Carbon Majors Database, only 100 companies are responsible for more than 70% of the world's greenhouse gas emissions since 1980. 78 According to the beneficiary principle, these companies should take more responsibility and make a greater contribution to preventing further climate change than other actors. 74 Ibid., 10. 75 Ibid., 25f. 76 Beardsworth (2015) argues that it is necessary to change our understanding from moral to political responsibility when discussing responsibility relations in a globalized age. Instead of just expanding the circle of actors by referring to moral responsibility, he emphasizes that, due to the structures of the political system, the political responsibility of states and the alignment of national interest and the solution of global problems is crucial (Beardsworth 2015, 86). 77 Bernstein 2020, 25. 78 The Carbon Majors 2017.
Following this argumentation, 79 actors would take responsibility proportionate to the benefits they have had from past and current activities that have contributed to climate change. Also, here, the CBDR principle is an existing example of applying the beneficiary principle by attributing responsibility based on historical culpability, referring to the structural responsibility between developed and developing states. While developed states have benefitted by and large from unregulated carbon dioxide emissions in the past, resulting extreme weather phenomena today affect in particular small island developing states and least developed countries. In fact, the countries least responsible for global carbon emissions (both in the past and today) suffer the most from climate change. An attempt to address this inequality is reflected in the decision of the Copenhagen Climate Change Conference in 2009 when developed countries agreed to provide US$100 billion annually to developing countries for their 'loss and damage' 80 in relation to climate change, as well as its inclusion in the articles of the 2015 Paris Agreement. 81 Negotiations at the COP 27 led to the 'historic decision' to establish a loss and damage fund for vulnerable states. However, many details, such as who should pay into the fund, still remain to be negotiated, which underlines how difficult it is to enforce the beneficiary principle in practice. Growing public awareness and a strong civil society might increase the pressure on industrialized countries as well as on companies to take responsibility for the future based on their profits in the past by referring to this ethical principle.

Discursive responsibility
Finally, consider situations in which there is neither a regulatory context nor a clear actor relationship (discursive responsibility). One of the pressing problems related to climate change is the social costs resulting from a transition to a green economy in the form of social exclusion or an increased inequality between countries that can more easily adapt to a green economy and others that are still mainly dependent on fossil fuels. The necessary changes will require such a profound transition that many jobs and even entire industries will be affected. 82 So far, however, the focus has been on how to reduce emissions rather than on how to address detrimental social impacts resulting from such changes (problem of ignorance). Assigning responsibility here proves difficult because the economy is globally interconnected and no treaty legally regulates how this transformation is to be accomplished. However, addressing climate change requires that we accompany the transition process with the creation of new jobs and industries and the training 79 Miller 2008;Page 2012;Caney 2010. 80 The term is usually used in the context of international climate negotiations. The term means different things to different actors: while some developed states try to relate it mainly to the concept of adaptation, developing countries refer to the mechanism of compensation. In the Paris Agreement two articles address the issue: Article 7 is on adaptation, Article 8 is on loss and damages (see UN 2015). While adaptation finance requires funding on short notice (due to weather extremes), funding structures that address loss and damages (e.g. desertification, salinization) require long-term investments. of an appropriately skilled workforce. The just transition approach is an attempt to address this challenge. According to the International Labour Organization (ILO), this requires 'greening the economy in a way that is as fair and inclusive as possible to everyone concerned, creating decent work opportunities and leaving no one behind'. 83 This approach assumes that social dialogue involving different actors at all levels of the economy (e.g. worker organizations, employers, and governments) is central to finding a just response to the transition needed in the face of climate change. Such a discursive approach seems necessary in light of a lack of regulation and diffuse actor relations.
Broadening our perspective by considering different ethical principles helps rethink responsibility relations. According to the connectedness principle, not only the respective states but also companies and each and every individual are asked to act. In this context, the relevance of the expectational environment is highlighted for identifying responsibility relations: processes of naming and shaming, as well as demanding more regulations, lead to states, companies, and individuals taking responsibility. Another framework within which it might be possible to assign discursive responsibility might be an 'environmental responsibility to protect'. 84 Such a framework would highlight the responsibility of states to protect the environment for the benefit of their citizens and vulnerable groups based on human rights. If a state is unwilling or unable to fulfil this duty, the responsibility shifts to the international community. This would empower individuals, help identify responsibility relations, and improve the regulatory framework for environmental protection.

Conclusion
Moral responsibility has always figured prominently as a reference point in political debates, and now also increasingly appears in IR theory. Yet while there is a general agreement that moral responsibility is required to tackle the global challenges of our time, there are divergent ideas about how this should be done. We have argued in this article that a more refined conceptualization of responsibility could help us overcome this shortcoming. Starting from a relational understanding of responsibility between subject, object, and authority of responsibility, we add the expectational environment in which this triangle is embedded in and which allows us to define responsibility conceptually as a dynamic process taking place in a social and regulatory context that can vary. We identify two dimensions along which the allocation of responsibility differs: the clarity of responsibility relations and the existence of a regulatory framework. The resulting typology allows us not only to identify different types of responsibility but also specific responsibility gaps that they highlight as well as ethical principles that could be employed to fill these gaps. Specifically, from this typology we are able to make recommendations for respective responsibility gaps. For instance, if we cannot identify a subject of responsibility due to various reasons, we might ask not who was responsible for an outcome but who would have the capacity to address it. Similarly, if no regulatory framework exists for 83 ILO. 84 Conca 2015; Bernstein 2020. the outcome of the object and subject of responsibility, the beneficiary principle could be applied asking the actor who has benefitted most from this regulatory context to take responsibility. In short, we argue for ethical pluralization. This also means that different ethical principles might be applied in the same situation.
While this ethical pluralization itself does not directly address inaction when it comes to global challenges, we contend that it is especially relevant in the first step of the policy-making process, that is, agenda-setting and policy debate about how to frame and define a problem. Put differently, we need to broaden the discursive space in the defining phase of policies to allow for alternative narratives based on which different policies can be suggested for implementation. In this phase, our conceptualization can help inform policy debates about how to go about tackling global challenges differently. Ethical pluralization can, however, also increase uncertainty or contribute to buck-passing. 85 But failing to attempt to address different concerns and enter into normative debates would not solve the problem either. As we can observe with the emergence of new forms of participatory democracy in the form of local governments or citizen assemblies, wider public debates are necessary to address ethically challenging societal problems.
Using the example of climate change to illustrate this ethical pluralization, the contribution of this article is twofold: policy-relevant and theoretical. First, with regards to policy, we have shown that global challenges cannot exclusively be framed in terms of who is causing a specific problem. When they cannot, we need to allocate responsibilities based on other ethical principles. The example of climate change has shown that ethical pluralization is already being applied. Even if there is still room for improvement, the application can serve as an example for other policy fields that lack a fixed regulatory framework or a clear constellation of actors. For example, the challenges of poverty and social inequality at the global level and the resulting devastating effects on communities and entire countries (e.g. human health, access to education, or the provision of clean drinking water) might be better addressed by reference to ethical principles other than the contribution principle (e.g. the beneficiary and capacity principles). The same applies to environmental problems such as the Pacific garbage patch, which is not caused by a clearly identifiable subject but by a multitude of actors (connectedness principle). Technological developments such as remote conduct of warfare by drones and other lethal autonomous weapons systems, as well as the use of AI in many other policy areas, also require answers to the questions of who is responsible for what and on what basis. When AI systems make decisions that have significant impacts on individuals and society as a whole (e.g. surveillance systems, social scoring, crime prediction), are software developers responsible, the organizations deploying the system, the users, or all of them together? Not everything is susceptible to regulation; we must therefore occasionally look for alternative solutions. Acknowledging that we cannot always assign responsibility in a clear-cut and regulatory way, our conceptualization can be seen as a call to initiate policy debates about responsibility beyond the contribution principle.
Second, the presented conceptualization of responsibility has theoretical implications for both mainstream IR and political theory. Regarding the first, we have 85 We owe this point to one of the reviewers.
shown how responsibility is an understudied analytical category even though it relates to other concepts such as compliance and legitimacy. Yet in its impoverished understanding of accountability, contribution, and (non-)compliance, it narrows our view to the question who has contributed to a certain harm and which sanctions are appropiate to address this. We have sought to show how responsibility as an analytical category makes it possible to consider different ethical principles such as the beneficiary, connectedness, or capacity principle, each of which allows allocating and attributing responsibility based on a mechanism other than how much one has contributed to a harm or outcome. By doing so, we have sought to show that normative debates are necessary in world politics and introducing responsibility as an analytical concept in this respect broadens our perspective to a 'negotiated governance' mode 86 in which dialogue rules and different perspectives are allowed to be brought forward in political debates instead of shying away from normative and ethical discussions. In addition, while we do not argue that the general inaction in case of global challenges absolves states of responsibility and shifts it onto the shoulders of other actors, our conceptualization of responsibility emphasizes the likelihood that more actors should share in it. 87 Political theory and ethical approaches have developed answers and approaches as well as ethical principles that can address responsibility gaps as described above. Our typology of responsibility contributes to this literature conceptually and points the way towards novel operationalizations. It provides a way to distinguish different types of responsibility, in turn (one hopes) making it easier to connect political theory to mainstream IR. 88 Both developments potentially have policy implications. A fresh take on responsibility, in short, has the potential to deepen and enhance our theoretical and empirical understanding of world politics.