Responsibility for structural injustice

ABSTRACT Following Iris Marion Young, Catherine Lu allocates responsibility for transforming unjust global structures to the agents who participate in perpetuating and reproducing those structures. She also adopts Young’s qualitative distinction between the ‘liability’ and ‘social connection’ models of responsibility, reserving the first for interactional injustice where identifiable victims and perpetrators are involved, and the second for structural injustice where unjust outcomes emerge without any identifiable wrongdoers. This article’s argument is that Young’s and Lu’s specific allocation of the burden for transforming unjust global structures makes sense only if we reject the notion of a qualitative distinction between two models of responsibility and acknowledge instead that there is continuity in the conceptual tools available for thinking about responsibility for both interactional and structural injustice.

Structural injustice; Iris Marion Young; responsibility; causal responsibility; outcome responsibility; remedial responsibility; Bernard Williams; liability model; social-connection model; David Miller In her book Justice and Reconciliation in World Politics, Catherine Lu argues that injustice in the international and transnational context cannot be adequately conceptualized within an 'interactional' frame that focuses on victims and perpetrators, be they individuals or groups. For such a frame assumes just background conditions and regards injustice simply as the violation by individuals or groups of a just baseline. This frame, she argues, does not account for the type of injustice that is baked in the background conditions, so to speak. The structures that mediate the actions of individual or group agents can yield unjust outcomes even when everyone acts within an accepted baseline of behaviour. Drawing inspiration from the work of Iris Marion Young, Lu proposes that we supplement the interactional frame with a 'structural' one and offers an illuminating analysis of colonialism as a form of structural injustice. In her examination of the 1919 Versailles peace process, she points to how the interactional frame's focus on victims and perpetrators obscures the degree to which that process reproduced and perpetuated the colonial structural injustice of the global order. She argues persuasively that the structural frame allows us to see colonial injustice in a new light, making visible some aspects of it that persist until today.
The distinction that Lu relies on between the interactional and structural frames has practical implications for how we conceptualize and organize practices to achieve justice and reconciliation. Adopting the structural frame, for instance, reveals the CONTACT Farid Abdel-Nour abdelnou@sdsu.edu Political Science Department, San Diego State University, 5500 Campanile Drive, San Diego, CA 92182-4427, USA limitations inherent in practices of reparation for historical injustices and practices of accountability for war crimes (Lu 2017, 267). These practices, the significance of which Lu does not deny, suffer from limitations that are endemic to the interactional frame, most notably the focus on individual or group perpetrators. For instance, when relying on the interactional frame, one tries to deal with historical injustices, one invariably seeks to establish continuity between historical agents and contemporary ones. This immediately raises serious conceptual and moral difficulties. If in contrast, one were to adopt the structural frame in considering historical injustices, the focus would shift towards the need to 'transform persisting unjust contemporary social structures' (253). Once colonialism is understood as a form of structural injustice, attention can be focused on transforming its unjust structures that persist until today. Some of the transformations that Lu proposes include 'limiting the coercive rights of states to control the application of territorial borders to transboundary groups' (269), ending 'public recognition and valorization of historic figures who were active perpetrators of historic wrongs' (272), and 'pluralizing the agents that can have political standing in international and transnational institutions and structures' (277).
Lu recognizes that a focus on transforming persistent unjust structures requires allocating the responsibility for bringing about the needed transformation to identifiable agents. In this article, I focus on how she allocates this responsibility. It is a central part of her argument. Lu writes in the third chapter that 'how to think about responsibility for structural injustice is the focus of the rest of this book' (Lu 2017, 101 n. 42). Viewing colonial injustice through a structural frame, as Lu does, means that responsibility for it cannot be meaningfully concentrated on perpetrators, but must be allocated more broadly. For, within the structural frame, injustice does not involve wrongful action by perpetrators against victims, but structures and baselines of action that yield unjust outcomes. Lu diffuses the responsibility to transform unjust structures by allocating it to all the 'agents involved in producing and reproducing structural injustice' (235). 1 On her account, all those who participate in perpetuating and reproducing the unjust structures are responsible for transforming them.
I find Lu's way of allocating the burden to transform unjust structures compelling. I think, however, that this allocation can benefit from a clearer explanation. In conceptualizing responsibility for structural injustice, Lu relies on Young's work and as a result imports some of its difficulties, most notably the claim that responsibility for structural injustice is qualitatively distinct from responsibility for interactional injustice. Following Young, Lu draws a qualitative distinction between what Young calls the 'liability' and 'social connection' models of responsibility, the first of which she aligns with the interactional frame and the second with the structural frame. I argue, however, that adopting Young's distinction clouds matters and obscures the shared conceptual tools for thinking about responsibility that apply across the interactional and structural frames. To clarify matters, I point instead to two familiar ideas of responsibility that are implicitly at work in Young's and Lu's arguments, but that do not map onto the distinction between the liability and social connection models, or the interactional and structural frames. The two broad ideas of responsibility I discuss work in conjunction with one another across the above distinctions. Coming to terms with these two ideas of responsibility and their relationship to one another clarifies why Young and Lu allocate the burden for transforming unjust structures to those who participate in perpetuating and reproducing them. Lest there be any misunderstanding, I agree with Young and Lu that responsibility for transforming unjust structures falls on the shoulders of those who participate in perpetuating and reproducing those structures. All I seek to offer here is a path towards making sense of this allocation.
Before proceeding with my argument, it is helpful to consider an alternative way of broadly allocating the burden for transforming unjust structures, in order to bring out in relief a particular strength of Lu's argument, one that it shares with Young's. Lu briefly evokes the possibility of 'extend[ing] John Rawls's concept of a "duty of assistance"' (Lu 2017, 235), thereby pointing to a direction of argument that she does not herself pursue. In that direction, the burden to transform unjust structures would be allocated on the basis of a universal duty or obligation. Such an argument implicitly taps into a particular type of motive for action, namely, that of doing one's duty or obligation to make the world less unjust. Lu's and Young's direction of argument, however, taps into an additional motive for action. In allocating the burden of transformation to those who participate in perpetuating and reproducing structural injustice, their arguments implicitly tap into the motive of seeking to make good what one participates in making bad. By implicitly tapping into this motive for action, Lu's and Young's direction of argument does not diminish the significance of the motive to do one's duty or obligation to make the world less unjust. It supplements that motive with an additional one. This is a strength of that direction of argument.
However, the motive of seeking to make good what one participates in making bad is inextricable from a broad idea of responsibility that stands in the background of Lu's argument but that she does not make explicit. 2 Whereas she offers a discussion of responsibility in the third chapter (Lu 2017, 102-111), Lu does not in that discussion bring into view the idea that serves to explain the allocation of responsibility for structural injustice to those who participate in perpetuating and reproducing it. This is where her reliance on Young's qualitative distinction between the so-called 'liability' and 'social connection' models of responsibility clouds matters. Young sees her liability model as appropriate for legal and moral contexts in which we focus on the interactions between specific agents, identify perpetrators or wrongdoers, and seek to ascribe blame and liability to those agents for injustices associated with their wrongdoing. In contrast, Young sees her social connection model as appropriate for the context of political relations in which we are faced with structural injustices that emerge out of unjust background conditions and that do not require any agent to violate recognized action norms. In the latter model, blame and liability are not at issue. All that is at issue is the amelioration and elimination of the structural injustice (Young 2011, 95-122). The liability model maps onto Lu's interactional frame, and the social connection model maps onto her structural frame.
Young goes through great pains to insist that the conceptual tools appropriate for thinking about responsibility in these two models are qualitatively different (Young 2011 92, 97, 104). She does this because she is concerned about the possible incursion of the vocabulary of blame and liability into contexts of structural injustice. She worries that such an incursion would be counterproductive because it would bring the blame game into politics and would distract from the task of transforming unjust structures (115-119). For these reasons, Young insists on a qualitative distinction between the two models of responsibility, so that the only conceptual tool that she explicitly admits for thinking about responsibility within her social connection model is one that deals with remedying, ameliorating, or eliminating injustice. When considering structural injustice in other words, she bans all other conceptual tools with which we might think about responsibility within the liability model. I argue that Young is mistaken in insisting on this qualitative distinction. Important as it is to avoid the blame game when faced with structural injustice, Young goes too far in her attempt to guard against this possibility. Indeed, if we are to make sense of her own allocation of responsibility for structural injustice to those who participate in perpetuating and reproducing it (Young 2011, 104-110, 142, 146-147, 153), we must acknowledge the continuity in conceptual tools that she herself implicitly relies on across her two models. When Lu imports Young's insights about responsibility (Lu 2017, 104, 155, 170, 258), she imports as part of the package the assumption that the appropriate conceptual tools for thinking about responsibility within the interactional frame must be qualitatively different from those that are appropriate for thinking about it within the structural frame. But the insistence on such a qualitative distinction makes it difficult to understand her allocation of responsibility for structural injustice to those who participate in perpetuating and reproducing it. To understand that allocation, we must turn to a distinction between two broad and familiar ideas of responsibility that do not map onto the distinction between the interactional and structural frames. Rather, these two broad ideas operate in conjunction with one another and together help answer questions of responsibility that arise within both frames. As Lu points out, 'responsibility is a complex and "slippery" concept' (102). Nothing would be gained here from cataloguing the many uses of the term and all the analytic distinctions they imply. 3 For the purposes of clarifying Lu's (and Young's) argument further, it suffices to focus here on the two broad ideas of responsibility I discuss below.
The first idea of responsibility deals with agents' involvement in the causal process that produces a bad state of affairs. The second idea deals with agents' burden for correcting or improving a bad state of affairs. David Miller captures the two ideas I am focusing on here when he distinguishes between the concepts of causal and outcome responsibility on the one hand and the concept of remedial responsibility on the other (Miller 2007, 86-90, 101-102, 98-100). 4 Joseph Raz captures them when he distinguishes between what he calls responsbility 2 and responsibility 3 (Raz 2013, 228). My purpose here is not to offer a precise definition of two concepts of responsibility, but instead to point roughly to two intuitive ideas and to highlight a particular way in which they are related to one another. Once the relationship between them is brought into view, the direction of argument pursued by Young and Lu for conceptualizing responsibility for structural injustice becomes clearer. We are then able to better understand the allocation of the burden of responsibility for transforming unjust structures to the participants who perpetuate and reproduce those structures.
The first idea of responsibility comes into view most clearly in Bernard Williams' work. Tony Honoré, the first articulator of the concept of 'outcome responsibility', draws inspiration from Williams (Honoré 1999, 128 n.17, 125-138), as does Raz in his articulation of responsibility 2 (Raz 2013, 232-237). According to Raz, 'responsibility 2 for a state or an event presupposes that one brought it about, or contributed to its coming about or being preserved, or that one allowed it to continue… one is responsible 2 for a state or an event … in virtue of responsibility 2 for an action that brought it about, caused it, preserved it, or allowed it to stay in existence' (228 n.3). 5 The broad idea of responsibility to which Raz alludes here is that an agent's involvement via her actions in the causal process of a bad state of affairs has normative weight. 6 Williams captures this idea in nuce in a famous example in which a lorry driver who runs over a child through no fault of his own experiences agent-regret (Williams 1981, 28). 7 That example makes clear that the idea of responsibility under consideration here is so broad as not to have any necessary connection to either legal or moral liability or blame. Not only would we in Williams' example not think it meaningful to subject the driver to legal prosecution, we would also not think it meaningful to subject him to moral blame or finger-pointing. Rather, we would be inclined to feel sorry for him and to try to comfort him. Nonetheless, something would be wrong with the driver if he did not experience the agent-regret that marks the fact that he was an agent, and not a mere spectator, in the causal process that led to the running over of the child (Williams 1981, 29). As Williams' example illustrates, the intuitive idea under consideration need not evoke legal liability or moral blame. 8 At its core, this broad idea of responsibility amounts to an awareness by oneself and others of one's agency in the causal process of a bad state of affairs (Abdel-Nour 2016, 266).
On this idea of responsibility, thinking back to Lu's argument, those who perpetuate and reproduce a structural injustice bear responsibility. This means that they have reason to take to heart their agency in the outcomes of the unjust structure. On the moral plane, they might indeed be indistinguishable from spectators because it is only their bad luck that might have placed them as agents in the structural causes of injustice. There is a plane, however, one that is not devoid of normative significance and weight, in which they are clearly distinguishable from spectators. As participants 5 Both Honoré and Raz develop this concept for the purpose of deploying it in legal contexts. Miller develops it for the purpose of deploying it in moral debate beyond legal contexts. But, its underlying idea transcends both contexts. 6 Here, I am using the notion of involvement in the causal process broadly, to encompass Raz's bringing about, causing, preserving, or allowing to continue. 7 Williams' theory of responsibility is developed in Shame and Necessity (Williams 1993) but is prefigured in Moral Luck (Williams 1981, 28-29) which contains that example. For a brief account of Williams' articulation of this broad idea of responsibility, see Abdel-Nour (2016, 265-266). 8 Whereas it is possible for questions of strict liability to arise in this type of situation, they are premised on prior assumptions about who will be held liable if something goes wrong, or on public authority's purely pragmatic need to identify someone to respond. (Williams 1993, 57, 64) In other words, strict liability is not based on the normative weight that Williams brings into view when he focuses on the driver's experience of agent-regret.
in perpetuating and reproducing unjust structures, they, at a minimum, have good reason to experience something akin to the lorry driver's agent-regret. On that plane, something would be amiss if they did not have any such experience of having participated in making something in the world bad. 9 Williams' example artificially isolates for us this minimal idea of responsibilityone that is beyond law and beyond moralityso that we can see it without the interference of other related ideas. This broad idea is in the background of Young's and Lu's arguments. It must in fact be what underlies the move to single-out the participants in perpetuating and reproducing structural injustice for special attention. In the context of both of their arguments, the participants find themselves functioning as perpetuators and reproducers of injustice by virtue of their bad luck. And that might suffice to take them off the hook of legal and moral blame. 10 It does not, however, negate the agency they exercise in the causal process of structural injustice, which exercise Young and Lu focus on when they allocate responsibility to them. Outside the confines of Williams' example, the above broad idea of responsibility usually appears in conjunction with another quite different idea, to which I now turn. The second idea of responsibility I am interested in underlies the concepts of remedial or corrective responsibility, or what Raz calls responsibility 3 . According to Raz, this 'is like having a duty' to do something (Raz 2013, 228). When applied to a bad outcome or an injustice, this second idea of responsibility focuses on correcting the outcome and ameliorating or eliminating the injustice. This idea belongs on a continuum with the ideas of duty, obligation, and assigned or role-associated tasks. As Young points out, following Joel Feinberg and Henry Richardson, responsibility in this sense is simply a more flexible notion than duty. Bearing responsibility on this idea is like having a duty, except that one has more flexibility in whether and how one does what needs to be done (Young 2011, 91, 143). But how do agents incur even such a flexible burden to do what needs to be done? This second idea of responsibility does not by itself furnish the criteria for selecting who the agents are who will bear the burden. Those who are to be burdened with correcting, ameliorating, or eliminating the injustice are pre-selected. In other words, the allocation of the burden to particular agents rests on an antecedent step. There is no one justification for bearing the burden of responsibility to correct, ameliorate, or eliminate injustice. Rather, there are a number of potential justifications. 11 When faced with an injustice that we wish to see corrected or remedied, we speak of the responsibility for doing so. But whenever we apply this idea of responsibility in a particular case, we necessarily evoke the need for a selection criterion that identifies the agents whose business it is in this instance to correct the injustice. Such a criterion is necessary. Without it, the second idea of responsibility does not specify the agents who are to do what needs to be done. One 9 In this vein, Jade Larissa Schiff emphasizes the centrality of 'experiencing' responsibility. Although she does not mention Williams' work, they are both focused on capturing the same general intuition (Schiff 2015, 1-49). 10 When discussing the failure of the participants in an unjust structure to do their part to transform it, Lu is willing to blame them for such a failure (Lu 2017, 258-259), whereas Young is only willing to criticize them (Young 2011, 143-144). I discuss this later. Note that in both cases, the criticism or blame is not for participating in the unjust structure but for failing to do one's part to transform it once one becomes aware of one's participation. 11 Miller lists six potential justifications. We might identify those who have capacity to do what it takes, those who benefit from the injustice, those who are outcome responsible, those who are causally responsible, those who are morally responsible, or those who have special communal ties to the ones who suffer the injustice (Miller 2007, 101-104).
selection criterion that we can use (by no means the only one) is the first broad idea of responsibility discussed earlier. 12 Where an injustice is involved, each of the two broad ideas of responsibility I have discussed is associated with a distinct question, and the two questions stand in a specific relation to one another: the answer to the first can, but need not, serve as the answer to the second. The first idea is associated with the question: who, if any, are the agents through whose actions the injustice is brought about, preserved, or allowed to stay in existence? The second idea is associated with the question: whose business, if anyone's, is it to ameliorate or eliminate the injustice? Of the number of potential answers to the second question, one involves pointing to those who benefit from the injustice. 13 Another answer calls on everyone to shoulder a share of the burden by invoking a universal moral duty or obligation to ameliorate injustice in the world. 14 Yet another option, the one I am interested in here and that Lu and Young both use, is that of treating the answer to the first question as the answer to the second. This means that we select as bearers of the burden to correct, ameliorate, or eliminate an injustice those agents who participate in bringing it about, preserving it, or allowing it to continue in existence. When we adopt this option, we open up the possibility of tapping into a particular motive for action, namely, these agents' motive of seeking to make good what they participate in making bad.
The option of using the answer to the first question as the answer to the second is not specific to either the interactional or structural frame for conceptualizing injustice. It is available regardless of which frame we use. If the injustice is interactional, we look for wrongdoers as we do in a court of law or in a moral debate, the issue of blame and liability comes up, and we allocate the burden accordingly. If the injustice is structural, we look to those who perpetuate and reproduce it (even if they do so innocently, so to speak) and turn to them to ameliorate or eliminate the injustice. The questions associated with both ideas of responsibility, as well as the possible conjunction of answers are all part of the conceptual toolkit we deploy regardless of whether we adopt an interactional or structural frame for conceiving of injustice. 15 When Lu selects the agents who participate in perpetuating and reproducing unjust structures, and allocates to them the task of transforming those structures, she deploys both ideas of responsibility and uses the answer to the first question as the answer to the second. In doing this, she follows what Young does, not what Young says that she is doing. Contrary to Young's insistence, the conceptual tools of responsibility in the liability model and interactional frame are not qualitatively distinct from those in the social connection model and structural frame. 12 In terms of the six justifications proposed by Miller, this would entail selecting those who are causally responsible and those who are outcome responsible. 13 That is an answer that Daniel Butt offers and Lu rejects (Lu 2017, 150). 14 This is the answer that Lu evokes briefly but does not adopt, when she discusses Rawls' duty of assistance (Lu 2017, 235). 15 Blame and liability are not unavoidable considerations that somehow taint the entire conceptual toolkit. In an interactional context, we might ask about whether an agent's involvement in the causal process of a bad state of affairs reaches the minimal threshold that would render her blameworthy. If it does not, that does not mean that she is now no different from a spectator of that state of affairs. She is still an agent in its causal process and that makes her and other similarly situated agents potentially appropriate targets for the allocation of the burdens of remedy and correction. This can be the case even if there is no judgement of blameworthiness or liability.
When Lu allocates the burden for transforming structural injustice to those who participate in its perpetuation and reproduction, she marks them as appropriate bearers of the burden to bring about structural transformation. She does not select them because they did anything wrong, but because their non-blameworthy participation in perpetuating and reproducing an injustice, has a weight of its own. This is a weight that should at a minimum matter to them. Much as in the case of the lorry driver, something would indeed be amiss if they were indifferent to this weight. Contrary to Young's worry, that weight is not an attenuated form of blame or self-blame (Young 2011, 103-104). It is simply the weight we all bear by virtue of being agents in the world. From Williams' articulation of the first idea of responsibility presented earlier, we learn that participating in bringing about a bad state of affairs, even if one does not intend to do so, or even if one does not know that that is what one is doing, has a weight of its own. When an agent becomes aware of her involvement in the causal process of injustice, she becomes cognizant of this weight, and this can motivate her to seek to make good what she participates in making bad. Arguments like Young's and Lu's that allocate the burden of ameliorating and eliminating a structural injustice to those who participate in perpetuating and reproducing it tap into this motive for action, above and beyond any motive the agent might have to do her universal moral duty to reduce injustice in the world. The added motive they tap into, however, is only visible if we explicitly acknowledge the centrality of the first idea of responsibility I discussed earlier in that allocation of the burden.
Young's fear of sliding down a slippery slope that would inevitably lead to the blame game led her to insist on a qualitative distinction between the liability model and the social connection model of responsibility, which in turn gave Lu reason to mistakenly consider the second idea of responsibility sufficient for her argument. (Lu 2017, 153-154) But Young's insistence on that qualitative distinction obfuscated the common conceptual toolkit available to us for thinking about responsibility in legal, moral, and political contexts. The direction of argument that Lu pursues has a motivating force above and beyond the force of the motive of doing one's universal moral duty. This is because the agents her argument targets are expected to have an awareness of the normative weight of their non-blameworthy participation in perpetuating and reproducing structural injustice. Those who participate in perpetuating and reproducing unjust structures might not acknowledge this weight, or might acknowledge it in theory but not be motivated by it to do their part in transforming those structures. In such instances, they would open themselves up to being criticized by Young and blamed by Lu for failing to discharge their responsibility (Young 2011;144;Lu 2017;259). They would not be criticized or blamed for participating in the reproduction and perpetuation of the structural injustice (which is usually attributable to their bad luck), but for their failure to seek to make good what they participate in making bad. For Young and Lu, merely being a participant in an unjust structure does not make one the legitimate target of criticism or blame. But if after finding out that one is, one remains unmoved to do one's part to transform that structure, one does become a legitimate target for criticism and blame.
The first idea of responsibility I discussed earlier is central to Lu's argument. Indeed, her way of allocating responsibility for structural injustice relies on the specific relationship I outlined between that idea and the second idea of responsibility. Yet, all this is left implicit. Once it is made explicit, the reasoning behind her allocation of the burden for transforming unjust structures becomes clearer. Contrary to her suggestion and Young's, the second idea of responsibility does not by itself suffice to make sense of the important direction of argument that they both pursue. When they target those who participate in perpetuating and reproducing unjust structures, they both implicitly rely on the insight that there is a weight we bear for participating in making something bad, one that might not in itself render us legally or morally blameworthy, but that anchors a normatively significant motive for action. Lu and Young tap into that motive of seeking to make good what we participate in making bad, which is part of the appeal of the direction of argument they pursue. In doing so, however, they both implicitly acknowledge that there is continuity in the conceptual tools needed for thinking about responsibility in the interactional and structural frames, and not a qualitative distinction.

Disclosure statement
No potential conflict of interest was reported by the author.