1Iris Marion Young defines structural injustices as harms that occur when “social processes put large groups of persons under systematic threat of domination or deprivation of the means to develop and exercise their capacities, at the same time that these processes enable others to dominate or to have a wide range of opportunities for developing and exercising capacities available to them.”1 While this definition is historically general, Young herself repeatedly frames the importance of addressing structural injustices as a function of the specific organization of modern society. Here are two characteristic passages:
Structures are properties of social systems or collectives. In the modern world these social systems involve millions of people connected with one another through multiple systems of communication, and within complex layers of institutions that often have long-distance effects on one another.2
- 3 Young, 2011, p. 108. The role of complexity in Young’s social ontology is most developed in her ref (...)
[I]t requires a wider and longer-run reflective point of view to understand how the many people who act within these accepted rules and practices interact in complex ways to produce the outcomes that many agree are unjust.3
2In the first quote, Young claims that while social structures may characterize human society in general across historical epochs and different forms of social cooperation, they are nevertheless “properties” of actual social systems and collectives. As a result, the form and importance of social structures changes depending on the organization of a given society. She then characterizes modern societies as “complex,” a characterization she repeats in the second quote, where she explicitly connects the complexity of modern social systems with the difficulty of assigning responsibility for structural injustice.
- 4 Young, 2011, p. 110. Young makes the same point elsewhere in the chapter, see Young, 2011, pp. 99-1 (...)
3“Complexity” has a common, non-theoretical meaning. In this usage, something is complex because it is intricate and difficult to understand. Throughout her works, Young regularly uses the term complexity in this colloquial way. But here, I argue, Young uses the concept of complexity in a more theoretically laden manner. She claims that structural responsibility must be collective rather than individual in complex societies because “it is not possible for any of us to identify just what in our own actions results in which aspects of the injustice that particular individuals suffer.” The complexity at stake here is not merely that which confounds ordinary understanding, but a specific form of social organization that eliminates the possibility of determining individual causal contribution to social outcomes.4
4Young’s appeal to social complexity, however, introduces a problem for theories of structural responsibility. Traditionally, we feel comfortable assigning moral and political responsibility when we can identify a clear causal connection between a harm and an agent who has caused (or helped to cause) that harm. Moral and political responsibility are then grounded in the attribution of causal responsibility. But if the complexity of modern societies inhibits the assignment of causal responsibility for the systemic social processes and effects that Young identifies, then traditional models of moral and political responsibility cannot address responsibility for structural injustice.
- 5 The literature here is vast. McKeown, “Structural Injustice” offers a helpful overview of the field (...)
5In response to this challenge, many theorists of structural injustice have abandoned causally grounded models of responsibility, in favor of forward-looking theories which assign responsibility to repair social injustices without reference to agent’s previous misdeeds.5 While casually grounded blame may be useful for motivating the assignment of forward-looking responsibility, a causally grounded approach alone risks assigning responsibility to repair social injustices to actors who are unable or unwilling to do so. When this occurs the theory of responsibility becomes moralistic rather than justice-oriented, focusing on who did wrong rather than on how we can make the world better.
6In contrast, theories of structural injustice focus not on what agents have done in the past, but on what they are able to do in the future. Unquestionably, ability to repair is of paramount importance when considering structural responsibility, and the blossoming of causally ungrounded forward-looking theories has significantly expanded our resources for addressing structural injustice. But the move away from causality comes at a cost. Causally grounded responsibility is recommended by its clear intuitive basis: you have done something wrong, so you are responsible for repair. If we sever this connection, it becomes all too easy for agents who are able to address structural injustice to shirk responsibility by pointing to their innocence. Why should I, simply because I am able, sacrifice my time, energy, or resources to fix a problem that I did nothing to create? At the extreme, causally ungrounded forward-looking responsibility can seem supererogative—an encouragement to charity for those able to provide it—rather than something that one has an obligation to discharge.
7Many proponents of forward-looking responsibility explicitly refer to Young as a source of theoretical inspiration. However, despite Youngs’ clear appreciation of forward-looking responsibility, she does not herself unground it from causal contribution. Explaining what she calls the Social Connection Model, she writes:
- 6 Young, 2011, p. 105. Emphasis mine.
The social connection model of responsibility says that individuals bear responsibility for structural injustice because they contribute by their actions to the processes that produce unjust outcomes. Our responsibility derives from belonging together with others in a system of interdependent processes of cooperation and competition through which we seek benefits and aim to realize projects. Within these processes, each of us expects justice toward ourselves, and others can legitimately make claims of justice on us.6
8Young leaves us, therefore, with a puzzle. On the one hand, she asserts that causal contribution is an important criterion for the assignment of forward-looking responsibility for structural injustice. But in tension with this, she asserts (quoted above) that the complexity of modern social systems makes the attribution of causal responsibility difficult if not impossible in many contemporary cases of social injustice.
9In this paper, I attempt to chart a course between these horns by drawing on scientific theories of social complexity that were not available to Young at the time of her writing to distinguish structural and systemic injustice and responsibility as distinct forms of social injustice. To do so, I turn in the next section to Sally Haslanger’s systems theoretic ontology of complex dynamic social systems. I then introduce Robin Zheng’s criticism of and proposed alternative to Young’s theory of structural responsibility, which I find to be one of the most compelling positions in literature to date. I argue that while Zheng rightly recognizes that causal contribution is not a necessary condition of forward-looking responsibility for structural injustice, she fails to appreciate the role of complexity in motivating Young’s position, and thus fails to consider the limits that social complexity places on role-based theories of responsibility. Instead, I argue, we should distinguish between two forms of social injustice—structural and systemic—assigning forward-looking responsibility for structural injustices through a causally ungrounded role-performance based model similar to Zheng’s, while addressing systemic injustices through a causally grounded participation-based model inspired by Young. Together, these two models of forward-looking responsibility expand our toolkit for addressing social injustice, each excelling where the other meets its limits.
- 7 Haslanger defends a variety of realist semantic externalism about the objectivity of social kinds ( (...)
10Feminist philosopher Sally Haslanger has developed a robust social ontology that combines the tradition of structural theory exemplified in Young and her respondents with ongoing developments in the natural and social sciences broadly grouped under the heading of “complexity theory.” Haslanger’s project is motivated by much the same problem as Young: how to understand the objectivity of social groups and structures, the ways in which they are reproduced, and the forms of injustice that attend to them.7 Simultaneously, Haslanger’s own approach was developed in dialogue with Young, so that her social ontology can be seen as a development of the same problematic and ideas.
- 8 Young identifies Anthony Giddens’ Structuration theory as her source at (Young, 2011, pp. 60-61). S (...)
11Haslanger draws on the resources of systems theory to revise the distinction between system and structure, taken from Anthony Giddens, on which Young relies.8 As Haslanger defines the two:
- 9 S. Haslanger, 2022b, p. 3.
A system, broadly construed, is a set of things working together in a way that forms a whole. There are different kinds of systems and different ways to think of the relationship between systems and structures. I find it useful to think of systems as historically particular, concrete, dynamic processes; structures are the networks of relations that hold between the parts.9
- 10 S. Haslanger, 2022b, p. 14.
12Applied to society, social systems are collectives of people cooperating to “coordinate around the production, management, [and] disposal of things of (positive or negative) value.”10 Social systems (like natural systems) are themselves composed of integrated sub-systems, down to the level of individual persons who may themselves be analyzed as complex biopsychic systems. Sub-systems are integrated when they depend on each other for their functioning, so that the behavior of one sub-system influences the behavior of another.
13Consider a university. As a system, it is composed of pedagogical, administrative, custodial, transportation, security, information technology, and many other sub systems. For the university as a whole to function, these sub-systems must cooperate in a manageably stable fashion. The IT department must ensure that the computers are functioning so that the administrators can properly register faculty and students, who are then able to obtain identification that enables them to pass through the security system, and so on. If any one of these sub-systems falters or breaks down, it will impose stress on the sub-systems integrated with it, requiring those sub-systems to compensate or risk breakdown themselves. And if a significant number of sub-systems fail simultaneously, it may lead to the failure of the university as a whole. For example, if a wave of illness strikes the custodians leading to a disfunction in the custodial sub-system, this may lead to a build up of trash that makes it difficult for the faculty to teach. Teachers and students may compensate by cleaning up their own classrooms, but if the failure is severe enough, the other members of the university may be unable to effectively counteract the disfunction, leading to a shutdown of the university as a whole.
- 11 S. Haslanger, 2022b, pp. 6, 14.
- 12 S. Haslanger, 2022b, p. 4.
14Every social system, Haslanger argues, has a structure defined by set of social roles that determine the “choice architecture” of their occupants in relation to people who occupy the same or different social roles.11 Importantly, while social structures are abstract in that “the same relation [or role] can be instantiated by different individuals,” they are not therefore ideal.12 While systems are composed of specific concrete individuals, distinct systems may instantiate the same structure. Haslanger uses the nuclear family structure as an example. Two families may differ in the exact persons who occupy the parent and child roles, while nevertheless instantiating the same kind of parent-child relations, and thus the same structure. The nuclear family structure is therefore abstract, in that it is not defined by any specific family, but it still exercises a real causal force. Whether a family has a nuclear structure or is organized by a different set of social roles and relations plays an important causal role in determining how members of the family will be treated. For example, the nuclear family structure is organized around the eventual separation between parents and children, meaning that houses built for nuclear families are not designed to accommodate multiple generations living together over their entire lifespans. Elder care can, therefore, become especially costly and difficult within these homes, since it is not easy for adult caretakers to live alongside the elders they would care for.
15Haslanger argues that the ontological distinction between social systems and structures corresponds to a distinction between two kinds of injustice:
Structural injustice occurs when the practices that create the structure—the network of positions and relations—(a) distort our understanding of what is valuable, or (b) organize us in ways that are unjust/harmful/wrong, e.g., by distributing resources unjustly or violating the principles of democratic equality.
- 13 S. Haslanger, 2022b, p. 22.
Systemic injustice occurs when an unjust structure is maintained in a complex system that is self-reinforcing, adaptive, and creates subjects whose identity is shaped to conform to it.13
16Structural injustice is caused by the ways that social “positions and relations” misrepresent what is valuable and thereby encourage individuals who instantiate those positions to perform unjust actions. For example, a vegetarian concerned with structural injustices associated with industrialized agriculture, might argue that our culinary structures incorrectly characterize cows as a type of food, thereby causing individuals to harm cattle by factory farming and consuming them. As we will see, this is very similar to the theory of role-based responsibility that Robin Zheng proposes as an alternative to Young’s model of structural injustice.
17In contrast, systemic injustices are reproduced by the complex dynamics of modern social systems. Going beyond Young, Haslanger presents a theoretically precise notion of complexity that draws on the resources of systems science. She writes:
- 14 S. Haslanger, 2022b, p. 7.
A simple system (very roughly) is one in which the behaviours of the whole can be explained or predicted by reference a sequence of regular (linear) operations on its parts… Complex systems, in contrast, are not straightforwardly decomposable into independent parts, the operations of the parts are not necessarily linear, and they are self-organizing and stable due to feedback loops.14
18(Non)Linearity is the operative concept in Haslanger’s definition, explaining when and why systems are complex and determining whether and how they are predictable, decomposable, self-organizing, and stable. A system is linear when the behavior of the integrated sub-systems which compose it adds up to the behavior of the whole. Consequently, a change in the behavior of the system as a whole can be explained by a proportional change in each of the sub-systems which compose it.
19A simple production process can model a linear system. In the most basic case, different subsystems are isolated from each other. If eggs, milk, sugar, and flour are required to make a cake, and each of these ingredients is provided by a separate sub-system (e.g. separate chicken, dairy, sugar cane, and wheat farms) then the total number of cakes produced will be a direct function of the number of raw ingredients provided by each farm. But the behavior of linearly integrated subsystems does not need to be entirely unrelated. Suppose that the cake factory contains two sub-systems, one which mixes the ingredients and outputs a batter and a second which takes that batter as an input and bakes it into cake. Here, the behavior of the baking sub-system will depend on the behavior of the mixing sub system, so that a disruption to mixing will affect baking. Nevertheless, because the integrated system is linear, the overall behavior of the cake factory could be determined without considering the relation between the mixing and baking system. If we know the number of batters input into the baking system, we can calculate the number of cakes baked, without any reference to the mixing system itself. While the second system depends on the first, they relate to each other in a direct way that enables their decomposition. As we will see, this is not the case in non-linear systems.
- 15 For more examples of non-linear dynamics see Mitchell, 2009, pp. 39-41; Mobus & Kalton, 2014, pp. 1 (...)
- 16 For a definition of the liability model see Young, 2011, p. 97.
- 17 Young, 2011, p. 96.
20A system is non-linear when its sub-systems reciprocally influence each other, in which case we say that those sub-systems are co-integrated. For example, some sub-systems may be formed of the same components or some sub-systems may take each other’s outputs as inputs in a circular manner.15 The non-linear dynamics that produce social complexity prevent us from explaining social outcomes as the result of the actions of specific agents, whether individual or collective. While we may be able to identify the agents who trigger these outcomes, as, e.g. the person who screams fire may trigger a stampede—and while this identification is often a proper basis for the attribution of certain kinds of backward-looking responsibility (what Young calls “liability”16)—we cannot, in principle, fully explain the social outcome through those agents’ actions alone. Instead, to causally explain social outcomes, including social injustice, in complex systems we must make reference to the relations between the many individuals who make up the system. In a literal sense, in complex systems, “all those who contribute by their actions to structural processes with some unjust outcomes share [causal] responsibility for the injustice.”17
21Returning to the cake factory example, if the revenue from selling finished cakes from the baking sub-system was used to purchase more ingredients for the mixing sub-system, then the overall cake factory would be a non-linear system. The behavior of the mixing and baking sub-systems would be co-integrated so that a failure in the mixing sub-system would disrupt the baking subsystem, which would in turn reduce the number of cakes sold, which would then decrease the number of ingredients purchased, which would further reduce the output of the mixing sub-system, and so on. Here, it would not be possible to decompose the overall system into isolated sub-systems, because the performance of each sub-system is a function of its own behavior as mediated through the other sub-systems. And in this particular example, we can see the emergence of positive and negative feedback loops that may lead the cake factory away from equilibrium towards rapid growth or spiraling decline.
- 18 My definition differs from Haslanger’s definition of systemic injustice because the connection to s (...)
22With this definition of complexity on board, we can distinguish systemic from structural injustice. Systemic injustice occurs when some form of social injustice is maintained by the causal relations that reproduce a complex system.18 As this definition indicates, systemic justice may arise even where the structures that define the roles and relations between individuals or groups are not themselves constitutively unjust. Systemic injustice is, therefore, an independent form of social injustice that, I will show, requires a distinct theory of forward-looking responsibility.
- 19 See, e.g. Mau, 2023; Postone, 1996; Roberts, 2016; Vrousalis, 2022.
- 20 As this example demonstrates, recognizing the distinction between systems and structures reveals th (...)
23For an example of systemic injustice, consider the market socialist argument that public ownership would, by abolishing the power of private capital, abolish the structural relation between owner and worker that Marxists argue is unjust. Market socialism would have an economic structure whose roles and relations were not constitutively unjust. In response, several value-theoretic political economists argue that in a sufficiently complex market system the causal dynamics of the price system would nevertheless produce unjust forms of exploitation and domination.19 These injustices would be properly systemic rather than structural, even if responding to them would require transforming the economic structure of society.20
24Recall that Young abstracts from the specific roles that individuals occupy in social structures, attributing responsibility to repair solely because individuals causally contribute to the reproduction of social injustice within in a complex system. Considered in light of Haslanger’s distinction between system and structure, the kind of social injustice at issue in the Social Connection Model includes the systemic injustice that is produced by the complex relations between individuals in modern societies, and not merely the structural injustice that results from the way social roles are related.
25Systemic injustice is produced in a different way than structural injustice and therefore requires a different kind of response. To demonstrate this point, I now turn to Robin Zheng’s account of structural responsibility. In developing a robust theory of structural responsibility based on role performance rather than causal contribution, Zheng goes beyond Young and advances our understanding of social injustice and forward-looking responsibility. However, in ignoring Young’s comments about complexity, Zheng leaves an important dimension of the Social Connection Model behind, and consequently lacks a theory of systemic injustice and responsibility. Putting the two together, I conclude, yields a more comprehensive and robust theory of social injustice and forward-looking responsibility than either alone.
26In a series of papers, Robin Zheng criticizes Young’s Social Connection Model and proposes an alternative model based on role performance independent of causal contribution. Zheng bases her critique on the distinction between two forms of responsibility, which she calls “attributability” and “accountability.” Of the difference she writes:
- 21 Zheng, 2016, pp. 62-63.
Responsibility as “attributability” depends on the notion that actions are expressions of our agency. We are morally responsible for our actions in this sense only when they reflect who we are as moral agents—that is, when they are properly attributable to us as manifestations of our ends, commitments, or values. Responsibility as “accountability,” however, depends on the social and institutional practices governing the distribution of duties and burdens across different roles and positions within a moral community. We are morally responsible for our actions in this second sense when it is appropriate for others to enforce certain expectations and demands on those actions—in other words, when it is appropriate for others to hold us accountable for them.21
27Zheng argues that attributability is not appropriate to structural injustices, where showing that “an agent caused harm knowingly and voluntarily” is difficult and contentious. For the same reason, she also rules out backward-looking theories of responsibility for structural injustices. Instead, she argues, that a theory of structural responsibility should be based on accountability for structural injustices, which looks forward to who in the community is best positioned to redress the harm.22
28Having severed the connection between injustice and causal contribution, however, she must reestablish a basis on which individuals bear responsibility as accountability. She writes:
Being assigned such a burden is not grounded in a person’s possessing the capacity to exhibit good or bad agency. Instead, it is grounded in the fact that oppressor, oppressed, and bystander all participate in social structural processes, albeit in different ways.23
29In the second sentence, Zheng claims that forward-looking responsibility for structural injustice is grounded in an agent’s “participation” in unjust “social structural processes”. She clarifies her use of participation here subsequently, when she contrasts the SCM with her proposed alternative, the Role Ideal Model.
- 24 Zheng, 2018, p. 874. The specific way that Zheng thinks about role ideals and their role in normati (...)
On the SCM [Social Connection Model], individuals are responsible for unjust outcomes because of their causal contributions to structural processes. By contrast, the RIM [Role Ideal Model] maintains that individuals are responsible because their role performances are what constitute unjust structures. This distinction between causation and constitution is subtle, but important.24
30Rather than basing accountability in an agent’s causal contribution to structural injustice, Zheng argues that it arises from the fact that individuals constitute social structures through the performance of the roles that define them. Zheng’s distinction between two ways in which agents participate in the reproduction of social injustice is analytically insightful. Causally, structures may incentivize individuals to act in unjust ways, as when quotas encourage workers to prioritize productivity over safety. But structures only exist and have power insofar as individuals accept the roles they have been given and act in accordance with them. Even where these individual actions do not directly cause injustice, they nevertheless constitute the structure, reinforcing its influence over society. Constitution through role performance thereby provides a basis for the assignment of forward-looking responsibility by identifying a non-causal relation of dependence between an agent’s actions and structural injustices. But as Zheng argues, role performance cannot establish backward-looking responsibility for specific structural injustices, instead, it shows that an agent has some power over the reproduction of the structure, so that it establishes their forward-looking ability to reform structures and make them just. As a result, Zheng argues that the performance of social roles can ground an agent’s accountability for structural injustice independent of any causal contribution thereto.
31I find Zheng’s argument compelling. Forward-looking accountability-oriented responsibility based on role performance is not well captured by the Social Connections Model due to Young’s persistent claim that the SCM depends on causal contribution. And I agree with Zheng that role-based theories of responsibility are better suited to specifically structural injustices. I do not, however, believe that we should therefore abandon the SCM in favor of Zheng’s Role Ideal Model or any model of specifically structural injustice and responsibility. Instead, I claim that Young was attempting to stake out a middle ground between individualistic backwards-looking theories of responsibility of the kind Zheng calls attributive and causally unmoored forward-looking theories of role-based accountability like Zheng’s.
32The unique non-linear causal dynamics that produce systemic injustice complicate the sharp dichotomy between having caused and being able to repair that motivates Zheng’s ungrounding of forward-looking responsibility from causality. As we have seen, in complex social systems it is in principle impossible to assign causal responsibility for systemic injustices to specific individuals considered in isolation. At the same time, systems theory shows us that individuals do contribute to the emergence of systemic injustices through the non-linear causal contribution their actions make to the emergence of complex dynamics. Systemic responsibility is assigned because of an agent’s causal contribution to injustice, but this contribution is not linear as Zheng presupposes it must be.
33Because systemic responsibility is not assigned on the basis of role performance, it does not, properly speaking, assign responsibility for structural injustice. But like structural responsibility, systemic responsibility is collective and forward-looking. Because agents’ causal contributions to systemic injustice cannot be isolated in a complex system, the responsibility one bears for causing systemic injustice cannot be discharged through a responsibility to repair a specific wrong for which one is individually guilty. Instead, to discharge systemic responsibility, agents must work together to transform the dynamic organization of the system in which they participate. And this requires them to look forward to their and other’s abilities to repair systemic injustices through collective action rather than backward to assign blame.
34Agents can work together to transform an unjust system in two ways. First, they may collectively organize to change the structure of the system. Complexity emerges from the non-linear relations between agents whose behavior is also shaped by structural roles and relations. Changing a system’s structure can thereby reshape the non-linear dynamics that give rise to complexity so that they do not also reproduce injustice. This strategy may be effective even when the system as a whole is not structurally unjust, i.e. where the roles and relations that define the system do not themselves encode injustice (as for example, where the same structure would be just in a simpler social system).
- 25 These strategies for collectively responding to systemic injustice are inspired by S. Fard, 2022; S (...)
35Second, agents can coordinate to collectively change their behavior even in the absence of structural change. While performing the same roles, agents may consciously choose to ensure that they act in ways that inhibit the formation of the non-linear dynamics which give rise to systemic injustices. For example, agents who benefit from self-reinforcing first-mover preferences that produce systemic inequality (as in many social media algorithms) may choose to collectively redistribute the resources they receive from these dynamics (like exposure). While a single agent redistributing their resources would be unlikely to affect systemic change, if enough agents follow the same redistributive strategy, they may collectively undermine the feedback loops that produce and exacerbate inequality.25
36This explains why Young emphasizes the importance of “organizing with others.” Though systemic responsibility is grounded in the fact that an individual agent causally contributed to injustice, this grounding relation only obtains because other agents acted in concert as part of a complex dynamic system. Because systemic injustices cannot be produced alone, neither can they be addressed alone. A systemically responsible agent must therefore look forward to their ability to cooperate with others to collectively transform the complex dynamics which emerge from their co-integrated actions.
37So is Young a theorist of causally grounded or ungrounded responsibility? Both and neither. Young was a creative and daring thinker, who drove theories of social justice to new and uncharted terrain. In between the clear poles of demonstrable individual causal responsibility for specific harms and social accountability to rectify them she found the subtle space of complexity, where important but easily overlooked forms of non-linear causality preserve the intuitive connection between responsibility and causal history while moving beyond the individualism that limits the usefulness of the liability model.
38To conclude, I introduce an example in which systemic and structural injustice come apart. Consider the racial history of housing provision in America. Decades of racist redlining perpetuated by banks following the prohibition of even more egregious discrimination through racial covenants has left us with a legacy of residential segregation. And in a country where social services and opportunities are directly based on or strongly correlated with zip-code, this discrimination has become a powerful engine of continuing racial injustice. This is a clear situation of structural injustice: the social roles and relations which make up the structure of the American housing system encode relations of inequality, oppression, and subordination that target Black homebuyers.
- 26 Of course, there may be important non-racial reasons for thinking such a lottery would be structura (...)
39So let us imagine a bold structural reform: the establishment of a robust public initiative that builds and provisions quality housing to individuals regardless of income or wealth through a lottery system. For simplicity’s sake, assume that existing housing stock is destroyed wholesale and replaced with egalitarian public housing projects (whether these be individual houses, townhouses, an apartment complex, a co-op, etc.) and that individuals are all assigned new residences. Insofar as lotteries are indifferent to race, the results of the reassignment would not encode historical racial injustices into the structure of the new housing distribution. The roles of provider and resident that make up the housing structure would make no reference to race, nor to income or wealth as a proxy for it. The new housing structure would, therefore, be structurally just from a racial perspective.26
40However, the lottery system itself could encourage the emergence of racialized systemic injustices. Black people constitute a minority and White people a majority in American society. As a result, a random lottery distribution would be likely to construct residential communities in which Black people were surrounded by relatively few other Black people and a majority of White people. It’s certainly possible that the lottery randomly constructs a majority Black neighborhood, in which case the same systemic dynamics discussed below would disadvantage the White minority, but given the demographic distribution of American society, this is improbable. Black, but not White people, are likely to be racially isolated by the lottery.
- 27 This example and argument are inspired by O’Connor, 2019, pt. 1.
41If we assume that racial isolation is not itself unjust, then there is nothing structurally wrong with this situation. Racial isolation may, however, lead to the emergence of unjust dynamics within residential communities. Theorists of complex social systems have shown that in societies with feedback loops between the outcome of cooperative exchanges and one’s initial bargaining position in subsequent interactions, small inequalities of bargaining power between groups can, over time, become self-reinforcing. Modeling helps us to understand and explain how these dynamics arise in concrete detail, but the basic idea is straightforward: If power differentials are fed by positive feedback loops then small initial differences can be reinforced to become larger and larger, with each transfer of power from the subordinate towards the dominant group making it easier for more power to be transferred in the next cycle and harder for any to be reclaimed.27
42These same theorists have argued that we do not need to presuppose any conscious or unconscious prejudice or wrongdoing to generate the appropriate kind of power difference. The mere existence of preferences for in-group relations are sufficient. In-group preferences can arise without any kind of obviously unjust bias. Familial or friendship ties, cultural commonalities, shared experiences and traditions, and much more may all give rise to in-group preference. And these preferences are all that’s required to jumpstart the complex dynamics that give rise to systemic injustices. The housing example will show us how this can lead to the emergence of racialized systemic injustices without any initial racial biases.
43Assume that individuals prefer to cooperate with members of their own group. In this case, just by virtue of being in the majority group, an individual will be presented with more bargaining options. As a result, members of the majority are less likely to say yes to a bad bargain because they are more likely to be able to substitute other ones. Consequently, members of the majority group approach each other with relatively strong bargaining positions, encouraging more egalitarian cooperative relations. Conversely, members of the minority group have relatively scarce choices just by virtue of being in the minority. As a result, they are more likely to be forced to accept worse bargains. Applied to housing, this could reflect the help a person gets from their neighbors in performing home repairs, decorating, or even just moving furniture. Members of the majority group will be more likely to find someone to help them repaint their apartment or pull in that really nice but ridiculously heavy armoire from the curb than members of the minority. Over time, this will make them more likely to have nicer homes.
44As a result, when a member of the majority bargains with a member of the minority, they will be less willing to accept deals which are not fair for them. Conversely, members of the minority, having fewer and worse options, will be more likely to accept a bad deal in order to obtain what they want from the bargain. This asymmetry enables members of the majority group to use the better bargaining position deriving from their majority status to force members of the minority group to accept unfair deals. This may result in exploitation even without any intention to do wrong on the part of the majority. Each bargainer is simply seeking the best possible trade for themselves. And in the extreme, this kind of exploitation can easily lead to domination. From “I take more than I give” the road to “I command and you obey” is easily walked.
45Consider two individuals who currently live in lottery housing but want to move to a nicer apartment they received in the latest round. The first is a member of the White majority. By virtue of their majoritarian status, they are more likely to have friends, family members, or communities who share their interests in the area and who would provide them with help packing and moving. Still, the White resident may decide they would prefer not to impose on their community and choose instead to hire a moving company. Nevertheless, when they bargain with a moving company, the fact that they could ask their friends for help will give them a distinct advantage. If the moving company charges them an exorbitant rate, treats their possessions without care and respect, or otherwise offers them an unfair bargain, the White individual will be able to refuse and accept the social costs of leaning on their friends.
46Compare this with the situation of a randomly distributed Black person. They will be less likely to have family in the area or to know people through their friendship network. Insofar as Blackness affects cultural identity, there will likely be fewer people who share their hobbies or taste in music, food, dance, celebration, and the like, and so cultural community will also be scarcer. As a result, a Black person will be more likely to be forced to use a moving company and therefore more likely to have to accept a worse offer from the company, whether that means paying more or receiving lower quality service.
47Over time, this dynamic can easily become entrenched. If the ease of one’s move affects future cooperative prospects, then the gap between the two individuals is likely to increase. The White person, having spent less time and money on the move, will be able to invest more in spending time with their neighbors, going to local cultural events and getting to know their new community. Conversely, the Black person will be bogged down with the move and working to recuperate costs. As a result, the White individual is more likely to grow their support network, further strengthening their bargaining position in the future, while the Black individual will have a harder time doing so. This can cascade and lead to more explicit injustices as above, for example if the two individuals subsequently engage in a bargain where the White person, intentionally or not, uses their better bargaining position to impose unfair terms on the Black person.
48These examples show that the random distribution of the lottery system combined with the demographic distribution of American society can give rise to racialized systemic injustices even in the absence of racial structural injustices or even implicit racial bias. The roles that make up the structure of the housing system make no reference either explicitly or implicitly to race, nor do they encode relations that structurally subordinate racialized groups (like wealth or education). There is nothing about the roles of participating in a lottery system or being a member of a minority group that are in themselves racially unjust, nor do they constitute racial injustice in their structural relations to other roles like resident, majority group member, and the like. Instead, these injustices arise from the non-linear processes that shape the dynamics of a complex housing system. They are an independent form of systemic injustice.
49Finally, consider the assignment of forward-looking responsibility for this systemic injustice. On the one hand, we can clearly identify how members of the White majority have causally contributed to the emergence of systemic injustice. They have willingly participated in the lottery system, actively preferred to socialize with people like themselves, and chosen to rely on their support networks to strengthen their bargaining position. Considered in isolation from the broader systemic context, none of these actions are themselves harmful and therefore unjust or blameworthy. White individuals did not intend to disadvantage or harm their Black neighbors, nor did they perform actions that could reasonably be expected to do so when considered in isolation. Only when considered together and in the context of the complex dynamics that arise from a lottery system can we clearly understand how the White majority acts to cause the emergence of systemic injustice. As a result, we are able to preserve the intuitive idea that, having caused injustice, the White majority is responsible for helping to rectify it, while avoiding moralistic or blame centered approaches.
50At the same time, the fact that White people (partially) compose the housing system means that they are able to help repair the injustice. The community may opt to self-organize with the goal of restructuring the lottery system itself and replace it with a more systemically just process of housing distribution. Or, White individuals may choose to collectively rethink who they choose to connect with and support in their communities. Alone, a single white person building relationships and offering help to their Black neighbors is unlikely to undermine the emergence of systemic injustice. However, if enough White people decide to do so, then, even in the absence of changes to the structure of the housing system, the feedback loops that feed to emergence of systemic injustice may be overcome.
51Without a theory of systemic injustice, the lottery system, by virtue of its randomness, may seem to be a socially just way of distributing housing. Our example shows, however, that the structural justice of the lottery system does not prevent it from reproducing systemic injustices. The same argument could be extended to other proposed lottery systems, e.g. for jobs, political representation, resource distribution, and the like. And the positive feedback loops that fuel systemic injustice in the lottery system are just one form of potentially unjust complex dynamic. First mover biases, scale-free or power law distributions, bifurcation dynamics, path dependence, hysteresis, and homeostasis are but a few of the many different complex dynamics that may all give rise to systemic injustices under the right circumstances. Recognizing systemic injustice and responsibility broadens both our horizon of possible harms, allowing us to recognize injustices that are invisible to pure theories of structural responsibility, and expands our toolbox for thinking about how we can respond to them.
52I have argued that the distinction between systemic and structural injustice implies a distinction between systemic and structural responsibility and I have argued that Young’s Social Connection Model should provide a model of the former. Our appreciation of the subtleties of social injustice and of the different possible vectors along which we can attack it is expanded by the recognition of systemic injustice and responsibility as distinct from both individual and structural justice and responsibility. If we fail to recognize systemic responsibility, our proposals for structural transformation are likely to be self-defeating at best and downright pernicious at worst. If we fail to recognize structural responsibility, our proposals for systemic transformation will be inadequate and lack orientation. Only by approaching problems of social justice with a dual vision can we propose solutions adequate to the problems of social justice that we face in modern complex societies.