Question (1): What is a Social System?

Very roughly, social systems are ways in which human interaction is organized and structured. To get more precise, it's worth making a distinction between two kinds of social systems: institutions and social structures.

While this is not the only way to understand how these notions relate to each other (see Fleetwood Reference Fleetwood2008 and Porpora Reference Porpora1989), it suffices for our purposes to distinguish between systems that are more formal and those that are less formal. Let's explore what this distinction amounts to.

Start with institutions. Institutions are formal organizations insofar as they are governed by explicit rules, laws, by-laws or policies. The formal rules etc. that govern institutions enjoy widespread recognition in the society, ranging from enthusiastic endorsement to hostile acknowledgement (Searle Reference Searle2010). The sorts of institutions we will focus on are made up of various positions that individual people can occupy, for example, a central bank, a sports team, a religion. These various positions ‘constrain and enable’ their occupants (Hardimon Reference Hardimon1994; Hodgson Reference Hodgson2006; Ásta Reference Ásta2018). Constraints and enablements should be understood broadly to include rights, obligations, powers, privileges, prohibitions, sanctions, etc., really anything that allows or disallows human action, in a broad sense. These positions help institutions function by providing people with roles or jobs in the institution. What distinguishes these positions from one another is the different ways in which they constrain and enable their occupants. A CEO, CFO, board member, shareholder, manager and employee, for example, all have different roles in the functioning of a corporation. Each role comes with prohibitions, freedoms and requirements. It's important to note, too, that institutions don't just constrain and enable those internal to the institution. They also constrain and enable those outside it – users, benefactors and victims – who are affected by the institution. Amazon customers, recipients of mail delivered by the United States Postal Service, and lobbyists for the coal industry all play an important role in the functioning of different institutions. Despite being ‘external’ to institutions, those who play these roles still occupy a position that comes with constraints and enablements.

What about social structures? In the social sciences, the notion of a social structure has a long history. There are many different conceptions of social structure, but here is one general way of thinking about the notion. Like institutions, social structures serve to organize and structure human interaction in a society. They involve various social positions that relate people to one another and to material resources (Haslanger Reference Haslanger2016, Reference Haslanger2022). These positions also come with different constraints and enablements that organize our behaviour and thought. But rather than explicit rules, laws or policies, these positions are characterized by more informal constraints and enablements, for example, norms, expectations, ideals, schemas, meanings and stereotypes. For example, they may provide ideals for us to live up to; they present certain courses of action, such as pursuing a certain career, as open or closed to us; they may offer censure or praise for different behaviours. The social relations that make up social structures can often be hierarchical and unequal. Some positions advantage their occupants while other positions disadvantage their occupants. Social structures function to organize human life by prescribing a ‘division of labour’ that opens or closes various forms of behaviour, thought and opportunity (Haslanger Reference Haslanger2016).

Think back to when you were in high (secondary) school. There were various groups of people who occupied different positions: nerds, jocks, band kids, goths, emo kids, hipsters, whatever. All of these positions came with a set of expectations, demands and boundaries. They prescribed certain kinds of behaviour, dress and language. They also came with different forms of reward and censure, for example, the more obscure the music the better for hipsters. All this can be more or less hidden from us. We may not see this structure in much detail, even if we feel its pressure on us to conform.

Unlike institutions, social structures don't typically need the widespread recognition that institutions do to operate. All they need to persist is the repetition of certain kinds of social interaction. Elizabeth Barnes (Reference Barnes2017) likens social structure to a rut made by the travel of wagon wheels. The rut is created by the repeated patterns of travel. But, once created, the rut directs where the wheels go, helping to explain why wheels travel as they do. By analogy, social structure is created by patterns of human social interaction. But once social structure is created, it exerts pressure on subsequent human interaction, helping to explain why humans act as they do. This pressure need not be formally enforced or recognized since it can arise, even unintentionally, from the ways in which humans interact with each other.

So, here is our answer to question (1): social systems – whether institutions or social structures – are more or less formal ways of organizing ourselves that involve positions or roles that constrain and enable us.

It is crucial to note that institutions and structures never exist in isolation from each other. All social systems exist entangled with others: they rely on each other, they perpetuate each other, they undermine each other, and so on. A university, for instance, is an institution, but it has embedded in it a social structure as well. ‘Professor’, while a formally recognized position in the institution, comes with informal constraints and enablements: professors wear certain kinds of clothes, they tend to be male and white, and can be ‘absent minded’ in ways that other university employees are not permitted to be (thanks to Elanor Taylor for this example).

Question (2): How Can a Social System be Racist?

It is worth, first, clarifying our second question. Our question is not whether individual or systemic factors are primary in explaining racism. It is reasonable to acknowledge that all forms of racism involve a complicated interplay between individuals and the systems they live within. Our question is rather, granted that racism can be systemic, what is it for racism to be systemic? And what forms can systemic racism take?

Systemic racism is a feature of social systems, not of the individuals who help compose the system (though they play a role, as we will see). For a social system to be racist, roughly, is for it to cause, perpetuate or amplify the oppression of a certain racial group (Haslanger Reference Haslanger2012: 317). Systems can contribute to racial oppression in a variety of ways.

The first form we'll consider begins with the observation that social systems are partly built from individuals, their beliefs, practices and actions. That is, individuals help constitute social systems; they are occupants of these systems and without them the systems would not exist. Given this, it makes sense to trace the racism of a system to the individuals within a system:

This analysis makes systemic racism continuous with our ordinary way of thinking about racism as a matter of the ‘heart and mind’ (see Garcia Reference Garcia1996) since the racism of the system originates from the individuals in the system. Imagine a white supremacist intermural football team. All the members of the team have racist attitudes and beliefs, and perform racist actions. I don't think we'd be blamed for calling such a team racist. It's important to note, however, that this sort of racism isn't first and foremost systemic or structural. Yes, the individuals contribute to racial oppression (to whatever extent) in their capacity as members of the team. But the rules of football, the by-laws of the league and the tournament structure aren't racist, only the members of this particular team. (Though the league may rightly be criticized for tolerating such a team.) If one were to rid the league of all its racist players – its ‘bad apples’ – the system would, arguably, no longer be racist at all. While we shouldn't reject (SR1) as one sort of systemic racism, it is not all we mean by ‘systemic racism’.

Above we considered the different ways in which systems could be organized and maintained. If we want to understand systemic racism, it makes sense to look at the nuts and bolts of the systems themselves. Institutions and structures, we've said, operate by putting various constraints and enablements on those who occupy positions in the system. Hence,

Constraints and enablements can themselves be racist when they encode a racist ideology (Shelby Reference Shelby2002). To encode a racist ideology is for the norms, policies, laws and rules of a system to express or entail commitment to the inferiority and subjugation of members of a race. But systems don't just encode and express values, ideologies, etc. They also serve certain functions; they have real effects in the world. To the extent that racist constraints and enablements determine the effects of a system, that system contributes to racial oppression.

For instance, Jim Crow laws in America after Reconstruction (1865–77) were specifically designed to express and establish white dominance. Bank policies that denied black people loans and real estate policies that disallowed black people from buying houses in certain neighbourhoods similarly expressed a commitment to white supremacy and had the effect of disadvantaging black people (Rothstein Reference Rothstein2017).

These sorts of racist systems not only required their original designers to have racist beliefs, but also involved explicitly racist rules, laws, policies and norms. However, in one sense, the present-day individuals within these systems are not vital to its functioning: so long as the system has occupants following its rules, it will function to bring about racial domination. It is easy to see how institutions can be racist if the formal rules of the institution are racist. But fewer institutions these days have overtly racist policies built into them. Of course, some still do. In America, law enforcement institutions have chosen (and been incentivized) to target communities of racial minorities in the war on drugs, with tragic consequences: black men are twenty-one times more likely to be killed by the police than white men; the prison population has exploded, overrepresenting black and brown people (see Gabrielson, Sagara and Grochowski Reference Gabrielson, Sagar and Grochowski2014).

Social structures are made up of social relations. They function by positioning individuals into roles with different and unequal constraints and enablements. This often gives rise to unequal power differentials among individuals and groups. A social structure that builds in unequal norms for different positions would be racist if those norms disadvantaged (socially, economically, materially, etc.) people of colour and advantaged white people. Think of the following stereotypes that generate norms and expectations:

• • Black criminality.

• • Black women as the ‘mammy’ or ‘jezebel’ or ‘welfare queen’ (Hill Collins Reference Hill Collins2009).

• • Asians as the ‘model minority’, for example, respectful to authority, excelling in science and maths.

• • Lighter skin as more beautiful (Taylor Reference Taylor2016).

These stereotypes function to keep people ‘in their lanes’. That is, they establish (sometimes contradictory) expectations and norms for behaviour, ideals for people to live up to, and they enforce punishments for deviation. Insofar as these stereotypes function to socially benefit some racial group over others, such a structure is racist.

Because social structures are more informal than institutions, this raises the possibility of systemic racism without individuals with racial hatred or disregard. (The possibility of ‘racism without racists’ is explored in Bonilla-Silva Reference Bonilla-Silva2013.) Social structures are rooted in regularities in social interaction. Because the rules that make up these structures need not be formally enforced or acknowledged, people – typically privileged people – can live and move within these structures without noticing them. Recall Barnes's example of the wagon wheel and the rut: ruts form by the repeated movements of wheels, but once the rut is created it directs all subsequent wheels into the rut. The path of the rut will be the most natural and comfortable for those riding in the wagon. Similarly, the social organization provided by structures may appear to us to be natural, normal and comfortable. Going with the flow of a social structure requires no effort, only that people do not resist that flow. Therefore, the operation of social structures that result in the dominance over people of colour don't require overtly racist individuals, only compliant passengers. Of course, oppressed groups don't have the luxury of being compliant passengers. As the ones being crushed by the wheels of our social structures, they are often the most vocal and clear-eyed critics of these structures.

It's worth working through an example of this point. Many institutions of higher learning were founded by individuals with racist beliefs and attitudes. These beliefs influenced the established policies, rules and norms of these institutions, for example, the exclusion of non-white students, faculty and administrators. These policies are no longer formally recognized or enforced. But the social structure embedded in many universities perpetuates the effects of past constraints and enablements. The individuals in these structures – professors, administrators, students and staff – don't need to share those racist beliefs to perpetuate the system. They need only perpetuate the patterns of interaction of the institution. Professors need only teach the curriculum they always have and to let stereotypes guide pedagogy: Asian students are mainly interested in science and maths; black students are not interested in highly theoretical disciplines, etc. Hiring committees need only focus on hiring the best ‘talent’. Administrators simply need not be aware of the unique obstacles facing students of colour on campus. Students need only to maintain the same segregated friend groups, all white sororities/fraternities, etc. Social structures can be racist simply by maintaining expectations and norms that serve to position racial groups differently and unequally.

We said above that structures also relate us to material things. The current American race structure positions different races hierarchically. It does this through the unequal constraints and enablements it puts on members of different races. But this is not simply a matter of norms or expectations. It's also a matter of how members of different races relate to aspects of their material environments: housing segregation, proximity to industrial pollution, access to city services, proximity to work, access to quality public schools (Massey and Denton Reference Massey and Denton1993, Sundstrom Reference Sundstrom2003, and Homer Reference Homer2017). Certainly, these differences can be explained historically by racist policies and laws of institutions. But the material features of cities function to differently constrain and enable members of different races in the present. In other words, these material features are part and parcel of the current race structure that functions to privilege white people and disadvantage people of colour. This feature of structures shows that not all racism resides in the attitudes, beliefs and actions of individuals. The system is maintained and perpetuated by material features of our social environments (Haslanger Reference Haslanger2022).

There is at least one more form that systemic racism may take.

(SR3) does not locate the racism of the system in its occupants or the system's constraints and enablements. Rather it locates the racism of the system outside it, in its relations to other systems.

This might strike you as problematic. Systems that function to cause, perpetuate or amplify racial oppression but which don't contain racist individuals or encode a racist ideology would seem to have these effects accidentally. The racist effects would merely be unintended consequences, from otherwise well-intentioned systems, so it is implausible to call such systems racist (Garcia Reference Garcia1997: 24). In the abstract, this is a fine worry. But practically, the connection between racism and the systems operative in America and other western countries is no accident. The kind of anti-black racism we see in the history of America was purposeful and many institutions/structures were founded on this racism. Given that all systems in America are interconnected in some way, it is a mistake to evaluate systems in isolation. Fixating on whether a specific system has ‘accidental’ or ‘purposeful’ discriminatory effects on a racial group ignores the interconnections between systems, past and present. Our evaluation of these systems needs to move beyond what their occupants believe and the stated aims of these systems to see the mechanisms by which racial disparities are perpetuated.

Consider, first, the way in which a system can have racist effects because of how it is related to racist systems of the past. Most of the present-day institutions in the US are no longer constituted by explicitly racist policies, rules and laws. Jim Crow, legal segregation and redlining are in the past. Yet there is a clear sense in which these current systems bear the marks of these past systems. We know these systems have resulted in massive inequalities between white and black people (Coates Reference Coates2014). Around 2016, in the US the net worth of the average white family ($171,00) was about ten times as much as that of the average black family ($17,150) (see Feagin Reference Feagin2006 and McIntosh et al. Reference McIntosh, Moss, Nunn and Shambaugh2020). Once these past systems had their effects, all that is required to perpetuate this inequality is for our current systems of banking and real estate to maintain the status quo. This doesn't require any racist individuals in these systems.

Or consider recent voter ID controversies in US states like North Carolina. Although the stated reason for requiring certain forms of ID was to prevent voter fraud, the effect was to disenfranchise black voters with ‘almost surgical precision’, as one federal appeals court judge remarked (see N.C. State Conference v. McCrory 2016 and Newkirk II Reference Newkirk II2018). The creators of these voter ID laws may have had no racial animus; perhaps they were merely attempting to benefit their political party. Even so, the effect of these laws is to disenfranchise voters of colour, a form of subordination that has a long and well-known history in the US. These laws, whatever the intentions of their creators, perpetuate, sustain and amplify practices that subordinate certain racial groups.

Second, systems can rightly be called racist because of how they are related to other racist systems in society. Consider the following example: suppose there is a court system that operates according to completely race-neutral laws, statutes and rules. The judges, prosecutors and defence lawyers follow the race-neutral laws and do not show any racial favouritism. But now imagine that the police system of the society is racist by having individual racist officers and racist policies against people of colour. The effect of this system is going to be that the people who show up in the court system will be disproportionately people of colour. So, the court system will be racist in virtue of the way it is related to the police system – as a conduit of racial oppression – but not because of anything inherent in the court system itself. Obviously, this example is not accurate with respect to the current American system (see Alexander Reference Alexander2010), but it shows that even if we were to make the courts completely race neutral, they would not stop being racist because of how they are connected to other racist systems.

What makes this use of ‘racism’ seem objectionable is that we are used to thinking of racism as something internal or intrinsic to those individuals and systems to which we ascribe it. While the possibility of (SR3) is rooted in the extrinsic, relational features of systems, the internal aspect of racism is not entirely missing. Indifference to the plight of victims of racial oppression can also be a source of racism. Systems that are indifferent to the ways in which they perpetuate or amplify racial disparities should not escape the charge of racism just because the origin of the racism is located outside them (cf. Taylor Reference Taylor2013: 34).

We now have an answer to question (2). Systemic racism is a feature of social systems that cause, perpetuate or amplify the oppression of a certain racial group (Haslanger Reference Haslanger2012). A system can be racist when it contributes to such oppression by (i) being composed of racist individuals, (ii) being constituted by racist constraints and enablements, and (iii) its interconnections to other racist systems. On the view presented here, a social system can contribute to racial oppression even if its occupants are not themselves racist and even if the system expresses no racial hatred or racist ideology.

Let me close by addressing a common worry about letting ‘racist’ apply to many sorts of systems, even those that don't purposefully target any racial groups for subjugation. The concern is that the charge of racism is a morally serious charge. Therefore, we don't want to lessen that charge by applying it too widely or in less than clear-cut cases involving racial hatred or assertions of racial inferiority (Blum Reference Blum2002, Case Reference Case2019). We should agree that the accusation of racism is a serious one. It's serious because of the evil and harm of racism. All the more reason to root out racism in all its nooks and crannies. By pinpointing the ways in which systems – unwitting or not – contribute to a society in which people of colour continue to face subjugation, we get a better understanding of how to change this society. There is no reason to assume, moreover, that all systems bear the same level of responsibility for the perpetuation of racism. That is why we need to be clear about the various ways in which systems contribute to racial oppression. The worry that a wide use of the term ‘systemic racism’ will water down the moral gravity of the charge is misplaced. The greater danger is that we would be blinded to the many ways, often subtle and mundane, in which racial subjugation is promoted and perpetuated