The recent cycle of antiracist struggle has driven increased interest in racial capitalism, both the specific concept as well as the general problematic to which it refers: the relationship between the reproduction of capitalism and the reproduction of racism. Julian Go’s recent essay points to some productive ambiguities in the concept, one of which is this: is the relationship between capitalism and racism merely an historical one, or is there some structural connection between the two?

My article with David Calnitsky in the Du Bois Review speaks to this question, affirming the notion that the relationship between capitalism and racism is structural rather than merely historical. It is not just that capitalism emerged in an already-racialized context, and so developed in a racialized way. It is also that racism can be explained by its functional role in stabilizing capitalist class relations. This view has occasionally been described as “class reductionist.” In our view, this does not itself amount to a fatal criticism, but it does associate our argument with others that we in fact reject.

What are we not saying? First, we are not saying that class is more central than race to the formation of people’s subjectivities. There is simply no good argument that class is the “true” identity. People’s identities are just what they think they are. Second, we reject the idea that class takes moral priority over race as a form of oppression. Our argument neither depends on nor implies the idea of class having a special moral status. Third, while we accept the premise that racism causally depends on class oppression, we reject the conclusion that race is a mere epiphenomenon of class. Race does mask deeper social relations, of course, but it does more than that: it has real effects on those relations, including class.

What we are saying is that, again, race can be explained by its class-stabilizing effects. Our argument is a functionalist one. Functionalism basically explains a thing by appealing to its effects on some larger system. It’s a strange idea – effects follow causes, so how could something be caused by its effects? For many social scientists, functionalism is discredited. Can it be defended?

We begin by distinguishing between two types of social construct: “belief-dependent” constructs and “structure-dependent” constructs. Belief-dependent constructs have to do with shared discursive and cognitive categories, symbolic universes, and ways of making meaning. Structure-dependent constructs consist in the distribution of resources that impact power and bargaining, irrespective of people’s beliefs. Race is a belief-dependent construct; class is a structure-dependent construct.

The fact that both are constructs does not mean that they are co-equal in a causal account of the world. Pressures exist for belief-dependent constructs to correspond to the underlying inequalities in bargaining power – structure-dependent constructs – and it is hard to retain beliefs that totally contradict material circumstances. A thought experiment by the “What Is Politics?” podcast expresses the idea nicely: Imagine a pastoralist patrilocal society with strong patriarchal norms, in which someone has created a drug that wipes sexist beliefs from the mind. She gets the drug into the water supply, whereupon all are convinced that women and men are equals deserving the same respect and liberty. But the material circumstances remain the same: men own their land and herds, and women live in the villages of their husbands’ families. Soon, as inevitable conflicts emerge between husbands and wives, they find that the distribution of their bargaining power is unchanged: men own everything and are surrounded by their own families, and women have no resources of their own and are more socially isolated. Before long, men come out on top in these domestic conflicts, and the men as a group will defend their prerogatives against the designs of the women – not because they are ideologically committed sexists, but because they are in a better position to get what they want. As they seek to rationalize the asymmetry, sexist beliefs will find fertile ground once again.

So, structure-dependent constructs like class (or patrilocalism) explain belief-dependent constructs like race (or gender norms) rather than the other way around. But none of this means that belief-dependent constructs have no empirical effects, nor that they are “less important” in some general sense; race often turns out to be significant as an independent variable in regression models, for instance. In fact, our argument – that race is explained by its effects – fundamentally depends on the independent effects of race. But it is important not to conflate statistical and explanatory models.

Note again that ours is not a claim about history per se. Many who argue, as we do, that racism is explained by capitalism, take the genealogical tack: they turn to the historical record to find the first instances of racism, and find that these arise in contexts of capitalist development. We take a different approach, focusing less on emergence than on reproduction over time. Discovering the first religious person would not itself explain religion; we need to understand why it sticks. This is what functionalism considers, and why a phenomenon’s effects are so important to its explanation. It isn’t enough to show that racism and capitalism emerged around the same time, nor even to show that racism was beneficial for capitalism. In fact, there are two main criteria for a successful functionalist explanation that we consider: first, there must be a causal feedback loop, a selection mechanism, connecting racism’s effects to its reproduction; and second, the stability-producing effects of racism must be unintentional and unrecognized by the beneficiaries.

To clarify the functionalist logic, consider the concept of a “local maximum”: taking for granted a given context, we can ask which traits or features best increase the chances of reproducing some structure. Why is some level of hazing a regular feature of tight-knit groups, like fraternities or gangs? Hazing exists because it improves group cohesion, which enhances the group’s odds of reproduction over time. Too light a level of hazing fails to produce the effect; overly severe hazing may dissuade joiners. Groups with the right levels of hazing last; those who get it wrong dissolve. This is the selection mechanism – we observe the hazing trait so often because groups without it have low odds of survival relative to those with it.

But if the gang implements a hazing ritual with the intention of enhancing group cohesion, and then keeps it when they see that it works, then what we have is a standard intentional explanation: hazing exists because gang leaders deliberately institute it in pursuit of the desired effect.

We think we can relax the intentionality criterion a bit for two reasons. First, functional effects may be intentional and recognized in one generation, only to be naively carried on by subsequent ones. Secondly, and more importantly, the crucial idea for functionalism is what G. A. Cohen referred to as the “dispositional facts” of a system. Functionalist explanation assumes the existence of a systemic problem: for gangs, their cohesion is not assured, because members may drift away or join other similar groups, and hazing emerges as a problem-solver. Whether or not hazing in fact emerges in a particular gang, it is a “dispositional fact” about gangs that they would be better off if they adopted hazing rituals. For capitalism, the problem is that the exploitation inherent to the system generates conflict, and conflict generates instability. Whether or not capitalists intend to create racism, the fact that racism has beneficial effects for them is not a feature of capitalists’ consciousness but of the social system.

Still, we can identify mechanisms by which capitalists unintentionally reproduce race when they simply intend to pursue profits. This is what the idea of “statistical discrimination” drives at, for instance: capitalists are not attempting to “divide and conquer,” but using an applicant’s race as a kind of proxy for the cost of hiring them. Another mechanism is the “kill switch”: when interracial solidarity does emerge, increasing wages and workers’ power, capitalists will tend to redirect investments into labor-saving technology or into unorganized fields. To the extent workers are successful in overcoming the racial divide, they are met with stagnation or even layoffs.

But why focus on the question of race, here? There are lots of mechanisms for reducing class conflict. Rising incomes may produce contented complacency, and all kinds of social divisions or even simple collective action problems can divide the discontented. We think the kind of social divider that race is makes it particularly well-suited to intraclass division: race sharpens skill divisions with kin divisions; it is presumed to be biological thus immutable and intergenerational; and it is usually readily visible and unconcealable.

Finally, we conclude by outlining the limits to our argument with empirical examples of racism that have political-cultural dimensions without clear economic ones and propose some provisional defenses. To wrap, we believe that a causal class reductionism is a crucial aspect of any good theoretical account of social division, and that undoing racial oppression is an undertaking with a fundamentally materialist character.

Michael Billeaux Martinez is an historical sociologist studying race, class, and politics in North America, especially the Great Lakes region. He teaches at Madison College and is a member of AFT Wisconsin.

This article is based on Michael Billeaux Martinez and David Calnitsky, “A Class Functionalist Theory of Race,” DuBois Review: Social Science Research on Race, doi:10.1017/S1742058X22000224

Image: Charles Zimmerman speaks at a civil rights rally in the New York Garment District on 38th Street near 7th Avenue in New York City. Signs include ‘Labor Opposes Discrimination’ May 17, 1960” by Kheel Center, Cornell University Library via Flickr.