It’s easy to think of human problems as cumulative. If Bob lives in a Mississippi flood plain, Phil lives near the San Andreas fault, and you live in an area that’s prone to both earthquakes and floods, you should cumulatively have both Bob’s problems and Phil’s problems. Social problems, however, don’t always accumulate that way. The problems of being an African-American woman are not simply the sum of the problems faced by the average African-American and the average woman. Similarly, the problems of being a poor immigrant are greater than the cumulative sum of the problems faced by the average poor person and the average US immigrant. The term intersectionality was coined to describe this phenomenon.
The term interaction effects could have been used—it’s a non-ideological, non-partisan term that had been in use since the 1930s. But, in the early stages, intersectionality scholars weren’t using statistics. In fact, the term intersectionality originated in legal scholarship. It is only in the last ten years that intersectionality has spread from the legal arena, and become a political football at American universities. Progressives tend to argue that everything must be intersectional, and conservative skeptics counter that intersectionality is bunk. My goal in this blog post is to explain how progressives and conservatives are both mistaken about the promise of intersectionality. Progressives have adopted an overambitious model of intersectionality in which everyone lies on axes of oppression, and I will explain this model’s three flaws. Conservatives generally believe that intersectionality is useless, but I explain how intersectional scholarship can be useful to researchers, regardless of whether they are liberal, centrist, conservative, libertarian, or eclectic.
In 1976, Emma DeGraffenreid, a black woman employed by General Motors, sued her employer for discrimination. Along with several co-workers, she accused GM of discriminating against black women by making it impossible for them to get a certain class of jobs. At GM, women were only considered for a certain set of jobs such as secretarial ones. Blacks were only considered appropriate for another set of jobs, primarily manual factory work. The combined effect was to limit black women’s opportunities for advancement in a unique way that was experienced neither by black men nor by white women.
Sadly, DeGraffenreid’s lawsuit failed. The court believed that GM could be sued for sexism for problems that affected all women employees, and that it could be sued for racism for problems that affected all Black employees. Plaintiffs, however, could not combine racism and sexism. If only Black employees of a certain gender, or only female employees of a certain race suffered discrimination, GM was not liable for racism nor sexism. The court stated:
The initial issue in this lawsuit is whether or not the plaintiffs are seeking relief from racial discrimination, or sex-based discrimination. The plaintiffs allege that they are suing on behalf of black women, and that therefore this lawsuit attempts to combine two causes of action into a new special sub-category, namely, a combination of racial and sex-based discrimination. The Court notes that plaintiffs have failed to cite any decisions which have stated that black women are a special class to be protected from discrimination. The Court’s own research has failed to disclose such a decision. The plaintiffs are clearly entitled to a remedy if they have been discriminated against. However, they should not be allowed to combine statutory remedies to create a new “super-remedy” which would give them relief beyond what the drafters of the relevant statutes intended. Thus, this lawsuit must be examined to see if it states a cause of action for race discrimination, sex discrimination, or alternatively either, but not a combination of both. [italics mine]
In a law journal article, Kimberle Crenshaw, a young law professor at the time, wrote about how the black female plaintiffs in DeGraffenreid v General Motors and two similar cases were at the intersection of female and Black, as in a Venn diagram, and coined the term intersectionality.
Psychologists and biologists had been studying intersections since the 1930s or earlier, but they used the label interaction effects. There was no “interaction-effects theory” because it would have been pointless to bundle all of the psychological and biological domains in which you find interaction effects; those domains have nothing in common. Interactions is also a statistical term, and Crenshaw’s argument was not statistical. It was an evidentiary set of case studies showing that discrimination against black women could not incur a penalty in the extant legal framework, because it did not violate laws against sex-based or race-based discrimination.
Intersectionality had progressive roots but it was not laden with ideology. Thus, academic intersectionality could have remained rooted in empirical research, with findings and arguments that could potentially persuade people of any ideology. However, intersectionality started to merge with progressive ideology when Patricia Hill Collins, a black feminist sociologist, moved the theory from grounded specificities to ungrounded generalities. In the long run, this deeply ideological turn has been bad for academic scholarship, because it started to decouple intersectionality theory from research.
Collins moved the discussion to general axes of oppression: classism, racism, sexism, heterosexism and the like. In Collins’ theory, the intersection of any two axes creates a unique multiplicative problem that cannot be predicted from additively processing, say, racism and sexism. Unhelpfully, Collins also treats intersectionality as an additive approach where the goal is to prevent any axis from being neglected. For clarity, I’m going to stay with the first definition, involving unique problems that are faced by people who are at the “bottom” end of two or more axes of oppression.
The Problems with “Axis of Oppression”
There are three problems with the axes-based approach. First, it posits consistent, unambiguous axes of oppression. Racism for instance must consistently take the form of one consistent oppressor class, namely Whites, and one victimized class, namely, people of color. For this to be true, we should have no evidence of pro-Black discrimination. However, in a specific educational scenario, we have robust evidence of pro-black discrimination. This does not mean we should retire the idea that black people are oppressed, or that pro-black and anti-black discrimination are comparable. That would be wrong-headed. But we shouldn’t oversimplify things either.
Second, intersectionality assumes that every intersectional effect is negative. The assumption here is that discrimination is always worse than the sum of its parts. Yet men who are both gay and black are perceived as more likeable than men who are both gay and White (see this study too). This does not mean it is easy to be a gay Black man in America, but does show one positive characteristic that gets attributed to gay Black men. Thus, not every intersectional effect is negative. For other examples, see here and here.
Third, intersectionality ignores the influence of the situation. Intersectionality has been more heavily influenced by sociology than social psychology. Sociologists typically study big categories—a sociologist might study education and race, but is unlikely to test variability across small educational situations. For instance, being African-American might be an advantage when you apply to an undergraduate college, but it might be a disadvantage when you apply for a summer internship. It might again be an advantage when you apply to an honors society, but a disadvantage when you apply for a research assistantship. In these cases, the intersectional effect is not about two identities, but rather an identity and a situation. I am not knocking sociologists for focusing on broad levels, but as social psychologists have been showing for decades, we need to consider the person-by-situation interaction. In fact, considering that oppression requires both agents (subjects) and patients (objects), we need to consistently consider the agent-by-situation and the patient-by-situation interaction.
More complex interaction effects may occur. To use a contrived example, research may show that when applying to a science (vs. arts) job in a liberal arts college (vs. a research university) in the Northeast (vs. all other regions), gay Asian women are treated more favorably. This is an arbitrary interaction, but such interactions could be real and discoverable. And tweaking either of the situations or the identities could lead to a different bias. Here a bunch of situational intersections and an identity intersection meet to form a six-part intersection.
When you bring in situational interactions, it becomes even more difficult to neatly classify people as privileged or oppressed along an axis of oppression. In fact, when you have information about the situation, it’s fruitless to ignore the interaction between the situation and the target person.
These problems aren’t fatal for intersectionality. By discarding the construct of “axes of oppression,” scholars could describe general trends, while also discovering intersectional effects—or their absence—by studying the people who lie inside and outside the focal intersection instead of making assumptions based on axes of oppression. Otherwise, we will go back to an era when Black woman are presumed to cumulatively bear the burdens of Black men and White women, but have no unique problems.
Crossing the Political Divide
Given the empirical status of intersectionality, for whom we have Crenshaw to thank, it would be difficult for a scholar of any ideological background to discard the theory in its entirety. However, intersectionality will continue to become yoked to ideological battles if liberals neglect the problems with intersectionality theory, and if conservatives assume that intersectionality is only about progressive causes.
Is everything intersectional, as some progressives say? That’s an empirical question. In the case of voting rights, there was an era when voting rights were denied to all women, which was direct discrimination; there was no intersectional effect by race or sexuality. There was also an era when voting rights were denied to all African-Americans in the South, which was also direct discrimination; there was, to my knowledge, no intersectional effect. A perusal of history would probably uncover similar cases where a consideration of interactions would miss the point. The idea that intersectionality is important can stand on its own empirical legs, without the need to append the axiom that intersectionality is pervasive.
Should conservatives continue to dismiss intersectionality, or may it benefit them to consider its uses? Assume that the paucity of black conservatives in academia needs an explanation. A progressive sociologist might argue that there are few black conservatives in academia because of their low base rate in the larger population. A conservative might counter the liberal by stating that although academic hiring discriminates against religious conservatives, there is also an intersectional hypothesis to consider. A hiring committee is perhaps more likely to hire a black liberal than a white liberal, but less likely to hire a black conservative than a white conservative. The liberal might aver that no such interaction effect exists. An empirical study could settle the dispute. In this case, the conservative is making an intersectional argument, one that matters for them because conservatives typically want liberals to acknowledge the opinions of black conservatives.
What this shows is that intersectionality theory needs repair. The fewer flaws in the theory, the more likely non-liberals will be to take it seriously. Simultaneously, if conservatives and other non-liberals come to understand that intersectionality spans issues that are relevant to people of every ideology, new and useful research on intersectionality may be conducted.