The Scapegoat Reflex: Crisis, Contagion, and the Psychology of Blame

The views expressed in this piece represent the analytical conclusions of the author and are intended to provoke serious examination of documented behavioral patterns, not to minimize the harm those patterns produce.

In the spring of 1349, the Black Death reached Strasbourg. It had been moving across Europe for roughly two years by then, killing somewhere between a third and half of every population it touched, and the residents of Strasbourg knew it was coming. They had watched it consume city after city. They had no accurate theory of how it spread, no treatment that worked, and no institutional capacity to stop it.

What they did have, within weeks of the outbreak's arrival, was a target.

The Jewish community of Strasbourg — several hundred families who had lived in the city for generations — was accused of poisoning the wells. There was no evidence for this. There could not have been: the disease was caused by Yersinia pestis, a bacterium transmitted by flea bites, a mechanism that would not be understood for another five centuries. None of that mattered. On February 14, 1349, approximately nine hundred Jewish residents of Strasbourg were burned alive. The plague continued.

Strasbourg was not an outlier. It was a data point in a pattern so consistent that historians and epidemiologists now treat it as essentially predictable: major outbreak, short interval, identification of outsider group, organized violence. The interval between outbreak and scapegoating in documented medieval cases averages weeks, not months. The selection of targets follows a logic that has nothing to do with epidemiology and everything to do with social structure — whoever occupies the most precarious position in the existing hierarchy becomes the most available vessel for collective terror.

The Mechanism, Described Plainly

Social psychologists have studied scapegoating under a variety of experimental conditions, and the findings are not flattering to the species. The core mechanism appears to involve the interaction of two well-documented processes: threat-induced attribution bias and outgroup homogeneity effect.

When individuals experience threats they cannot control — illness, economic collapse, natural disaster — they show a measurable increase in the need for causal explanation. Uncontrolled threat is psychologically intolerable in a way that explained threat is not. A disease that kills randomly and invisibly is more frightening than a disease with an identifiable source, even if identifying the source does nothing to reduce exposure. The mind, under stress, will accept a wrong explanation over no explanation because wrong explanations at least imply the possibility of a response.

The outgroup homogeneity effect — the well-replicated tendency to perceive members of other groups as more similar to each other and more different from oneself than they actually are — provides the targeting mechanism. Under stress, outgroup members become interchangeable representatives of their category. A Chinese-American family in San Francisco in March 2020 was not, to the person who attacked them, a specific family with a specific history. They were a category, and the category had been associated, in public discourse, with the origin of a frightening disease.

The result is a system that is, in a cold analytical sense, highly efficient. It converts uncontrolled threat into explained threat, identifies a target for collective action, and provides a temporary reduction in anxiety — all within a timeframe measured in days to weeks. That the explanation is wrong and the target is innocent are not, from the perspective of the mechanism, relevant variables. The mechanism does not optimize for accuracy. It optimizes for the feeling of restored agency.

The Medieval Template in Modern Dress

The Stop AAPI Hate reporting center, established in March 2020, documented 3,795 hate incidents against Asian Americans in its first year of operation. The Center for the Study of Hate and Extremism at California State University, San Bernardino, found that anti-Asian hate crimes in major American cities increased by 149 percent in 2020 compared to 2019. The timeline tracks with near-clinical precision: the first significant American COVID-19 coverage ran in January 2020; the first documented surge in anti-Asian harassment began in February.

The specific language used by those who committed these acts — verbal harassment, physical assault, vandalism — frequently referenced China, Chinese food, or disease transmission directly. The targets were not selected by any logic related to actual viral exposure. They were selected by visible membership in a category that public discourse had linked to the outbreak's geographic origin. The mechanism was identical to the one that operated in Strasbourg in 1349. The technology was different. The psychology was not.

It is important to be precise about what this comparison does and does not claim. It does not claim that anti-Asian violence in 2020 was equivalent in scale or institutional organization to the pogroms of medieval Europe. It was not. It does claim that the underlying cognitive process — threat activation, attribution need, outgroup targeting — was the same process, operating on the same timescale, producing proportionally similar results given the different social and legal constraints of the two environments. The constraints matter enormously. The process underneath them is the same.

The Uncomfortable Predictive Argument

Here is where this analysis becomes genuinely difficult to sit with.

If scapegoating in crisis is a stress response rather than a failure of education or information, then the standard interventions — public awareness campaigns, media literacy programs, appeals to shared humanity — are likely to be insufficient as primary prevention tools. This is not because those interventions are worthless. It is because they operate on the deliberative, reflective cognitive system, and scapegoating activates before that system fully engages.

The medieval populations who burned their neighbors were not, by the standards of their time and culture, unusually ignorant or unusually cruel. Many of them were literate, religiously observant, and embedded in communities with robust norms of mutual obligation — toward members of their own group. The crisis did not create their capacity for violence. It activated a targeting mechanism that was already present and redirected it with extraordinary speed.

Contemporary Americans are not exempt from this architecture. The behavioral record does not support the comfortable assumption that education, democratic institutions, or cultural sophistication reliably suppress the mechanism under sufficient stress. What those factors do — and this is not nothing — is raise the threshold of stress required to activate it, provide competing frameworks that can slow the process, and create legal and social consequences that constrain its expression. But they do not eliminate it.

The predictive implication is stark: the next major crisis in the United States — the next pandemic, the next catastrophic weather event, the next economic collapse — will produce scapegoating. The target group will be determined by who is already most marginal in the social hierarchy at the moment the crisis arrives, and by whatever associations public discourse has most recently activated between that group and the nature of the threat. We do not know who that group will be. We know the process that will select them.

What Useful Preparation Looks Like

If the mechanism cannot be educated away, it can be anticipated and structurally constrained. Medieval European communities that had pre-existing legal protections for minority residents, or that had strong institutional leaders willing to intervene early, showed lower rates of organized violence during plague outbreaks — not zero, but lower. Pope Clement VI issued a bull in 1348 explicitly condemning the persecution of Jews and threatening excommunication for those who participated; in regions where his authority was strong, the violence was somewhat reduced.

The analogy for contemporary institutions is not subtle. Early, explicit, authoritative public communication that refuses to link a disease or disaster to an ethnic or national group — and that actively counters such linkages when they emerge — reduces the speed at which the targeting mechanism completes its cycle. Legal consequences for bias-motivated violence that are swift and publicly visible raise the cost of acting on the impulse. Community organizations with pre-existing relationships across group lines provide competing social pressures.

None of this is guaranteed. All of it helps. The behavioral record of five thousand years does not offer utopian solutions. It offers, instead, a clear-eyed account of what humans do under pressure, and the modest but genuine possibility of designing our institutions with that account in mind rather than in spite of it.