It’s hard to make sense of witch hunts. Many people of early modern Europe and colonial America seemed to have genuinely believed that witches posed a serious threat. But if witch trials—like the ones in Salem, Massachusetts, and in European communities between the 1400s and 1700s—escalated out of control, with no clear beneficiaries, then why did they happen? In a new paper, philosopher Maarten Boudry and historian Steije Hofhuis argue that witch hunts weren’t “coordinated intelligent strategies with underlying goals,” even though it often looks like they were. In other words, they weren’t motivated by a desire to, for example, oppress the lower classes or women, and weren’t a result of powerful economic interests.
One reason they find those theories unconvincing is that there appears to be no evidence in historical documents of anyone explicitly organizing against particular groups. It is, Boudry and Hofhuis write, “hard to grasp how the witch-hunters could have developed such a shrewd hidden functional purpose, if they did not discuss this with each other.” Instead they offer a Darwinian explanation for witch hunts, involving selfish memes. Put simply, the idea of witches propagated because it was good at propagating, even though no individual really wanted it to. “We argue,” they write, “that witch persecutions form a prime example of a ‘viral’ socio-cultural phenomenon that reproduces ‘selfishly,’ even harming the interests of its human hosts.” Although I’m skeptical, the theory is fascinating and is worth thinking about.
The self-reinforcing character of witch-hunting practices does not, in itself, demand memes.
The idea of selfish memes originates from evolutionary biologist Richard Dawkins. In his view, famously expounded in The Selfish Gene, genes are “selfish” because replicating takes top priority—genes are “interested” in their own fitness. This supersedes its other effects: To replicate might come at some cost to the organism a gene is in, and yet it persists. Dawkins thought the same principle was at work in cultural evolution. He proposed that bits of culture—various beliefs and behaviors—should be thought of as variants in the same way that there are different variants of the same gene across human populations.
His idea is controversial. What usually concerns critics of meme theory is its usefulness: the degree to which it captures cultural phenomena and explains their emergence. It has been criticized, for instance, for not being of much help in understanding history. But Boudry and Hofhuis disagree, and argue that witch hunts are the sort of historical event for which meme theory does provide some explanatory value.
Many beliefs about witches and practices in witch trials were self-reinforcing—trying and convicting one person for witchcraft often led to trying many more. For example, evidence for being a witch included naming other witches, and accused witches were frequently interrogated by torture. Unsurprisingly, then, successful witch trials often resulted in many more witch trials, and so on. Boudry and Hofhuis propose that selfish memes help to explain the emergence of these witch hunts: Witch hunts escalated because the memes associated with them were selfish. “The evolutionary scenario that we propose is as follows,” Boudry and Hofhuis write, “Ideas that accidentally triggered larger persecutions were cumulatively preserved precisely because of that effect, in repeated rounds of variation and selection.” Witch memes arose by accident in a normal process of cultural variation, and they stuck because they are good at reproducing themselves.
Convinced? Not me. The self-reinforcing character of witch-hunting practices does not, in itself, demand memes. Boudry and Hofhuis note the self-reinforcing character before they even get to talking about memes: “Witchcraft therefore became known as a crimen exceptum; an extraordinary crime requiring extraordinary means of investigation… Various forms of physical torment were recommended to make alleged witches confess to their evil deeds and to make them name accomplices. Unsurprisingly, the likelihood of suspects pleading guilty significantly increased.” When they draw these connections between some aspects of witch hunts and selfish memes, what we are left with is basically an analogy—a Darwinian vocabulary for saying that those beliefs and practices are self-reinforcing.
To call a belief a metaphorically selfish meme is just to restate its property of self-reinforcement: They are selfish because they are self-reinforcing; they are self-reinforcing because they are selfish. So, while the analogy is not necessarily wrong or inaccurate, selfish meme vocabulary does not help researchers understand or explain witch hunts in any greater depth.
What does explain them, then? Perhaps conformity bias. It is worth considering that maybe only a few people genuinely believed witches were among them, and that apparent beliefs about witches might have spread, contrary to the selfish-meme view, by no inherent property of their own, but by other mechanisms of social influence. Witch panics are popularly called instances of mass hysteria—and one way that scientists make that idea more precise is to compare them to psychological experiments on how people conform to groups. In the 1950s, the psychologist Solomon Asch showed evidence of a conformity bias in his subjects. In one of his most famous experiments, he asked subjects to match the length of a vertical line with one of three others—one option was in fact the same length, while the others varied. Seven actors posing as fellow subjects coordinated to deliberately select a line that did not match. One-third of experimental subjects went along with the majority, and reported that those lines matched, even though they clearly did not. Conformity bias separates behavior from belief: People can act in accordance with beliefs they do not truly hold, as a consequence of context.
While many people at the time probably believed in the existence and dangers of witchcraft, the explanation by conformity bias would only require a few to believe in the presence of witches in their own community, and in the legitimacy of the methods used to interrogate and convict them. Then, the beliefs would appear to spread, but not because the beliefs themselves have some property of selfishness—simply because the beliefs appear to be held by others. Conformity bias may not be a complete explanation for witch hunts, but it could be a powerful factor, especially considering that the majority of people, in the case of witch hunts, were not strangers to each other, as in Asch’s experiments; they were known and presumably trusted members of the community.
Another possible explanation builds off of Boudry and Hofhuis’ framework: They briefly discuss the idea that memes can be compared to viruses—replicating themselves using humans at their expense—and that this comparison might be more than just a metaphor. “The similarities between biological viruses and outbreaks of witch panics are not just superficial or curious coincidences without any theoretical significance,” they write. “On the contrary, the analogy arises from the fact that both phenomena underwent Darwinian selection processes.”
If it is in fact accurate to think of witch trial beliefs as viruses, maybe it would be helpful to study their spread the way scientists study the spread of viruses: using an epidemiological model. This would take into account that the witchcraft beliefs are “encouraging” their own persistence and would yield a pattern for how they might spread. This incorporates logic of natural selection that Boudry and Hofhuis stress in their selfish meme framework, but it goes a step further by allowing researchers to obtain more specific patterns for the spread of ideas.
Several studies have shown that fairly basic epidemiological models can model the spread of ideas. Researchers begin with an established mathematical model for the spread of disease and carefully adapt the variables to instead represent the spread of ideas. Some researchers take Dawkins’ memes as a starting point, but it is worth pointing out that they’re not necessary: If “selfishness” is meant to describe how easily an idea spreads, researchers can capture that with a variable called the “incidence rate”—the number of new “cases” of the idea divided by the “susceptible” population for some interval of time. Viruses that are exceptionally good at rapidly reproducing themselves and getting out to other hosts have a higher incidence rate. If they are genuine beliefs, the spread of beliefs about witchcraft can be at least partially explained by epidemiological models, which, like meme theory, rely on replication and transmission. But they don’t actually make use of the meme framework itself.
Neither of these alternative explanations entirely solves the puzzle of witch hunts, either. If conformity bias is responsible, it is still unclear why those apparent beliefs in particular were so ripe for spreading. If epidemiological modeling is the correct approach, then the causes for high incidence rates are still hazy. What makes these ideas different from other self-reinforcing ones—or why the cycle of self-reinforcement is so hard to break—is not totally obvious from any of these possible explanations.
Margaret E. Farrell is a Ph.D. student in the Department of Logic and Philosophy of Science at the University of California, Irvine. Her research is focused on the history and philosophy of biology.