“Evolutionary Psychology: Predictively Powerful or Riddled with Just-So Stories?
Laith Al-Shawaf, Ph.D. is a researcher and Assistant Professor of Psychology at the University of Colorado. He has taught and conducted research internationally, been a Visiting Fellow at the Institute for Advanced Study in Berlin, and is an academic adviser at Ideas Beyond Borders. His research (with collaborators) has been featured in outlets such as the BBC, Washington Post, The Atlantic, Psychology Today, Slate, World Economic Forum, and Time, and his essays for general audiences have appeared in Areo and PopMatters. In 2019, the Association for Psychological Science (APS) named him a Rising Star.
This essay is part of a series on the value of evolutionary approaches to psychology.
Part 1 clears away seven key misconceptions.
Part 2 shows why evolution is necessary for a complete science of the mind.
Part 3 (this essay) illustrates how evolutionary thinking leads to new discoveries.
They do not need to be read in order.
Acommon refrain in the social sciences is that evolutionary psychological hypotheses are “just-so stories.” Amazingly, no evidence is typically adduced for the claim—the assertion is usually just made tout court. The crux of the just-so charge is that evolutionary hypotheses are convenient narratives that researchers spin after the fact to accord with existing observations. Is this true?
Do Evolutionary Approaches Lead to New Predictions? What About New Discoveries?
In reality, the evidence suggests that evolutionary approaches generate large numbers of new predictions and new discoveries about the human mind. To substantiate this claim, the findings in this essay were predicted a priori by evolutionary reasoning—in other words, the predictions were made before the studies took place. They therefore cannot be post-hoc stories concocted to fit already-existing data.
Consider the following evolutionary predictions about disgust, all of which were made a priori: 1) people’s disgust will be more strongly triggered by objects that pose a greater risk of infection, 2) women will be more disgusted during the first trimester of pregnancy compared to the second and third trimesters, 3) people who grow up in regions of the world with higher levels of infectious disease will be less extraverted, less open to new experiences, and less interested in short-term mating than their counterparts who grow up relatively pathogen-free, 4) cross-cultural differences in pathogen prevalence will predict cross-cultural differences in individualism-collectivism, 5) those with a stronger proclivity for short-term mating will be less easily disgusted, 6) experimentally triggering disgust will reduce interest in short-term mating, 7) people will feel less disgust toward their own offspring and their offspring’s bodily waste compared to the offspring of others, and 8) presenting people with the threat of disease will cause a host of psychological and physiological changes that reduce the likelihood of infection, including a) releasing pro-inflammatory cytokines, b) behaviorally withdrawing, c) temporarily becoming less open to new experiences, and d) reducing one’s desire to affiliate with people. All of these predictions were generated before the fact on the basis of evolutionary reasoning, and all were subsequently supported by the data.
Note that some of these findings could probably have been predicted without evolutionary reasoning. For others, it would have been harder. And for others still, it would have been nearly impossible.
A final example of the predictive power of evolutionary thinking comes from Error Management Theory, a theory about the evolution of cognitive biases. Error Management Theory suggests that in decision-making scenarios, you can make two possible kinds of error: a Type I error (a false positive) or a Type II error (a false negative). If one error is more costly than the other, and this cost asymmetry recurs over evolutionary time, then the species in question will evolve neurocognitive mechanisms that are adaptively biased toward the safer error. In other words, animal brains operate according to a similar logic as humanly engineered smoke alarms: they are built to be biased toward the less costly error because this minimizes the likelihood of the more catastrophic error.
This simple evolutionary theory leads to new discoveries in areas such as social cognition, visual and auditory perception, and immune function. For example, the theory predicts that when people look down at the ground from a high vantage point such as a steep hill, they will systematically overperceive their distance to the ground, because this is safer than underperceiving the distance to the ground, which could lead to a lack of caution and a lethal fall. This prediction is verified by the data—as is the supplementary prediction that this height estimation bias will be attenuated when people are looking up to a precipice from below (because it is not as dangerous when you are at the bottom), as well as the remarkably precise a priori prediction that the height overestimation bias will apply to environmental verticality, but not retinal verticality (because only environmental verticality is related to falling risk). We owe our knowledge of these fascinating discoveries to the evolutionary reasoning that led to these predictions—predictions that didn’t exist before researchers thought to approach the problem from an explicitly evolutionary perspective.
The logic of Error Management Theory also predicts that heterosexual women will exhibit an on-average “commitment skepticism bias.” The idea is that, on average, overestimating a suitor’s commitment intent was more costly for our hominin female ancestors than underestimating it—so the theory predicts that modern women will exhibit an on-average bias toward erring on the side of underestimating potential mates’ commitment intent. This a priori prediction is confirmed by the data—as is the supplementary prediction that postmenopausal women will not exhibit the bias. More data are needed to test this prediction in different cultures and to figure out which contexts upregulate and downregulate the bias (or annul or reverse it), but initial findings seem promising so far.
Next, Error Management logic predicts that we will exhibit an auditory looming bias. Specifically, the theory suggests that we will perceive approaching sounds to be closer than they actually are, and to be arriving more quickly than they actually are. This is because the safer error is to be prepared for an oncoming danger too early rather than too late. Indeed, studies show that humans do exhibit this auditory looming bias—as do monkeys.
Studies also confirm that, as predicted, we perceive approaching sounds as both starting and stopping closer than equidistant receding sounds.
Finally, less physically fit individuals need longer to escape an oncoming threat, so they have a more pronounced auditory looming bias than fitter individuals—exactly as predicted by the theory.
By now the reader has doubtless noticed that many of these findings are counterintuitive, and not the kind of result you could predict using common sense. Some, maybe even most, would have remained undiscovered were it not for the evolutionary reasoning that generated the hypotheses in the first place. And even if somehow that statement is incorrect, what is completely unambiguous is this: these hypotheses were generated a priori and then led to new discoveries about how the mind works. They decidedly did not involve working backward from existing data to convenient stories.
For example, we could have discussed how evolutionary thinking leads to new predictions about pride, shame, hunger, gratitude, jealousy, political preferences in leaders, universality in mate preferences, cultural differences in mating strategies, reputation, punitive sentiment toward criminals, volunteering for charity, support for economic redistribution, moralizing people who opt out of public goods, the “erasure” of race, our ability to solve mathematical problems that are framed in terms of frequency versus probability, what kinds of conditions improve our statistical inferences, our ability to detect violators of social contracts, whom newborn babies are said to resemble, what psychological features might accompany illness, and theoretically predicted cultural variation in the extent to which people value physical attractiveness—to name a few.
We might reasonably want to ask why evolutionary approaches to psychology are so successful with respect to predictive power. A brief and incomplete accounting suggests that it is partly because evolutionary thinking reduces the search space by insisting on consilience with biology, thereby ruling out hypotheses that violate the basic principles of evolutionary theory; partly because evolutionary theory has been worked out in sufficient detail that deriving predictions from the theory is easier than it is from less well-specified theories; and partly because evolutionary approaches offer researchers useful conceptual-methodological tools such as “task analysis”, which is well suited for generating novel predictions about human psychology and behavior.