Critics of Evolutionary Psychology Say It’s All Just Storytelling. Here’s Why They’re Wrong.
This criticism is based on a faulty understanding of science
Darwin’s theory of evolution by natural selection is one of the most powerful and elegant theories in science. Without it, as Theodosius Dobzhansky noted, nothing in biology makes sense. With it, scientists have a tool to classify, conceptualize, and understand the vast and variegated world of life from simple single-celled organisms to unfathomably complicated creatures such as elephants and whales.
Natural selection is such a crucial component of biological thinking that biology textbooks are framed around it. Outside of creationists, few would criticize this approach. If you want to understand why birds build nests, why spiders spin webs, why primates clash for dominance, or why ungulates leap into the air at the sight of a predator, then you need to understand natural selection.
Humans, like other animals, are the product of millions of years of evolution by natural selection. So, one might imagine that scholars and intellectuals would approach the study of humans in the much same way they approach the study of bees, fish, frogs, or squirrels. Yet in disappointingly many cases, one would be wrong.
In principle, most thinkers will say, “yes, of course humans are evolved animals.” But in practice, many exhibit what can be called methodological dualism: They object to applying the logic of natural selection to the human mind. The liver? Sure. The spleen? You bet. The mind? Hold on a minute now!
Of course, these thinkers are too sophisticated to appeal to a spirit, a soul, a mysterious something that makes the human mind inexplicably immune to natural selection, so they use a different strategy, a strategy one might describe as selective skepticism. The most powerful weapon in the selective skeptic’s armamentarium is the just-so story accusation. This accusation has become so potent that it often ends serious discussion as the just-so storyteller attempts to convince others that he or she is not hopelessly naive. And, perhaps worse, it has caused confusion and serious misunderstanding about how science actually works.
The just-so story accusation has a long history. From what I can tell, it was first used against sociobiology by the eloquent and indefatigable critic Stephen Jay Gould in his 1978 article, “Sociobiology: The Art of Storytelling.” In that article, Gould railed against what he saw as the imperial ambitions of the neo-Darwinian (adaptationist) paradigm. According to Gould, this paradigm was in danger of becoming dogmatic and unproductive because it relied more on compelling storytelling than on testable hypotheses.
Scholars, blinkered by their Darwinian enthusiasm, forced all observations onto a procrustean bed of “Panglossian-like” adaptationism. Cute stories such as “I’ll bet women are more attracted to high-status men because they need resources to raise their children,” substituted for rigorous hypotheses and data. Gould called these “just-so stories” in reference to Rudyard Kipling’s children’s book of the same name. In that book, Kipling offered many whimsical explanations for animal anatomy, such as that the camel got its hump because it was punished by a djinn (genie) for refusing to work. The comparison of sociobiology to just-so storytelling, of course, was meant to be highly unflattering.
Since Gould’s attacks on sociobiology and evolutionary psychology, many others have raised the just-so story accusation. Some who forward the argument are clearly perturbed by what they see as the politically unpalatable consequences of evolutionary psychology. Many predominantly leftist scholars, for example, believe that evolutionary psychology simply observes the current status quo and then attempts to justify it with just-so stories explaining that it is natural and inevitable. Others who forward the argument are legitimately disappointed by what they see as evolutionary psychology’s methodological sloppiness.
For these scholars, the silliness of some evolutionary storytelling detracts from what could be a powerful research enterprise. These arguments have trickled into the popular media—so much so that one can hardly forward an evolutionary hypothesis without being accused of fanciful storytelling.
However, the just-so story criticism rests upon a flawed and unnecessarily austere vision of what real science is. According to this vision, science is comprised of a series of directly testable hypotheses. Falsify a hypothesis, and the theory is tossed overboard as dead weight into a sea of failed theories. But this strong version of falsification hasn’t been popular since the 1960s (if it ever was) because it just doesn’t describe how science actually works. Most science is much more like solving a murder mystery than it is dropping a litmus paper into a beaker and declaring “I have refuted your theory thus!”
Inference to the Best Explanation
Imagine that you wake up in the morning, walk to the living room, and find your furniture chewed and torn. What happened? There are many possible options, but only a few are plausible, and only one or two strongly coheres with the observed evidence and background conditions.
You bought a terrier from the shelter the day before. And he seems fond of chewing and destroying just about anything he gets his paws on. At first, it seems very likely that he had a blast at the expense of your new couch while you were sleeping. Suppose, however, that you check his crate and the door is locked and your dog is inside sleeping. This might cast some doubt on your theory.
Suppose you observe that the back door is open and that there are small tracks leading from the woods outside into your kitchen. The tracks look like raccoon paws, and your neighbors have complained that raccoons have been getting into their garbage. What seems likely now is that you had one gin drink too many, left the door open, went to bed, and while you were sleeping an army of raccoons came into your house, gobbled some food, tried to eat your couch, and then ran back into the woods.
Your hypothesis is not deductive, not the result of the syllogism:
- If there are bite marks in the couch, then there was a raccoon in your house.
- There are bite marks in your couch.
3. Therefore, there was a raccoon in your house.
It is also not inductive in a straightforward way.
In this example, we infer that raccoons are the best explanation for the destroyed furniture because that hypothesis is more coherent and parsimonious than the theory that the dog got out of the crate, ate the furniture, went back into the crate and locked the door and that raccoons also came into the house, walked around, left the furniture unperturbed, and went back into the woods.
Philosophers have compiled lists of the virtues of a best explanation, i.e., virtues that determine which inference is, in fact, the best explanation. For our purposes, a few of these virtues are worth mentioning (see table below).
First, the best explanation is coherent with background information. It is consistent with what we know about the world and does not require significant alterations to our current web of knowledge.
Second, the best explanation is parsimonious. It is the simplest theory or hypothesis — the one with the fewest ad hoc hypotheses or auxiliary assumptions — that explains the given data.
Third, the best explanation is fruitful. It offers a framework that can be extended to a large body of phenomena and that can generate many plausible hypotheses.
Inference to the best explanation is a better description of what most scientists do, and what they should do, than other accounts.
Scientists examine evidence with a certain set of background assumptions and attempt to offer the best, simplest, and most coherent explanation for the evidence. They then infer, provisionally, that this explanation is closer to the truth than alternative explanations. Of course, they update their confidence in the explanation as new data are discovered. If the new data appear inconsistent with the theory, for example, then they might become skeptical of it and seek to revise or replace it. Or, they might be skeptical of the new data because they believe that the theory is stronger and more plausible than the source of the new information.
For example, suppose that Bob the ill-informed ranger claims that there are no raccoons near your house. This is new evidence, but it isn’t likely to reduce your confidence in the raccoon hypothesis much. However, if a wildlife expert also claims that there are no raccoons in the area, then it might cause more skepticism of your hypothesis (see table below). (This is one reason naive falsifiability is an incorrect description of science; contradictory data rarely lead to an immediate falsification of a theory. Confirmatory and falsifying data are two inputs in the theory-building system and cause Bayesian updating, but often not radical upheavals to the theory.)
Now let’s consider a real example from the history of science: Dalton’s atomic theory. In the late 1800s, natural scientists had accumulated evidence that made an atomic hypothesis plausible. First, Lavoisier (and others before him) had forwarded the principle of the conservation of mass. Put simply, this states that matter is neither created nor destroyed. Put more specifically, this means that reactants have the same mass as the products of a chemical reaction. And second, Proust had confirmed the law of definite proportions. This means that substances always combine or are comprised of the same proportion of their constituents, i.e., the mass of water is 8/9 oxygen and 1/9 hydrogen.
Dalton theorized that these principles (or laws) followed directly from an atomic theory of matter, whereby each chemical element is composed of indestructible particles that combine in predictable ways to create various substances (e.g., oxygen combines with hydrogen to produce water). The theory was coherent, parsimonious, and very fruitful. However, it was also, by most uses of the term, a just-so story because it was impossible to test directly in Dalton’s day. In fact, it wasn’t until after Einstein devised equations based on atomic theory to explain Brownian motion that we had strong, direct, confirmatory evidence of atoms (see table below).
If one examines the history of science, from Newton to Darwin to the Quantum generation, one sees a similar pattern: Most theories are inferences to the best explanation; and although they make testable predictions (as do evolutionary hypotheses), the relation between the theory and the experiment is rarely so straightforward as in idealized accounts of science, simplified to adhere to the “all science is falsifiable” slogan.
Einstein’s theories of Relativity (Special and General) were both eloquent and parsimonious inferences to the best explanation. They solved outstanding puzzles with small, although often counterintuitive, adjustments to the web of knowledge in physics. And Einstein was explicitly more of a scientific holist than a positivist, believing that theory is often underdetermined by evidence and that theory choice is therefore motivated by other important concerns such as simplicity and fruitfulness.
Critics of evolutionary psychology often berate it for its scientific softness, comparing it unfavorably with physics or chemistry. In fact, physics and chemistry and psychology follow the same basic hypothesis-building procedures. The chief difference is not that physics eschewed “storytelling,” but rather that it can produce mathematically precise predictions. Such predictions are great, of course, but they are not the sine qua non of a mature science. Psychological theories also make predictions, but they are often unavoidably statistical.
Despite the often furious denunciations it provokes, evolutionary psychology is nothing more than the application of Darwinian principles to the human mind.
The evolutionary psychologist accepts the three basic premises of natural selection—variation, differential reproduction, and heredity—and uses them to understand human cognitive and emotional propensities.
One of the most important concepts in this quest is adaptation. The word has a few meanings, which often creates needless ambiguity, but as used in evolutionary psychology, an adaptation is a trait or characteristic that evolved because it increased the inclusive fitness of an organism. In other words, it is a trait that serves a function for an organism that, on average, increased the organism’s ability to survive and reproduce (and also help kin reproduce). The human stomach, for example, is an adaptation that allows human to consume, store, and break down food. A human without a stomach would not live long enough to reproduce; therefore his or her genes would not last for long.
The project of discovering and assiduously studying adaptations in the plant and animal kingdoms is called the adaptationist program. Evolutionary psychologists employ the adaptationist program to study the human mind. The logic is simple. The human mind is a product of the human brain (or, perhaps, is the human brain). The human brain is a product of natural selection. Therefore, the human brain is likely comprised of myriad adaptations — mechanisms that, on average, increased the inclusive fitness of human ancestors.
For example, the human brain is likely equipped with machinery that helps to pick out potential fertile mates, pursue them, and copulate with them. This might seem obvious. Of course humans desire sex with other humans (usually of the other sex)! But such an apparently trivial pattern of behaviors and desires is made possible by a sophisticated suite of adaptations. Humans, after all, could be sexually attracted to frogs, to trees, even to the sun. And they could have no interest in sex, instead contenting themselves with a chaste and detached contemplation of beauty (some philosophers believe this is precisely how humans ought to behave).
This project is often difficult because natural selection also produces byproducts (spandrels), which are characteristics associated with an adaptation that do not serve an adaptive function. Blood is red, for example, because it is composed of hemoglobin (and further because of chemical interactions between oxygen and iron), but the red color presumably serves no adaptive function per se.
Furthermore, humans are cultural creatures who learn to adapt to local environments through instruction and observation; therefore, many adaptive behaviors are not the direct result of specific biological adaptations. There is not, to our knowledge, an adaptation for writing film screenplays, even though being a successful writer is an adaptive behavior which often increases a person’s fitness (by increasing one’s attractiveness in the mating market). Screenwriting ability certainly builds from specific biological adaptations such as the ability to see, to engage in folk psychology, to tell stories—so the challenge, here, is to ascertain the adaptations that allow humans to write successful screenplays.
Making the adaptationist quest even more arduous, evolutionary analysis is inevitably historical. Even when we spot an adaptation, we are often clueless about how exactly it works, when it arose, and how it increased fitness incrementally across time. Human pair bonding is a good example. The desire to form pair bonds is universal and appears designed to increase inclusive fitness, but how do we know when it first arose or what selection pressures led to it? The best we can do is combine a variety of methods, explore all available data, and forward an inference to the best explanation.
Hypothesizing About Adaptations
George Williams and others have adduced criteria for identifying an adaptation:
- Does the putative adaptation increase the organism’s fitness?
- Does it show evidence of special design?
- Is it universal, subtle, complicated?
According to Williams, one should call a trait an adaption if, and only if, it enhances fitness and appears specially designed.
Although I largely agree, I should add the caveat that not all biological adaptations currently increase fitness. There can be a mismatch between an adaptation and the environment such that the adaptation no longer increases an organism’s ability to survive and reproduce. Suppose, for example, that frogs have green skin as an adaptation to a largely green habitat. And suppose that an artistically flamboyant Martian has decided to turn all green plants on earth red for its aesthetic pleasure. Green might now be conspicuous, drawing attention to relatively helpless frogs and decreasing their fitness. In humans, who constantly alter their environment — and often in dramatic ways — this is rather important to bear in mind.
Like physicists, chemists, geologists, and other scientists, evolutionary psychologists use both previous theory and empirical evidence to guide them. Let’s consider Darwin’s theory of sexual selection, for example. For a long time, Darwin was perplexed by the often gaudy ornaments of animals. How could a peacock’s garish and rather cumbersome tail help it to survive? Then Darwin hit upon a brilliant insight. If peahens prefer to mate with elaborately plumed peacocks, then the tail would be favored (accepting certain conditions), because it would increase the mating success of both the peacock and of the peacock’s offspring (who would inherit the flamboyant feathers). Mating success is, after all, the ultimate arbiter of evolutionary success. A bird that has two hundred partners before perishing at five years old has higher fitness, on average, than a bird that lives until 25 but has only two partners.
This observation, inter alia, formed the scaffolding of what became known as sexual selection theory. For our purposes, one important insight (had by Darwin and elaborated by others) that follows from the basic principles of sexual selection theory is that the sexes often faced different pressures related to mating. Males and females have different sized gametes (anisogamy) such that male gametes are smaller (and often more motile) than female gametes. This generally leads to a predictable difference between male and female behavior.
In humans (like other mammals), this difference is also affected by a required eight to nine months of internal gestation. Women can only get pregnant once every nine months (at the most, and this is rather implausible). Men, on the other hand, can impregnate perhaps four or five women a day (also implausible, but not impossible). One way to think about it is as follows.
Imagine a group of 50 men and 50 women. Now, suppose that all the men were killed except one. The lone man could impregnate all the remaining women (siring 50 or more children a year). However, if the situation were reversed (all but one woman is killed), the surviving woman could only get pregnant once in the year (siring perhaps two or three children at most). Therefore, men often stand to gain more from sexual variety than do women, because men’s reproductive success is largely a function of obtained sexual partners, whereas women’s is not.
This theory suggests a straightforward hypothesis: Men possess psychological adaptations that lead them, on average, to desire a greater number of sexual partners than women. This means that if we observe such a difference empirically, and we do, then it is reasonable to conclude that it is a result of psychological adaptations and not, say, the result of expected sex roles, pornography, or other cultural influences — although these certainly might affect sexual behavior (see table below).
Of course, at this point the hypothesis is a just-so story in the Gouldian sense. But so are competing hypotheses such as that male desire for sexual variety is socialized or that it is a byproduct of another adaptation. Furthermore, it is a more plausible story — a better story — than these other hypotheses because it is coherent, parsimonious, and fruitful.
The byproduct hypothesis is unspecific. What is it a byproduct of? Perhaps a greater libido in general. But why would men have a greater libido than women if it did not evolve specifically for sexual variety? And how do we explain the so-called Coolidge effect whereby men experience a rejuvenated libido if introduced to a novel sexual partner? The socialization hypothesis is reasonably parsimonious and fruitful, but it doesn’t cohere with theory.
If men stood more to gain (in fitness terms) across evolutionary history from copulating with a variety of partners than women, then why wouldn’t they have evolved propensities to do so? At this point, we have different degrees of confidence in these hypotheses; my contention is that we should have the most confidence in the adaptive hypothesis because it is the most coherent explanation.
Evolutionary psychologists don’t stop their analysis after forwarding a theoretically coherent explanation, i.e., they don’t simply tell a charming, superficially compelling story and cease further exploration. They seek evidence from sundry sources, including archaeology, anthropology, biology, medicine, primatology, and psychology. Like a good physicist or chemist, they think:
Okay. Here is my hypothesis. What would I expect to see in the world if it is correct? And what would I not expect to see? What would cast significant doubt on my hypothesis?
In other words, they seek confirmatory evidence, but also look for potentially falsifying evidence.
Consider some predictions that follow straightforwardly from the hypothesis that men possess adaptations that make them, on average, more desirous of sexual variety than women. It’s important to note that these predictions do not follow straightforwardly from the competing hypotheses.
First, men should report desiring more sexual diversity than women in most, perhaps all societies. Second, men should report having lower standards than women for short-term mates. Third, men who have many sexual opportunities should have more sexual partners than women who do. Fourth, men should have more variance in reproductive success than women because some men, those who have high status or are otherwise desirable to women, will have many sexual partners, whereas those who are not so desirable will have have few sexual partners. And fifth, men should be more likely to pay for sexual opportunities than women. (Of course, there are many more predictions that follow from the adaptationist hypothesis, but these are a generous offering.)
Evolutionary psychologists have found strong support for each of these predictions (see table above). With each successful prediction, reasonable people should increase their confidence in the adaptive hypothesis. Of course, we are never 100 percent certain that the hypothesis is correct. It could turn out that the theory is misguided and there are better explanations for the data. So, the best we can do is to say, “As of now, the adaptive hypothesis is the best explanation for this pattern of evidence.” If tomorrow scientists uncover strong evidence that in most societies for most of history, women preferred more sexual variety than men, or if they discover compelling evidence that socialization plays a strong role in sexual desire, then we would have to decrease our confidence in the adaptive hypothesis.
A common retort, at this point, is that such adaptive hypotheses are still just stories because researchers cannot forward the exact genetic or neurological mechanisms that give rise to the putative adaptations.
Of course, it is virtually impossible to forward such mechanisms for any complicated human (or nonhuman animal) behavior. Although our knowledge of human genetics is expanding, we still haven’t uncovered the genetic variants “for” many behaviors (often these traits are polygenic, making the task arduous). But, that does not vitiate the explanatory power of an adaptive hypothesis, which serves as a guide for further exploration. Good hypotheses, well supported by theory and evidence, don’t require a perfect understanding of mechanisms (or of latent variables).
Socialization hypotheses are also often forwarded without an understanding of the underlying mechanism(s). The crucial tests of such hypotheses, for now, do not include an articulation of the neurological or genetic mechanisms that cause the behaviors. They include theoretical coherence, explanatory breadth, simplicity, and fruitfulness. This holds for most psychological theories, latent variables, and hypotheses, such as g (the general factor of intelligence), social dominance theory, social identity theory, cognitive dissonance theory, etc. It is simply a misunderstanding of how social science works to demand that researchers offer a molecule by molecule causal description of their latent constructs.
Of course, some evolutionary inspired hypotheses (“just-so stories”) will turn out false and others will exist in a limbo of uncertainty for many years. This is how science works. Researchers advance plausible hypotheses and then scrutinize them. Most fail and are replaced by better hypotheses. For just one example, there is reason to suspect that an ovulatory shifts hypothesis, whereby women become more attracted to particularly masculine men while they are ovulating potentially to secure “good genes,” is incorrect. It is important to note, however, that the hypothesis of ovulatory shifts was interesting and theoretically plausible. And it served as a fruitful guide to research. This is precisely what we should desire of our hypotheses. And, in fact, the failure of the ovulatory shifts hypothesis actually illustrates that adaptive hypotheses are not just-so stories, but are rather inferences to the best explanation that are often revised or replaced as evidence is collected.
I have argued that the just-so story criticism of adaptive hypotheses is based on a flawed understanding of science. At the most extreme, this view of science supposes that theories are only legitimate if they are directly falsifiable (naive falsifiability). And that science progresses by proposing and falsifying theories in succession. But science is actually richer and more complicated. I maintain that inference to the best explanation most accurately describes how science is (and ought to be) practiced. According to this description, scientists forward theories and hypotheses that are coherent, parsimonious, and fruitful. Theories that fit these criteria earn confidence; theories that don’t, don’t. And then scientists seek out all kinds of evidence to strengthen or weaken their confidence. Adaptive hypotheses are no different.
Of course, it is perfectly legitimate to criticize adaptive hypotheses. But it is time to bury the just-so story criticism. It is unspecific, unhelpful, and likely survives only because it scores rhetorical points. It should be replaced with a “I don’t think that hypothesis is plausible because X, Y, Z…” critical model. If history is a good guide, this probably won’t happen soon. But I hope that intelligent critics force me to update my confidence in this hypothesis.