How are evolutionary psychology hypotheses tested?
On the face of it, the job of evolutionary psychology seems impossible: to elucidate on the evolutionary history of brains and behavior, neither of which fossilize or necessarily leave any physical evidence behind. Indeed, many have argued that it is impossible, or nearly so.
In his famous paper The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme , biologist Stephen Jay Gould argued that many hypotheses in evolutionary biology were “just so stories”, convenient and plausible-sounding explanations of the kind offered Dr. Pangloss in Voltaire’s Candide. He insinuated that biologists (and later sociobiologists) leapt to their conclusion without adequate consideration of alternative explanations. That was in 1979. Today, more than three decades later critics of evolutionary psychology make the very same arguments. Here’s an excerpt from Wikipedia’s page  on such criticism:
A frequent critique of [evolutionary psychology] is that its hypotheses are difficult or impossible to adequately test, challenging its status as an empirical science. As an example, critics point out that many current traits likely evolved to serve different functions than they do now, confounding attempts to make backward inferences into history. Evolutionary psychologists acknowledge the difficulty of testing their hypotheses but assert it is nevertheless possible.
Critics argue that many hypotheses put forward to explain the adaptive nature of human behavioural traits are “Just-so stories”; neat adaptive explanations for the evolution of given traits that do not rest on any evidence beyond their own internal logic.
Where Gould was alleging insufficient rigor, others simply state flatly that such hypotheses cannot be tested, period. I will explain a few methods used in evolutionary psychology (hereafter EP) but first, a few basics.
It is quite correct that some hypotheses we might construct are untestable. This is true in all sciences. For example, when did our ancestors first use tools? Since it is likely that the first tools were fashioned from wood and plant material which decays too quickly to leave evidence, we can only make educated guesses. EP researchers are well aware of such limitations, and reminders from critics fail to astonish. Similarly, some hypotheses are very difficult, but not impossible to assess. Is female orgasm an adaptation, or just a side-effect of the male adaption for orgasm necessary to copulation? Book-length discussions have been written on both sides of that question.
Lastly, I believe that when conceived, our hypotheses pretty much are “just so stories”, at least in the sense that they are explanations for phenomena wrought in our own imaginations. It is true in EP, but it’s just as true for most new scientific hypotheses. Agostino Bassi conjectured that tiny invisible living bits caused disease. His just-so-story sounded crazy when he proposed it in 1835, but he (and others after him, such as Louis Pasteur) collected the evidence, and so we accept it as factual. The essence of the criticism is not that hypotheses are narratives congenial to our personal viewpoints, but what happens next: are they tested properly, or just assumed to be correct after going through the motions? I’ll refrain from commenting on the temerity of the latter charge and answer it instead.
Methods in EP
In his book, Evolutionary Psychology: The New Science of the Mind , prominent EP researcher David Buss lists 5 important ways that an EP hypothesis can be tested:
1. Compare different species
2. Compare males and females
3. Compare individuals within a species
4. Compare the same individuals in different contexts
5. Experimental methods
And that empirical data used in the above come from
1. Archaeological records
2. Data from hunter-gatherer societies
5. Life-history data and public records
6. Human products
The first item on his list of methods is the one that I am the most familiar with, as it is the method I used in my paper on sex differences in spatial cognition. The logic of the cross-species comparison is that if a phenotypic trait(any observable aspect of organism or its ecology) X is caused by selective pressure Y, then we should see a trend toward X whenever we can observe or infer that Y is present. The more species in which the correlation holds true, the more confident we might be that Y is inducing X. For example, if I hypothesize that arctic-dwelling foxes and hares are white in color because camouflage is important to the survival, and thus fitness, of predators and prey, then it is predicted that many predator and prey animals should be white in color if they live in snowy locales. Indeed, this is what we do find. Species diversity makes the conclusion stronger. Birds, lemmings, and polar bears all fit the predictions of the hypothesis, which vastly reduces the chance that the species are similar by way of being closely related.
In one relatively well-known example, researchers have hypothesized that testes size in animals may correlate with the degree of intrasexual competition for mates. That is to say, in species where a female may copulate with more than one male, the males tend to have larger testes to produce more sperm which compete with the sperm of other males. Thus, mating system may be predicted by relative testes size. This has been found to be the case in primates; gorillas which control harems through mate-guarding and face no sperm competition have tiny testicles (relative to body size) while promiscuous chimpanzees have enormous testicles. The hypothesis has even been tested in other kinds of animals. J.D. Rising found that testes size correlated with mating system among several closely-related sparrow species . Humans, when compared to other primates, have medium-sized testes which may indicate a moderate level of female promiscuity in our evolutionary past.
I will cover this type of method in great detail in December, concurrent with the release of my paper. I will speak to the other 4 methods listed above in up-coming posts.
Common potential confounds
Any hypothesis about human psychology rooted in biology faces a number of methodological challenges. Most commonly, in EP we wish to know if a phenotypic trait is an adaptation. The first challenge then is to make sure the trait is not purely an environmental, social or cultural product. Americans, for example, are familiar with the “thumbs-up” gesture. Although it is an observed behavior, it is not understood universally and so cannot be an adaptation. EP’s compare many cultures in order to find out if a trait is a universal human characteristic. I am presently working on a project in which data will be collected in over 20 countries including a pre-state hunter-gatherer society.
A trait might also turn out not to be an adaptation, but a consequence of adaptions as well as random processes such as genetic drift. In humans, the belly button is not adaptive. It is the scar left by a detached umbilical cord. Blood is red, but not because it was selected for redness; it just happens that a protein good at holding oxygen, hemoglobin, is red in color.
We could also get the adaptive significance wrong. Charles Darwin thought that the crinkly shape of human ears was vestigial, lacking a functional explanation. It was later discovered the auricle, or external ear, is well-designed for catching sounds from many directions. We can forgive Darwin, for he lived in a time before acoustic science existed, but we should be on our guard lest we make a similar error.
These are some of the criticisms levied broadly against the entire discipline of evolutionary psychology. Such complaints, without focus on specific examples, are misguided. While my last paper was under peer-review (and pre-review discussions among myself and co-authors) these are precisely the possible problems we discussed at length, and in peer review, had to answer to the satisfaction of the reviewers. These issues are part and parcel of doing EP work. They are part of the earliest discussions researchers have when designing a study. They are integral to the peer-review process. They become sections of almost all published EP papers involving human behavior and ecology, including my own. This does not mean we never get it wrong, but it does mean thoughtful criticism must spell out how, based on actual examples of research, and not merely be armchair-quarterback speculation.
1. Gould, S. J., & Lewontin, R. C. (1979). The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proc R Soc Lond B Biol Sci.
3. Buss, D. M. (2008). Evolutionary Psychology: The New Science of the Mind. Pearson/Allyn and Bacon.
4. Rising, J. D. (1996). Relationship between Testis Size and Mating Systems in American Sparrows (Emberizinae), 113(1), 224–228.