Evolution: In 1992, Leda Cosmides and John Tooby published The Adapted Mind, an edited volume whose introduction ran to over a hundred pages of dense theoretical argument.[9]
The question at its center was deceptively simple and, as they framed it, deeply unsettling for most prevailing assumptions in psychology: is the human mind a general-purpose learning device, shaped primarily by culture and experience, or does it consist of specialized psychological mechanisms - cognitive adaptations - shaped by natural selection during the millions of years of hominid evolution in the Pleistocene?
Are humans, in other words, blank slates who could in principle be raised to be anything, or do we carry in our skulls a Stone Age mind, a set of evolved programs calibrated to the specific challenges of life as a hunter-gatherer in small kin groups on the African savanna?
The question mattered because the answer changed everything about how psychology should be conducted. If the mind is a general learning device, then the mechanisms underlying behavior are largely acquired from culture, and the relevant scientific questions are about learning, socialization, and social structure.
If the mind consists of evolved specialized mechanisms, then those mechanisms have specific inputs, specific computational procedures, and specific outputs that were designed for specific adaptive problems - and studying behavior without understanding those mechanisms is like studying digestion without knowing about enzymes.
Cosmides and Tooby's answer, backed by experimental evidence that became some of the most influential in cognitive science, was that the specialized-mechanism view was correct.
The evidence they cited most prominently came from an obscure logical puzzle called the Wason selection task - and what happened when you changed its framing from abstract logic to social exchange.
Participants who reliably failed to solve the abstract version of the task - if a card has a vowel on one side, it has an even number on the other - solved it easily when the same logical structure was presented as a social contract rule: if a person is drinking beer, they must be over 21.
Cosmides's experiments showed the improvement was not due to familiarity, concreteness, or prior experience with the specific content.[2] It was specific to social exchange contexts, and particularly to detecting cheaters - people who take the benefit without paying the cost.
The logic of natural selection suggested why: ancestral humans who could not reliably detect exploitation in social exchange would have been exploited relentlessly, and would have left fewer descendants. Those with an evolved detection mechanism would have fared better.
The experiment was evidence of a mind shaped by selection to solve a specific adaptive problem.
"Natural selection is the only process known to us that creates complex functional organization in biological systems. It is reasonable to suppose that the human mind is no exception to this principle." - Leda Cosmides and John Tooby, The Adapted Mind (1992)
Key Definitions
Evolutionary psychology: the study of how natural selection and other evolutionary processes shaped the psychological mechanisms underlying human cognition, emotion, and behavior.
Environment of Evolutionary Adaptedness (EEA): the ancestral environment - approximately the Pleistocene, from 2.5 million to 12,000 years ago - in which human psychological mechanisms were shaped by selection; not a specific place but a statistical average of selective pressures.
Evolved psychological mechanism (EPM): a specific information-processing system proposed to exist in the mind, shaped by natural selection to solve a particular adaptive problem; has defined inputs, computational procedures, and outputs.
Modular mind: the view that the mind consists of a large number of specialized functional systems rather than a single general-purpose reasoning engine.
Social brain hypothesis: Robin Dunbar's theory that the expansion of the primate neocortex was driven primarily by the cognitive demands of managing complex social relationships, not ecological challenges.
Dunbar's Number: the predicted maximum stable human social group size of approximately 150, extrapolated by Dunbar from the primate neocortex-ratio-to-group-size regression.
Hamilton's rule: the inclusive fitness condition for the evolution of altruism: an altruistic behavior evolves when rB > C, where r is genetic relatedness, B is the benefit to the recipient, and C is the cost to the actor.[5]
Reciprocal altruism: Robert Trivers's 1971 concept that cooperation can evolve between unrelated individuals when benefits are exchanged over time and cheating can be detected and punished.
Sexual selection: selection arising from differential success in obtaining mates; includes intrasexual competition (male-male competition for access) and intersexual selection (female choice of mates).
Spandrels: Stephen Jay Gould and Richard Lewontin's term for features of organisms that are byproducts of other adaptations or structural constraints rather than direct products of selection; a methodological caution against pan-adaptationist reasoning.
Prepared learning: Martin Seligman's concept that organisms are biologically prepared to acquire certain associations more easily than others; humans show prepared fear conditioning for evolutionarily ancient threats.
Gene-culture coevolution: the process by which cultural practices create selection pressures that drive genetic change, and genetic change enables new cultural practices; lactose tolerance is the standard example.
Key Evolutionary Psychology Concepts
| Concept | Proposed By | Core Claim | Supporting Evidence | Key Critique |
|---|---|---|---|---|
| Adapted Mind / Modular mind | Cosmides and Tooby, 1992 | Mind consists of specialized evolved mechanisms, not a general learning device | Wason selection task: social contract framing dramatically improves performance | Pan-adaptationist risk; spandrels exist |
| Inclusive fitness / Hamilton's rule | W.D. Hamilton, 1964 | Altruism evolves when rB > C; organisms maximize gene propagation, not individual survival | Kin nepotism across species; Westermarck effect | Hard to quantify r and B in complex environments |
| Reciprocal altruism | Robert Trivers, 1971 | Cooperation between non-kin evolves when cheating can be detected and punished | Axelrod tit-for-tat tournaments; iterated Prisoner's Dilemma results | Requires memory, reputation tracking - may overestimate ancestral computation |
| Social brain hypothesis | Robin Dunbar, 1998 | Neocortex expansion driven by social complexity demands, not ecological intelligence | r = 0.78 between neocortex ratio and group size across 36 primate genera | Correlation; not necessarily causal |
| Sexual selection | Darwin, elaborated by Zahavi | Traits favored by mate choice (honesty signals) or intra-sex competition | Cross-cultural mate preference studies (Buss 1989, 37 cultures) | WEIRD sampling; large cultural variation in effect sizes |
| Prepared learning | Martin Seligman | Organisms acquire some associations more easily than others - not blank slates | Snake/spider fear conditioning vs. arbitrary stimuli | Hard to falsify mismatch hypotheses |
The Central Question: General Learning or Specialized Mechanisms?
The Standard Social Science Model, as Cosmides and Tooby labeled the prevailing view, held that the human mind is characterized by extreme plasticity: a general learning capacity that absorbs cultural content and constructs behavior from it.
On this view, the mind's content is largely arbitrary - what people believe, fear, value, and desire reflects their cultural environment rather than any innate specification.
The extreme version of this view, associated with behaviorism and some strands of anthropology, holds that culture can write almost any program on the blank slate of the human mind.
The evolutionary psychological alternative does not deny learning. It argues that learning is not neutral - that evolved mechanisms determine what kinds of information are processed preferentially, what kinds of associations are formed easily, what kinds of problems are solved efficiently.
This is the concept of prepared learning: organisms are not equally capable of learning all associations.
Humans learn to fear snakes, spiders, and heights very easily; we learn to fear electrical outlets and cars very slowly even though the latter are objectively more dangerous to modern humans than the former. The fear system is calibrated to ancestral threats, not to objective modern danger.
The Wason selection task experiments, Cosmides's most important contribution, made the case with unusual clarity.[1] Conditional reasoning - "if P then Q" - is straightforwardly taught in introductory logic courses, and adults who have been through formal education generally understand the rule.
Yet most people, in experiments run thousands of times since Peter Wason first designed the task in 1966, fail to solve the abstract version correctly. They tend to look for confirming evidence rather than disconfirming evidence, missing the logical necessity of checking the P-not-Q case.
When the same logical structure is embedded in a social exchange scenario - detecting whether someone has received a benefit without meeting the cost - the failure rate drops dramatically.
Cosmides's systematic experiments showed this was specific to social contract framing: precaution schemas (checking for environmental dangers) showed improvement too, but social contract specifically improved performance on detecting violations.
The conclusion: the mind does not contain a general-purpose conditional reasoning system of the kind formal logic describes.
It contains specialized systems for social contract reasoning and cheating detection that happen to implement the same logical structure efficiently, while the general form is processed poorly.
Hamilton's Rule and Kin Selection
Before evolutionary psychology existed as a named field, William D. Hamilton's 1964 paper on the genetical evolution of social behavior transformed the theoretical foundations of how biologists thought about altruism.
The puzzle was fundamental: natural selection favors traits that increase individual fitness. How, then, can self-sacrificing behavior evolve? Why do animals help each other at all?
Hamilton's solution was elegant. Natural selection does not maximize individual reproductive success; it maximizes inclusive fitness - the propagation of genes into future generations. Copies of your genes exist not only in your own body but in the bodies of your relatives, in proportion to your degree of relatedness.
A gene that causes you to sacrifice yourself to save more than two siblings (who each share 50% of your genes) can spread by natural selection, because the gene is more likely to persist in the population than if you had not sacrificed yourself.
This is Hamilton's rule: altruistic behavior evolves when rB > C, where r is the coefficient of genetic relatedness, B is the fitness benefit to the recipient, and C is the fitness cost to the actor.
Hamilton's rule explains nepotism across species, including humans: all else equal, individuals help genetic relatives more than non-relatives, and they help closer relatives more than distant ones.
Human kin psychology is sophisticated: people use multiple cues to estimate kinship, including facial similarity, known biological relatedness, childhood co-residence (the Westermarck effect), and maternal perinatal association.
The Westermarck effect - the finding that people raised together from early childhood develop sexual aversion toward each other regardless of biological relatedness - is an elegant demonstration of an evolved kin-detection mechanism: co-residence in infancy was a reliable cue to kinship in ancestral environments, so the system uses it to prevent inbreeding without requiring explicit knowledge of genealogy.
Reciprocal Altruism and the Evolution of Cooperation
Kin selection explains cooperation among relatives but cannot explain cooperation between unrelated individuals, which is extensive in human societies and present in other species. Robert Trivers's 1971 paper on reciprocal altruism provided the theoretical framework for cooperation between non-kin.[6]
Cooperation between unrelated individuals can evolve when interactions are repeated, when the benefits of cooperation exceed the costs, and when there are reliable mechanisms for detecting and punishing cheaters.
Under these conditions, a strategy of cooperation contingent on partners' past cooperation - a form of conditional altruism - can invade and stabilize in a population of unconditional defectors.
Robert Axelrod's famous computer tournaments of the 1980s, in which different strategies competed in iterated Prisoner's Dilemma simulations, found that a simple strategy of initial cooperation followed by immediate reciprocation (tit-for-tat) defeated more sophisticated strategies.
The cognitive requirements for reciprocal altruism are demanding: remembering past interactions with specific individuals, tracking reputation over time, identifying cheaters, and calibrating cooperation based on partners' history.
These are precisely the kinds of cognitive tasks that Dunbar's social brain hypothesis proposes drove neocortex expansion.
A mind shaped by the demands of managing a complex web of reciprocal relationships - tracking who has cooperated, who has defected, who owes whom - would look exactly like what Cosmides and Tooby's modular model predicts.
Sexual Selection and Mate Preferences
Charles Darwin recognized that natural selection, which selects for survival, could not by itself explain many conspicuous animal traits that seem to reduce survival chances - the peacock's tail, the elk's antlers, the elaborate songs of birds.
He proposed sexual selection as a separate process, driven by the competition for mates rather than the competition for survival.
Sexual selection operates through two mechanisms: intrasexual selection (competition between members of one sex, typically male, for access to the other sex), and intersexual selection (choice by members of one sex, typically female, of mating partners).
The two mechanisms produce different kinds of traits. Intrasexual selection favors size, strength, and weaponry for competitive confrontation. Intersexual selection favors traits that reliably signal genetic quality or resource acquisition to choosy mates.
Zahavi's handicap principle explains how honest signaling can evolve under sexual selection: only genuinely high-quality individuals can afford the cost of a truly costly signal.
The peacock's tail is costly to produce and carry; only healthy peacocks can grow extravagant tails, making tail extravagance a reliable signal of health.
This logic applies to human signals as well: conspicuous consumption, artistic ability, athletic performance, and humor have been analyzed as costly signals in human mate attraction contexts.
David Buss's 1989 study of mate preferences across 37 cultures, recruiting over 10,000 participants in countries ranging from the United States to Zimbabwe to China, found cross-cultural patterns predicted by evolutionary theory: men placed greater relative weight on physical attractiveness and youth cues in potential mates; women placed greater relative weight on resource acquisition capacity, status, and ambition.[4]
These results were statistically consistent across cultures, though the magnitudes of sex differences varied substantially.
Critics note that cross-cultural consistency is compatible with both biological and socialization explanations, and that social and economic equality between sexes shifts these preferences significantly in predicted directions. The debate over what these patterns mean remains active.
Language as Evolved Capacity
The evolution of language is simultaneously one of evolutionary psychology's strongest cases and one of the most contested problems in cognitive science.
The case for language as an evolved biological capacity rests on multiple lines of evidence: language is universal across human cultures; children acquire it without explicit instruction during a critical developmental period; it has a dedicated neural substrate; it follows universal structural principles across languages (Universal Grammar, as theorized by Chomsky); and it has a specific genetic component.
The FOXP2 gene provides the clearest genetic evidence. Mutations in FOXP2 in the KE family, a British family in which a dominant mutation affected several generations, produce a specific deficit combining impaired speech motor control and grammatical processing difficulties.[7]
The human FOXP2 variant differs from the chimpanzee version at two amino acid positions, and these changes show signatures of recent positive selection - appearing to have spread through the human population within the last 200,000 to 400,000 years.
FOXP2 is not a "language gene" in the sense of a single gene for all language; it is expressed in motor circuits involved in complex sequential movements and in basal ganglia circuits involved in procedural learning. But the evidence that specific genetic changes contributed to human language capacity is convincing.
The theoretical debate pits Chomsky's view - that language evolved primarily for thought rather than communication, with the communication function as secondary exaptation - against Pinker's language instinct argument that language is a complex adaptation shaped specifically for communication.
Michael Tomasello's account adds a third position: the cognitive foundation of language is shared intentionality - joint attention, the ability to direct and follow others' attention to third objects, and understanding communication as a cooperative meaning-making act.
Tomasello argues that this capacity for shared intentionality is the key human-specific prerequisite, and that its evolution preceded and enabled language rather than being a product of it.
The Social Brain Hypothesis and Dunbar's Number
Robin Dunbar's social brain hypothesis, published in a landmark 1998 Evolutionary Anthropology paper, found that neocortex ratio - the proportion of total brain volume dedicated to the neocortex - correlates strongly (r = 0.78) with mean social group size across 36 primate genera.[3]
The ecological intelligence hypothesis, which attributes encephalization to foraging complexity and home range size, showed weaker correlations with neocortex size in Dunbar's analysis.
Extrapolating the primate regression to the human neocortex ratio predicts a natural human group size of approximately 150, a value Dunbar has documented in diverse human social contexts: the typical size of hunter-gatherer bands in the ethnographic record, the size of Neolithic village sites in the archaeological record, the splitting threshold for Hutterite communities, the size of typical personal social networks, and the approximate size of effective military units across cultures.
The 150 figure represents the limit of relationships maintainable through direct personal knowledge and reciprocal investment; beyond this number, social cohesion requires formal hierarchy and institutional enforcement.
Dunbar extended the hypothesis to language: he proposed that language evolved as vocal grooming, a more time-efficient mechanism for maintaining social bonds at group sizes larger than those supportable through physical grooming alone.
Physical grooming is limited by the requirement of pair-wise contact; language allows one individual to maintain bonds with multiple others simultaneously, and to maintain them at a distance.
Critically, the content of human language use supports this interpretation: Dunbar's research on conversation content finds that approximately two-thirds of conversation time is devoted to social topics - relationships, reputation, who did what to whom - rather than to factual information about the physical world.
Critiques: Spandrels, Falsifiability, and WEIRD Science
The critiques of evolutionary psychology from within science are serious and deserve full engagement.
Stephen Jay Gould and Richard Lewontin's 1979 "spandrels" paper established the most enduring methodological challenge.[8] Gould and Lewontin argued that evolutionary biologists were too quick to assume that every feature of an organism is a direct adaptation shaped by selection for its current function.
Spandrels in architecture - the spaces between arches that appear in any vaulted construction - are not designed features but structural byproducts.
Biological features can similarly be byproducts of other adaptations, developmental constraints, or genetic drift rather than direct products of selection.
The evolutionary psychologist's habit of constructing post-hoc adaptive stories - "this behavior exists because it was adaptive for our ancestors in this way" - is difficult to falsify when the ancestral environment is poorly specified and alternative hypotheses are not rigorously tested.
Gould's related concept of exaptation - the co-option of existing structures for new functions - adds another complication: traits shaped by selection for one function can be recruited for entirely different functions without new selection for the new function.
Human reading ability is not an adaptation for reading, which has existed for only a few thousand years; it is built from evolved visual systems, pattern recognition, and language circuits that were shaped by selection for other purposes. The adaptationist framing, applied carelessly, would generate a false story.
The WEIRD problem is a sampling critique: most evolutionary psychology studies, including the most influential ones, have been conducted on Western, Educated, Industrialized, Rich, Democratic populations.
Cross-cultural generalizability is an empirical claim that requires cross-cultural data, and the evidence is mixed: some proposed universals (basic emotion recognition, kin favoritism, cheater detection) appear robustly cross-cultural; others (specific mate preference patterns, social comparison norms) show substantial cultural variation.
Behavioral genetics provides a complementary and sometimes corrective perspective: genetic factors are real and important contributors to psychological variation, but they operate through development and gene-environment interaction in ways that make simple adaptationist accounts of specific behavioral phenotypes often oversimplified.
Prepared Learning and the Mismatch Hypothesis
Martin Seligman's concept of prepared learning - that organisms are not blank slates for conditioning but are biologically prepared to acquire some associations more easily than others - fits neatly within the evolutionary psychological framework.
Human infants show preferential attention to faces from birth. Young children show enhanced fear learning for snakes and spiders relative to arbitrary fear stimuli, even before any personal experience of these animals.
Social fear learning - acquiring fear after observing others' fear reactions - is particularly strong for evolutionarily relevant threat stimuli.
The mismatch hypothesis extends this logic: many modern psychological problems are the outputs of evolved mechanisms operating in environments radically different from those that shaped them. Sugar and fat cravings, evolved when calorie-dense food was scarce, drive overconsumption in environments of cheap food abundance.
Status competition, evolved for small groups of 50-150 people, now operates against a global comparison class of billions, driving anxiety through social media's provision of constant upward comparison.
Fear responses calibrated for predators and hostile humans now fire chronically in response to abstract economic and social threats, producing generalized anxiety disorders.
The mismatch hypothesis must be applied with methodological care: it is possible to construct a mismatch narrative for almost any modern psychological condition, which makes the hypothesis unfalsifiable without independent evidence about the ancestral environment and the operation of specific evolved mechanisms.
Connections
For the mechanisms of natural selection that underlie all evolutionary psychology, see how evolution works. For the role of evolved emotional systems in human psychology, see why we have emotions. For why disgust in particular has been recruited for moral purposes, see why we feel disgust.
Sources & Further Reading
- Cosmides, L., & Tooby, J. (1989). Evolutionary psychology and the generation of culture, Part II: A computational theory of social exchange. Ethology and Sociobiology, 10(1-3), 51-97. DOI: 10.1016/0162-3095(89)90013-7
- Cosmides, L. (1989). The logic of social exchange: Has natural selection shaped how humans reason? Cognition, 31(3), 187-276. DOI: 10.1016/0010-0277(89)90023-1
- Dunbar, R. I. M. (1998). The social brain hypothesis. Evolutionary Anthropology, 6(5), 178-190. DOI: 10.1002/(SICI)1520-6505(1998)6:5<178::AID-EVAN5>3.0.CO;2-8
- Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12(1), 1-14. DOI: 10.1017/S0140525X00023992
- Hamilton, W. D. (1964). The genetical evolution of social behaviour I and II. Journal of Theoretical Biology, 7(1), 1-52. DOI: 10.1016/0022-5193(64)90038-4
- Trivers, R. L. (1971). The evolution of reciprocal altruism. Quarterly Review of Biology, 46(1), 35-57. DOI: 10.1086/406755
- Lai, C. S. L., Fisher, S. E., Hurst, J. A., Vargha-Khadem, F., & Monaco, A. P. (2001). A forkhead-domain gene is mutated in a severe speech and language disorder. Nature, 413(6855), 519-523. DOI: 10.1038/35097076
- Gould, S. J., & Lewontin, R. C. (1979). The spandrels of San Marco and the Panglossian paradigm: A critique of the adaptationist programme. Proceedings of the Royal Society of London B, 205(1161), 581-598. DOI: 10.1098/rspb.1979.0086
- Tooby, J., & Cosmides, L. (1992). The psychological foundations of culture. In J. H. Barkow, L. Cosmides, & J. Tooby (Eds.), The Adapted Mind: Evolutionary Psychology and the Generation of Culture (pp. 19-136). Oxford University Press.