RESEARCH DOSSIER
Adaptation, Selection, Drift, and Evolutionary Explanations of Behavior
Abstract. Evolutionary explanations of behavior require the identification of phenotypes, heritable variation, population histories, ecological conditions, and processes responsible for change across generations. This fundamental research monograph examines the development of evolutionary behavioral theory from Darwin and modern population genetics to ethology, behavioral ecology, human evolutionary psychology, and contemporary quantitative and social evolutionary research. A critical narrative synthesis of original studies, formal models, field manipulations, longitudinal research, comparative methods, and methodological reviews distinguishes historical adaptations from current utility, genetic drift, exaptation, correlated responses, developmental constraints, phenotypic plasticity, and cultural transmission. Empirical case studies include experimental selection on industrial melanism, Darwin's finches during environmental fluctuation, and guppy life-history responses under altered predation. The analysis assesses multivariate fitness models, phylogenetic nonindependence, measurement validity, causal inference, social genetic effects, cross-cultural generalization, and the limits of molecular evidence for behavioral origins. Its synthesis establishes how evolutionary hypotheses gain support through discriminating predictions and independent tests of competing processes. A final comparative discussion specifies when artificial adaptation provides a useful model and when biological evolution requires additional inheritance and population conditions. This work is a critical narrative research synthesis; it reports no original experiment, registered systematic review, or meta-analysis.
Keywords: evolutionary explanation of behavior; biological adaptation; natural selection; genetic drift; exaptation; behavioral evolution; fitness; behavioral ecology; evolutionary psychology; developmental constraint; phenotypic plasticity; correlated selection; quantitative genetics; phylogenetic comparative methods; behavioral variation; human evolutionary psychology; gene–culture interaction; Artificial Era
Introduction: What an Evolutionary Explanation Must Explain
Behavior is an observable sequence of activities generated by organisms whose bodies, nervous systems, developmental histories, social relationships, and ecological circumstances jointly determine what they do. An evolutionary explanation asks why the capacities and dispositions producing those activities have acquired particular distributions and forms across generations and lineages. This question is historical and population-based. It becomes scientifically answerable when investigators specify the behavioral phenotype, the population and time scale, the pattern of inheritance or reliable transmission, and the proposed processes capable of producing the observed distribution. Darwin's account of descent with modification provides the historical framework; Wright's population-genetic analysis supplies a framework for competing population processes; and Tinbergen's analysis locates evolutionary questions alongside proximate ones rather than allowing one to replace the other. Darwin (1859) Wright (1931) Tinbergen (1963)
The present dossier examines the standards by which researchers distinguish adaptation by natural selection from drift, exaptation, correlated response, genetic and developmental constraints, ecological plasticity, and cultural transmission. Its unit of analysis is the quality of an evolutionary explanation of a behavioral phenotype. The subject therefore spans comparative psychology, population genetics, behavioral ecology, quantitative genetics, and research methodology, while maintaining a coherent question: what observations justify one account of behavioral origins rather than another? Explaining current behavioral function alone is insufficient to identify the historical mechanism. Equally, documenting heritability, cross-population similarity, or a statistically significant behavioral association does not establish that a particular feature was selected for its currently proposed function. The principles established in this inquiry apply to human and nonhuman organisms, with special attention to the distinct evidentiary difficulties posed by human psychological traits. Gould & Lewontin (1979) Schmitt & Pilcher (2004) Nielsen (2009)
This work uses a critical narrative synthesis of identifiable primary theoretical contributions, field experiments, longitudinal studies, comparative analyses, and later reviews. It reports no new participant data and does not claim the search, screening, and risk-of-bias procedures of a systematic review or the pooled estimates of a meta-analysis. Original research is evaluated in terms of design, causal identification, historical inference, construct validity, and independent corroboration. Foundational contributions are treated as historical evidence of the development of scientific concepts; their claims are considered separately from subsequent empirical support. The synthesis uses evolutionary equations as models of processes under stated assumptions, rather than as direct measurements of any particular human or animal psychological adaptation. Mayr (1961) Lande & Arnold (1983) Bateson & Laland (2013)
The Explanatory Architecture of Evolutionary Behavioral Science
An explanation of a behavioral event may concern the immediate sensory information that provokes it, the learned expectations that shape it, its developmental emergence, the differential survival and reproduction associated with the disposition, or the history of the lineage in which it appears. These questions demand distinct instruments and time scales. Tinbergen formulated major problems of behavioral biology in terms of causation, development, survival value, and evolution; Bateson and Laland subsequently examined how the framework should be used in light of changes in genetics, development, and ecology. A complete account of a trait is often assembled from several studies, each establishing a different part of the chain. Claims about evolutionary function become stronger when they are linked to, but carefully distinguished from, mechanistic and developmental accounts. Tinbergen (1963) Bateson & Laland (2013)
Consider a hypothetical avoidance response. Behavioral observation can establish when an organism withdraws; a conditioning experiment can show how prior experience alters that response; a developmental study can track its emergence; a field study can test whether response variation predicts survival under particular threats; and a phylogenetic analysis can evaluate whether the trait covaries with ecological exposure across lineages. Even a convincing combination of these results leaves open whether the response arose specifically through selection for threat avoidance, was inherited as part of a more general arousal system, or is a plastic response expressed by an older mechanism. Such explanations predict different patterns of specificity, covariation, historical distribution, and reaction to environmental manipulation. The research task is to construct informative contrasts and collect evidence capable of discriminating among them. Mayr (1961) Felsenstein (1985) Schmitt & Pilcher (2004)
The scientific meaning of fitness is central. In an evolutionary model, fitness denotes an expected contribution to future generations in a specified population and environment, not an organism's perceived well-being, moral merit, or cognitive sophistication. An apparently efficient problem-solving response may carry metabolic or opportunity costs. A trait that improves immediate survival may reduce mating opportunities; an action beneficial to recipients may be costly to an actor. For social traits, models must identify whose reproductive interests are counted and how relatedness, assortment, and direct and indirect effects enter the hypothesis. Clarifying these quantities prevents the routine conversion of every desirable behavioral outcome into a purported evolutionary advantage. Hamilton (1964) Price (1970) West et al. (2007)
Historical Formation: From Darwinian Behavior to Population Genetics
Darwin's 1859 formulation connected adaptation to differential preservation of heritable variants within populations over generations, and his 1871 treatment of sexual selection extended the explanatory problem to competition and mating preferences. The 1872 study of emotional expression explored continuities in expressive behavior between humans and other animals and provided hypotheses concerning inherited behavioral patterns and their transformation. These works are historically consequential because they encouraged the study of mental and behavioral traits as biological characteristics with histories. They are not modern validation studies of particular neural mechanisms, and a plausible Victorian functional account should not be mistaken for an independently demonstrated history of selection. Reading Darwin in his original intellectual context reveals both the extraordinary reach of his proposals and the additional empirical commitments imposed by later genetics and experimental design. Darwin (1859) Darwin (1871) Darwin (1872)
The early twentieth-century synthesis between Mendelian inheritance and evolutionary theory transformed broad historical explanations into explicit models of changing gene frequencies. Wright analyzed mutation, selection, migration, mating systems, and random sampling in finite populations in his 1931 paper. The resulting conceptual space made it possible to ask whether a behavioral difference could have emerged through differential reproduction, stochastic fluctuation, dispersal, or interactions among these processes. Subsequent quantitative-genetic approaches treated traits influenced by multiple loci and separated the effects of direct selection from changes carried along through genetic correlation. A behavioral phenotype can therefore evolve even when its measured current function is not the property under direct selection. Wright (1931) Lande (1976) Lande & Arnold (1983)
The development of ethology, behavioral ecology, and evolutionary psychology further expanded the range of questions. Tinbergen emphasized the combined investigation of mechanisms, development, adaptive significance, and history. Hamilton's analysis of social behavior introduced a rigorous account of indirect fitness effects under explicit assumptions. Later criticisms of the adaptationist program challenged the habit of treating each visible trait as an individually optimized solution; the proposed category of exaptation provided language for the recruitment of a feature to a role different from that for which it arose. These conceptual changes mattered empirically because they enlarged the set of testable alternatives that a behavioral researcher must consider. Tinbergen (1963) Hamilton (1964) Gould & Lewontin (1979) Gould & Vrba (1982)
Defining Adaptation and the Evidence for Historical Design
An adaptation, in the historical sense used here, is a heritable feature whose form was shaped by past natural selection because of effects that contributed to differential reproductive success in the relevant ancestral environments. An exaptation is a feature currently used for a function that was not the selected-for function responsible for its origin. A current beneficial effect is evidence of utility, while the claim of adaptation concerns a historical causal process. The origin and subsequent modification of traits can involve several episodes of selection and co-option, so the history of a complex behavioral system may contain adaptive and exaptive elements together. Gould and Vrba's distinction gives investigators a language for reconstructing such mixed histories without forcing a single-origin story. Gould & Vrba (1982) Gould & Lewontin (1979)
Evidence for adaptation is cumulative and comparative. A credible account specifies a recurrent challenge, the information available to the organism, the phenotype that could be favored, its predicted organization, and alternative causal explanations. Evidence of condition-sensitive performance, specialized information processing, recurrence across populations, developmental canalization, or a fitness benefit can be relevant, but none functions as a universal stand-alone test. Schmitt and Pilcher proposed integrating different evidentiary forms rather than asserting adaptation from one conspicuous fit between a trait and a proposed function. The historical prediction is strongest when independent lines of inquiry converge, including comparative distribution, genetic architecture, ecological experiments, and theoretically discriminating trade-offs. Schmitt & Pilcher (2004) Felsenstein (1985) Kingsolver et al. (2001)
The inference remains asymmetrical: observing an adaptation can support an evolutionary hypothesis, but a highly elaborate phenotype by itself does not identify the exact agent, location, or historical duration of selection. Similarity across species may reflect inherited ancestry; similarity across populations may reflect common development or shared culture; and an apparent efficient design may be produced by learning within a genetically widespread response system. Gould and Lewontin highlighted the explanatory importance of organismal integration and structural byproducts; Nielsen later argued that even molecular signals associated with selection, combined with present-day function, do not automatically establish the intended history of adaptation. The appropriate practice is explicit model comparison supported by independent tests. Gould & Lewontin (1979) Nielsen (2009) Schmitt & Pilcher (2004)
Natural Selection, Differential Fitness, and Behavioral Variation
Natural selection is a population process in which variation associated with differential reproductive contribution changes the distribution of inherited characteristics across generations. Its operation requires variation, some relevant transmission across generations, and a covariance between transmitted characteristics and realized reproductive success. Price's covariance formulation shows how changes in a population mean can be decomposed into components associated with differential reproductive contribution and transmission. This identity is general; interpreting its terms requires a causal model, and an observed selection differential does not by itself identify the ecological agent of selection. Behavioral studies therefore need to measure what individuals actually do and connect that variation to demography under stated assumptions. Price (1970) Lande & Arnold (1983)
Direct selection on a behavioral variable is different from a correlated response. Suppose exploratory behavior covaries with body size, condition, age, or a neuroendocrine trait that is itself under selection. Selection may alter the mean behavior even when the direct causal contribution of exploration to fitness is negligible. Lande and Arnold's framework addresses multivariate selection by relating relative fitness to several phenotypes and estimating gradients conditional on their statistical relationships. This approach improves inference when the relevant traits have been measured, but it does not make observational fitness regressions immune to omitted variables, measurement error, nonrandom survival, or genetic confounding. In particular, phenotypic covariance should not automatically be interpreted as genetic covariance. Lande & Arnold (1983) Lande (1976)
Field estimates are environmentally contingent. Kingsolver and colleagues reviewed estimates of selection in natural populations, providing an empirical basis for studying the variety of selection strengths reported in the field while drawing attention to variation in measurement and publication practices. The same action may be favored in one season, carry costs in another, or interact with density and social environment. Short-term changes in a behavioral response can track a contemporary selective regime without implying that the response originated in that regime. Successful evolutionary explanation therefore links measured traits, realistic fitness components, inheritance, ecological context, and evidence accumulated over appropriate generations. Kingsolver et al. (2001) Bolnick et al. (2011) Grant & Grant (2002)
Drift, Neutral Change, and Finite Population Histories
Genetic drift denotes stochastic changes in allele frequencies resulting from finite sampling across generations. Its expected importance depends on population structure, effective population size, inheritance, and the strength of selection; neutral variants can rise in frequency or disappear without a reproducible fitness advantage. Wright's models placed these stochastic effects inside the same population-genetic analysis as selection and migration. Kimura's neutral theory, developed chiefly for molecular substitutions, demonstrated why a substantial fraction of molecular change could be understood without attributing each substitution to positive selection. The relevant lesson for behavioral science is conceptual discipline: historical change and heritable difference do not, in themselves, demonstrate adaptive optimization. Wright (1931) Kimura (1968) Gillespie (2004)
A distinction is essential between neutral genetic variation and behavioral neutrality. A neutral molecular change may occur in a genomic region connected to a behavioral trait, but the genetic neutrality result does not directly show that the behavior is neutral. Conversely, a behavioral difference with no detected fitness association in one study may have effects that were unmeasured, rare, developmentally contingent, or present in a previous environment. Inferring drift as the dominant explanation for behavioral divergence requires an explicit null model, estimates of population history and genetic architecture where possible, and comparison with expected divergence under selection and migration. The uncertainty is greatest for psychological constructs whose genetic bases and cross-population measurement invariance are poorly established. Lande (1976) Gillespie (2004) Schmitt & Pilcher (2004)
Founding events, isolation, bottlenecks, nonrandom dispersal, and changing effective population sizes all complicate historical inference. A difference between present populations may reflect inherited stochastic changes, a common response to different environments, differential selection, cultural transmission, or combinations of these mechanisms. Researchers can use pedigrees, genomic data, common-garden designs, longitudinal sampling, and model-based historical reconstruction to narrow this space; each instrument has its own assumptions. Drift and selection are components of a causal history that can operate simultaneously. Treating drift as the default whenever an adaptive story is unconvincing would be as unwarranted as treating selection as the explanation whenever behavior appears useful. Wright (1931) Lande (1976) Nielsen (2009)
Experimental Natural History: Selection Under Environmental Change
The strongest field evidence for selection ties a defined environmental perturbation to differences in survival or reproduction associated with measurable traits. In the Galápagos, Boag and Grant documented nonrandom survival of Darwin's finches during drought, when the distribution of available seeds changed and birds with particular body and beak characteristics were more likely to survive. Their study showed selection acting in an identifiable natural episode. Its target variables were morphological and ecological, and the behavioral mechanism of seed handling should be evaluated as part of the functional chain rather than assumed directly from beak measurements. The study's significance lies in the connection between environmental conditions, measurable trait variation, and differential survival. Boag & Grant (1981)
Grant and Grant's longer research program revealed why a documented episode of selection does not license a timeless account of evolutionary direction. Across decades, changes in ecological circumstances altered the direction and intensity of selection; hybridization and correlated morphological traits also affected trajectories. The investigators distinguished the relative predictability of short-term response given selection and inheritance from the challenges of predicting future environments and longer-term outcomes. A psychological analogue would examine whether behavioral sensitivity to food availability improves success in specified contexts, how the response is inherited or learned, and whether the selective regime remains stable enough for cumulative evolutionary change. Historical contingency is part of the explanation, rather than experimental noise that can be averaged away. Grant & Grant (2002) Lande & Arnold (1983)
Industrial melanism in moths provides an earlier example of experimentally investigating differential visibility and predation under different environmental backgrounds. Kettlewell's 1955 experiments are historically important for linking environmental appearance, predators, and phenotype frequencies, while the broader history of this model illustrates why methods of release, resting position, background, and exposure must be evaluated rather than converting a textbook narrative into an unqualified claim. For evolutionary behavioral research, the generalizable logic is the combination of a discriminating ecological mechanism with directly tested survival consequences. A finding of current camouflage benefit remains different from reconstructing every historical step through which the phenotype became common. Kettlewell (1955) Nielsen (2009)
Field Perturbations, Guppy Life Histories, and Causal Inference
Reznick, Bryga, and Endler studied guppies experiencing different predation regimes, using deliberate changes in the predators encountered by a natural population to test predictions about life-history evolution. Their Nature report described changes in the predicted direction after a documented period spanning many generations. The research provided an instructive causal advance over cross-sectional comparisons because the investigators manipulated a major selective feature of the ecological environment, followed descendant populations, and assessed life-history traits expected to respond. This is evidence for evolutionary change under altered predation, while the most direct measured phenotypes concern maturation, reproduction, and life history rather than a complete map of cognitive mechanisms. Reznick et al. (1990)
The interpretive strength of this design depends on recognizing what changed together. Predation alters mortality risk and can covary with habitat, food, and population density; introduced populations can also undergo stochastic changes associated with founding. Earlier guppy introduction work explicitly discussed changes in environmental conditions alongside the altered predator setting. The experiment therefore increases causal leverage without abolishing the need to investigate ecological confounders, genetic variation, and replicate populations. An evolutionary explanation gains credibility when different cohorts, populations, and manipulations support the same mechanism and the measured changes are transmitted beyond the immediate plastic response to a novel environment. Reznick et al. (1990) Lande (1976)
Behavioral science can adapt this logic by combining field observations of avoidance, foraging, mating, or exploration with controlled measurements and long-term data on reproductive consequences. The first causal question is whether behavior affects fitness under a defined environment; the second is whether the response itself is heritable or is a plastic consequence of ancestral mechanisms; the third concerns the historical sequence that produced it. Genetic, demographic, and environmental records support these questions differently. An experiment that changes behavior within individuals can establish a mechanism of flexibility, but it is not by itself an experiment demonstrating evolutionary change in a population. Tinbergen (1963) Pigliucci (2001) Kingsolver et al. (2001)
Sexual Selection and the Evolution of Behavioral Preferences
Sexual selection concerns differences in reproductive success associated with mating competition and mate choice. Darwin treated these processes as potentially distinct from selection mediated by survival, opening research on display behavior, competition, preference, and reproductive strategies. Such hypotheses require clear measures of opportunities to mate, realized mating success, parentage, and the ecological and social conditions that affect reproductive outcomes. A preference observed in one laboratory task may be developmentally plastic, shaped by local norms, or contingent on the options offered. Identifying it as an adaptation requires further evidence about its origins and historical selection, not simply evidence that one sex currently reports stronger preferences in a particular sample. Darwin (1871) Schmitt & Pilcher (2004)
Later work on female sexual selection, including Clutton-Brock's review, illustrates the range of possible mating systems and the limits of generalized accounts of competition. Variability in ecology, operational sex ratio, parental investment, social organization, and mating opportunities changes the costs and benefits of courtship, aggression, choice, and display. Behavioral phenotypes also covary with morphology and endocrine physiology, so a shift in behavioral frequency may be correlated with selection on another characteristic. Comparative analyses should therefore model population and phylogenetic structure rather than reading every observed sex difference as a fixed species-wide trait. Clutton-Brock (2009) Felsenstein (1985) Lande & Arnold (1983)
For human psychological research, the social measurement problem is especially consequential. Preferences expressed in a questionnaire, choices in a constrained experimental task, and behavior in ongoing relationships are distinct outcomes; each may be affected by institutions, culture, opportunity, and self-presentation. Claims about universal ancestral mating problems become testable when their predicted direction, specificity, developmental conditions, and cultural moderators are specified before analysis. The strongest research program compares these predictions with competing learning, social-structural, ecological, and correlated-trait explanations, and reports where data support more than one account. Henrich et al. (2010) Schmitt & Pilcher (2004)
Inclusive Fitness, Cooperation, and the Units of Selection
Social behavior presents an apparent puzzle when an action benefits another organism while imposing costs on the actor. Hamilton's 1964 analyses showed how genetic relatedness and fitness effects can enter the evolutionary conditions for social traits. The familiar inequality commonly summarized as rB greater than C is a compact statement of the balance of relatedness-weighted benefits and costs under relevant assumptions; it is not a universal field equation applicable without defining actor, recipient, fitness effects, population structure, and how relatedness is estimated. Price's covariance approach supports decompositions of change across organizational levels, while interpretation of these partitions still requires substantive causal modeling. Hamilton (1964) Price (1970)
West, Griffin, and Gardner emphasized the importance of unambiguous terminology in work on altruism, cooperation, mutual benefit, kin selection, and group selection. A behavioral act may benefit a partner, provide a direct return to an actor, or increase indirect reproductive success through relatives. The everyday category of kindness does not determine which evolutionary mechanism is implicated. Similarly, group-level correlations can be decomposed in different ways, but the presence of variation between groups does not independently establish that historical selection acted at the group level as the principal explanatory cause. The researcher must measure costs, benefits, population assortment, and the scale at which competition and reproduction occurred. West et al. (2007) Price (1970)
Social learning adds an additional transmission route. A population may retain cooperative norms because individuals learn from others, enforce expectations, or preferentially interact with known partners, even if a genetic model alone does not capture these proximal processes. The coexistence of biological and cultural transmission does not erase the need to identify selective consequences and mechanisms. It increases the number of possible causal pathways and the importance of longitudinal and cross-population data. The same cooperation pattern may therefore have distinct developmental and evolutionary explanations in different settings; general claims should retain the specificity of the populations, resources, institutions, and experimental tasks in which an effect was established. West et al. (2007) Laland et al. (2015) Henrich et al. (2010)
Behavioral Plasticity, Reaction Norms, and Learning
Phenotypic plasticity is the capacity of a genotype or developmental system to produce different phenotypes in response to environmental conditions. Behavioral flexibility is often a rapid expression of plasticity through learning, attention, motivation, or regulation. A reaction norm formalizes the relationship between environment and phenotype for a given genotype or genetic background. The existence of flexible performance demonstrates that a phenotype depends on environmental information; it does not specify whether selection favored the flexibility, the particular response, a correlated developmental feature, or another property of the organism. Pigliucci's synthesis placed plasticity in a framework of ecological genetics and development, clarifying the distinct questions of proximate production and evolutionary origin. Pigliucci (2001) Tinbergen (1963)
Plasticity can modify selective environments as well as outcomes. Foraging choices affect resource availability, social responses alter partner behavior, and repeated exposure to hazards changes learned avoidance. These activities may influence which phenotypes subsequently experience favorable survival or reproduction. Extended evolutionary approaches investigate how development, niche construction, and inherited environments contribute to the generation and maintenance of variation. Such perspectives remain compatible with empirically estimating selection and drift, provided that causal propositions are specified rather than replacing process explanations with a generalized vocabulary of organism–environment interaction. Laland et al. (2015) Bolnick et al. (2011)
A powerful design compares the same genotypes or families under multiple controlled environments and then examines whether reaction norms predict survival and reproduction in natural settings. Repeated measures permit assessment of habituation, learning, temporary state, and stable individual differences; genetic designs can assess transmissible variation. Behavioral reliability across contexts should not be presumed solely because a score is statistically consistent within a single test environment. Generalization requires measuring the ecological range over which a response is expressed and evaluating whether changes across environments are themselves adaptive, historically contingent, or constrained. Pigliucci (2001) Kingsolver et al. (2001) Schmitt & Pilcher (2004)
Behavioral Measurement and the Construction of Evolutionary Phenotypes
Evolutionary claims concern phenotypes, yet psychological research often measures performances that are mixtures of capacity, motivation, sensory access, training, opportunity, and experimental apparatus. Inhibitory control scores, for example, may depend on food preferences, motor requirements, attentional engagement, or comprehension of a procedure. The research phenotype must therefore be specified at the level actually observed, and the causal interpretation should identify which latent mechanism is supported by the design. A high test score cannot be equated with reproductive advantage or with an inherited adaptation until those relationships are established independently. Comparable labels across species, cultures, and tasks do not guarantee comparable measurement. Tinbergen (1963) Bateson & Laland (2013)
A measurement program should distinguish behavioral state from stable individual variation, within-individual plasticity from between-individual differences, and phenotypic correlation from genetic covariance. Longitudinal observation and repeated testing estimate stability and reaction norms; parent-offspring or genomic designs assess aspects of inheritance; field demography links phenotypes to fitness components; and experiments evaluate the contribution of specific ecological mechanisms. Each method answers only part of the evolutionary question. A stable behavioral disposition may be produced by stable environments or early learning, while an inherited difference may have no demonstrated adaptive function. Multimethod evidence is needed to establish both the reliability of a construct and the evolutionary interpretation of its variation. Pigliucci (2001) Lande & Arnold (1983) Kingsolver et al. (2001)
Individual differences also matter ecologically. Bolnick and colleagues reviewed how within-species trait variation can alter ecological interactions and population dynamics that are obscured by reliance on population means. Behavioral variation can change which prey, risks, competitors, or social partners individuals encounter, with consequences for the selective regime itself. Studies should document the distribution of behaviors rather than reporting only average performance and should assess whether rare strategies, frequency dependence, or context-specific trade-offs contribute to outcomes. An explanation founded on one mean score per population often erases precisely the variation on which evolutionary processes operate. Bolnick et al. (2011) Price (1970) Lande & Arnold (1983)
Quantitative Genetics and the Selection–Drift Comparison
Quantitative behavioral traits can be influenced by many genetic variants and by environmental and developmental inputs. To reason about their evolution, quantitative-genetic models describe additive genetic variance and covariance, selection gradients, and changes in population means under specified assumptions. Lande's analysis of selection and random genetic drift in phenotypic evolution showed the importance of comparing the observed magnitude and structure of divergence to plausible stochastic expectations. Lande and Arnold subsequently addressed the measurement of selection on correlated characters. These approaches permit more precise alternatives than an informal choice between 'adaptive' and 'accidental,' but their empirical application requires careful measurement of phenotypes, inheritance, fitness, and demographic history. Lande (1976) Lande & Arnold (1983)
A common comparison asks whether divergence in a quantitative trait exceeds an expectation derived from neutral genetic differentiation. Such reasoning can be informative, but it is sensitive to the estimation of genetic variance, the identity of neutral comparison markers, population structure, and the relevant historical demographic model. Behavioral traits pose additional difficulties when environmental effects are large, genetic architecture is incompletely known, or measured phenotypes are not invariant across populations. An apparent excess of phenotypic divergence may reflect environmentally induced plasticity, while similar population means can conceal divergent mechanisms. Inference should therefore be based on uncertainty intervals and sensitivity to demographic alternatives rather than a single threshold interpreted as a universal test of adaptation. Lande (1976) Gillespie (2004) Pigliucci (2001)
Neutral molecular evolution supplies a useful conceptual benchmark without becoming a direct account of psychological change. Kimura's proposal concerned the prevalence of neutral substitutions and the explanatory power of random fixation at the molecular level. The pathway from variants to developmental systems and observable behavior is mediated by genetic interactions, regulation, context, and learning. Consequently, evidence of a molecular selective sweep, high differentiation, or neutral evolution must be connected to a measured behavioral phenotype before it can bear on a psychological explanation. Even when genetic selection is established, the precise behavioral feature selected and its historical function remain additional questions. Kimura (1968) Nielsen (2009) Schmitt & Pilcher (2004)
Cultural Evolution, Social Learning, and Biocultural Feedback
Behaviors can spread across a population because individuals copy, teach, imitate, conform, or respond to institutions. These processes transmit information at time scales that can differ from genetic inheritance and can interact with ecological or biological selection. A pattern of cultural persistence should therefore be investigated in terms of learning mechanisms, social network structure, access to models, and transmission fidelity. The extended evolutionary synthesis discussed by Laland and colleagues highlights forms of ecological and social inheritance that reshape the developmental environments encountered by later generations. Its relevance is clearest when researchers specify which transmission mechanism is operating and how it alters the predicted distribution of behavioral variation. Laland et al. (2015)
A behavior common in many populations might be supported by genetic ancestry, repeated individual learning in similar environments, widely shared cultural instruction, or convergent social practices. Universality alone does not distinguish these routes. Conversely, substantial cultural differences do not exclude the presence of inherited capacities for learning or social responsiveness. Research should specify which level is claimed to vary: underlying capacity, eliciting cues, acquired content, performance norm, or institutional incentive. Experimental and ethnographic comparisons can measure these components, while cross-generational data and phylogenetic reconstruction can evaluate historical pathways. Conflating a transmitted norm with a genetically selected disposition produces causal overreach. Henrich et al. (2010) Laland et al. (2015) Felsenstein (1985)
Behavior also changes its own selective context. Cooperative practices can alter the distribution of resources and risk; communication strategies can modify partner choice; and technology can change which cognitive activities provide immediate economic or reproductive advantages. These feedbacks motivate explicit models of niche construction and reciprocal causation, but they remain empirical hypotheses requiring operational definitions. A repeated cultural pattern is not proof of genetic evolution, and a shift in rewards following an institutional innovation is not proof that a new biological adaptation has arisen. A high-quality biocultural account traces the mechanisms and time scales linking transmitted environments, phenotypic expression, demographic effects, and potential population-genetic change. Laland et al. (2015) West et al. (2007) Pigliucci (2001)
Human Evolutionary Psychology and Its Evidence Requirements
Research on human behavior has particular strengths in controlled experimentation, developmental observation, historical documentation, and large-scale population studies, while facing severe difficulties in reconstructing ancestral environments and measuring evolutionary fitness. Psychological adaptation hypotheses often propose specialized responses to recurrent ancestral problems. The crucial scientific move is to state what architecture, constraints, cue sensitivity, cross-cultural distribution, and fitness-relevant effects would be expected if the hypothesis were true, and then to compare these expectations with plausible alternatives. Schmitt and Pilcher emphasized integrating evidence across disciplines rather than using apparent efficiency or a single universality claim as a decisive criterion. Schmitt & Pilcher (2004) Darwin (1871)
The sampling basis of a human generalization matters. Henrich, Heine, and Norenzayan documented how behavioral science has frequently treated populations drawn from Western, educated, industrialized, rich, and democratic settings as if they were representative of humanity. Such sampling threatens both the description of the target phenotype and claims about universality. A response observed in one national or institutional context may be strongly influenced by local constraints, education, language, material resources, and measurement practice. Cross-cultural research strengthens inference when it deliberately samples variation in the hypothesized causal conditions and tests measurement comparability, while recognizing the independence and diversity of historical and social contexts. Henrich et al. (2010)
Retrospective accounts of human psychological evolution also face the problem that a behavior may have multiple present functions and may have changed through cultural evolution much more recently than its biological substrate. Ancestral reconstruction should not presume a homogeneous ancestral setting or a single enduring challenge. The investigator should specify uncertainty about past environments, identify potentially competing historical trajectories, and distinguish observations supporting a mechanism today from evidence of its selection history. A good adaptive hypothesis generates risky, discriminating predictions; a story that can accommodate any behavioral result after the fact has weak explanatory value. Gould & Lewontin (1979) Nielsen (2009) Schmitt & Pilcher (2004)
Evidence Hierarchies and Causal Research Designs
Evolutionary explanation benefits from an evidence matrix organized by claim rather than by prestige of method. A laboratory manipulation can demonstrate cue-sensitive behavior under controlled conditions; a longitudinal field study can estimate relations between phenotypes and survival; a common-garden experiment can help separate environmental induction from inherited differences; pedigrees and genomic data can inform transmission; and phylogenetic comparisons can test historical recurrence. No single approach is sufficient for all propositions. The evidence needed to establish current function differs from the evidence needed to reconstruct origins, and a rigorous monograph should report those differences explicitly. Tinbergen (1963) Felsenstein (1985) Reznick et al. (1990)
Observational associations are particularly vulnerable to confounding. More cautious animals might survive longer because of avoidance, because they occupy safer territories, or because physical condition simultaneously influences risk-taking and mortality. Randomized ecological manipulations provide leverage where feasible, but experimental settings may themselves alter the behavior being studied. Repeated observations, individual fixed effects, natural experiments, and explicit multivariate models can reduce particular sources of bias, while none erases all uncertainty. Statistical significance should be accompanied by interpretable effect sizes, uncertainty intervals, selection procedures, and sensitivity analyses where the source design permits them. Lande & Arnold (1983) Kingsolver et al. (2001) Reznick et al. (1990)
The strongest historical explanations join independent components: a plausible developmental mechanism, reliable phenotype measurement, documented ecological relevance, estimates of fitness consequences, a transmission model, and evolutionary patterns consistent with the mechanism. Repeated convergence among these components increases evidentiary support, especially when alternative hypotheses make different predictions. Findings that contradict a favored account should be reported because they reveal whether the hypothesis distinguishes evidence from convenient post hoc adjustment. This is an integrative evidentiary standard, not a claim that every individual study must perform every measurement. Schmitt & Pilcher (2004) Gould & Lewontin (1979) Nielsen (2009)
Testing Adaptive Hypotheses Against Explicit Alternatives
A useful evolutionary hypothesis begins with a contrast. For a putative vigilance adaptation, the competing accounts might involve selection for predator detection, a general arousal mechanism, developmental exposure to danger, learned cultural conventions, or correlated variation with movement ability. Each predicts a different relationship among cue specificity, developmental history, physiological response, phylogenetic distribution, ecological conditions, and fitness. A strong test identifies observations that would favor one model over the others and specifies what would count against the preferred explanation. This contrasts with the adaptationist tendency criticized by Gould and Lewontin, in which every observed trait is readily supplied with a post hoc benefit. Gould & Lewontin (1979) Schmitt & Pilcher (2004)
Model comparison is most informative when alternatives are developed sympathetically and carry independent empirical commitments. Drift predicts a distribution of divergence conditional on population history, stochastic sampling, and demographic parameters. Plasticity predicts environmental responsiveness under defined developmental conditions. Constraint models predict covariance and limits on feasible phenotypic combinations. Selection models predict relationships among inherited variation, fitness differences, and directional changes conditional on ecology. Exaptation models predict evidence of an older origin of the relevant substrate before its current use. These are not mutually exclusive ingredients of natural history, but they can be compared as accounts of the particular difference the investigator seeks to explain. Wright (1931) Lande (1976) Pigliucci (2001) Gould & Vrba (1982)
A methodological warning follows from this structure. The ability to produce a plausible functional narrative is nearly unlimited once outcomes are known. Predictive value is demonstrated when a model anticipates patterns, survives testing outside its development sample, and correctly distinguishes environments in which the trait should or should not confer advantages. Strong prior theory can guide measurement and manipulation, but it must not determine the conclusion before alternative pathways are examined. Researchers should report uncertainty about ancestral states and causal variables rather than forcing a binary label of adaptation versus nonadaptation on every complex behavioral system. Nielsen (2009) Schmitt & Pilcher (2004) Felsenstein (1985)
Molecular Evidence and the Limits of Genetic Adaptation Tests
Population genomic evidence can test whether a genetic region bears patterns compatible with historical selection, demographic change, or neutral evolution. The conceptual distinction between selection and drift was established in classical population genetics and later refined through molecular models. A detected genomic signature can sharpen evolutionary hypotheses by supplying independent evidence about lineage history and inheritance, but it does not specify a psychological phenotype or its ancestral function without additional linkage. Genotype-phenotype mapping, developmental mechanisms, functional assays, and behavioral validation become necessary bridges from molecular evolution to claims about cognition or motivation. Wright (1931) Kimura (1968) Gillespie (2004)
Demography is a major alternative explanation for genomic structure. Bottlenecks, population subdivision, migration, and linked selection can alter genetic patterns that might otherwise be interpreted as evidence of local adaptation. Moreover, a genetic variant can affect multiple traits, and a fitness advantage may be mediated by an unmeasured correlated phenotype. Lande's analysis of polygenic evolution and the later method for correlated-trait selection explain why an identified genetic association does not necessarily establish direct selection on the named psychological characteristic. A well-designed genomic study therefore needs an explicit demographic null model, careful population sampling, and independent phenotypic evidence. Lande (1976) Lande & Arnold (1983) Gillespie (2004)
Nielsen's assessment of adaptationism in genomics makes an important methodological point: even functional information joined to evidence of selection can be insufficient to reconstruct the exact historical target and benefit. A molecular marker may be close to a selected locus without being the selected target, and a behavioral effect measured today may differ from the earlier consequence that influenced reproductive success. Claims about human psychological adaptations are especially exposed to this distinction because many measured constructs are broad and affected by institutions, development, and environment. Cross-disciplinary triangulation is the appropriate remedy: molecular and behavioral data should constrain the same causal model rather than being treated as interchangeable proofs. Nielsen (2009) Schmitt & Pilcher (2004) Henrich et al. (2010)
Replicability, Publication Bias, and Generalization
A finding that behavior covaries with an ecological predictor in one sample may be specific to the measurement procedure, season, local demography, or a particular evolutionary lineage. Independent replication should therefore aim to reproduce the causal contrast, not merely the exact surface stimulus or the same statistical significance threshold. Field ecology requires repeated seasons and populations; experimental psychology requires construct-valid tasks and sampling diversity; and phylogenetic work requires sensitivity to alternative trees and models. Replication should be planned around the level of generalization claimed. An adaptive mechanism expected only in a narrow ecological condition should be evaluated in that condition, while a claim of species-wide universality requires broader sampling. Kingsolver et al. (2001) Henrich et al. (2010) Felsenstein (1985)
Publication and selection biases can affect the apparent evidence base. Studies yielding striking support for specialization or large effects may be more visible than reports of weak or contradictory associations. Selective reporting of outcome definitions, environmental contrasts, or population subsets can also create an appearance of coherent adaptive design. A critical review should identify how many independent research programs support a conclusion, whether controls and analyses were preplanned, and whether counterevidence has been incorporated. Meta-analytic aggregation is useful only when outcomes, models, and populations are sufficiently comparable; otherwise a narrative synthesis with explicit heterogeneity may be more scientifically informative. Kingsolver et al. (2001) Schmitt & Pilcher (2004)
Generalization to human evolution is constrained by unequal inclusion of cultures and populations. Henrich and colleagues showed that common research samples cannot automatically represent the psychological distribution of the species. Species comparisons face analogous biases because captive populations, charismatic taxa, and convenient model species are studied intensively while other lineages remain sparsely sampled. The remedy is a transparent sampling frame, cross-setting measurement checks, and reporting of the precise inference supported. The strength of a claim depends on the population, trait, and environment represented in the evidence—not simply on the number of published papers endorsing a preferred evolutionary narrative. Henrich et al. (2010) Felsenstein (1985) Bolnick et al. (2011)
Adaptationism, Neutralism, and Explanatory Pluralism
Adaptationism remains productive when it generates mechanistically and historically discriminating predictions about how behavior should covary with reproductive challenges. It becomes less informative when every characteristic is interpreted as an optimized solution and failures of prediction are explained away after the fact. Gould and Lewontin argued for treating organisms as integrated systems shaped by ancestry, architectural dependencies, and developmental constraints; Nielsen's later analysis showed the continuing relevance of this concern in molecular studies. Their critique does not undermine natural selection as an established evolutionary process. It demands that selective histories be investigated rather than inferred from plausible purpose alone. Gould & Lewontin (1979) Nielsen (2009)
Neutral theory and genetic drift establish that historical change can be produced without adaptive explanation for each variant. Kimura's proposal primarily concerns molecular evolution, while Wright and Lande offer frameworks relevant to population and quantitative trait evolution. These developments do not imply that psychological traits are generally neutral; they clarify the burden of proof for attributing a specific change to a particular process. A drift model must be fitted to credible demography and inheritance, just as a selection model must be linked to measured or reconstructed fitness differences. The two processes can interact, and their relative importance can differ among loci, traits, populations, and times. Kimura (1968) Wright (1931) Lande (1976)
Exaptation, developmental constraint, plasticity, and niche construction expand the explanatory vocabulary. An inherited capacity may have a selected origin, acquire new uses through cultural practice, and vary across developmental environments without a single new genetic adaptation arising. The extended evolutionary synthesis literature highlights constructive and reciprocal processes, but its conceptual proposals should be assessed against specific empirical predictions and compared with what existing population-genetic models already explain. Explanatory pluralism is scientifically valuable when every proposed mechanism is operationalized and subjected to testing. It is weaker when pluralism becomes a list of possibilities without identifiable observations that would narrow them. Gould & Vrba (1982) Pigliucci (2001) Laland et al. (2015)
From Evolutionary Fitness to Contemporary Psychological Outcomes
Evolutionary fitness and present psychological well-being are different constructs. A motivational mechanism that historically increased reproductive success need not promote contemporary happiness, health, or social harmony; conversely, a present-day protective behavior may be achieved through education and institutions without having a dedicated inherited origin. An evolutionary account can illuminate why some responses are sensitive to risk, reward, affiliation, or uncertainty, but a causal historical hypothesis is not a clinical assessment or a policy prescription. Behavioral experiments must measure the psychological phenomena they claim to explain and should avoid substituting conjectures about ancestral reproductive consequences for evidence about current outcomes. Mayr (1961) Schmitt & Pilcher (2004)
Mismatch hypotheses are a particular application. They propose that inherited dispositions operating in contemporary environments generate outcomes different from those characteristic of historical environments. The hypothesis becomes testable when it identifies the relevant trait, reconstructs environmental differences as precisely as evidence allows, and predicts measurable performance or health consequences. It remains sensitive to variation among past environments, individual developmental calibration, and social change. A mismatch narrative drawn from a stylized ancestral scene can be suggestive but should not be presented as a verified causal account without historical and contemporary tests. Darwin (1859) Pigliucci (2001) Henrich et al. (2010)
In psychological practice, evolutionary ideas are best used as one level of explanation alongside behavioral mechanisms, development, interpersonal context, and social institutions. They can stimulate studies of recurring vulnerabilities or benefits of behavioral flexibility, but established intervention efficacy requires its own clinical evidence. The scope of this dossier is the explanation of behavior by evolutionary processes, not diagnostic classification or personalized treatment. This distinction protects the evidentiary meaning of fitness, adaptation, and functional design and allows behavioral research to remain specific about which claims are tested and which remain historical inferences. Tinbergen (1963) Bateson & Laland (2013)
Artificial Systems as Comparative Models, Not Biological Lineages
Artificial adaptive systems can modify behavior through reinforcement learning, optimization, memory, or iterative selection of candidate policies. Their observable functional improvement provides opportunities to study general questions about feedback, variation, search, and environmental contingencies. The term adaptation in such systems can name changes that improve performance under a specified training or deployment distribution; an evolutionary explanation in the biological sense additionally requires a defined process of inheritance across reproducing units and a population history. The distinction matters when researchers transfer conceptual tools between organismal evolution and artificial learning. Functional similarity alone does not establish homologous origins or a shared selection mechanism. Darwin (1859) Wright (1931) Bateson & Laland (2013)
Within the Artificial Era, research on human–artificial relations may investigate how artificial agents reshape learning environments, access to information, human decision-making, and the distribution of social opportunities. Such changes can be measured as psychological and cultural phenomena in the present. Their possible long-term effects on human evolution constitute hypotheses requiring generation-spanning demographic and inheritance evidence. The frameworks of ecological inheritance and niche construction make these questions conceptually tractable, while their empirical application requires measured variables and explicit time scales. Studies of Artificial agency should be distinguished from studies of biological fitness, and any proposed shared model should disclose exactly which variables are comparable. Laland et al. (2015) Pigliucci (2001)
The broader historical implication is that behavioral sciences now compare organizations of adaptive activity across biological and designed systems. This comparison can be scientifically fruitful when investigators keep mechanism, function, developmental learning, and historical origin separately identifiable. A trained artificial policy may change during operation through error correction; a biological behavioral trait may change within a lifetime through learning; a population may evolve through differential reproduction over generations. These processes can share mathematical structures while remaining different empirical objects. The analysis of adaptation, selection, and drift supplies an exacting language for the comparison, and it prevents metaphor from being mistaken for evidence. Price (1970) Tinbergen (1963) Laland et al. (2015)
Contemporary Research: Long-Term Dynamics and Selection on Socially Embedded Behavior
Recent work has increasingly emphasized the value of continuous observation across ecological events and successive generations. Long-term field studies allow investigators to relate the same lineages and populations to repeated climatic disturbances, changing competitors, selective episodes, and demographic shifts. A review by Stroud and Ratcliff (2025) situates the distinctive contribution of extended time series in evolutionary biology: short observations can capture a selection episode, while longer records reveal variation in its direction, the role of environmental recurrence, and the capacity of different evolutionary processes to shape outcomes over time. For behavioral research, this invites the integration of individual-level learning histories and performance with reproductive trajectories instead of treating one season's performance as evidence about evolutionary origins. Stroud & Ratcliff (2025) Grant & Grant (2002)
Socially mediated variation complicates the conventional division between an individual's genes and its environment. Indirect genetic effects occur when genes expressed by social partners influence the focal individual's phenotype. Santostefano and colleagues (2025) synthesized empirical evidence for these effects and discussed their contribution to the genetic variance available to selection, with particular attention to behavior. The relevant interpretation is that the social environment can carry genetically influenced inputs that change a behavioral phenotype even when the focal organism's genotype is held constant. Research needs designs that identify social partners, distinguish direct from indirect genetic influences, and examine how stable the social configuration is across ecological settings. This expands the causal architecture of selection without making every social correlation heritable. Santostefano et al. (2025) Lande & Arnold (1983) West et al. (2007)
Recent evolutionary work on personality and individual differences has likewise framed behavioral consistency and flexibility as empirical questions about variation, trade-offs, and environmental conditions. Volk and Puchalski (2025) reviewed links between evolutionary theory and personality, an area in which inferences from individual differences to fitness histories remain demanding. Variation may persist through fluctuating selection, frequency dependence, developmental constraints, or genetic and environmental heterogeneity. The research priorities are to separate descriptive personality scores from their proximate causes, establish measurement reliability across contexts, and connect individual variation to the appropriate fitness components. Population-level selection may act on behavioral plasticity as well as on mean behavioral tendencies. Volk & Puchalski (2025) Bolnick et al. (2011) Pigliucci (2001)
A Comparative Evidence Matrix for Evolutionary Behavioral Claims
Evidence should be assembled according to the precise claim. If the claim is that a behavior responds to environmental cues, within-individual manipulations and longitudinal repeated measures are directly relevant. If the claim is that behavioral differences predict survival or reproduction, field data with defined fitness components and suitable adjustment for confounding are needed. If the claim is that variation is inherited, pedigrees, common-environment breeding, or genomic approaches become relevant. If the claim concerns evolutionary change, repeated-generation data must link transmissible variation to changes in population distributions. And if the claim concerns ancestral adaptive origin, historical and phylogenetic evidence must address alternatives such as exaptation, correlated response, and drift. The empirical strength is therefore multidimensional rather than a simple rank ordering of study labels. Tinbergen (1963) Lande (1976) Felsenstein (1985)
The finch and guppy research programs illustrate these distinctions. A particular drought episode can generate observable differential survival and a shift in traits, while subsequent decades reveal changing selective regimes, inheritance, and demographic contingencies. A guppy translocation or ecological perturbation changes the predation environment and permits descendant phenotypes to be studied over multiple generations. Neither study directly establishes the adaptive history of a general human cognitive construct, but both illustrate how rigorous evolutionary explanation builds an observable chain from ecology through phenotype to differential survival, inheritance, and subsequent population change. Extrapolation to psychology should preserve the structure of that chain while developing measures appropriate to cognition and behavior. Boag & Grant (1981) Grant & Grant (2002) Reznick et al. (1990)
The evidence matrix also exposes the limits of apparently persuasive single findings. A positive association between behavioral flexibility and reproductive output may arise through body condition or habitat quality; a cross-species association may reflect shared phylogenetic history; a difference between populations may be culturally or developmentally induced; and a genomic selection signal may have acted on a correlated physiological feature. Researchers should identify which inferential gap remains after each test and seek another design capable of closing it. Consistent movement across independent methods offers greater support than repeated use of one confounded measure. The aim is an explanation whose assumptions and vulnerability to falsification are visible. Felsenstein (1985) Lande & Arnold (1983) Schmitt & Pilcher (2004) Nielsen (2009)
Open Research Problems and Priorities for New Evidence
One outstanding challenge is reconstructing the evolution of complex cognitive phenotypes whose laboratory scores combine several latent processes. Many psychological traits are measured through tasks with limited ecological validity, and the relation between task performance and fitness can be unknown. Greater explanatory precision requires studies that connect mechanistic measurement with naturally occurring behavioral variation and consequences over time, ideally while tracking learning, age, sex, physiological state, and opportunities to express the behavior. Genetic and cultural pathways need to be estimated in the same population where feasible. Such work will make it easier to distinguish an inherited adaptive capacity from an acquired strategy or an incidental correlate of another trait. Tinbergen (1963) Pigliucci (2001) Henrich et al. (2010)
A second challenge concerns identifying the historical balance of selection, drift, migration, and developmental bias in traits with sparse ancestral records. Behavioral phenotypes rarely fossilize directly; phylogenetic reconstruction can depend on uncertain character coding and incomplete species sampling. Genomic analyses face linked selection, demography, and pleiotropy, while experiments on living populations may observe only current function. Progress will therefore come from explicit simulation and sensitivity analysis, improved comparative sampling, multigenerational field studies, and theories that make different predictions under plausible historical scenarios. Evidence should be reported as conditional on model assumptions rather than converted into an unconditional ancestral narrative. Felsenstein (1985) Nielsen (2009) Gillespie (2004) Stroud & Ratcliff (2025)
A third challenge is generalization in social organisms. Human institutions, nonhuman social networks, cultural learning, and indirect genetic effects all change which environments and interactions individuals experience. Research programs should connect genetic and cultural transmission models to measured social dynamics and evaluate whether outcomes replicate across distinct groups, ecological settings, and historical conditions. Artificial systems can serve as carefully defined comparative testbeds for learning and optimization, while their mechanisms and timescales remain separately specified. The scientific horizon of the Artificial Era widens the comparative questions available for behavioral research, and the integrity of that comparison depends on preserving the distinction between biological population change and functional improvement within a designed system. Santostefano et al. (2025) Laland et al. (2015) Henrich et al. (2010)
Scientific Synthesis: Conditions for a Defensible Evolutionary Explanation
An evolutionary explanation of behavior is defensible when it identifies a behavioral phenotype, the relevant historical population and environment, a causal process capable of generating the observed distribution, and evidence that differentiates this process from meaningful alternatives. Natural selection describes covariance between transmitted variants and reproductive contribution; drift describes stochastic change in finite populations; plasticity explains environment-sensitive expression; exaptation identifies a change in use relative to selected origins; and constraints and correlations shape the phenotypes available for selection and their joint responses. These processes can coexist across stages of the same trait's history. Their analytical separation allows studies to specify causal responsibility rather than apply a single label to a multifactorial result. Wright (1931) Price (1970) Gould & Vrba (1982) Pigliucci (2001)
Different research designs support different levels of inference. Ecological manipulation and field demography can establish contemporary selection pressures; common-environment and breeding designs can support inheritance claims; phylogenetic models can characterize historical covariance; multivariate analyses can estimate correlated selection; genomic and demographic models can assess aspects of historical population processes; and developmental research can explain the production and flexibility of behavior. Convergence across these methods strengthens a particular account, while inconsistencies identify areas where its causal structure needs refinement. No analysis of reproductive fitness alone establishes the detailed cognitive mechanism, and no laboratory demonstration of information processing establishes a historical selective origin. Reznick et al. (1990) Felsenstein (1985) Lande & Arnold (1983) Bateson & Laland (2013)
The central scientific result of this synthesis is a research logic: explanatory claims become stronger as the investigator shifts from plausible functional stories toward predictions that distinguish adaptation from other routes to a present phenotype. Such reasoning is suited to human psychology, comparative cognition, behavioral ecology, population biology, and carefully delimited comparisons with artificial learning systems. It allows historical questions to guide experiments while maintaining independent standards for behavioral measurement, developmental mechanism, population change, and causal identification. Evolutionary psychology earns explanatory precision through those links, and future research will be judged by the discriminating evidence it produces. Schmitt & Pilcher (2004) Gould & Lewontin (1979) Nielsen (2009)
Conclusion
Adaptation, selection, drift, and evolutionary explanation describe a connected set of historical problems concerning the origins and distributions of behavioral phenotypes. Natural selection can produce adaptive organization when transmitted variation contributes to differences in reproduction; stochastic drift can change inherited variation in finite populations; correlations and developmental constraints shape which phenotypes evolve together; plasticity and learning determine how behavioral capacities are expressed; and exaptation captures the historical recruitment of inherited features to new uses. Each process has a distinct evidentiary signature and a particular domain of uncertainty. A present beneficial function establishes a question about evolutionary origins, and historical attribution requires additional evidence about inheritance, ecology, fitness, and population change. Darwin (1859) Wright (1931) Gould & Vrba (1982)
The enduring achievement of evolutionary behavioral science is its ability to join natural-history observation, controlled experiments, quantitative-genetic models, phylogenetic reasoning, and psychological mechanism into causal accounts that can be challenged by evidence. Long-term studies of changing ecological selection, experimental research on altered predation regimes, and formal analyses of correlated characters provide methods for investigating how evolutionary change actually occurs. Their lessons also constrain claims about human cognition and social behavior, where ancestral conditions, cultural diversity, and construct validity demand explicit treatment. The scientific task is to identify what an observation reveals and which historical propositions remain untested. Boag & Grant (1981) Reznick et al. (1990) Lande & Arnold (1983) Henrich et al. (2010)
Future research will be most consequential where psychological variation, social transmission, ecological contingencies, and multigenerational outcomes can be measured together. Such studies can clarify the relative roles of selection, drift, development, and cultural change in the evolution of behavior. The emergence of artificial adaptive systems adds new comparative cases, provided their training histories and optimization processes are defined independently of biological reproduction. The resulting framework supports an integrated science of adaptive behavioral organization across systems while preserving the causal specificity necessary for credible explanation. Evolutionary accounts remain powerful precisely where their mechanisms, evidence, and uncertainty are made explicit. Stroud & Ratcliff (2025) Santostefano et al. (2025) Tinbergen (1963)
References
- Bateson, P., & Laland, K. N. (2013). Tinbergen's four questions: An appreciation and an update. Trends in Ecology & Evolution, 28(12), 712–718. https://doi.org/10.1016/j.tree.2013.09.013
- Boag, P. T., & Grant, P. R. (1981). Intense natural selection in a population of Darwin's finches (Geospizinae) in the Galapagos. Science, 214(4516), 82–85. https://doi.org/10.1126/science.214.4516.82
- Bolnick, D. I., Amarasekare, P., Araújo, M. S., Bürger, R., Levine, J. M., Novak, M., Rudolf, V. H. W., Schreiber, S. J., Urban, M. C., & Vasseur, D. A. (2011). Why intraspecific trait variation matters in community ecology. Trends in Ecology & Evolution, 26(4), 183–192. https://doi.org/10.1016/j.tree.2011.01.009
- Clutton-Brock, T. (2009). Sexual selection in females. Animal Behaviour, 77(1), 3–11. https://doi.org/10.1016/j.anbehav.2008.08.026
- Darwin, C. (1859). On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life (1st ed.). John Murray. https://darwin-online.org.uk/content/contentblock?basepage=1&itemID=F373&viewtype=text
- Darwin, C. (1871). The descent of man, and selection in relation to sex (1st ed., Vols. 1–2). John Murray. https://darwin-online.org.uk/content/contentblock?basepage=1&itemID=F937.1&viewtype=text
- Darwin, C. (1872). The expression of the emotions in man and animals (1st ed.). John Murray. https://darwin-online.org.uk/content/contentblock?basepage=1&itemID=F1142&viewtype=text
- Felsenstein, J. (1985). Phylogenies and the comparative method. The American Naturalist, 125(1), 1–15. https://doi.org/10.1086/284325
- Gillespie, J. H. (2004). Population genetics: A concise guide (2nd ed.). Johns Hopkins University Press. https://doi.org/10.56021/9780801880087
- 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. Series B, Biological Sciences, 205(1161), 581–598. https://doi.org/10.1098/rspb.1979.0086
- Gould, S. J., & Vrba, E. S. (1982). Exaptation—a missing term in the science of form. Paleobiology, 8(1), 4–15. https://doi.org/10.1017/S0094837300004310
- Grant, P. R., & Grant, B. R. (2002). Unpredictable evolution in a 30-year study of Darwin's finches. Science, 296(5568), 707–711. https://doi.org/10.1126/science.1070315
- Hamilton, W. D. (1964). The genetical evolution of social behaviour. I. Journal of Theoretical Biology, 7(1), 1–16. https://doi.org/10.1016/0022-5193(64)90038-4
- Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2–3), 61–83. https://doi.org/10.1017/S0140525X0999152X
- Kettlewell, H. B. D. (1955). Selection experiments on industrial melanism in the Lepidoptera. Heredity, 9, 323–342. https://doi.org/10.1038/hdy.1955.36
- Kimura, M. (1968). Evolutionary rate at the molecular level. Nature, 217, 624–626. https://doi.org/10.1038/217624a0
- Kingsolver, J. G., Hoekstra, H. E., Hoekstra, J. M., Berrigan, D., Vignieri, S. N., Hill, C. E., Hoang, A., Gibert, P., & Beerli, P. (2001). The strength of phenotypic selection in natural populations. The American Naturalist, 157(3), 245–261. https://doi.org/10.1086/319193
- Laland, K. N., Uller, T., Feldman, M. W., Sterelny, K., Müller, G. B., Moczek, A., Jablonka, E., & Odling-Smee, J. (2015). The extended evolutionary synthesis: Its structure, assumptions and predictions. Proceedings of the Royal Society B: Biological Sciences, 282(1813), Article 20151019. https://doi.org/10.1098/rspb.2015.1019
- Lande, R. (1976). Natural selection and random genetic drift in phenotypic evolution. Evolution, 30(2), 314–334. https://doi.org/10.1111/j.1558-5646.1976.tb00911.x
- Lande, R., & Arnold, S. J. (1983). The measurement of selection on correlated characters. Evolution, 37(6), 1210–1226. https://doi.org/10.1111/j.1558-5646.1983.tb00236.x
- Mayr, E. (1961). Cause and effect in biology. Science, 134(3489), 1501–1506. https://doi.org/10.1126/science.134.3489.1501
- Nielsen, R. (2009). Adaptionism—30 years after Gould and Lewontin. Evolution, 63(10), 2487–2490. https://doi.org/10.1111/j.1558-5646.2009.00799.x
- Pigliucci, M. (2001). Phenotypic plasticity: Beyond nature and nurture. Johns Hopkins University Press. https://doi.org/10.56021/9780801867880
- Price, G. R. (1970). Selection and covariance. Nature, 227, 520–521. https://doi.org/10.1038/227520a0
- Reznick, D. A., Bryga, H., & Endler, J. A. (1990). Experimentally induced life-history evolution in a natural population. Nature, 346, 357–359. https://doi.org/10.1038/346357a0
- Santostefano, F., Moiron, M., Sánchez-Tójar, A., & Fisher, D. N. (2025). Indirect genetic effects increase the heritable variation available to selection and are largest for behaviors: A meta-analysis. Evolution Letters, 9(1), 89–104. https://doi.org/10.1093/evlett/qrae051
- Schmitt, D. P., & Pilcher, J. J. (2004). Evaluating evidence of psychological adaptation: How do we know one when we see one? Psychological Science, 15(10), 643–649. https://doi.org/10.1111/j.0956-7976.2004.00734.x
- Stroud, J. T., & Ratcliff, W. C. (2025). Long-term studies provide unique insights into evolution. Nature, 639, 589–601. https://doi.org/10.1038/s41586-025-08597-9
- Tinbergen, N. (1963). On aims and methods of ethology. Zeitschrift für Tierpsychologie, 20(4), 410–433. https://doi.org/10.1111/j.1439-0310.1963.tb01161.x
- Volk, A. A., & Puchalski, A. (2025). Evolutionary theory and personality. Current Opinion in Psychology, 65, Article 102069. https://doi.org/10.1016/j.copsyc.2025.102069
- West, S. A., Griffin, A. S., & Gardner, A. (2007). Social semantics: Altruism, cooperation, mutualism, strong reciprocity and group selection. Journal of Evolutionary Biology, 20(2), 415–432. https://doi.org/10.1111/j.1420-9101.2006.01258.x
- Wright, S. (1931). Evolution in Mendelian populations. Genetics, 16(2), 97–159. https://doi.org/10.1093/genetics/16.2.97