Key Takeaways
Treat human thinking as cultural technology, not a genetic blueprint
Minds have a history. Heyes calls distinctively human ways of thinking cognitive gadgets: mechanisms assembled during development through social learning, rather than detailed mental programs ready-made in the genes. Genetic inheritance still matters, but chiefly by supplying general learning capacities and social tendencies that culture can build upon.
Literacy shows how this works. Writing appeared only about five to six thousand years ago, too recently for a specialized reading instinct to evolve. Yet teaching children to read reorganizes brain activity and creates specialized reading processes from older neural systems. The lesson is not that culture builds minds from nothing. It can repurpose existing parts into new machinery, then pass the conditions for rebuilding that machinery to another generation.
The technology analogy is useful because it distinguishes a mechanism's materials from its design history. A brain area that reliably responds to print does not, by itself, prove an inherited reading module: learning can make old circuitry perform a new job. The same caution applies to any apparently specialized mental ability. Still, cultural construction is not a claim that biology is irrelevant. It relocates biology from providing finished blueprints to providing learning systems that can be repeatedly reshaped.
Separate inherited genes, personal experience, and inherited culture
Three information sources. Heyes rejects the idea that a trait must be either natural or learned. Cognitive development reflects genetically inherited information, information acquired directly from the environment, and cultural information passed between people through social interaction. These sources interact, but identifying their separate contributions still matters for explaining development.
Her teleosemantic approach distinguishes mere correlation from information with an evolutionary role. Tree rings correlate with a tree's history, but they do not guide its development. A developmental influence counts as biologically informative in this stronger sense when an inherited structure has been selected because it reliably carries information relevant to the outcome. Researchers can then ask whether a cognitive capacity changes with learning opportunities, and whether those opportunities are personal or culturally transmitted. No single test settles the question; converging evidence is needed.
This framework avoids two easy errors: treating genes and experience as competing causes, or treating their interaction as a reason to stop asking causal questions. It also highlights a practical distinction. Direct experience can teach a child about the surroundings currently present, while culture can transmit solutions accumulated before that child existed. The teleosemantic test clarifies evolutionary explanation, but it does not replace ordinary causal analysis of development. A factor can matter to a child's learning even if it was never selected to provide that particular information.
Build on humans' social starter kit, not imagined cognitive blueprints
Small beginnings, large consequences. The genetic starter kit, in Heyes's account, is not a set of complete programs for language, imitation, or mindreading. It includes social tolerance, motivation to engage with others, attention to faces and voices, and powerful general learning and control processes. These features help children find teachers and absorb what those teachers know.
Some foundations appear very early. Newborns prefer face-like arrangements, and infants show a broad preference for speech sounds. Experience then sharpens those preferences toward familiar human faces and voices. Associative learning and executive function also contribute: they are found in other animals, but may be especially powerful in humans. The proposal is that these ordinary components, combined with a richly social childhood, can produce abilities that look far more specialized in adults.
The argument turns attention away from a false choice between a blank slate and a miniature adult mind. A small bias toward faces can change what a baby learns, while social motivation changes how much learning is available. Such developmental cascades offer a plausible route from modest differences to large outcomes. Their importance is also a reason for caution: the account must distinguish genuinely inherited capacities from effects that emerge quickly because infants encounter highly structured social environments.
Distinguish attention-driven copying from learned rules about expertise
Selectivity has two routes. People and animals often learn more from one model than another, or rely more on social information in some situations than others. Heyes argues that many such biases arise because attention is drawn toward salient, rewarding, or previously reliable cues. The learner may copy a model more without consciously applying a rule about whom to trust.
A study of vervet monkeys illustrates the distinction. Monkeys paid more attention to female than male models opening a food box, and later followed the females' example more often. Since females tend to remain in their group and know more about its local conditions, their behavior may have become a more useful predictor. Adults sometimes do use explicit rules about expertise, such as seeking help from someone with relevant experience. These metacognitive strategies are learned socially, rather than being universal instincts.
The distinction matters because the same visible behavior can have different causes. A learner who follows an expert may be consciously judging credibility, or may simply have learned to attend to that person. Separating the routes helps researchers test what makes cultural transmission unusually precise. It also guards against over-crediting deliberation: human choices often feel like explicit judgments even when attention and past reward do much of the work.
Explain imitation through learned links between seeing and doing
Solving the mismatch. Imitation means reproducing the form of another person's movement. The puzzle is how someone can copy an action they cannot see themselves perform, such as putting their hands behind their back. Heyes calls this the correspondence problem.
Her associative sequence learning model proposes that experience links sensory representations of an action with motor representations of the same action. Seeing and performing movements close together, as when an adult imitates a baby or people move in synchrony, can build these links. They allow the observed action sequence to activate a matching motor sequence. Evidence includes training studies that changed adults' automatic imitation, even reversing which finger movement they made after observing another, and a large infant study that found no reliable imitation of tested gestures in the first nine weeks, apart from tongue protrusion.
This account makes imitation less mysterious without making it trivial. It replaces an unspecified inborn matching device with a mechanism that makes testable predictions about experience and error. The evidence that training can alter imitation is especially informative, though changing a capacity experimentally does not alone prove how it normally develops. The broader case is stronger because developmental patterns and limits on copying also fit the learning account.
Teach mindreading through talk about thoughts, feelings, and reasons
Mental concepts are taught. Mindreading is the ability to represent thoughts and feelings in oneself and others, and use them to interpret behavior. Heyes argues that children learn this skill through conversation and instruction, much as they learn to read print. Talk gives children words for mental states, explanations of how states relate to actions, and conventions for explaining their own behavior.
Evidence comes from several directions. In a study of more than a thousand twin pairs, identical and fraternal five-year-old twins had the same reported correlation in mindreading performance, 0.53, suggesting a strong shared-environment influence. In Nicaragua, users of an emerging sign language who lacked mental-state vocabulary showed weaker false-belief understanding than a later cohort whose language had developed more of those signs. In Samoa, where talking about mental states is discouraged, children reached this understanding several years later than children in Europe and North America.
Conversation does more than attach labels to pre-existing ideas. Explanations can teach children which kinds of inner states count as causes, and how to use those states to make sense of action. The cross-cultural examples make development look like a collective achievement, not a private discovery. Yet timing differences do not establish a single cause: language, schooling, and local expectations travel together. Careful comparisons are needed to identify which parts of that social environment do the teaching.
Question whether quick predictions really represent another person's beliefs
Prediction is not always mindreading. Infants, apes, and adults under pressure sometimes behave as if they know what another person believes. Heyes calls the alternative explanation submentalizing: general processes of attention, learning, and memory can predict behavior without representing mental states.
For example, in a study asking adults to judge how many dots a person could see, responses slowed when the person faced a different view. The same effect appeared when the person was replaced by an arrow, suggesting that attention to directional cues, rather than a belief about another mind, could explain the result. In infant studies, distraction may make a child miss an object's movement, producing a look that resembles false-belief understanding. The argument does not deny genuine mindreading. It reserves that term for more explicit, effortful reasoning about mental states.
This is a useful warning about interpreting behavior: an action that fits a sophisticated explanation may also arise from simpler processes. Comparing social scenes with matched nonsocial cues can reveal whether a task truly depends on representing minds. The challenge is that simple and sophisticated processes may operate together, and looking-time studies provide indirect evidence about what infants represent. A cautious interpretation preserves the possibility of early social understanding without treating every successful prediction as proof of a theory of mind.
Treat language as culturally shaped, while its genetic origins remain contested
A serious alternative to grammar instincts. The genetic account proposes that children inherit language-specific grammatical knowledge, often called Universal Grammar. The cultural account instead emphasizes general sequence learning, memory, and the language children hear. On this view, languages change in ways that make them learnable and usable by human minds.
Several findings make the cultural account plausible. Language activity is spread across many brain areas rather than confined to a single language center. Skill at learning artificial sequences is associated with some grammatical abilities, and children diagnosed with specific language impairment can also struggle with sequence learning beyond language. Adults correct some children's grammatical errors, giving them information that older poverty-of-the-stimulus arguments assumed was unavailable. None of this settles the debate: defenders of Universal Grammar can treat such evidence as concerning language use, rather than underlying grammatical knowledge, making the theory difficult to test against cognitive data.
The chapter's strongest contribution is not a verdict that genes play no role, but a change in the burden of explanation. Specialized grammatical knowledge should not be assumed merely because language is complex or acquired early. At the same time, general learning mechanisms do not automatically explain every structural property of language. The distinction between underlying competence and observable performance can protect theories from contrary evidence, but it also raises a legitimate methodological question: what finding would count against the theory?
Cultural evolution can reshape thinking mechanisms, not just tools and customs
Culture changes the mills. Cultural evolution is often used to explain how skills, beliefs, or artifacts spread. Heyes extends the idea to cognitive mechanisms themselves. A mechanism can vary between groups, be reconstructed through social interaction, and become more common when it helps people or groups achieve their aims.
This is cultural group selection: cognitive variants become more frequent through culturally inherited differences between groups, not only through differences in genes or individual reproduction. A group with a more effective imitation mechanism, for example, might coordinate ritual movements more reliably and transmit useful techniques more accurately. The archaeological record shows long periods of modest tool change and then a more uneven emergence of complex practices. Heyes suggests that changes in population size and connections between groups may have enabled cultural evolution of both accumulated knowledge and the mental machinery for acquiring it.
The extension from artifacts to cognitive mechanisms is the book's most ambitious move. It makes cultural evolution relevant not only to what people know, but to how they learn, remember, and interpret. It also creates a demanding research agenda: scholars must measure variation in mechanisms, show how social interaction transmits it, and identify the sorting process that changes its frequency. Group success alone cannot demonstrate group selection unless those inheritance and selection pathways are established.
Treat human nature as historically changeable, and vulnerable to cultural loss
Mental capacities can spread or disappear. If cognitive mechanisms are partly products of cultural evolution, human nature is not a fixed package delivered anew by genes. Literacy is a striking example: it arose thousands of years ago, and reading mechanisms are now present in an estimated 80 to 90 percent of people. Education can deliberately reconstruct such capacities in new generations.
The same dependence creates vulnerability. War, epidemics, or the collapse of social networks can remove not only practical knowledge, but also the people and practices that help children develop ways of thinking. The genetic starter kit may support learning, but it does not guarantee that every complex capacity will reappear. This view also makes minds adaptable: new social and technological environments can reshape cognition rather than merely challenge an unchanging ancient design.
Plasticity and fragility are two sides of the same theory. If education can build cognitive capacities, institutions become part of the infrastructure of human development. If those institutions and networks fail, some abilities may not be recoverable from genetic inheritance alone. The claim should not be read as saying that catastrophe erases all cognition: the starter kit remains, and learning can begin again. But the speed and completeness of recovery depend on what knowledge, teachers, and social practices survive.
Analysis
Heyes's central move is to shift the explanatory target from cultural products to the cognitive machinery that generates and learns them. The usual account gives culture a substantial role in shaping behavior while treating the mechanisms of learning as genetically installed. Cognitive Gadgets challenges that division: culture can construct parts of the learner as well as fill the learner with information. Literacy supplies a powerful proof of possibility, because a recent cultural invention can reorganize neural processing without a dedicated inherited reading module.
The strength of this argument is methodological as much as theoretical. Heyes repeatedly asks what mechanism could produce an observed behavior, then compares explanations through training, development, cross-cultural differences, and animal studies. This is especially persuasive where the instinct account relies on a black box, as in imitation, and where alternate explanations can be tested with specific controls. Her insistence that adaptation does not prove genetic design is a valuable correction to evolutionary storytelling.
The book's most important qualification is also its unresolved burden. Showing that experience shapes a mechanism does not establish that cultural selection shaped it across populations. Developmental plasticity, individual learning, and cultural group selection are distinct claims. The last requires evidence for variants, reliable transmission, and differential spread. Heyes proposes plausible pathways, but acknowledges that direct measurement of the inheritance of cognitive mechanisms remains underdeveloped. Similarly, the language chapter makes a compelling case that cultural explanations deserve serious consideration, but the competence-performance distinction leaves the genetic alternative difficult to falsify.
The resulting position is best read as a research program, not a completed replacement for evolutionary psychology. Genes provide powerful general learning capacities and social dispositions; cultural environments may organize those resources into historically variable mechanisms. This view preserves biological continuity while explaining why human cognition can be both unusually flexible and dependent on social systems. Its broader implication is consequential: education, conversation, ritual, and social networks are not merely influences on what people know. They may help determine what kinds of thinking people can do.
Review Summary
Cognitive Gadgets by Cecilia Heyes presents a novel perspective on human cognitive evolution, arguing that many distinctively human traits are culturally learned rather than genetically inherited. While some readers found the book thought-provoking and well-argued, others struggled with its academic style and dense content. Critics praised Heyes' challenge to established theories but noted the book's complexity might deter general readers. Overall, reviewers appreciated the book's contribution to debates on cultural evolution and human cognition, despite its occasionally challenging prose.
Glossary
Cognitive gadgets
Culturally built thinking mechanismsDistinctively human cognitive mechanisms that Heyes argues are assembled during development through social interaction and cultural inheritance, rather than specified in detail by genes. Genetic inheritance supplies general learning capacities and social tendencies that help build them. The term emphasizes that culture can shape how a mind works, not only what a person knows or does.
Cultural evolutionary psychology
Cultural evolution of cognitionHeyes's research framework for explaining the development and evolution of human cognitive mechanisms. It combines cognitive science and evolutionary theory, while proposing that culturally inherited information can shape mechanisms of thought. It extends cultural evolutionary explanations beyond behaviors and artifacts to the mental processes that produce and acquire them.
Teleosemantic conception of information
Selected developmental informationA way of identifying biologically meaningful information in development. A structure carries information about an outcome, in this sense, when its states correlate with that outcome and the structure was selected because of that correlation. This separates information with an evolutionary function from facts that merely correlate with development, such as tree rings recording a tree's history.
Cultural learning
Social learning specialized for cultureIn Heyes's framework, social learning assisted by processes specialized for cultural inheritance, such as mechanisms that improve the fidelity with which information is passed between people. It is distinguished from social learning that merely benefits from another person's presence, attention, or behavior. The distinction concerns how learning occurs, not simply what the learner acquires.
Associative sequence learning (ASL) model
Learning account of imitationHeyes's model explaining imitation through learned links between sensory and motor representations of the same actions. Seeing and performing an action in a predictive, closely timed relationship strengthens these links. The links can connect perceptual learning of an observed movement sequence to motor learning, allowing an observer to reproduce movements without an inborn mechanism that calculates their similarity.
Matching vertical associations
Links between seen and done actionsExcitatory connections between a sensory representation of an action and a motor representation of that same action. The sensory representation captures what the action looks like, while the motor representation relates to performing it. Heyes argues that correlated experience of seeing and doing builds these links, which help solve the mismatch between observing another body and sensing one's own movements.
Submentalizing
Mindreading-like general processingA term for cases where general cognitive processes, such as attention, learning, or memory, produce behavior that resembles mindreading without representing mental states. For example, directional cues or distraction may lead someone to predict another person's action. The term does not deny genuine mindreading; it distinguishes behavioral effects that can be explained without it.
Cultural group selection
Cultural spread between groupsA selectionist account in which culturally inherited variants become more or less common as groups differ in outcomes and influence. A cognitive mechanism can spread when it helps its bearers or their group, and when social learning transmits it to new members. The process is cultural rather than genetic because group membership and social inheritance shape which mechanism individuals develop.
Genetic starter kit
Inherited foundations for learningThe set of genetically inherited characteristics that Heyes argues supports the later development of distinctively human cognition. It includes social tolerance and motivation, attention to faces and voices, and general learning, memory, and control capacities. These are not complete programs for complex abilities; they provide resources that development and cultural experience can elaborate.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.