Book Summary: “Cultural Evolution: How Darwinian Theory Can Explain Human Culture” by Alex Mesoudi

Cultural Evolution

Title: Cultural Evolution: How Darwinian Theory Can Explain Human Culture and Synthesize the Social Sciences
Author: Alex Mesoudi
Scope: 5 stars
Readability: 3 stars
My personal rating: 5 stars
See more on my book rating system.

Topic of Book

Mesoudi makes the case that human culture can best be understood by applied Darwinistic biology.

Key Take-aways

  • Darwinists believe that:           
    • Variation exists between individuals
    • Individuals must compete for limited resources
    • Characteristics are inherited from parents via reproduction           
  • Humans have evolved biologically to use culture to learn and adapt
  • Cultural learning is faster and more flexible than biological evolution.
  • There are many means by which humans learn from one another (see list below).

Important Quotes from Book

Darwin’s genius—the reason he is still celebrated today—was to provide the first coherent and workable scientific explanation for these phenomena. According to Darwin, the diversity and complexity of the natural world could be explained using a handful of simple principles, all of which could be demonstrated empirically to be operating in nature: first, that variation exists between individuals; second, that a

“struggle for existence” occurs due to limited resources, such as food, nesting space, or mates, and ever-increasing population sizes, such that not every individual has an equal chance of surviving and reproducing; and third, that characteristics are inherited from parent to offspring during reproduction. The consequence is what Darwin called “natural selection”: those characteristics that increase an individual’s chances of surviving and reproducing are more likely to be inherited by the next generation, and such characteristics will increase in frequency within a population. Over time, beneficial characteristics gradually accumulate and combine to generate the eyes, wings, and so on that had previously been attributed to the actions of a creator. Darwin’s work has stimulated a century and a half of enormously productive research in evolutionary biology, which has been devoted to working out the details of Darwin’s principles, such as the genetic basis of inheritance, of which Darwin was only dimly aware.

In this book I survey a growing body of scientific research that is based on the fundamental premise that cultural change—by which I mean changes in socially transmitted beliefs, knowledge, technology, languages, social institutions, and so on—shares the very same principles that Darwin applied to biological change in The Origin a century and a half ago. In other words, culture evolves.

I argue that all the research that I have discussed here is indicative of a coming “evolutionary synthesis” for the social sciences.

Humans are a cultural species. We acquire a multitude of beliefs, attitudes, preferences, knowledge, skills, customs, and norms from other members of our species culturally, through social learning processes such as imitation, teaching, and language. This culturally acquired information affects our behavior in quite fundamental ways.

As a result, any explanation of human behavior that ignores culture, or treats it in an unsatisfactory manner, will almost certainly be incomplete. Yet a large number of social and behavioral scientists—many psychologists, economists, and political scientists, for example—either implicitly or explicitly downplay or ignore cultural influences on human behavior.

Other social scientists—many cultural anthropologists, archaeologists, sociologists,

and historians, for example—do  acknowledge the importance of culture, yet their methods and approaches often lack the scientific rigor and precision needed to satisfactorily explain how and why culture is the way that it is and how it affects behavior in the way that it does. Consequently, while the natural and physical sciences have made huge progress in the last century or so in explaining the hidden mysteries of life, matter, and the universe, the social sciences have failed to provide a unifying and productive theory of cultural change. The different branches of the social sciences remain fractionated, each speaking their own, often mutually unintelligible, languages and holding assumptions and theories that are mutually incompatible.

The definition that I will adopt in the rest of this book is that culture is information that is acquired from other individuals via social transmission mechanisms such as imitation, teaching, or language.1 “Information” here is intended as a broad term to refer to what social scientists and lay people might call knowledge, beliefs, attitudes, norms, preferences, and skills.

The vast majority of between-group behavioral variation in humans simply cannot be explained by genetic differences. Behavioral geneticists, by comparing identical twins (who are genetically identical) and fraternal twins or siblings (who share on average half their unique genetic variation), estimate that most behavioral and cognitive traits, such as IQ, personality, and psychopathology, have a heritability of around 40–50 percent. That is, around half the variation in behavior between people living in the same society can be attributed to genes. That leaves around 50–60 percent for culture. But the crucial point is that these differences are within societies, not between societies: for the kind of between-society differences discussed above, genetic influence will be substantially lower. Recent worldwide analyses estimate that the vast majority of human genetic variation (93–95 percent) is found within populations and just a tiny proportion (5–7 percent) between populations, and that this between-population genetic variation is far too small to explain documented behavioral variation in customs, practices, and languages.

More proximate research on the learning habits of children supports this conclusion. Children seem to be predisposed to rapidly and automatically acquire huge amounts of information from other people. They are, in a sense, “cultural sponges,” soaking up knowledge from those around them.

Evolutionary anthropologist Michael Tomasello has argued that it is our capacity to rapidly and accurately acquire huge amounts of information culturally that sets humans apart from other species.

A more theoretical line of support for the claim that humans are a cultural species comes from a set of theoretical models that show that it is often genetically adaptive for individuals to acquire information culturally.  In other words, it often in our genes’ interests (metaphorically speaking) to forego direct control over behavior and let culture take over.

The Origin as “one long argument.” This argument can be seen as comprising three elements, or preconditions: variation, competition, and inheritance.

It is reasonable to conclude that there is huge variation in human culture, on the order of millions to billions of variants, and that this variation can be documented and quantified. We can therefore state with some certainty that the first of Darwin’s preconditions—variation—is present in culture.

We therefore see competition in culture both at the psychological level, in the form of competition for space in memory, and also the effects of that competition, in the form of the extinction of various cultural practices and forms. Cultural traits, like biological organisms, take part in an endless struggle for existence.

Human culture therefore also exhibits the last of Darwin’s three preconditions for evolution, inheritance. Cultural variants can be passed faithfully from one individual to another, just as genes are passed from parent to offspring in biological evolution. Moreover, this cultural inheritance is of sufficiently high fidelity that it can successfully support the gradual accumulation of modifications, just as Darwin observed for lineages of biological organisms.

Spencer’s theory views species as a group of homogenous individuals that all share the same essential qualities. Evolutionary change occurs when one species abruptly steps up to the next rung of the evolutionary ladder to become a new, more “complex” species. The same was true of Spencerian theories of cultural evolution, except instead of species there are societies that move up a stage-like ladder. Darwin’s theory, on the other hand, focuses instead on the variation within populations (e.g., in beak size) and how this variation gradually changes over time…. within the population, rather than propelled by an externally triggered shift. This change in thinking, from essentialist, ladderlike thinking to variation-focused, Darwinian “population” thinking, has been described as one of Darwin’s major contributions to science.

Whereas genetic inheritance is particulate, cultural inheritance in many cases appears to be nonparticulate. As anthropologist Maurice Bloch puts it, “culture simply does not normally divide up into naturally discernable bits.”..

There is also evidence that cultural traits blend when transmitted.

Whether cultural evolution can properly be described as Lamarckian depends on how one defines the equivalent of the genotype phenotype distinction in culture. The ideational definition of culture given in chapter 1 implies that the cultural equivalent of the genotype is the information stored in people’s brains that represents their beliefs, attitudes, values, skills, knowledge, and so on. The cultural equivalent of the phenotype is the expression of that information in the form of behavior, speech, and artifacts. It is the latter—the phenotype equivalent—that is copied during cultural transmission: we do not directly acquire neural patterns of activation in people’s brains; we copy people’s behavior, we listen to what they say, and we read what they write. If we then modify the acquired beliefs, knowledge, and skills in some way before transmitting them to someone else, we can be said to be engaging in Lamarckian cultural inheritance.

Whereas cultural evolution does not appear to resemble neo-Darwinian evolution, with its strict assumptions of blind mutation and particulate, non-Lamarckian inheritance, cultural evolution can still be described as Darwinian, given the evidence reviewed above that it exhibits the basic Darwinian properties of variation, competition, and inheritance.

What is needed is a theory of Darwinian cultural evolution that explicitly incorporates non-neo-Darwinian microevolutionary processes such as blending inheritance, Lamarckian inheritance of acquired characteristics, and nonrandom variation, as well as other processes that may have no parallel whatsoever in biological microevolution.

Processes that cause changes in cultural variation over time



  1. Vertical: Transmission from biological parents (either uniparental or biparental)
  2. Oblique: Transmission from unrelated members of the parental generation
  3. Horizontal: Transmission from unrelated members of the same generation


  1. One-to-one: Face to face learning from one individual to another
  2. One-to-many: One individual influencing many others via mass education or mass media


  1. Blending: Adopting the average value of a continuous trait from more than one model
  2. Particulate: All-or-nothing transmission of discrete cultural traits


  1. Cultural mutation: Generating innovations entirely at random
  2. Guided variation: Individuals modify acquired information according to individual cognitive biases


  1. Content biases: Preferentially adopting traits based on their intrinsic attractiveness
  2. Model-based biases: Preferentially adopting traits based on characteristics of the model, e.g. their prestige, age, or similarity
  3. Frequency-dependent biases: Preferentially adopting traits based on their frequency, e.g. conformity (copying the most popular trait)

CULTURAL DRIFT: Random changes in cultural trait frequencies due to cultural mutation, random copying and sampling error

NATURAL SELECTION: Cultural traits spread due to their effect on biological survival and reproduction


  1. Demic diffusion: Cultural traits spread as their bearers move between groups
  2. Cultural diffusion: Cultural traits spread across group boundaries due to cultural transmission

Cavalli-Sforza, Feldman, Boyd, and Richerson showed that it is possible to build a useful theory of cultural evolution that is fully Darwinian, yet differs in many of its underlying microevolutionary assumptions from biological evolution.

See page 59 for Long-term population-level consequences of different cultural evolutionary processes.

This “diffusion of innovations” literature has been summarized by sociologist Everett Rogers, who identified several characteristics of successful innovations. In order to successfully spread, novel innovations must (1) have some relative advantage over existing practices or technologies, (2) be sufficiently compatible with existing practices or technologies, (3) be simple enough to use such that a potential adopter can quickly and easily understand or operate them, (4) be easily testable by adopters such that their relative advantage can be discerned, and (5) be observable to others, thus facilitating their spread. Each of these represents a broad class of content bias that will cause some traits (e.g., advantageous, compatible, understandable, testable and/or observable traits) to spread at the expense of other traits.

[Statistical analysis of diffusion show that] From this model, Henrich showed that guided variation did not generate S-shaped curves. Guided variation instead generated r-shaped curves, which show a rapid and steady initial uptake followed by a final slowdown.

Content bias, on the other hand, did generate S-shaped curves… With content bias, there is no individual trial and-error learning. Individuals are assumed to periodically sample the behavior of another randomly chosen member of the population, and if that person’s behavior is intrinsically more attractive than their existing behavior, then they copy it…  the slow initial uptake occurs because there is little variation in the population and few opportunities for people to encounter the favored trait. The accelerating increase that follows occurs because more and more models are available to learn from. The final slowdown occurs because of a lack of variation again, this time because everyone already has the favored trait.

Boyd and Richerson constructed models to explore this intuition more formally. They confirmed that a general prestige bias is indeed a good way of acquiring adaptive behavior compared to individual learning and unbiased transmission (random copying). However, this depends on the extent to which indicators of success (e.g., number of golf tournaments won) correlate with the traits that are copied.

Anthropologist William Durham has suggested that branching is maintained in cultural evolution because of “transmission isolating mechanisms,” or “TRIMs” for short, which act as barriers to the cultural transmission of information between groups… What might constitute a TRIM in cultural evolution? Language is probably a good candidate, given that people who speak different languages may not be able to communicate ideas, beliefs, and knowledge effectively, if at all. Ethnocentrism may be another potent TRIM. Ethnocentrism describes the widespread tendency for people to identify with members of their own groups (the ingroup) and avoid, derogate, or act aggressively toward members of other groups (outgroups). Ethnocentrism has been documented by anthropologists in virtually every society ever studied, as well as under controlled conditions in the psychology lab. Ethnic boundaries, then, may not only define but also create different social groups by preventing anything being transmitted across them.

Tehrani interviewed Iranian tribal women living in southwestern Iran, asking them from whom they learned weaving techniques and the specific patterns that they used. The women reported that they learned weaving techniques exclusively from their mothers, indicating vertical cultural transmission. This is consistent with the treelike pattern detected in the phylogenetic analysis. The textile patterns, on the other hand, showed a mix of vertical and horizontal transmission, as patterns learned from the women’s mothers were supplemented with patterns learned from other members of their community. Crucially, however, this horizontal transmission was predominantly restricted to within communities.

Postindustrial societies are, superficially at least, characterized by much greater intergroup transmission than traditional societies, given the consequences of mass communication technology such as television and the internet. As a result, modern technological evolution can probably seldom be described as a branching process, although this has yet to be empirically addressed.

So high innovation rates, in the form of intergroup transmission, lead to both high within-group diversity and low between-group diversity. Conversely, low innovation rates produce homogenous groups that are different from one another.

DNA analyses of modern humans show that African populations have the greatest genetic diversity, and that genetic diversity steadily declines the further one moves away from east Africa (in order of decreasing genetic diversity: the Middle East, then Southwest Asia and Europe, then Southeast Asia, then Oceania, and finally the Americas). Indeed, 85 percent of the variation in genetic diversity of modern human populations can be predicted from their geographical distance from East Africa.

In quantitative terms, a full 50 percent of the variance in hand-axe diversity could be explained in terms of geographical distance from east Africa.

Indeed, there are so many striking parallels between linguistic and genetic evolution that language may be particularly amenable to evolutionary analysis.  Words and grammatical rules appear to be transmitted with a fidelity close to that of genes: there is a rather high probability that you speak the same language as your parents, and that they spoke the same language as their parents, and maybe even back a couple more generations. Consequently, compared to, say, scientific knowledge or clothes fashions, languages change very slowly, perhaps at a similar rate to many species.

Lenski’s studies have revealed that when a genetically identical colony is placed in a novel environment (e.g., from a glucose to a maltose solution), fitness tends to increase rapidly at first, before plateauing to a constant level.

Lenski’s experiments suggest that evolutionary trends are to some degree repeated, but with small historical deviations.

Lenski found that the more E. coli there were in the colony, the faster was the increase in fitness. At very small colony sizes, little increase in fitness occurred at all. This suggests that colony size is an important determinant of adaptation.

The second major finding was that skills were more likely to be acquired by vertical transmission than knowledge. Reyes Garcia et al. suggest that this is because skills are more difficult to acquire, requiring repeated exposure and practice, compared to more easily acquired knowledge. Consequently, it may be easier to acquire knowledge from a range of sources through oblique transmission, while costly skills must be acquired from parents who are more accessible and more willing to spend time teaching their children.

Nelson and Winter outlined an evolutionary theory of economic change…

The first element of their theory was what they called “routines,” defined as well-learned and automatically executed sequences of behaviors carried out by workers or managers in a firm… Routines are transmitted to new workers when they arrive at a firm and are transmitted from managers to workers when new strategies are implemented.

An economic theory built on culturally transmitted routines rather than individual rational calculation can explain hitherto puzzling aspects of economic change.

Routines provide the inheritance in Nelson and Winter’s evolutionary theory of economic change… Variation comes in the form of new technological innovations that result from firms’ research and development (R&D) efforts… Competition might occur at the level of routines, with more-effective routines replacing less-effective routines. More commonly, however, competition is modeled at the level of the firm


Steven Klepper’s model starts with the appearance of some new technological or scientific innovation, which is potentially exploitable by firms. Yet because this innovation is unfamiliar and novel, firms are not aware of how best to exploit it (given their limited foresight). Many different firms appear, each one exploiting the innovation in a different way. Eventually one or more firms hit upon a product design that is particularly effective, or at least one that consumers become familiar with. This design comes to dominate the industry and firms that do not adopt this design go bust. Once product design has converged on this single dominant form, surviving firms can invest in R&D to develop the dominant design further. As a result, new companies that try to enter the market will be at a disadvantage compared to the incumbent firms, which form an oligopoly.

One potential explanation centers on the theory of cultural group selection, as developed by anthropologists and economists including Robert Boyd, Ernst Fehr, Herbert Gintis, Joseph Henrich, and Peter Richerson.  These researchers argue that the tendencies to cooperate and punish free riders arose in our evolutionary past as a result of gene-culture coevolution. Specifically, cultural groups in which people cooperate with one another and punish selfish free riders would have, during human evolutionary history, outcompeted cultural groups that were less internally cooperative and allowed free riders to exploit collective rewards. This intergroup competition might be through direct conquest, because internally cooperative groups are more effective in intergroup warfare (e.g., members of cooperative groups are more likely to sacrifice themselves for the rest of the group and punish deserters or cowards). Or it might be through more indirect means, such as when people are more likely to migrate to groups that exhibit prosocial norms (e.g., providing welfare for the poor or sick). Whether direct or indirect, this cultural group selection may then have favored a set of genetically specified psychological dispositions such as those exhibited in the experiments and field studies discussed above, such as a tendency to cooperate with other members of one’s group, and a sense of fairness that motivates people to punish selfish free riders.

These genetically specified psychological dispositions would have then facilitated the cultural evolution of various large-scale cooperative institutions. At first, these large-scale institutions took the form of egalitarian hunter-gatherer societies. Then larger social groups emerged, such as the empires.

Culture in Nonhuman species:

Moreover, cross-species comparisons have identified three different aspects of culture—one-to-one social learning, cultural traditions, and cumulative cultural evolution—that have distinct characteristics and appear to be underpinned by different underlying mechanisms. Yet they have often been conflated in humans, because we, unlike other species, have all three.

Countless field and laboratory experiments have shown that numerous species have the capacity for social learning, defined in minimal terms as where one individual acquires information from a second individual nongenetically, as a result of exposure to the second individual’s behavior.

We should perhaps not be surprised that social learning is so widespread in the animal kingdom. Social learning provides a quick and cheap shortcut to the acquisition of adaptive information: what to eat and where to find it, who to mate with and how to attract those mates, what predators to avoid, and so on, without going through the time-consuming and potentially costly process of individual trial-and-error experimentation. Why run the risk of eating something poisonous if you can copy what other individuals are eating?

While one-to-one social learning is necessary for any kind of culture, a distinctive element of human culture is the presence of group-specific traditions. These occur when all or most of the members of one group exhibits one behavior, while all or most of the members of another group exhibits a different behavior, and these group-level differences can be explained in terms of social learning rather than as the product of genetic differences between the groups, or of individual learning in response to different external conditions.

This group-specific aspect of culture is not necessarily the same as the one-to-one social learning outlined previously. In order for a behavior to become traditional, it must be transmitted not only from one individual to another, but across all members of a social group without any significant loss of fidelity. It must also persist for long enough such that distinct group-level cultural traditions can be clearly identified. This would likely be a minimum of one biological generation.

One important way in which human culture appears to differ from the culture of any other species is that human culture is cumulative. The products of human culture are in many cases the result of the gradual accumulation of many successive modifications, with each incremental modification increasing the effectiveness of the cultural trait in question.

Tomasello uses the metaphor of a ratchet to describe human culture.

In contrast, none of the examples of nonhuman cultural traditions given above appear to be the product of cumulative culture. The examples in fish of culturally acquired schooling site preferences and migratory routes are in no real sense accumulated in that they cannot be separated out into constituent subcomponents that could potentially have been built up over time. Moreover, a single fish could easily discover a particular schooling site on their own. Birdsong, while showing clear change over time, does not accumulate changes of increasing functional effectiveness. Indeed, the fact that birdsong fits the expectations of neutral drift models shows that different song syllables are functionally neutral. And while the culturally transmitted tool-using behaviors of chimpanzees, orangutans, and dolphins are impressive, they too do not appear to be composed of a series of smaller components that could have accumulated over time in the way that a car or computer does.

This distinction between cumulative and noncumulative culture is hugely important, because only the former constitutes the gradual evolutionary change that Darwin termed “descent with modification.” In other words, only humans exhibit a fully Darwinian process of cultural evolution.

Cumulative culture may depend not only on the ability to imitate others but also on the compulsion to imitate others. If children compulsively imitate adults’ behavior, then cultural modifications are more likely to be preserved and accumulate over successive generations.

Other recent studies with chimpanzees and children have presented an alternative explanation for the former’s lack of cumulative culture. Rather than a lack of transmission fidelity, it may be that other species do not possess the ability to easily switch to a newly encountered, better way of doing something. Instead, they show a preponderance to stick with their existing solution to a problem. This “stickiness” prevents nonhuman species from cumulatively building up successively more effective cultural elements.

Another suggestion for the key to cumulative cultural evolution is teaching.

Csibra and Gergely argue that teaching is particularly useful when what has to be learned is “cognitively opaque,” that is, when it is not immediately obvious what function other people’s actions serve. Much of cumulative culture is likely to involve the transmission of cognitively opaque actions.

So just as modularity facilitates biological evolution by reducing the need to independently evolve repeating body parts, so too modularity facilitates cultural evolution by reducing the need to independently evolve repeating technological components, a prediction that is supported by computer simulations of artifact learning.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s