Book Review: The Secrets of Our Success by Joseph Henrich

Why are human beings so much more successful than any other species on earth? 

Why are we able to build rocket ships when our nearest primate cousins struggle to use basic tools like sticks?

For a long time, I thought the answer had to do with raw intelligence. That somewhere during the gradual genetically-driven expansion of brain size amongst primates, our lineage hit some raw intelligence threshold that caused the phenomenon of human sentience to emerge fully formed; that an analogous process to what produced, for example, the abnormally long neck of the giraffe, was responsible for the unusually large brain of Homo sapiens.

But I recently came across two key ideas which threw this narrative, and indeed many of my basic beliefs about what it means to be human, into question:

1. Cultural learning ability, not general intelligence, is our superpower. Raw intelligence, i.e. our brain’s ability to hold and process information, can’t be the explanation for our success, because (counterintuitively) we’re not even that great at it. There’s evidence that, even today, if stripped of culturally acquired tools and prebuilt solutions, our brains are not much innately better than other primates except when it comes to our social learning abilities. This is the one area where we clearly excel other species. (Homo sapiens, “thinking man”, appears to be a misnomer — Homo imitans, “copying man”, is closer to reality).

2. Culture, not DNA, is the driving force in human evolution. What extraordinary natural intelligence we do possess is largely a byproduct of gene-culture coevolution, a process whereby cultural adaptations became so advantageous to human survival and reproduction that the drive to be able to acquire, retain, and pass on cultural information began to exert genetic selection pressure on our brains to become larger — i.e. capable of holding and processing more cultural information — triggering a positive feedback loop. 

Such, at least, is the central argument put forth in the seminal 2017 book by Harvard human evolutionary biologist Dr. Joseph Henrich: “The Secret of Our Success: How Culture Is Driving Human Evolution, Domesticating Our Species, and Making Us Smarter”. 

Henrich’s “Cultural Intelligence Hypothesis” implies that it wasn’t crossing some IQ threshold that made humans able to stuff like build fire and fashion stone tools. In fact, Henrich argues, there’s good evidence that even educated modern people could not figure out those things by virtue of their intelligence alone.

Instead, the true inflection point in our evolutionary past seems to have been reached when we first became capable of cumulative cultural learning — passing down cultural practices we pick up from one mind to another and from one generation to the next. 

As soon as we became capable of this, it allowed the evolutionary process to go to work on a new substrate for the first time — instead of selecting for biological information carried by genes, it started selecting for cultural information carried by human minds like ideas, know-how, technological innovations, social norms, rituals, etc. And this new cultural evolution proved faster and more powerful than what had come before.

Henrich calls this moment crossing the Rubicon, as it was a critical moment in our history as a species from which there was no turning back.

Cultural Evolution: A Simple Idea with Far-Reaching Implications

In the few years since The Secret of Our Success was written, I’ve read and reread it. I’ve highlighted so much text in every chapter that highlighting stuff became a pointless exercise. I took no less than 30 pages of notes. 

Why? Because I think the concepts it contains are nothing short of revolutionary. 

They represent a paradigm shift in the way we think about the origins and nature of humanity, psychology, culture, and intelligence. It has far-reaching implications for every discipline from anthropology to economics to history to philosophy.

If we truly are a culturally-determined species, a “new kind of animal” with two systems of inheritance, genes and memes, instead of just one (genes) like every other species, it implies that the old nature/nurture dichotomy, one that has been the cause of endless debate for millennia, has been missing a key piece of the puzzle.

What the nurture side misses is that the content of nurturing (i.e. culturally transmitted information) is the product of the same evolutionary process that drives genetic evolution. As such, the idea that you can consciously design someone’s nurture to lead to a particular outcome is simply untrue. It vastly underestimates the complexity and density of implicit cultural information.

What the nature side misses is that cultural differences are more important to differences between individuals and populations than generic ones. What’s more, genetic change itself is powerfully driven by the pressure to make optimal use of fitness-enhancing cultural adaptations.

To whatever extent genetic differences between people matter, it is almost certainly the case that cultural differences matter much more. 

Therefore, meaningful differences between populations do exist, but they are culturally acquired, not innately genetic. Further, we are neither fully “self-programmable”, nor is very much of human nature innate. 

And crucially, just like life itself, culture evolves as part of an undirected process of trial and error, it is not designed by any intentional intelligence, whether that be of a deity or of human beings. We can choose what cultures to expose ourselves and others to, but we cannot design cultures from the top down. 

If culture is an evolutionary product that underlies both our individual natures and our societal traits, then it’s unlikely that intentional design can improve upon it as easily as we may intuitively think. As we’ll see, there’s reason to believe that cultural evolution is much smarter than we are. 

Every section in the essay to follow has two parts:

1. Foundational concepts of the cultural evolutionary perspective from The Secret of Our Success. Here, I’ll compress Henrich’s most compelling explanatory models on questions like what makes us intelligent, how humans evolve, how innovation happens, and why social norms spread. 

2. Important implications of these Henrich’s models, as I see them. I’ll discuss how cultural evolution relates to many of the issues that dominate public discourse today, including race, intentionality, social designs, and systems thinking. To clarify where I’m contributing original ideas, as opposed to where I am channeling Henrich’s, I’ve used the subheader “implications” as a signpost throughout the essay. 

In no way do I think the essay to follow is an adequate substitute for reading Henrich's work, nor I am not the first person to review The Secrets of Our Success and grasp its profound significance (for example, see this popular review by Scott Alexander over at Slate Star Codex).

But I do think I can add something to the conversation. This essay is somewhere in between a cover and a remix, an homage and a summary. 

10 Key Concepts in “The Secret of Our Success”

If you want cool new technology, is it better to be social or to be smart?

Your intuition probably tells you that it’s better to be smart. In school, for example, we probably don’t expect the hyper-social people to go on to be great inventors and technologists. But this general observation, which may apply for individuals, likely does not apply for populations. 

Henrich poses the following thought experiment to explain why:

Consider two very large prehuman populations, the Geniuses and the Butterflies.

Suppose the Geniuses will devise an invention once in 10 lifetimes. The butterflies are much dumber, only devising the same invention once in 1000 lifetimes. So, this means that the Geniuses are 100 times smarter than the Butterflies.

However, the Geniuses are not very social and have only 1 friend they can learn from. The Butterflies have 10 friends, making them 10 times more social. 

Now, everyone in both populations tries to obtain an invention, both by figuring it out for themselves and by learning from friends. Suppose learning from friends is difficult: if a friend has it, a learner only learns it half the time… do you think the innovation will be more common among the Geniuses or the Butterflies?

Well, among the Geniuses a bit fewer than 1 out of 5 individuals (18%) will end up with the invention. Half of those Geniuses will have figured it out all by themselves. Meanwhile, 99.9% of Butterflies will have the innovation, but only 0.1% will have figured it out by themselves. Keep in mind that the Geniuses were 100 times smarter than the Butterflies whereas the Butterflies were only 10 times more social. 

Bottom line: if you want to have cool technology, it’s better to be social than smart.”  - The Secret of Our Success, pg. 213

This is one example of how applying population thinking and a cultural-evolutionary lens can lead to some surprising and counterintuitive conclusions. I’ve chosen to highlight the 10 concepts below from The Secret of Our Success because they all share that trait. 

1. Human Beings Are Butterflies, Not Geniuses

As humans, we certainly feel like we’re the most important life form on the planet. But maybe we only think this because we’re biased. What can the evidence tell us about our own success as a species?

Henrich begins his book with some impressive stats about our ecological dominance:

• Humans take up 100x more biomass than any large species that has ever lived, including the dinosaurs.

• Together with our pets and domesticated animals (sheep, cows, etc.), we account for 98% of the biomass of land-bound animals with a spine.

• If you put every human being in existence on a scale and weighed us against other species, only ants would even remotely be in the same ballpark. But there are 14,000+ species of ant, and only one species of human.

Thus we’re a clear outlier in terms of ecological success. We’re also unequalled in terms of our control over — and unintended impact on — our environment. For example, we’ve altered more than one-third of the earth’s land surface to serve our agricultural or industrial needs. No other species could conceivably travel to other planets, nor split the atom. The question is, why?

As mentioned earlier, the intuitive answer is that we’re simply that much smarter than other animals because the right mix of environmental factors in our past led to the genetic development of bigger brains in our lineage than in any other species.

But what exactly caused us to have these big brains, and whether it is brain size alone that accounts for our intelligence, is the source of considerable controversy among scholars. They’ve come up with a large number of competing “intelligence hypotheses”, none of which appears to have reached universal consensus.

Henrich is one of the foremost advocates of the “Cultural Intelligence Hypothesis”. The claim here is that information transmitted culturally, which accumulates over generations via a blind evolutionary process (and is thus “cumulative”), produces useful adaptations that drastically improve our odds of survival and reproduction.

As discussed in greater detail in sections 3 and 4 below, this cultural process created genetic selection pressure that improved our memory and information processing power in order to more effectively store and retransmit cultural information, and also produced mental tools and pre-built solutions that augment our intelligence considerably.

Cultural adaptations themselves take various forms — everything from physical technologies like how to light a fire or build a bow and arrow, to social technologies like ritual dances or sacral rites that increase social cohesion within a tribe, to mental tools like terms for “left” and “right”. These adaptations are selected for because of their capacity to improve the survival and reproduction odds of human groups who hold them.

The capacity to learn and pass on such cultural adaptations (cumulative cultural learning) both explains our success, Henrich argues, and drove the threefold expansion in our brain size from cm3 to 1350 cm3 in the last ~5 million years. To the extent that we have developed greater innate intelligence than other animals, it was primarily driven by the need to learn, process, and pass on cultural adaptations — not vice versa.

In other words, raw human general intelligence wasn’t responsible for the origin of culture and cultural evolution. It’s the other way around!

To illustrate that cumulative cultural learning is the unique advantage/superpower of human beings, not brain size or raw intellect, Henrich brings up various points:

• Brain size is the strongest predictor of cognitive abilities (raw intelligence) in primates. Neanderthals had bigger brains than us, and therefore were likely smarter innately, but were much less successful, both ecologically and technologically. 

• In a laboratory showdown between chimps, orangutans, and young kids without the advantage of culturally-acquired mental tools (like algebra), the only way in which humans far outperformed other apes was in their social learning abilities. With respect to spatial, quantitative, and causal reasoning, humans were not much innately better (findings by Michael Tomasello, Esther Hermann, et. al. published in 2007 & 2010).

• In a test comparing the working memory and information processing speed between chimps and humans conducted by Japanese researchers Sana Inoue and Testuro Matsuzawa, the humans were only slightly better when it comes to working memory, while the chimps dominated in terms of information processing speed.

• In a strategic game called matching pennies played between chimps and humans, chimps were far faster at zeroing in on a game theory optimal strategy, while the humans struggled. 

Therefore, there is little compelling evidence that, even now, humans stripped of culturally-acquired tools have more raw intelligence than other apes or extinct hominids (Neanderthals, who seem to have exceeded us in brain size, may even have been innately smarter). Instead, our real superpower is cultural learning. 

Implication: Human Intelligence Is Not Uniquely Rational

I have to confess that I found Henrich very persuasive here, despite the fact that my beliefs prior to reading his work were directly in opposition to what he’s saying. I’ve always basically aligned with Plato’s theory of the tripartite mind, which implies in modern terms that our large neocortex, where higher reasoning occurs — Plato’s “λογιστικόν” (logistikon) — is what separates us from the animal world and makes us intelligent.

There are several versions of this narrative of intelligence, not all attributed to Plato, which put higher-order reasoning at the center of what defines human intelligence. For example, Aistotle, spoke of man as being distinguished by “having reason” (λόγον ἔχον) in the Nicomachean Ethics.

I’ve found that this “Promethean script”, as Jonathan Haidt recently called it, is difficult to maintain against the strong case Henrich makes for the Cultural Intelligence Hypothesis. 

While we can still value human reason, the evidence seems to contradict the intuitive idea that our ability to reason — or to think ahead (the meaning of Prometheus’s name) — is what caused us to become intelligent and develop the ability to (for example) create fire. 

This matters because I think it redefines what it means to be human. What makes us unique, and even what makes us sentient, is that we alone of all life forms can transmit cultural information. This means we can transmit evolutionary adaptations by way of memes (Dawkins’ proposed term for the atomic unit of cultural selection), whereas all other organisms rely on genes to transmit such adaptations. 

Henrich says that “it’s not our intelligence”, nor even our reason, that makes us special. The explanation he offers instead is far more complex, counterintuitive, and fascinating.

2. Why Influencers Exist: Cultural Learning Is Preferential

So we’ve seen that the human brain is particularly geared toward cultural learning — i.e. being able to acquire, process, and retransmit information from others. If this is our superpower, what makes us so good at it?

First, we must realize that cultural learning doesn’t just mean copying everything other people do indiscriminately. Henrich is careful to emphasize that we are preferential learners - that is, we selectively choose who and what we learn from. We filter cultural learning with two key, seemingly innate cognitive abilities that have been observed cross-culturally:

• Model-based learning: we are attuned to certain cues of skill, success, prestige, age, and self-similarity that help us choose “models” that we preferentially pay attention to and learn from in any particular context. 

• Inferring intentionality: the innate ability to “mentalize”, or to infer the intentionality of people’s behaviors, allowing us to filter the signal from the noise in terms of what we should learn from them (and what we should ignore).

It’s also important to understand that our cultural learning abilities don’t imply that we use cultural learning for everything we do. Individual learning — i.e. figuring things out for yourself — is obviously common for humans.

But we do rely on cultural learning to define some of our most important mental architecture, including strategies, beliefs, motivations, tastes, and values — all acquired by cultural transmission, not innate. 

We also heavily rely on cultural learning in contexts of high uncertainty or when getting the right answer is critical to our survival: “the more important getting the right answer is, the more we rely on cultural learning” (The Secret of Our Success, pg. 35).

Putting it another way, we rely on cultural learning the most when we have no idea what we’re doing, or when something critical is at stake. It’s also the case that many instances of “individual learning” can only happen with the aid of culturally downloaded information.

Here’s an example I use to think about this: 

Say you’re a math student assigned the following problem: “2x+3 = 11. What is x?”

For whatever reason, you haven’t yet been taught how to balance algebraic equations, so your only choice is to plug in random variables for “x” and proceed on a trial and error basis. 

But then you have a great flash of insight — “if I just subtract 3 from 11, the equation becomes 2x = 8, and then the answer is obvious”. So you have “individually” learned how to balance an equation, relying on nothing but your own smarts!

Was this a genuine example of the awesome deductive intellect of the middle-school brain? It might seem to be — until you realize that things like variables, equations, rules of mathematical operation like subtraction and multiplication, and the base 10 counting system itself are all culturally-downloaded concepts, without which this insight could never have occurred. They are, as Henrich puts it, culturally-learned mental tools that fundamentally reshape our neurological infrastructure.

How likely is it that the student could have figured out all those concepts from scratch in the same manner that they recombined culturally-acquired concepts to figure out a novel algebra problem? Henrich would probably argue it’s not very likely, considering the thousands of years it took to invent them in the first place. Though a child can now learn and understand them, they were very difficult to discover for the first time. 

I take Henrich’s point to be that much of what we think of as individual problem-solving is actually just a case of the individual applying culturally-learned concepts and mental tools to come up with a solution. These mental tools are the product of a long line of cultural evolution, and they augment our natural intelligence in ways we don’t realize (more on this later).

To return to the subject of how preferential cultural learning works, let’s take a closer look at the cues we are innately attuned to, even as invents, that we use to figure out who to learn from. Henrich divides these cues, which he says are psychologically compelling across cultures, into several categories:

Success/skill cues

If someone is particularly successful or competent in a given domain, they are likely to become models for cultural learning by others. 

Henrich illustrates this by describing a laboratory study with MBA students. The study found that students who used success cues to copy others’ investment strategies ended up with investment portfolios that heavily outperformed a group that could only use individual learning. It was also found that (without being direct to do so) the group who could copy others usually chose to copy those with the most successful investment outcomes. 

This happens with kids as well: one experiment with German infants showed that even babies tend to preferentially copy models they judge as more competent.

Success/competence cues are both the most indirect and potentially the most useful from an evolutionary perspective since they aggregate the most information. In a complex world, it is hard to tell what makes someone successful. Is a tribesman a good hunter because he makes his bow a certain way, or because of how he tracks?

In practice, it’s too hard for a young learner to isolate the cause of success in a causally opaque and complex world. Therefore, evolution favors that they copy the successful people more or less blindly so that, even at the risk of doing a lot of unnecessary things, they will be more likely to do the one thing that helps them recreate success and become a good hunter.

Prestige cues

Prestige bias inclines us to culturally learn who to culturally learn from.

That is (unsurprisingly), one of the main ways we learn who to model behavior on, and who to copy ideas from, is by reference to others — especially our peers. Even in infancy, we automatically notice who people tend to listen, watch, or pay attention to, and preferentially learn from them.

For example, in one experiment conducted by Henrich and his team, children watching a prestige-cued model (who bystanders preferentially watched) play with a toy were 13 more times more likely to play with that toy in the same manner than how an alternative model played with it. Critically, when interviewed after the experiment the children had no conscious awareness of how they’d been affected by prestige cues — just as we rarely are conscious of similar effects on us in the real world. 

Prestige is especially powerful because it’s cross-domain. People are influenced by what prestigious individuals say and do, even if the domain they’re commenting on isn’t directly relevant to the area of expertise that made them prestigious in the first place. This is probably the evolved psychology that underlies the success of everything from influencer marketing, to celebrity political candidates, to social advocacy from athletes.

Self-similarity cues

Because characteristics like height, personality, gender, and cultural traits may all influence how successful we are in our endeavors in unpredictable ways, learners have evolved a predilection to pay preferential attention to those who are most like them across these different dimensions.

This would make evolutionary sense since it allows learners to filter for the most relevant cultural information to them, and has been shown to apply in numerous ways. 

Cultural self-similarity is also important. For example, kids prefer to learn from those who speak a similar dialect or language, even if they are speaking complete nonsense, rather than someone who speaks nonsense in a different dialect.

One reason for this is that dialect is often a cultural marker that signals a host of other correlated cultural traits, like beliefs, ideas, practices, values, and tastes — in a word, ethnicity (more on this later). All else being equal, learning from someone in the context of a shared cultural package signalled by dialect makes more sense than someone who isn’t, especially since many cultural adaptations are mutually reinforcing and synergistic. 

For example, an American boy who learns piano from a German classical pianist may have some difficulty when he tries out for his high school orchestra program and finds that he has inadvertently picked up musical terms that no one else understands, such as “violinschlüssel” instead of “treble clef”.

Age cues

In the small-scale societies of our evolutionary past, and even in some places today, reaching old age was seen as an accomplishment. Without the benefit of modern food production techniques, medication, etc., more people died before growing old.

Therefore, the practices of the elderly contained relevant survival information for young learners. Say you had a group of 100 hunter-gatherers living in a tropical environment, all in their 20s. 40 of them use chilli peppers regularly when they cook meat, while 60 don’t. If the antimicrobial properties of chilli peppers increase the probability of surviving past 65 from 10 to 20%, then 57% of this group of 100 will be chili-pepper eaters by the time they’re 65.

In other words, natural selection will have selected for the cultural practice of chilli pepper eating in this cohort. The next generation of learners, using age-biased learning, will preferentially copy those older than them, meaning that more than 40% will use chili peppers in their meat, increasing the survival rate of the tribe as a whole.

On top of this, in small-scale societies without the benefit of literacy, the elderly were frequently repositories of vast bodies of cultural information that younger members of the tribe had not been able yet to master. Henrich makes the argument that the adaptive benefit of having access to this repository is one of the reasons why human beings live so long past reproductive age.

Inferred intentionality

Finally, in addition to being attuned to various cues, our evolutionary psychology has also equipped us to infer the mental states of other people: their motivations, goals, intentions, etc. Henrich argues that the primary evolutionary purpose for this, and the reason we get it so early on in life (as early as eight months old), is that it aids our ability to learn from others. We can only truly acquire information from others if we can put ourselves (somewhat) in their shoes and infer lessons based on their behavior.

Thus, if a child sees their father shaving, they infer intentionality — they assume that he’s doing it to achieve some goal, and are therefore equipped to ask one of the many, many questions young children are wont to ask about everyday things like “dad, why do you shave?” But notice that a child isn’t similarly inquisitive about everything a parent or some other model does — more likely those things which display some level of deliberation, intentionality, or goal-seeking. 

So we now see that (1) we have learning biases deeply built into our cultural learning psychology which “fire up” on certain cues, and (2) that we are hyper-attune to intentionality — we mentalize others and try to figure out their motivations.

Implication: Our Approach To Education Should Change

Henrich’s explanation of how preferential learning works changed how I viewed education, or indeed any form of social influence. On the one hand, it has practical implications for how we can successfully teach others. It also brings into focus who is most likely to influence us, and why — whether it be prestige bias, age bias, or self-similarity. 

I believe this has tactical implications for educators and influencers as to who their likely audience of learners will likely be. It also presents a challenge for those of us who want to overcome the limitations imposed on us by the accident of birth and access ideas according to their true merit, rather than because they are held by cultural models defined qualities that may be arbitrary today (though they were evolutionarily advantageous at one time) like self-similarity, success, or even prestige. 

3. Cultural Evolution Is (Much) Smarter Than Us

So there’s clear evidence that we evolved a strong capacity for cultural learning, and that our natural affinity for cultural transmission is much more pronounced than our raw, individual intelligence.

From an evolutionary standpoint, the only reason we’d have developed such a remarkable talent for cultural learning is if it somehow produced much better solutions (adaptations) for the diverse and challenging survival situations we faced in our evolutionary past than individual intelligence ever could.

So if it can be shown that unconscious “dumb” process of cumulative cultural evolution — trial, error, and preferential copying of lucky improvements — produces better solutions over time than conscious, individual or group problem-solving, then that process can be said to be “smarter” than we are. 

And were that to be true, it would make follow that the selection pressure on our evolved psychology would be towards better enabling us to learn, process, and retransmit cultural adaptations (i.e. gotten really good at cultural learning) instead of amping up our individual intelligence by equipping us with innately more powerful working memories, processing speeds, and causal/spatial/quantitative reasoning skills than can be found in other species (which the evidence, as we’ve seen, doesn’t support).

But how can a blind, unintentional process outperform the intentional, conscious problem-solving of human beings? Doesn’t that go against all of our life experiences, where we are constantly solving problems requiring the use of intentionality and intelligence?

The idea of impossibly complex solutions being developed by evolution is not a new one. It is now accepted in biology, for example, that fantastically complex solutions to the problem of cellular metabolism have been developed without the aid of any intentional agent; not only that, but it is unlikely that even with the aid of the best AI now available to us, we’d be able to design a better system from scratch than what evolution has done. 

Here it’s helpful to recall that one of the biggest challenges to Darwin’s evolutionary theory mounted by intelligent design advocates was the argument that the existence of biological patterns of “irreducible complexity” (like the human eye) implied the existence of an intelligent designer. 

This view was eventually debunked — and now, at least when it comes to biology, it’s commonly accepted that complexity and order beyond the capability of “intelligent” human agents to consciously design can arise from simple, unintentional processes like natural selection.

This bottom-up view has been seen to apply in other domains as well. Something very similar was proposed by none other than Adam Smith, whose Wealth of Nations preceded the Origin of Species by 83 years. Smith’s famous “invisible hand” argument was another way of demonstrating the counter-intuitive concept that the unintended consequences of simple forces (in this case, market incentives as opposed to natural selection) can give rise to emergent order.

In more recent times, it has become an increasingly influential view that lucky trial-and-error processes, whereby small iterative innovations are selected for and propagated, are largely responsible for modern technological innovation, as argued in popular books like How Innovation Works by Matt Ridley and Antifragile by Nassim Nicholas Taleb. 

Henrich makes the same argument for pre-modern cultural innovation. 

He shows how the expansive repertoire of both cultural know-how and hunter-gatherer technology developed by small-scale societies over a slow, unconscious process of trial and error ends up with solutions more advanced and complex than any individual or group could possibly figure out in a lifetime — everything from how to hunt seals, to how to make a bow and arrow, to how to detoxify foods like manioc and corn.

These are solutions which modern scientists would have struggled to come up with, and that baffled modern people when they initially encountered them. 

Here are just some of the examples Henrich mentions to illustrate the point. 

The Impossibility of Arctic Seal Hunting

In June 1845, the HMS Terror & Erebus, two British ships, set out on an expedition to chart the Northwest Passage. Their ships ended up getting stuck in ice in a place called King William’s Island. The local Inuit tribe, the Netsilik, called this place “Uqsuqtuuq” meaning “lots of seal fat”. 

However, despite having enough provisions for 19 months, the crews of the Erebus & Terror eventually ran out of food, ended up resorting to cannibalism, and presumably all died (the same expedition now being dramatized in AMC’s television series The Terror). 

In all that time, why weren’t a group of 105 modern European explorers, equipped with the latest industrial-era technology, able to figure out how to live off the land? Why was a place basically known as “the land of milk and honey” by the locals so inhospitable to the explorers?

Henrich suggests that it’s because, in this novel environment, none of the cultural adaptations that made finding or growing food possible in England applied, and they had to start from scratch. Instead of hunting deer or pheasant (assuming some of the crew had learned this back in England), they had to figure out how to hunt seals, which was a completely different ball game. 

As we’ll see, the complex repertoire of cultural adaptations needed for something even as simple-sounding as Arctic seal hunting is so great that it’s basically inconceivable that a person or group of people could figure out how to do it relying on problem-solving skills alone during their lifetimes. 

Here are the necessary steps involved in seal hunting in the Canadian arctic, as described by Henrich:

1. First, you must somehow discover how to identify seal breathing holes in the ice. You must also figure out that the surrounding area has to be snow-covered, or the seals will hear you coming.

2. Open the hole & make sure it’s still being used by smelling it (you’ll have to figure out what seals smell like).

3. Measure the hole using a special curved piece of caribou antler (assuming you can figure out how to build with a composite recurve bow using nothing but driftwood and then use it to hunt and kill a caribou).

4. Learn how to cover the hole with snow, except for a small gap at the top for a down indicator so you’ll know when the seal enters the hole.

5. When the indicator moves, plunge your harpoon into the hole. Learn to do this with all your weight. Also, figure out that you need to build a 5 ft harpoon with a detachable tip tethered to the shaft by a braid of sinew line, and then learn how to build it. 

6. After you’ve harpooned the seal, you have to reel it in using the detachable tip, wrestle it onto the ice, and finish it off with a spike made of polar bear bone (so you’ll need to know how to find and kill polar bears before you can do this).

7. Now you have to cook the seal meat. Since there’s no wood in the area, for fire you have to carve a lamp from soapstone, render oil from blubber, and use a particular species of moss to make a wick. 

8. You also need water, which means you have to be able to identify old sea ice which has lost most of its salt by means of its color and texture.

Each step of the way, you need to figure out not only what to do but that you should even do it in the first place. How are you supposed to predict that you’ll need a polar bear spike to kill the seal after you’ve reeled it in? How are you supposed to know what soapstone looks like, or what the color and texture of old sea ice is — let alone that it is the best source of water?

And this “seal hunting package” just one of the many adaptations in a “cultural repertoire” which allowed tribes like the Netsilik to flourish in the Arctic, or indeed in any part of the world. 

What are the odds that such complex, co-dependent cultural adaptations arose through someone inventing them from scratch, as opposed to a slow process of trial and error? If 105 stranded explorers couldn’t figure out how to seal hunt when their lives depended on it and ended up resorting to cannibalism out of desperation, why would we expect it should be possible for a single, particularly inventive ancestor of the local Inuits? 

What’s more is that this story of Franklin’s Lost Expedition is just one of many in what Henrich calls the “Lost European Explorer” files, which are particularly illustrative historical accounts of literate people being stranded in novel environments.

In case after case, in diverse environments spanning from the Australian outback to the Gulf Coast of Texas, the story is the same: a group of European/American explorers that get stranded in a remote place, find themselves unable to adapt, and waste away from either disease or hunger, with most of the group eventually dying. 

In the few occasions where the explorers survive, it is usually with the help of the locals, who culturally transmit vital local adaptations to them. Thus, two expeditions that took place at different times and which also happened to get stranded on King William’s Island were able to survive with the help of the Netsilik. (The historical origins of Thanksgiving follow this same pattern; the settlers of Plymouth were rescued from starvation by a Patuxet Native American whose cultural know-how helped the Pilgrims learn how to catch eel and grow corn, among other things that they could not have figured out by themselves.)

Not only this, but Henrich gives an example of the opposite case: how a person stranded with the help of culturally transmitted local adaptations could survive. Juana Maria was a Nicoleño Native American who was stranded on an island off the coast of California for 18 years on San Nicolas Island after the rest of her tribe evacuated but accidentally left her behind. Despite her total isolation, she survived by relying on a repertoire of locally-adaptive cultural information. Her tale was later popularized by one of my favorite childhood books, the Island of the Blue Dolphins. 

Unintentional Innovation: The “Savannah Hunting Package” Thought Experiment

If cultural adaptations like “how to hunt seals”, and specific technologies like composite recurve driftwood bows weren’t invented single-handedly by some forgotten paleolithic version of Nikola Tesla, how can we explain their existence using cultural evolution? How is cultural evolution able to fashion solutions to problems that our conscious intelligence can’t?

Let’s consider Henrich’s demonstration of how gradual trial-and-error in early tool use may have led to complex “packages” of adaptations to the local environment.

To do this, he sets up a hypothetical. Imagine a small group of our ancestors — forest-dwelling primates. 

One of them figures out how to use a stick to take termites out of a termite mound and eat them (we know this ex nihilo innovation is possible since modern chimpanzees have been observed doing it). From this baseline, a series of lucky errors could lead to a complex package of cultural adaptations over the generations, as follows:

• Let’s call the initial generation of the termite-stick inventor Generation 0, and label the cultural trait of using a stick in this way as T.

• Suppose that in Generation 1, two offspring copy the termite-fisher. But in the process, one of the copiers mistakenly thinks the stick has to be sharpened first. This copying mistake, i.e. mutation, can be labelled T*. 

• At the same time, another member of generation 1 figures out how to use hollow reeds to drink water from the troughs of trees, which enables them to find water out in the open Savannah. This new innovation can be labelled T2. 

• Members of Generation 2 preferentially copy T* and T2 because members of generation 1 who practiced them survived to an older age & were more successful on average than others, leading to the spread of T* and T2 in generation 2.

• One member of generation 2 even copied both T* and T2, which made them particularly successful, and was thus the combination of both was preferentially copied by 3 members of Generation 3. 

• One day, a member of Generation 3 accidentally spears a rodent with his stick while termite fishing in an abandoned mound. The termite-fishing-stick now becomes a “hole-spear”, T**, which can be used for a wider range of food sources.

• Elsewhere, another member of Generation 3 notices a rabbit enter a hole during rain and observes the muddy tracks it leaves behind, leading to T3 (rabbit-tracking). 

• In Generation 4, T** (hole-spearing) spreads, and one member who is also taught T3 (rabbit tracking) manages to combine the two to hunt rabbits in their holes.

• In Generation 5, though no one invents anything, several members learn a combination of T** (hole-spearing), T2 (tree-drinking), and T3 (rabbit-tracking). This package allows them to spend more time hunting on the savannah instead of the trees — thus, the “savannah hunting package”.

Diagram from The Secrets of Our Success, pg. 56

Although this is just a hypothetical and not a real-world observation, I think it brilliantly describes the (likely) process by which cultural adaptations arise through a combination of happenstance, preferential learning, and cumulative memetic spread over the course of generations.

And the most fascinating, deeply counterintuitive thing about it is how little intentionality or conscious invention plays a role. The “savannah hunting package” thus illustrates in clear terms how complex behavioral adaptations can have simple, unintended causes. 

Implication: We Can’t Out-Think The Evolutionary Process

The concept of evolution being smarter than us has clear implications beyond simple paleolithic tool use. As discussed earlier, a similar logic placing combinatorial, iterative evolution as the driver of technological innovation has been gathering steam in the public consciousness of how modern technology advances.

In fact, the tech industry itself is becoming more and more cognizant of this concept, thanks to the publication of two influential books within the last decade or so drawing a direct parallel between technological innovation and biological innovation: The Nature of Technology: What It Is and How It Evolves (2009) by W. Brian Arthur of the Sante Fe Institute, and What Technology Wants (2001) by Kevin Kelly, founder of Wired magazine (products like TikTok have even begun building features that encourage users to engage in combinatorial innovation along the same lines, as described in this conversation between a16z’s Sonal Chokshi and Eugene Wei).

The central theme of these books and others like them — that technological innovation is the product of an unintentional, iterative, evolutionary process — is a more specific formulation of Henrich’s broader thesis around cultural evolution, which applies not only to technology but to all forms of adaptive cultural innovation including especially language, know-how, beliefs, ideas, customs/rituals, and other culturally transmissible patterns of thought and behavior. 

Just like technological innovation, these things arise within a framework of competition that tends to select the most adaptive option from an evolutionary standpoint. And because these adaptations accumulate over time, cultural evolution is able to produce solutions more advanced and wonderful than an individual could likely conceive, let alone figure out, relying on their conscious, raw intelligence alone.

The broader implication of this is that we should be very careful in trying to “out-think” organic processes of trial-and-error iteration. It’s very unlikely we’ll succeed. 

4. Cultural Evolution Makes Us (Much) Smarter

We have seen that our evolved psychology defines our capacity for cultural learning. Cultural learning, in turn, defines our intelligence.

For example, consider the gap between Generation 5 & Generation 1 in the savannah hunting package hypothetical. Wouldn’t it be fair to say Generation 5 is “smarter”, in that they use more complex and effective solutions to help them survive?

It’s notable that using the word “smarter” in this way doesn’t imply literal increases in brain size or innate intelligence — although because of gene-culture coevolution, there is evidence for genetic evolution in this direction as well (more on that later).

For now, though, let’s assume that innate intelligence is a fixed property and confine the discussion to how cultural “software” (as opposed to hardware) updates could have improved the Human OS. In other words, how is it possible for cultural information to make us better at thinking and solving problems, and what are some examples of this process at work?

Mental Tools: Intelligence-Enhancing Technology

The first way that cultural evolution makes us smarter is by equipping us with powerful mental tools that allow us to think and problem-solve in much more sophisticated ways than we could without them. 

Consider something as basic as counting. We take for granted our ability to count to an indefinite number, which allows us to assess relative quantities of any size. But this ability is far from universal.

On the one hand, all human societies have number words for “1”, “2”, and “many”. Some societies also have a word for “3”, and in New Guinea, there are examples of cultures that have developed counting systems that top out anywhere in the range from 22- 74.

But the base 10 system that allows us to keep counting no matter how high the quantity seems to have been a recent development. It’s a pre-built solution that we downloaded culturally and use daily without even realizing the enormous benefits because we weren’t born in the many societies (past and present) where it wasn’t widespread. So even though “10”, “20”, and “258,389” are just numbers to us, our understanding of them makes us smarter. 

Such concepts may seem obvious, but they’re not easy to come up with at all. There are plenty of other examples of culturally-evolved mental tools:

• The number zero, for example, was a concept invented maybe twice in all of human history. Even advanced civilizations like the Romans had no symbol for it (although their word nihil expressed a similar concept).

• The wheel was invented relatively late in human history, after agriculture and settled cities, but only in Eurasia. As a means of transport, it was never invented in the Americas, Australia, or Sub-Saharan Africa. 

• Color terms also vary between cultures. Many languages have no color terms, or only two which roughly translate to “light’ and “dark” (Homeric Greek had no word for blue, so the Iliad speaks of the “wine-dark sea”). Some have 5 color terms. English is supposed to have 11, although my artist friends seem to have picked up more and expect me to know what they mean when they say things like “mustard” and “oxblood”. Korean has 14 basic color terms — even more than English.

• Other examples: The Phoenician alphabet (Roman letters, which we use today, are a variant of this), the Julian/Gregorian calendar, clocks, fractions, right vs. left, north-south-east-west, pulleys, levers, shapes, angles, etc.

Concepts like these are innately very difficult to figure out for the first time, and are often the result of lucky errors; more rarely, they arise in “eureka” moments of genuine insight. Once they do arise, they spread quickly as a result of preferential cultural learning.

Equipped with such concepts, individuals become effectively smarter. What’s more, the next useful mental tool is found more easily, since the greater the toolkit, the faster innovation occurs. In other words, the more mental tools of this kind we have, the better we get at designing novel solutions purpose-built to new problems. 

Implication: We Can Culturally Acquire More Intelligence

I think there are a lot of fascinating implications to the fact that culturally-acquired adaptations can make us better problem-solvers.

For one thing, it explains the correlation between book-readers and intelligence. This may be a more directly causal relationship than I used to think — the analogy of weight-lifting for our mental muscles doesn’t do justice to this. Instead, we can think of reading as accelerated cultural learning, where we are retrofitting our minds with an upgraded suite of intelligence-enhancing concepts, mental models, and pre-built solutions — including new vocabulary to refer to those concepts in shorthand.

The books, teachers, classes, lessons, TedTalks, or YouTube videos we remember the best are often remembered precisely because they did this work. I will never forget my sophomore biology teacher Mr. Doyle for teaching me the fundamentals of natural selection as an explanatory model, for example. I’ll never forget going to Athens and hearing a recitation of Plato’s Apology on the Acropolis, because of what this Socratic dialogue taught me about the importance of natural inquiry. I now understand these events to not just have given me new information, but to have actually expanded my consciousness by giving me tools that made me smarter.

5. Population Size & Interconnectedness Accelerate Innovation

One of the big counterintuitive arguments that Henrich makes is that the technological sophistication of any given human population has little to do with the average innate intelligence of a population, and everything to do with its size and interconnectedness. He calls this phenomenon the “collective brain”.

Considering the following thought experiment posed by Henrich:

• Say that any individual trying to figure out an innovation like arrow-fletching will discover how to do it once every 1,000 lifetimes. 

• In a group of 100 people, at least 1 person will come up with the innovation in 10% of the time, so on average it will take 11 generations (275 years) to figure out how to fletch an arrow.

• In a group of 1,000 people, there’s a 63% chance of the discovery in one lifetime, so it will take 1.6 generations (40 years) to figure it out.

• Finally, in a group of 10,000 people, you will almost certainly have fletching within a single generation.

So the size of a population is directly related to the probability of any given innovation occurring. But interconnectedness matters too — if someone in a group of 1,000 figures out how to fletch an arrow, but they never interact with anyone else, then cultural learning can’t do its job and people won’t preferentially learn from the arrow-fletching innovator.

So we can then abstract the collective brain idea as a formula:

 Rate of Innovation ∝ Population Size x Interconnectedness

The implication here is that when there’s an increase in the network size and network density of human populations — from hunter-gatherer tribes to modern social networks — it accelerates the process of cultural evolution. This is especially visible with regard to technological innovation, because it’s possible to compare the toolkits of human populations of varying size and interconnectedness, and therefore toolkit comparisons below can help illustrate the principle. 

So as we’ll see, size and interconnectedness reliably correlate with both a) the ability of a given population to produce technological innovations and b) the spread and retention of those innovations. Innovation, seen in this way, is a direct effect of network growth.

Evidence For The Collective Brain

To show the importance of population size and interconnectedness to innovation, Henrich relates the findings of several studies that point to a strong positive correlation between the two:

The 1-model vs. 5-model treatment

• In one study conducted by Henrich, 100 undergraduates were recruited and divided into two groups, which were then further subdivided into 10 laboratory “generations”. 

• The groups were given the task of recreating an image using a computer program, with each “generation” passing along written instructions of what they’d learned to the next.

• One group received instructions from 5 people of the previous generation, while the other group only received 1 set of instructions (modelling 5x less interconnectedness).

• The results were stark: the group learning from 5 people vastly overperformed the group learning from just 1 person. By generation 10, even the least skilled person in the 5-model group was better than the most skilled person in the 1-model group.

The results of this experiment (above) have two important implications:

1. the spread of successful techniques and innovations via cultural learning relies on a degree of social interconnectedness.

2. variance in individuals’ natural ability didn’t end up mattering much besides the massive improvements caused by greater interconnectedness.

Fishing toolkit study (laboratory)

• In a French laboratory experiment conducted by Maxime Derex, groups ranging in size from 2 to 16 people constructed virtual fishing kits and used them to “fish” in the program. Their catch was measured in points.

• There were 15 rounds of fishing. After each round, the participants could look at other members in their group and copy their techniques, or rely on their own trial-and-error, causal models, and innate intelligence.

• By the end of the experiment, the performance of people in the largest group (16 people) was 50% better than the smallest group (2 people). 

Fishing toolkit study (real-world)

• To measure the effects of population size and interconnectedness on technological innovation in the real world, it’s necessary to be able to compare isolated populations. Comparing islands with hundreds of miles of ocean between them is a good way to do this.

• In a study conducted by Michelle Kline and Rob Boyd, fishing toolkits between islands in Oceania were evaluated for their size and complexity using detailed ethnographies.

• The result showed a clear correlation between population size and sophistication of fishing toolkits.

In addition to these positive correlations, Henrich also brings up examples where sub-populations cut off from a broader network by disasters or terrain changes causes the rate of innovation to not only slow down but actually reverse. 

In other words, if a network’s size is suddenly reduced, or if a population is suddenly cut off from a broader network it used to be a part of, the cultural know-how and technology of that population will slowly decline, as shown in the following examples discussed by Henrich:

• “The Tasmanian Effect”: When Europeans first encountered Tasmanians in the late 1700s, they had the simplest toolkit ever observed: leaky rafts, crude weapons, minimal clothing, and no fishing ability despite the abundance of local fish — a level of technology easily outstripped by the toolkits of paleolithic societies 40,000+ years ago, and consistent with the stone tools of Neanderthals. By contrast, Pama-Nyungan speaking Aborigines of Australia to the north had hundreds of specialized tools, many of which the Tasmanians themselves possessed until they were cut off from the rest of Australia by the Bass Strait, which formed as the seas rose about 12,000 years ago. After that point, the Tasmanians slowly began to lose their tools — there’s evidence, for example, that they once had the ability and tools to fish about 5,000 years ago, but lost it gradually over time.

• The Polar Inuit: Polar Inuits are the northernmost human population that has ever existed, dwelling in northwestern Greenland. In the 1820s, an epidemic killed many of the oldest members of the population. This sudden disappearance of an important reservoir of cultural knowledge led to the Polar Inuit losing its ability to make many tools — including kayaks, which cut them off from other Inuit groups and prevented them from re-learning lost technologies like bows and arrows. Their population began a downward spiral until 1862, when another Inuit group happened across them and the Polar Inuits rapidly relearned their lost technologies — including kayaks. 

• The Torres Islands: When the master canoe builders in the north of the Vanuatu archipelago suddenly died out, the inhabitants of the Torres Islands became isolated. They never regained the use of seaworthy canoes which they must have had at one point to get to the island in the first place, as attested by W.H.R. Rivers in just one of many examples of the Disappearance of Useful Arts.

In all these cases, the groups who lost technologies and know-how by a sudden reduction in their collective brain didn’t stop needing what they lost. There was a clear need, for example, for the Tasmanians to be able to fish — but despite wanting to, their collective brain had atrophied too much for them to overcome.

Henrich points out that this has an important implication: you cannot judge the innate intelligence of a human population by the size of its toolkit, as individual intelligence has a demonstrably negligible effect on the level of technological sophistication of any given society. 

Much more important are the size and interconnectedness of populations that form a society — and, as we’ll see later, “prosocial” norms that promote greater sociality between larger and larger groups are evolutionarily favored for being highly adaptive.

If you want to understand a culture’s capacity for innovation, then, a good place to start is by looking at how capable that culture is at promoting a high degree of interconnectedness amongst a large number of people. 

Implication: Network Structure Matters To Innovation

As someone who has spent the last 3 years researching and helping write essays about network effects, “the collective brain” strikes me as a familiar and related concept. 

Network effects (in economics) are an emergent phenomenon whereby the growth in the size and interconnectedness of a networked product leads to nonlinear value creation for its users. 

Henrich’s concept of the collective brain seems to apply a similar observation in the context of cultural evolution. Just as individual neurons networked together produce the phenomenon of consciousness inexplicable as a mere (linear) sum of its parts, so too individual minds networked together can produce innovation at a (superlinear) pace greater than what you’d expect by simply adding together the intelligence of each individual node in the network.

In his book Scale, Geoffrey West describes a similar phenomenon in relation to urban networks. He provides evidence for the astounding conclusion that the size of cities directly correlates with metrics of innovation like number of patents, startups, etc. Although approaching the subject from the standpoint of network science rather than cultural evolution, his conclusion is strikingly reminiscent of Henrich — at least to me. 

The larger point here is that once you reach a certain group size, the mathematical laws are so powerful that the creativity and intelligence of individuals start to matter much less than 1) how many of them there are and 2) how interconnected they are. 

6. Culture Drives Human Genetic Evolution 

As soon as cultural evolution became cumulative, the speed at which it was able to produce adaptations began to exceed the speed of genetic adaptations. This is Henrich’s “Rubicon moment”. 

After that point, genetic evolution didn’t stop. It is a common misconception that we stopped evolving genetically because of modern technology — implying that we’ve been stuck with basically the same genetic characteristics since the Paleolithic and before.

On the contrary, it’s more likely that if anything, genetic evolution gained speed. Only now, culture has taken the driver's seat. Several lines of selection pressure arose directly because of cultural evolution.

• Brain growth - the survival advantage of being able to acquire and store cultural information became so great that it drove a relentless genetic expansion in our brain size, which only stopped about 200,000 years ago because of the limits of fitting a baby’s head through a birth canal. Our brains comprise ~2% of our body mass but consume about 20% of our energy, as opposed to 8-10% in other primates and 3-5% in other mammals. 

• Muscle loss - to compensate for the heightened energy needs of the (culturally adaptive) brain expansion of our evolutionary lineage, our bodies traded muscle mass for less energy-expensive fat mass and became reliant on toolkits and know-how to acquire food with less time and effort than other primates. As a result, compared to other primates, we are extremely weak. However, cultural adaptations like water containers, tracking, and throwing spears enabled our bodies to adapt for hunting strategies utilizing distance running and accurate throwing. So while even the strongest human would have difficulty wrestling a much smaller chimpanzee, we are much better at running marathons and throwing balls.

To put a fine point on the importance of water containers as a driver of our genetic evolution into distance runners, Henrich points out that water-expensive sweat-based thermoregulation would only be possible after cultural evolution supplied us with water-container technology and the know-how to be able to find water in diverse environments — for example, Kalahari foragers have been observed being able to find underground water sources and using reed straws to suck the water from hollow tree trunks.

Water containers thus allow us to sweat to regulate our heat, which other animals can’t do — meaning we can run them down in the heat as a hunting technique. Thus, cultural adaptations like water containers kicked off over a million years of genetic adaptation towards persistence hunting in our lineage, producing big physiological differences between us and other apes like:

• Springy arches in our feet for shock absorption

• Longer legs and extended spring-like tendons for longer, energy-efficient strides

• A high proportion of slow-twitch leg muscle (up to 80%) for better aerobic capacity

• Reinforced joints, enlarged stabilizing muscles for our spines

• Upper back musculature that allows us to turn our heads, unlike other primates

• A nuchal ligament which secures our brains against shocks

To see culturally-driven genetic evolution at work, we don’t have to look back as far as ancient developments like the expansion of our brain size or adaptations for endurance running. Fairly recent genetic adaptations within the last 10,000 years resulting from new selection pressures driven by the agricultural revolution, like the spread of the gene for alcohol intolerance or lactase persistence, illustrate the point that genetic evolution has not ended and has likely even accelerated.

• The spread of ADH1B (alcohol intolerance):ADH1B is a gene that causes adverse reactions to alcohol consumption, reducing the probability of alcoholism by 2-9x. Its geographical spread can be accounted for by the earliness of the date at which rice agriculture began in a particular area. The earlier it started, the higher the frequency of ADH1B. Because pre-agricultural populations can’t easily create fermented beverages, ADH1B seems to have been a genetic adaptation to the early cultural prevalence of making spirits like rice wine.

• The spread of lactase persistence (the ability to drink milk into adulthood): 68% of adult humans are lactose intolerant. The reason for this is that, as with most mammals, they stop producing lactase, an enzyme that allows you to metabolize lactose, after childhood. But because in certain areas of the world in the last 12,000 years, humans have domesticated milk-producing animals like cows, sheep, camels, and goats, there has been a genetic selection pressure of cultural origin towards the production of lactase into adulthood (lactase persistence). However, crucially, if these same societies also produced cultural adaptations that reduce lactose content in dairy products (cheese, yogurt, etc.) early enough, the selection pressure for lactase persistence disappeared — meaning those populations continued to have a high prevalence of lactose intolerance.

Implication: We Haven’t “Stopped Evolving”. We’re Evolving Faster.

The fact that genetic adaptations like the spread of lactase persistence and alcohol intolerance can only be described in terms of gene-culture coevolution, a process by which our genetic adaptations interact with our cultural ones in a self-propelling, self-reinforcing feedback loop, means that:

1. Culture is likely to be the driving force in our genetic evolution in complex ways we’ve only begun to understand.

2. It is unlikely that human evolution has stopped or slowed down even as cultural evolution has picked up speed — more likely the reverse. 

Henrich is careful to note that more recent cultural accelerations like the industrial revolution are too recent for us yet to judge their effect on the direction of our genetic evolution, but that doesn’t mean it won’t end up having a massive impact over time.

 I often see people basing their arguments for various philosophical or political perspectives on the assumption that “human nature” is both static and universal. But given the importance of culture (as opposed to genetics) to our “nature”, it can hardly be thought of as universal — human beings manifest in very different ways depending on how they’re cultivated, as I’ll discuss later. 

And our genetic evolution almost certainly isn’t static. If cultural evolution, including technological innovation, is accelerating, then it is likely to exert powerful selection pressures. I agree with Henrich that it’s too early to tell exactly what the existence of something like massive apartment buildings or online dating will do to us genetically as a species, but it seems unlikely that “nothing” will end up being the answer. For organisms subject to evolutionary processes, change is the only constant.

7. The Origin of Cultures Was A Major Evolutionary Transition

We’ve seen that cultural evolution can explain a surprising range of human behavior and psychology. How and when did cumulative cultural evolution originally arise? When was Henrich’s “Rubicon Crossing”?

To begin with, it is helpful to return to the analogy Richard Dawkins drew between genes and memes as two types of “replicators” in The Selfish Gene. His argument was as follows:

• The universe is populated by stable things since unstable things don’t last.

• The more complex a form of matter it is, the less stable it tends to be.

• Therefore, in order for complex forms to exist, they have to be self-replicating.

• Forms that self-replicate are subject to evolutionary pressures — they mutate randomly, and the forms that most effectively self-replicate in a given environment will spread.

• Competition for scarce resources means that less effective replicators will die off.

• Genes were the original replicators. Their “substrate” was organic matter.

• Memes — the “atomic unit of cultural selection”, are a “new type of replicator”. Their substrate the mind, and the scarce resource they compete for is attention.

Dawkins defined the meme as an “atomic unit of cultural selection.” He described them as a new type of replicator that constitutes the building blocks of culture in the same way that genes are the building blocks of organisms. 

If this is an accurate framing, then culture must have arisen at the precise moment when memes began to pass from one mind to the next in a self-replicating way — just like how species first arose when the first complex organic molecules, i.e. genes, became self-replicating (although original biological replicators may have originally been RNA-based). 

Similar to the question of when genes (and therefore life) first arose on the planet, it’s tough to pinpoint this hypothetical moment of origin of memes. However, Henrich uses the following lines of evidence to “triangulate” the earliest days of cultural evolution and reconstruct the likely timeline for when human cultural transmission first became cumulative:

1. Studying apes most closely related to our lineage (e.g. chimps) with the nearest common ancestor, which we believe we diverged from 5-10 million years ago. Though capable of a degree of social learning and tool use, no other primates have significant cumulative culture, so it’s safe to say that our common ancestor with them probably also lacked it. Culture therefore likely arose more recently than 5 million years ago. 

2. There is circumstantial evidence, from brain size and an increase in tool use, that members of our evolutionary lineage called Australopiths were capable of significant social learning between 4 and 3.2 million years ago. However, there is little evidence that learning had become cumulative at this stage. 

3. 2.6 million years ago, the first crudely crafted stone tools emerged in the archeological record (the “Oldowan” toolkit) — for example, stones shaped to be choppers, scrapers, hammerstones, and awls to cut the meat off large animals like buffalos and giraffes, breaking bones to get at the marrow, cracking nuts, etc. Contemporary primates are not capable of making tools with similar sophistication. 

4. Early members of our genus, Homo, emerged about 2.4 million years ago. These showed early signs of being affected by gene-culture coevolution, such as the development of smaller jaws and cheek teeth (suggesting the genetic impact of food processing enabled by the Oldowan toolkit). However, it’s conceivable that many of these physical tools could be figured out without the aid of cumulative culture. Henrich speculates that at this point, we were “dancing on the threshold of cumulative cultural evolution”, without clearly having crossed it.

5. By 1.8 million years ago, Homo erectus, a new species in the Homo genus, had emerged. Erectus was characterized by bigger brains and greater dependence on food processing. They may even have had some control of fire, have learned to cook, and more used sophisticated tools than the Oldowan kit — tools which Henrich thinks would take modern humans hundreds of hours to learn to make, even with help. Therefore, this is the point at which humans likely crossed the threshold into cumulative cultural evolution — and memes were born.

Henrich describes the process being slow at first, and then accelerated from 1.6 - 1 million years ago — contrary to the prevailing wisdom that Homo erectus culture was in stasis. There is therefore good reason to think that the birth of memes and cumulative culture predates our own species, and may have even given rise to it. This would certainly make sense, given the degree to which cultural learning defines our nature, intelligence, and physical peculiarities (from our lack of muscle and hair to our abnormally large brains).

Implication: Humanity May Be a Unique Form of Life After All

Taken together, the rise of cultural evolution and the meme — a new type of replicator — seems to have given birth to what Henrich calls “a new kind of animal” (i.e., hominids). Our genus, and our species in particular, is defined by its dependence on the products of cumulative cultural evolution for survival, and in this way is distinct from all other known life.

This conclusion has a strange symmetry with Darwin’s original work on the Origin of Species and the Descent of Man. Whereas these works emphasized (and were very controversial for) the counterintuitive concept that mankind is no different from any other animal in terms of our common descent, the implication of cultural evolution, as interpreted by Henrich, is that humanity is rather unique despite sharing a common genetic ancestry with all life.

Maybe we’re unique after all.

8. Culture Is Evolved, Not Intentionally Designed

We have seen that much of what we do, how we act, and even what we think is culturally “programmed”.

It’s important here to re-emphasize that this is not the same as saying it is “self-programmed”, or “programmed by society” in any conscious or intentional way.

Yes, our preferences, ideas, beliefs, know-how, and even our motivation, down to what causes pleasure and pain at a neurological level, are all culturally programmable. As Henrich puts it, “cultural learning reaches directly into our brains and changes the neurological values we place on things and people” (The Secret of Our Success, pg. 45).

This, for example, is why it’s possible for people to culturally learn to enjoy very spicy food or fine wine, both of which babies find disgusting. 

Henrich even argues that copycat suicides show hints of being a culturally transmissible phenomenon — where the suicides themselves are often influenced by the prestige and self-similarity cues of a celebrity behavioral model. 

A consistent theme throughout the book are examples of little awareness we have of a) the origins of our culturally acquired behaviors, ideas, and values and b) the actual adaptive work being done by them.

 This is the origin, according to Henrich, of superstitions like the practice of interpreting animal bones or the flights of birds to make critical agricultural or military decisions. The reason why such practices are adaptive (that they help us simulate game-theoretically optimal random behavior in crop rotation or in military strategy) are not the conscious reasons why people do them (they think they’re doing the will of the gods). Even more fascinating to consider: if people did know the real reason why such practices were adaptive, they might be less motivated to do them… making them less adaptive!

Implication: Intentionally Reprogramming Culture Is Very Dangerous

The cultural ecosystem that produces this highly flexible “programming” of our minds has several important traits that we must notice. 

First, the elements of that ecosystem were designed by evolutionary processes, not by conscious agents. Second, it consists of multiple codependent ideas that, taken together, form an adaptive package. But isolating what makes that package adaptive is not an easy task.

Adaptivity is however the basic quality that explains why “social constructs” emerged in the first place, and it is important to understand that adaptivity is neither subjective nor arbitrary, even though it’s hard to pinpoint exactly how any given cultural package might be adaptive. Rather, adaptivity is an emergent property that arises from the constraints imposed by reality. Cultures, therefore, carry vast amounts of implicit information on what is and isn’t adaptive

Thus, attempting to redesign our culturally-acquired “social constructs” is like playing an existential game of Jenga. You can’t just remove a piece without destabilizing the entire structure, you can't judge a single cultural trait without having a full understanding of the context in which it lives, and you cannot easily predict which pieces can be removed or changed without harm and which pieces will send the whole edifice crashing down.

Nevertheless, it is the most common thing in the world for those of us who are first-order thinkers to try to design culture. Working under the false impression that someone designed it in the first place, we then try to rectify aspects of it that they don’t like — aspects, for example, that we might believe to be the root of social injustices — without realizing that a) the intended effects of trying to reprogram something as complex as culture rarely materialize, and that b) the unintended and unanticipated consequences of such attempts are frequently so harmful as to outweigh or directly negate the “positive” intent of the change. 

For example, it is now in vogue (again) to intentionally reprogram the use of language along ideological lines. Given that the English language is an evolved, adaptive system — arguably one of the most effective communicational repertoires the human race has ever developed if judged only by superficial metrics like number of speakers or literary production — the conscious attempt to design it will almost certainly fail, or succeed only in making it less usable and adaptive for its users.

Such was the warning of George Orwell, who had an instinctive awareness of this problem. It is also evident in the abject failure of created languages like Esperanto which, although interesting experiments, could not be designed to be more usable, learnable, or adaptive than organic languages, which were designed by cultural evolution to best fit our evolutionary psychology. This is not to say that language is unimprovable — it’s just to say that its evolution must be organic, iterative, and (you guessed it) unintentional.

9. Culture, Not Race, Explains Differences Between People

Recall from earlier that ants are the only land-bound animals can even come close to matching human biomass and, like us, have spread across the entire globe. But they could only do so with the help of a massive number of genetic adaptations — leading to the aforementioned 14,000+ different known species of ants.

Although we live in a similar diversity of environments, human populations vary so little genetically that there is still only one species of human. The concept of human “races'' bears little meaning from a biological perspective. Henrich points out, for example, that the differences in appearance that lead to race categories only capture 7% of the total genetic variation within our species, and genetic variation within traditional racial categories often exceeds genetic variation between “races”.

As for the relationship between superficial “racial” characteristics and human variability: at the end of the day, skin color can tell you almost nothing significant about a population — genetically or otherwise. Here are several examples of why skin color conveys little information mentioned in The Secret of Our Success:

• It doesn’t correlate with genetic categories: The prevalence of important genes, like the ability to produce lactase into adulthood (lactase persistence) varies considerably within racial categories. For example, in some parts of Sub-Saharan Africa, this gene is extremely rare (less than 10%) while in other parts it’s quite common (over 70%).

• It doesn’t correlate with region of origin: Paler skin and lighter eye color variations are the results of selection pressure from Vitamin D deficiencies for populations living at high latitudes. However, this selection pressure only applied for people who developed high-latitude agriculture early on, because they didn’t culturally evolve supplemental Vitamin D sources in their diet as did high-latitude hunter-gatherers like the Inuit, who consequently did not experience selection pressure for lighter skin. The emergence of very light skin among cereal-eating Baltic peoples was also extremely recent since high-latitude Baltic agriculture didn’t develop until ~6,000 years ago. Many other high-latitude populations never developed a similar genetic adaptation to Vitamin D deficiency, because they already had cultural (dietary) adaptations that did the same job. 

• Skin color doesn’t even describe common ancestry. Populations in New Guinea and Africa are at the opposite ends of our species’ genetic family tree, but have similar skin pigmentation and are classed as part of the same superficial skin color category. From the standpoint of actual genetic variation, these populations are much more different from each other than either one would be from Europeans, North Africans, or Central Asians of lighter complexions and superficially different features. 

So race is essentially a largely arbitrary division that conveys little real information about the members of its respective categories, genetic or otherwise. Yet even if racial categories were much more “accurate” in grouping people by their actual genetics, given the relative genetic homogeneity of our species, they would not be able to explain the ability of humans to adapt and thrive in a myriad of environments.

Implication: Cultural Differences Can Help Explain Societal Outcomes

From the Pala-Nguyen tribes of the Australian outback to the Inuit tribes in the Arctic Circle, once you realize that cultural adaptations — not genetic ones — are what allow our species to survive and thrive, it is impossible to explain the vast variability in human societies by any other means.

Nor are simple material conditions like environments and relative abundance of resources sufficient to explain the differences between societies. For one thing, we have seen how, lacking cultural adaptations, even resource-abundant environments like King William’s Island or the Gulf Coast can reduce people to starvation and cannibalism.

For another thing, culture evolution is self-propelling. That is to say, once a culture evolves in a certain direction (because it is adaptive in the context of the local environment), it tends to retain and build on those early patterns. 

There are many ways to structure a society, and there are innumerable possible cultural attitudes, beliefs, preferences, tastes, and rituals that cultural evolution could develop in response to local selection pressures. Thus, you can have two populations in similar environments with very different cultural packages including toolkits, languages, social organizations, beliefs, ideas, customs, and tastes. 

Lacking bow-and-arrow or compressed air technologies, for example, Australian populations developed other ways to hunt (boomerangs) whereas populations in similar environments elsewhere relied heavily on bows for hunting. The Hadza, an African hunter-gatherer group, rely on a heavily starchy diet — so much so that it caused genetic selection pressure to increase their number of copies of the starch-digesting AMY1 gene relative to other regionally proximate African hunter-gatherer and herder populations. There are many ways to culturally adapt to the same environment or material conditions. 

In some cases, similar geographies do produce convergent practices — as shown by the tendency of tropical populations to evolve a taste for spicy food in their local cuisines (spicy food has antimicrobial properties, and microbes are relatively more abundant in the tropics, although cultures with spicy cuisine don’t consciously spice their food for this reason). But convergent cultural adaptations are by no means universal, meaning that similar material conditions do not predictably produce the same patterns of behavior. 

Since there are many paths to adaptivity, then, the only reliable way to understand, analyze, and explain differences between humans is to first look at their cultural traits. Culture is as formative of societal conditions as it is of individual intelligence. In my view, cultural evolution is the engine of history. It is the primary force that underlies the values, preference, social norms, technology, and ultimately military, economic, and political organization of a society. Culture is the system at the heart of all other human systems. 

In other words, relying on racial categories is not a great way to understand human variability — and I would argue that excessive focus on these categories has obscured more than it has revealed. Even if racial categories were to be more accurately divided along lines of ancestry and genetic variation, the vast majority of human variability would still be better accounted for by culture — as would genetic selection pressures themselves. 

Culture, not race, therefore explains differences between people.

10. Social Norms That Promote Network Expansion Are Adaptive

According to Henrich, explanations of the human ability to live in large groups and cooperate are not sufficiently understood by relying on the explanation that we have slowly evolved innate psychological traits like reciprocal altruism. Innate proclivities don’t explain the strength, stability, and variability of social norms. 

Prosociality — our ability to cooperate in larger and larger societies — is better explained by the evolution of social norms as cultural adaptations. As with other cultural adaptations like bow-and-arrow technology or seal-hunting know-how, social norms are the result of unintentional processes of trial and error whereby societies slowly “learn” to organize themselves and regulate the behavior of their members. 

What’s critically important is that social norms are group traits, not individual traits, which means that group competition drives the selection pressure for groups to innovate prosocial norms, or adopt them from more successful other groups via preferential cultural learning. 

At the heart of Henrich’s argument here is the collective brain hypothesis. Recall that the size and interconnectedness of a population directly contributes to the pace of cultural evolution — the innovation and spread of new adaptations. 

As such, we can expect that societies that evolve social norms that promote greater network size and greater sociality between its members will be more successful in the long term than other societies, leading their social norms, ritualistic practices, and other sociocultural traits to spread.

Take marriage norms, for example. Marriage norms create greater prosociality by expanding kinship networks beyond direct genetic relatives. Kinship networks appear to be the building blocks of social organizations for small-scale societies. But Henrich points out that even in the smallest-scale societies observed, many people aren’t directly related. According to him, if you trace blood links back as far as five generations, 75% or more of people are not directly related even in the smallest-scale tribal societies. 

To expand the prosocial kinship psychology to fit larger groups, cultural evolution favored social norms that were able to promote similar prosocial behavior among those who don’t have any direct shared genetic ties. Henrich argues that in-laws dramatically expand kinship networks: with in-laws, kinship ties go from comprising, at most, a quarter of ties in a tribal society to up to two-thirds. Thus, by promoting prosocial behavior among larger groups, marriage norms are adaptive. Over half the links between adults in any tribal network are created by in-laws. 

This prosocial effect explains the existence of marriage norms across many different cultures, although the exact nature of marriage norms in different societies vary widely. Social norms, in this way, bridge the gap between our innate psychological drive for prosociality with direct genetically-defined kin, and our culturally-acquired drive to engage in prosocial behavior with culturally-defined kinship — i.e., ethnic groups. Social norms that build on our innate psychology to expand networks of cooperation, in other words, are selected for by group competition. 

Social norms (i.e. systems of morality) that lead to greater prosociality within a society are therefore more likely to spread and persist. But social norms can serve other functions as well. For example, meat taboos promote cooperative hunting and meat sharing, which paleoanthropologists believe to be important elements in human evolution. 

Other food taboos, like those described by Henrich that ban pregnant women on Yasawa Island in Fiji from consuming toxic marine species leading to conditions like ciguatera poisoning, confer direct survival advantages and are enforced as social norms. Crucially, when Yasawans are asked why their food taboos exist, they either said they didn’t know or made up a reason (like “it will cause the babies to be born with rough skin or smelly joints”). 

This is another key consequence of the unintentional aspect of cultural evolution — many people who benefit from culturally adaptive social norms practice them despite not knowing why they’re adaptive, or even adopting supernatural or false explanations for their benefits. Intentional, conscious understanding is not required for cultural adaptations to work — as a cultural learning species, causal explanations (understanding “why”) is sometimes beside the point. 

Implication: Morality Is An Emergent Evolutionary Adaptation

The basic idea of seeing social norms as cultural adaptations can be summarized as follows:

1. Social norms are like rules for how societies work, but they aren’t written consciously.

2. Instead, they’re the product of trial and error, where successful social norms and cultural values — much like technological innovations, food processing techniques, know-how, etc. — prove better for the survival of their societies than alternative systems of norms.

3. Social norms are selected, in part, for their ability to power greater prosociality, i.e. their ability to enable groups to become larger and more interconnected.

This view of the evolutionary adaptiveness of social norms, to me, is perhaps one of the most impactful concepts that a cultural-evolutionary lens gives us. I think it powerfully challenges the current status quo beliefs about morality in Western thought. Specifically, it has the following implications:

• It seems to me that most formal Western ethical systems, whether religious or philosophical in origin, do not sufficiently account for the emergent nature and adaptive function of social norms.

• Further, formal philosophical ethics do not sufficiently appreciate or account for the complexity and density of information contained implicitly in evolved social norms. They (often naively) think they can do better and design better systems from the top down using first-principles thinking and first-order reasoning.

• As with other cultural adaptations, there are multiple paths evolution can find to reach the same goal (greater adaptivity). But this does not imply that all paths are equally effective, that social norms are arbitrarily chosen (moral relativism) or chosen by powerful groups to oppress others (postmodernism). Quite the opposite.

• Social “constructs” like social norms are not arbitrary; they are better viewed as social technologies. While it’s true to say that we built them, it is necessary to appreciate that they were acquired at massive cost over long periods of trial and error. We cannot cavalierly cast them aside without risking massive harm to our societies. To improve, we must take an iterative, careful, and most of all an intellectually humble approach. 

• As Jonathan Haidt argues in the Righteous Mind, our evolved moral psychology is foundational to how we see the world and this would not be the case if social norms didn’t have profound implications for our ability to survive and reproduce.

Maladaptive Memes: My Biggest Criticism of Henrich’s Perspective

Although I agree with Henrich’s implication that social norms are selected by natural selection for their adaptivity, the really important question is — adaptive from whose perspective? 

Because social norms, as well as other ideas and beliefs, can spread between people and gain new converts (unlike genes), there are two separate ways in which they can spread among and between groups. One is by providing utility for the people who hold the social norm and thereby spreading to others by success-biased cultural learning. The other way is by evolving properties that make the social norm meme more viral, resistant to change, and better at keeping adherents.

Often, such social-norm memes arise originally because of their utility-to-the-host, but Henrich acknowledges very briefly that “cultural evolution can produce sticky social norms that are bad for everyone”, which stay around because they have qualities that provide utility-to-the-meme, even at the host’s expense. Henrich provides examples of this like female genital mutilation and the practice of consuming the brains of dead relatives, which can spread deadly prion diseases.

My biggest criticism of The Secret of Our Success is that it doesn’t sufficiently account for the importance of such maladaptive memes, nor sufficiently explain how and why they arise. There are a lot of insights that we can gain from applying a cultural evolutionary perspective to the subject of cultural pathogens, superviral memes, and misalignment of survival interests between genes, organisms, and memes — as I will describe in later essays (subscribe below if you’re interested in reading more about those topics). 

What’s Next?

In this essay, I’ve tried to recap the main ideas of The Secret of Our Success in a digestible way, place them in a larger context of related ideas, and explain why I think they matter by exploring some of their implications.

I hope you found it valuable. For me, writing this in-depth review has been an immensely rewarding project, insofar as it reinforced my understanding of Henrich’s cultural evolutionary perspective and forced me to clarify my own thoughts surrounding it. It was a learning project that has more than repaid the effort already. 

But I’m not stopping there. There is much more to cover on this topic. In particular, I think we can use the cultural evolutionary perspective to address questions like:

• How is it possible for ideas (memes) to spread independently to the truth or usefulness of those who hold them?

• Why don’t people change their minds as soon as “evidence” or “logic” indicates that they’re wrong?

• What explains the persistence of maladaptive (i.e.) harmful ideas and social norms?

• What patterns can we deduce from widespread cultures, worldviews, religions, and ideologies that explain their prevalence?

• How do particular cultural adaptations explain the differential success of societies historically? 

• How can we better inoculate ourselves and our societies to maladaptive or otherwise harmful cultural pathogens? 

These are questions I believe to be of profound importance to the times we live in, where cultural evolution has accelerated by orders of magnitude by the advent of the largest and most interconnected network in human history — the internet. 

If you’ve found this essay valuable and are interesting in reading my forthcoming essays on the topics above, subscribe below and <share> 

Acknowledgements & Other Readings

I want to be clear that not all of the ideas discussed in this essay are unique or original to Dr. Henrich’s book, which is one of many recent works on the topic of cultural evolution.

As he acknowledges, much of the groundwork for his approach was laid out in the 1985 book Culture and the Evolutionary Processby Peter Richerson and Robert Boyd. Since then, people from diverse fields have converged on and contributed to the cultural evolutionary approach, including (among others): 

• Psychologist Mike Tomasello in his 1999 book The Cultural Origins of Human Cognition

• Technologist & Microsoft Word creator Richard Brodie in his 2004 book Virus of the Mind

• The philosopher Daniel Dennett in his 2017 book From Bacteria to Bach and Back

• Songbird scholar & evolutionary biologist Nicole Creanza on How Culture Evolves

However, I chose The Secret of Our Success as the subject for this essay because it seems to present the ideas of cultural evolution in their most comprehensive and compelling form. I think it persuasively shows not only that human cultural evolution exists, but that it is the force most responsible for our success, our intelligence, and indeed, our very nature.

In my view, The Secret of Our Success thus represents the denouement of Dawkins’ original argument that ideas evolve, literally: that they mutate more or less randomly, are preferentially passed from one mind to another and from generation to the next, and compete for the limited resource of human attention and memory.

Although cultural evolution is distinct from memetics in important ways, they share in common the approach of applying Darwinian ideas to the cultural realm — an approach that I believe will soon change how we understand ourselves and our societies. The mission of Living Ideas is to advance this approach. 

In addition to paying homage to my intellectual influences, I want to acknowledge and thank the editors of this piece, Vishal Maini, Tarun Wadhwa, Étienne Fortier-Dubois, and Caleb Ontiveros. Without their long hours of patient editing and thoughtful feedback, this essay would not exist. I’d also like to thank Michele Coscia for his generous help recommending literature, including The Secret of Our Success, when I began research on memetics and cultural evolution two years ago.