Completing Darwin’s Unfinished Symphony: A TVOL conversation with Kevin Laland

by David Sloan Wilson & Kevin N Laland

Completing Darwin’s Unfinished Symphony: A TVOL conversation with Kevin Laland thumbnail

These are exciting times for the study of cultural evolution, with important books appearing regularly. One of these is Darwin’s Unfinished Symphony: How Culture Made the Human Mind, by Kevin N Laland, which won the British Psychological Society’s prize for the best academic book of 2017. Kevin has been featured several times on TVOL as a leading proponent of the Extended Evolutionary Synthesis (for example, here and here). He is a polymath who studies so many subjects that it’s easy to forget the many contributions he has made to the study of cultural evolution—in nonhuman species and computer simulation models in addition to our own species—which are ably summarized in his book. Our interview covers Kevin’s work against the background of the other exciting developments that are taking place.

 

DSW: Greetings, Kevin, and welcome once again to TVOL. Congratulations on your book and award from the BPS.

 

KNL: Many thanks David. It’s always a pleasure.

 

DSW: Congratulations on your book, which I am eager to discuss not only by itself but in relation to other exciting developments in the study of cultural evolution, such as Joseph Henrich’s The Secret of Our Success: How Culture is Driving Human Evolution, Domesticating Our Species, and Making Us Smarter. A lot of overlap might be expected on the basis of the titles, but in fact the two books are largely complementary. How would you describe their differences and what does their complementarity, rather than overlap, signify about the current study of cultural evolution?

 

KNL: Joe’s book is terrific. I admire Joe Henrich a lot, so when I learned a while back that he was writing a book on broadly the same topic this caused me a certain amount of angst. His book was scheduled to come out first. Would there be a market for two books on this topic? Would I have anything new to add on these issues? To maximize their distinctiveness, I resolved not to read The Secret of our Success until I had completed Darwin’s Unfinished Symphony. I wanted to tell my own story about human cognitive evolution, based on a personal reading of the literature, and drawing on my research group’s experiences of investigating these topics over 30 years. Naturally, when I had finished, The Secret of our Success was right at the top of my reading list, and I was very pleased (not to mention relieved) to see, as you have noted, how complementary, yet distinct, the two are.

 

Joe comes at these issues from a background in biological anthropology, psychology and economics, whilst my approach is rooted in animal behavior and evolution. Yes, we have both attacked the issues using mathematical models of cultural evolution and gene-culture coevolution, but in the main our methodologies are quite different. I rely much more on insights gleaned from comparative studies of animals, as well as statistical analyses of primate brain evolution, whereas Joe’s thinking his been heavily influenced by his ethnographic fieldwork in small-scale human societies. Perhaps Joe’s book tells a richer story of the last 10,000 years, whereas mine concentrates on earlier stages of our evolutionary history. However, while we come from different disciplines, deploy different methodologies, and focus on different timespans, we reach strikingly similar conclusions. In important respects, humans are cognitively distinct from other animals. Our uniqueness can be understood as arising through a process of gene-culture coevolution. The truly distinctive features of humanity – our language, our intelligence, our cooperation – are adaptive responses to our ancestors’ cultural activities. And culture is what makes us smart, with our species’ ecological and demographic success arising primarily from a remarkable ability to pool and build upon learned knowledge. Of course, it is tempting to take the fact that we have independently come to the same position as convergent evidence for our arguments.

 

DSW: One of your contributions has been to study the social transmission of behaviors in nonhuman species that are very distantly related to humans, such as stickleback fish. Why are these studies relevant to human cultural evolution?

 

KNL: Fishes are actually a terrific model system for investigating social learning because of the practical advantages they offer. It is no coincidence that fishes are so widely used by animal behavior and evolution researchers. The relevant issue here is that both traditions and the diffusion of innovations are group-level phenomena, and if they are to be studied reliably, scientists require not just replicate individuals, but replicate populations of individuals. Leaving aside some nontrivial ethical considerations, it would be financially ruinous for a researcher to establish large numbers of experimental populations of, say, chimpanzees or Japanese macaques, but it is straightforward and cheap to set up multiple populations of fishes in the lab. Of course, those practical advantages would count for nothing if fishes were poor learners, but there are now many experimental studies showing that is not the case. Fish behavior is constantly and flexibly adjusted to exploit information and resources in the environment, including information provided by other fish.

 

Let me give you an example. For some 15 years, the members of my lab and I have been investigating a fascinating difference in the social learning of two closely related species of fish – threespine and ninespine sticklebacks. ‘Public-information use’ is the capability of an animal to assess the quality of a resource, such as the richness of a food patch, through observation; that is, by monitoring the success and failure of other feeding fish. We discovered that ninespines exhibit this form of learning, but threespines don’t. This intrigued us, because in many respects the fishes lead very similar lifestyles, and even forage together in mixed-species shoals. Through a long series of experiments, which I describe in the book, we have made sense of this species difference as an adaptive specialization in social learning tied to differences in the fishes’ morphology. Threespines have large spines and heavy-duty plating on their bodies – robust physical defences that leave them, relative to the ninespines, comparatively invulnerable to predation. (Remarkably, threespines can even survive being eaten! Their spines get stuck in the throats of predators, which cough them up and they escape). This allows the threespines to explore their environment and sampling patches at relatively low cost, leaving them with little to gain from exploiting public information. Conversely, ninespines, which have weaker defenses, typically respond to the threat of predation by hiding in refuge, such as weeds. Seemingly, natural selection has fashioned their ability to use that time in refuge profitably by tracking the success and failures of other fishes at food patches, including the threespines, and then swimming directly to the richest patch when the coast is clear.

 

These experiments were an early clue that animals exploit information provided by others in a decidedly strategic manner. Our ninespines did not copy at every available opportunity, but instead were highly selective – for instance, utilizing social information when they had no relevant prior experience to rely on, or when the knowledge gained by that experience was unreliable, or out of date. We also found that ninespines were able to combine personal experience and social information effectively, to maximize foraging returns and minimize risk.

 

Once we had been alerted to the strategic nature of stickleback copying, we noticed that other animals copied very selectively too. My research group has studied the behavior of a lot of different animals, and in every species that we have worked with, without exception, the social learning observed is highly strategic and rule governed. Social learning researchers around the world have reached the same conclusion. Interestingly, many of the rules that predict the patterns of social learning of sticklebacks are also observed in other animals, including humans. In fact, the rule that accounts for the differences between our sticklebacks, which is sometimes labelled ‘copy when asocial learning is costly’, was first proposed by Boyd and Richerson (1985), very much with humans in mind. This means that the comparative work helps to set human social learning in context.

 

DSW: Right! The similarities are based on the functional requirements of learning in the context of particular ecologies, not phylogenetic similarity. This is an important lesson to learn because so much thinking about human cognition and culture tacitly assumes that humans are ”smartest”, chimps and bonobos are “next smartest” and so on. You were also among the first to build computer simulation models of cultural evolution. What were the specific questions that you were trying to address with these models?

 

KNL: Once we had learned that animal social learning was highly strategic we were naturally led to ask ‘What is the best strategy that an animal can deploy?’ However, this proved to be a tough question to answer. We did explore the issues using population genetic and game theory approaches, which allowed us to compare the relative merits of a small number of well-studied social learning strategies, such as conformity, or model-based copying. However, I was very conscious of the fact that there was an almost infinite set of hypothetically possible learning strategies that an animal could deploy. In principle, superior social learning strategies that nobody has yet considered could be implemented in the real world.

 

This problem troubled me for a long time. What we needed to make headway was a means to compare the relative merits of a very large number of social learning strategies. It struck me one day that the challenge confronting researchers in the field of social learning was similar to that faced by researchers such as Robert Axelrod in the 1970s investigating the evolution of cooperation. As many readers will know, Axelrod made great progress by organizing tournaments, based around the Prisoner’s Dilemma game, which led to the emergence of TIT-FOR-TAT. I wondered whether we might be able to provide a similar impetus to the field of cultural evolution by organizing a tournament to work out the best way to learn. I got a grant from the EU, which allowed my collaborators (notably Luke Rendell) and I to organize a competition based on a game of our own devising – the Social Learning Strategies Tournament. It was free to enter, open to everyone, and with a €10,000 prize for the winner. Each entrant proposed a learning strategy, and we pitted all the entries against each other in computer simulations to see which worked best.

 

The exercise was highly instructive. We discovered that the most successful strategies were those with the greatest efficiency. Such entries spent the minimal time learning, thereby maximizing the time they spent caching in on what they had learned. Social learning was found to be far more effective than learning through trial and error. We observed a strong positive correlation between the proportion of a strategy’s moves that involved copying, as opposed to asocial learning, and how well that strategy performed in the tournament. The most successful strategies did not play learning moves often, but almost always copied when they did. Yet, while among the top-performing strategies the more the strategy learned socially the better it did, among the poorer performing strategies we actually witnessed the reverse relationship. That told us something very interesting—social learning is not universally beneficial. Rather, copying only pays if it is done efficiently. This, combined with the observation that the winning strategy was the one that learned with greatest efficiency, implies that natural selection should favor accurate, high-fidelity and efficient forms of social learning. This proved to be a seminal insight, around which my book is structured.

 

DSW: All good. Now I’d like to focus on “Rogers’ Paradox”, named after the theoretical biologist Alan Rogers, who noted that social learning can be beneficial for individuals without necessarily being good for the group. Please describe this paradox in more detail and how it has been (partially) resolved by yourself and others.

 

KNL: Sure. Several theoretical analyses using evolutionary models have concluded that some mixture of social and asocial learning is usually necessary for animals to thrive in a variable environment. An intuitive way to see this is by analogy. Wherever some animals are able to find or produce food, other animals will typically come along and try to steal it from them, which leads to a frequency dependent balance of producers and scroungers. The same applies to learning. Some individuals solve novel tasks through trial and error, which makes them knowledge producers; such asocial learning is thought to be accurate but costly. Conversely, social learning can be thought of as information scrounging; individuals obtain information cheaply from others, but are vulnerable to acquiring outdated, inaccurate or irrelevant information. Consequently, theoretical studies predict a mix of social and asocial learning at equilibrium, where the fitness of the two types of learning is equal.

 

This finding – first pointed out by anthropologist Alan Rogers – is called “Rogers’ paradox” because it ostensibly conflicts with the commonly held assertion that culture enhances biological fitness. We know that the spread of technological innovations has repeatedly led to increases in human population size, which implies more individuals survive and reproduce. For instance, with the agricultural and industrial revolutions, birth rates and life expectancy increased. Against this backdrop, Rogers’ conclusions appear paradoxical, as they seem to challenge the observation that social learning explains our species’ success. Of course, when theory and data don’t coincide it can be highly informative. Rogers’ model had assumed that social learners copied at random. He demonstrates that unselective copying does not increase absolute fitness over asocial learning, which implies that, if social learning does truly underlie the human success story, our copying cannot be indiscriminate.

 

The tournament helped us to understand how Rogers’ paradox could be resolved. The winner, called DISCOUNTMACHINE, relied heavily on social learning, but had other neat features. To determine to what extent DISCOUNTMACHINE’s success could be attributed to its copying we produced a mutant version that was identical in every respect but that learned asocially. We then played off the two versions of DISCOUNTMACHINE across a range of simulation conditions as a means of exploring the relative merits of the two forms of learning, with surprising findings. Rather than the frequency dependent balance of social and asocial learning that we had expected, we found that copying beat asocial learning over virtually all plausible conditions. Unlike the analytical theory, in the tournament we were finding that social learning had higher fitness than asocial learning, which matched the empirical data.

 

What was different about the two modeling approaches? In Rogers’ analysis, social learners had been modeled as inflexible agents that continued to perform the same behavior even when the environment changed, which allowed pure asocial learners to be maintained in the population. Conversely, agents in our tournament possessed a repertoire of behaviors that they exploited flexibly. Following environmental change, successful strategies switch to the behavior in their repertoire with the next highest payoff, allowing both them and other copying agents to acquire a behavior with a reasonable return. Here, social learners were not stuck in a frequency-dependent relationship with asocial learners. Provided there was a small amount of copy error, copying generated enough behavioral diversity to allow social learners to track environmental change effectively. The tournament showed how we could resolve the apparent conflict between Rogers’ findings and our species’ demographic success. Simple, poorly implemented, and inflexible social learning does not increment biological fitness, but smart, sophisticated, and flexible social learning does.

 

DSW: Very well stated and more details are provided in your book. Rogers’ Paradox treats asocial learning as a kind of altruism and social learning as a kind of free-riding, which brings us into the realm of Multilevel Selection (MLS) theory. The inclusion of a large number of conditional strategies in your tournament doesn’t alter the fact that the successful strategies might evolve in part by within-group selection, in which case they would not be optimal from the standpoint of the whole group. When you say “we pitted all the entries against each other in computer simulations to see which worked best”, how exactly did you do this? Was it a single well-mixed population? Was it subdivided into local groups? As you know, the relative strength of within- and between-group selection bears critically upon these details.

 

KNL: That’s right. These are good questions. In our original social learning strategies tournament there was only one group, and hence no opportunity for between-group selection. We analyzed the entries in two phases, with the best performers in ‘pair-wise’ contests then being entered into a ‘melee’ where the top 10 performers competed together in a single, well-mixed population. You are absolutely right to draw attention to the fact that the strategy that wins out in such circumstances need not be optimal from the perspective of the whole group. In fact, to make sense of our findings we carried out a series of simulations, including establishing single-strategy populations, and we found an almost perfect inverse correlation between how well a top-10 strategy did on its own and how it performed in the melee. The most successful strategies were highly exploitative, drawing on the useful knowledge produced by others, but contributing little new knowledge themselves, which meant they typically did poorly on their own. The outcome might well have been different had there been multiple groups with competition between them. We have run a second tournament with multiple groups and migration between them, but groups and group sizes were fixed, so the results do not bear on this issue. However, I don’t think primary results of the tournament would change – namely, that it pays to copy, and that selection should favor accurate, efficient copying.

 

DSW: I look forward to a next generation of models that are more explicitly multilevel. Social learning that is a product of group selection will no doubt favor accurate, efficient copying, but will differ in many other respects from social learning that is parasitic in nature. This raises a larger issue with respect to your book as a whole. Even though I know you are comfortable with MLS theory and the concept of Major Evolutionary Transitions, these don’t seem to feature very strongly in your book. At the same time, you do rely upon kin selection in ways that strike me as potentially problematic. To begin, please describe the importance of kin selection, as you see it, for topics such as the evolution of teaching and language.

 

KNL: Perhaps I should premise my answer to this question by explaining how I think about these frameworks. My view is that for some problems kin selection is a more useful and intuitive way to think, whilst for other problems MLS is better. I personally use both perspectives, and I wouldn’t say that I favor either. In the book I present kin selection arguments for the evolution of teaching and language, and MLS accounts of the origin of large-scale human cooperation and complex societies.

 

At the heart of my book is a “cultural drive” mechanism, whereby selection for accurate, efficient information transmission shaped the evolution of the primate brain and intelligence. We have carried out a lot of comparative phylogenetic analyses to explore these ideas, and the data support this cultural drive suggestion. For instance, there are strong associations between reliance on social learning, innovativeness, and brain size in primates, and social learning also co-varies with rates of tool use, dietary complexity, and performance in laboratory tests of learning and cognition. There are other drivers (diet, sociality) but it would seem that selection for cultural intelligence was an important cause of brain evolution in the Great apes and perhaps one or two other primate groups. However, that raises a conundrum: if selection inexorably favors efficient social information transmission, then why is teaching – which could be defined as behavior that functions to enhance the efficiency of information transmission – rare in nature?

 

As you know, countless animals acquire skills and knowledge from others, but the “transmitter” of the information generally does not actively facilitate learning in the “receiver”. Indeed, for years it was thought that humans alone actively teach. However, recently, there have been a small number of reports of teaching in animals, but not amongst the species we had anticipated – ants, bees, meerkats, rather than chimpanzees, orangutans and dolphins. We wanted to understand why teaching was not more widespread in animals when it seemed so beneficial, and needed to make sense of the odd taxonomic distribution for teaching. So we devised a mathematical model in which the evolutionary fitness of individuals hung critically on whether or not they possessed a valuable learned skill, and explored the circumstances under which individuals carrying an initially rare genetic mutation that conferred an ability to teach the skill would have a fitness advantage over nonteachers.

 

Kin selection comes in because, as one might anticipate, we found the more closely related the tutor and the pupil were, the more likely teaching would be favored. Teaching evolved where its costs were out-weighed by the inclusive fitness benefits that resulted from the tutor’s relatives being more likely than other individuals to acquire the valuable information. This implies that teaching initially appears amongst close relatives. However, the analysis reveal tough conditions that must be met for teaching to evolve, as neither easy nor difficult to learn skills supported teaching. The teaching of easy to learn skills is rarely beneficial, since it is not adaptive to engage in what is by definition a costly means of information donation to ensure that your relatives acquire skills that they are likely to pick up anyway. This helps to explain why teaching is seemingly absent in “clever” animals: such species are typically good at imitation and trial-and-error learning, which makes teaching less likely to be economical. Yet difficult to learn skills also do not favor teaching since, being difficult to learn, few individuals acquire them, and thus the knowledge is not available to tutors to pass on to pupils. These criteria explain why teaching is rare in nature. What was particularly neat about this analysis was that it also explained how our ancestors were able to get past these constraints, by showing that teaching is more likely to evolve in a population with cumulative culture. The model implies that teaching and cumulative culture co-evolved in our ancestors, creating for the first time in the history of life on earth a species that taught their relatives across a broad range of contexts. This is the context in which I think language first evolved.

 

At the same time, I endorse the view (championed by yourself, Boyd, Richerson, Henrich and others) that group selection works at the cultural level through the selection of cultural traits, and in my book apply this argument to the spread of agriculture. There is ample evidence throughout history that those human societies that possessed more effective traditions, norms and institutions fared better in competition with other groups. Here I find kin selection explanations neither compelling (as there is often low relatedness within groups) nor intuitive (as one would have to twist the concept of relatedness beyond its biological meaning to make a credible model). I envisage that natural selection will have operated on groups of individuals because many, perhaps most, of the fitness benefits associated with agriculture derive from group-level activities. A lone farmer scratching a living solely through his own efforts, would not typically produce any more food or raise any more offspring than a hunter-gatherer. Only when practiced by groups of people engaged in shared enterprises to produce public goods that benefit the collective does agriculture generally start to become highly productive. Several hundred people are typically required to construct a decent irrigation system, corral or fish weirs. Burning the land, sowing the seeds, harvesting the crops have all traditionally been activities in which entire communities engaged. Such activities are infeasible for an individual farmer, but a group of agriculturalists working together could yield substantial dividends and those that cooperate in this manner generally outcompeted those that did not. The end result was not only the propagation of agricultural practices, but also of domesticated animals, city states, gods and countless other entities.

 

DSW: Actually, I just gave a talk on this topic at Harvard, hosted by Joe, based on this TVOL essay. Saying much more is beyond the scope of this interview, but when you model teaching as a relatively costly form of altruism, then a kin selection model will tell you that should be restricted to interactions among close relatives. However, a MLS model will identify other conditions whereby relatively strongly altruistic traits can evolve among non-relatives. This is a pretty big deal, because it means that genetic relatedness did not necessarily have to be very high in proto-human groups for teaching to have evolved. Moving on, two of the most distinctive human capacities are language and the arts. Please provide a summary from your book of how these evolved from species that lack these capacities or possess them only in extremely rudimentary forms.

 

KNL: In my view, the origins of language reside not in ancient forms of communication but, rather, in social learning and teaching. We all know that animals possess rich forms of communication, but most experts would concur that these do not exhibit the complex syntax and flexibility of human language. What I believe happened is that the cultural drive mechanism, described previously, favored rich cultural repertoires amongst our hominin ancestors. As a result, 1-2 million years ago, at the dawn of cumulative culture, our ancestors will have had to learn a large number of food types, foraging skills, hunting skills, scavenging methods, food preparation and processing methods, tool manufacturing techniques, fire making and control procedures, medicative treatments, gestures, and so forth. Many of these are behaviors that involve complex multi-step procedures that are difficult to learn through trial-and-error, and are the kind of skills that our theoretical analysis suggests should be taught. Given that an animal is teaching, any adaptation that reduces the costs of instruction without diminishing effectiveness, or enhances effectiveness without increasing costs, ought to be favored by selection. Language is such a character.

 

If you think about it, language is a really cheap way to teach. Telling someone where to find a food patch is far easier than taking them there. A simple “yes” or “no,” or “this way, not that way,” will allow a tutor to provide helpful guidance for a pupil acquiring a new skill at very low cost. Language is also an unusually accurate way to teach. There is a precision to information transfer through language that is difficult to achieve through other means. Simple utterances convey messages like “pay attention,” “dig here,” “like this,” “faster,” provide invaluable clues, and help learners to focus on what actions need to be imitated or where new skills need to be applied. Once an individual is committed to teaching, language is the most efficient means to do so.

 

We alone possess language because our ancestors alone created a cultural world sufficiently rich, rapidly changing, and replete with teaching opportunities, to require a learned means of communication to track it. In my book I list a series of criteria that a good theory of the origins of language ought to meet, and the hypothesis that language originally evolved to teach close relatives is the only account that meets all the criteria. It explains the honesty, cooperativeness, uniqueness, and “symbol grounding” of language, as well as how it got started, its power of generalization, and why it is learned.

 

While the initial selection for language was probably to aid the teaching of young by parents or siblings, early language subsequently spread to teaching more distant relatives, where the direct benefits of ensuring that relatives possess relevant skills compensated for the reduction in the degree of relatedness. Complex, coordinated actions are often difficult to carry out without a means to teach, or tell, individuals what their specific roles should be, so language would be a powerful coordination tool. With language, teaching could be extended to support other cooperative processes, such as mutualistic exchanges, indirect reciprocity, and group selection. Like Martin Nowak, I suspect the efficient functioning of indirect reciprocity probably requires gossip. Like Ernst Fehr, I believe that linguistically taught social norms allow humans to institutionalize the punishment of non-cooperative individuals, for instance, through policing or socially sanctioned retaliation, which further enhances cooperation. I think that Mark Pagel is correct in suggesting that language evolved to promote cooperation, but I maintain that the origins of language began with teaching, a very specific form of cooperation. Nonetheless, other cooperative contexts would have exploited a pre-existing linguistic capability, generating selection for enhanced linguistic skills. Such selective feedback would have made a big difference both to the scale of human cooperation, and to the potency of human language.

 

DSW: What about the arts?

 

KNL: Here I think the production of art owes a rarely-appreciated debt to imitation that goes far beyond the copying of styles, techniques, and materials. In fact, I would go so far as to assert that in the absence of a mind fine-tuned by natural selection for optimal social learning, art works simply could not be produced.

 

We engage in the arts only because are all descended from a long line of brilliant imitators. Through copying, our ancestors learned how to make digging tools, spears, harpoons, and throwing sticks, or how to build a fire. In my book I describe how hundreds of thousands, perhaps millions of years of selection for competent imitation has shaped the human brain, leaving it supremely adapted to translating visual information about the movement of others bodies into matching action from our own muscles and joints. These cognitive abilities allow us to learn new skills today, such as how to drive or cook – however they are also what permits actors to act and dancers to dance. The ancestral sharing of emotions in social settings – for instance, responding with anxiety to the fear of a child – helped shape the empathy and emotional contagion that makes watching movies a powerful experience. In the absence of those social learning capabilities we would watch movies like sociopaths, unmoved by the Psycho shower scene. Likewise, when it comes to the evolution of dance, our imitative competence explains things like why humans are capable of moving in time to music, how we are able to synchronize our actions with others, and how we can learn long sequences of movements. All these artistic competences are reliant on neural circuitry that evolved for efficient social learning.

 

DSW: I have always admired the breadth of your scholarship. More than most, you appreciate the diversity of intellectual traditions that exist within evolutionary science (e.g., your earlier book with Gillian Brown titled Sense and Nonsense: Evolutionary Perspectives on Human Behavior). For much of this interview we have stressed the zone of agreement between you and others studying cultural evolution. Not that I wish to end on a note of discord, but what are some of the most important disagreements that exist among the cognoscenti?

 

KNL: Thank you. A couple of things come to mind. One is the disparity between evolutionary psychologists like Steven Pinker and the cultural evolution community, as represented by say Boyd, Richerson and Henrich, in how they conceive of human intelligence. In recent years, evolutionary psychologists have tended to emphasize how humans can devise solutions to challenging problems “on the fly”, through a combination of individual learning, improvisation and mental simulation. Conversely, cultural evolutionists tend to emphasize how it is the pooling of our collective knowledge that makes humans smart, with solutions arising as cultural adaptations through a collective process. I myself lean towards the latter account, but nonetheless envisage an important role for individual-level problem solving in cultural evolution. My impression is that, these days, most people accept that both of these kinds of problem solving occur in humans, so hopefully we can move beyond polemical pronouncements to an era in which we evaluate which is more important through empirical research.

 

Another unresolved issue is whether humans possess adaptations for enhanced social learning, and what these might be. Some researchers, for instance, Oxford psychologist Cecilia Heyes, believe that the human capacity for imitation relies on an ancient associative learning capability, and that both the extent of our reliance on imitation, and our imitative proficiency, are largely socially constructed. I think the opposite – that it is more appropriate to regard the enhanced human capacity for asocial learning as a side effect of selection for proficient copying, rather than the other way around. As I detail in my book, there is extensive neural and genetic data suggesting that our learning capabilities, including our ability to see connections among events, to discern the consequences of our actions, and to adjust our behavior flexibly, have been substantially upgraded during recent human evolution.

 

More generally, many researchers believe that natural selection has fashioned competences in human minds dedicated to proficient copying. Our potent capacity for imitation, the capacity and motivation of humans to teach and be taught, the tendency to produce (and attend to) infant-directed speech, to follow gaze, to share experience through joint attention, are all candidate adaptations to cultural life. I expect that multiple psychological adaptations for social learning may eventually be uncovered in our species, although I anticipate that the unlearned roots of such cognition might be quite subtle. Social learning adaptations are likely to be expressed as complex products of development that build upon evolved motivational, perceptual, or cognitive biases through very general learning processes that respond sensitively to a culturally constructed, symbolically encoded environment.

 

DSW: As I said at the start of this interview, these are indeed exciting times, which means that much is in flux. Thanks for helping me take this deep dive into your book and best of luck with your current endeavors!

 

 

This post was originally published in This View of Life (TVOL), 10 October 2017

Back to top