The Social Learning Animal

Our intelligence, language, cooperation, and technology… are not adaptive responses to extrinsic conditions. Rather, humans are creatures of their own making. The learned and socially transmitted activities of our ancestors, far more than climate, predators, or disease, created the conditions under which our intelligence evolved. Human minds are not just built for culture; they are built by culture.1

Kevin N. Laland is an evolutionary neuroscientist, whose life’s mission has been to try to understand how human beings developed mental capacities far beyond those of other species. I am sympathetic to Laland’s emphasis on culture, described in his newly-released book, Darwin’s Unfinished Symphony: How Culture Made the Human Mind. Tyler Cowen has argued that the proper understanding of the relationship between culture and economic behavior is “the most important unsolved problem in economics and indeed in social science more broadly.”2

Still, I think that we struggle to define culture. For example, recall Joel Mokyr’s definition:

Culture is a set of beliefs, values, and preferences, capable of affecting behavior, that are socially (not genetically) transmitted and that are shared by some subset of society.3

Joseph Henrich, whose work overlaps with Laland’s, wrote,

By “culture” I mean the large body of practices, techniques, heuristics, tools, motivations, values, and beliefs that we all acquire while growing up, mostly by learning from other people.4

Laland writes,

By “culture” I mean the extensive accumulation of shared, learned knowledge, and iterative improvements in technology over time. Humanity’s success is sometimes accredited to our cleverness, but culture is actually what makes us smart. Intelligence is not irrelevant of course, but what singles out our species is an ability to pool our insights and knowledge, and build on each other’s solutions. (page 7)

I have suggested defining culture as socially communicated thought patterns and behavioral tendencies. This leaves out the physical manifestations of culture—tools and technology—that are included in Laland’s and Henrich’s definitions. I prefer to think of tools and technology as phenomena that are distinct from culture, although they help to shape culture and are products of culture.

Laland’s definition stresses the cumulative nature of culture, and he argues that this is what sets humans apart. To the extent that other species share knowledge, as when animal parents help train their offspring to hunt, this knowledge is limited and fixed. Among other species, innovation gets lost, because animals are unable to record the results of innovation and pass these on to others.

Why does only one species on earth develop a cumulative culture? Why are there no intermediate species, with less culture than ours but nonetheless with socially communicated thought patterns and behavioral tendencies that are being developed and passed on to subsequent generations? (Of course, one possibility is that all intermediate species became extinct once human beings took over.)

Laland writes,

Herein lies a major challenge… to work out how the extraordinary and unique human capacity for culture evolved from ancient roots in animal behavior and cognition…. We must first understand why animals copy each other at all… The circumstances leading to the evolution of the abilities to innovate, teach, cooperate, and conform must all be established. Also critical is knowing how and why humans invented language… Finally, and crucially, we need to comprehend how all of these processes and capabilities fed back on each other to shape our bodies and minds. (Laland, pages 10-11)

Laland argues that the genetic gulf between humans and other species is wide.

Among the sample of genes that do differ between humans and chimpanzees, a disproportionately high number are expressed in the brain and nervous system. Genes expressed in the brain have been subject to strong positive selection in the hominin lineage, with over 90% of such genes upregulating their activity relative to chimpanzees…. Human brains are more than three times the size of chimpanzee brains and have been structurally reorganized… the former have proportionally larger neocortices and more direct connections from the neocortex to other brain regions. (page 18)

Within the hominin species, brain size nearly quadrupled over the past three million years, and gene expression within the human brain has greatly upregulated. Laland believes that this is due to positive feedback between culture and evolution.

Laland draws many insights from a “social learning strategies tournament” that he and other researchers carried out. The article describing this tournament was published in 2010.5

The simulated environment for our tournament was a ‘multi-armed bandit’ (18), analogous to the ‘one-armed bandit’ slot machine but with multiple ‘arms’. In the tournament, the bandit had 100 arms, each representing a different behavior, and each with a distinct payoff drawn independently from an exponential distribution. Furthermore, we posited a temporally varying environment, realised by changing the payoffs…

Entered strategies had to specify how individual agents, in a finite population, choose between three possible moves in each round, namely INNOVATE, OBSERVE and EXPLOIT. INNOVATE represented asocial learning, that is individual learning stemming solely through direct interaction with the environment, for example, through trial-and-error. An INNOVATE move always returned accurate information about the payoff of a randomly selected behavior previously unknown to the agent…. An OBSERVE move returned noisy information about the behavior and payoff currently being demonstrated in the population by one or more other agents playing EXPLOIT…. Finally, EXPLOIT represented the performance of a behavior from the agent’s repertoire, equivalent to pulling one of the multi-armed bandit’s levers. Agents could only obtain a payoff by playing EXPLOIT.

Think of a manufacturing worker in Ohio in 1999. If you stick with your industry and location, you are playing EXPLOIT. If you check out what people with your skills are doing in other industries and other locations, you are playing OBSERVE. If you try to start your own business or study a new trade, you are playing INNOVATE.

In the tournament, social learning (playing OBSERVE) proved much more effective than asocial learning (playing INNOVATE). Only when the researchers altered the parameters of the simulation to have the environment change at an implausibly high rate was there any advantage to playing INNOVATE frequently. Laland writes,

… copying pays because other individuals filter behavior, making adaptive information available to copy… An animal does not need to be smart to benefit from copying, because a lot of the smart decision making has already been done for it by the copied individuals who have already prefiltered their behavior…. That explains why copying is widespread in nature. (pages 70-71)

Widespread copying serves to slow the rate at which new behaviors are introduced. However, it also allows behaviors to be passed along to new generations, which keeps behaviors from getting lost. Laland argues that this positive effect tends to outweigh the negative effect of slower innovation. In fact, to avoid loss of knowledge of a behavior, copying has to be sufficiently accurate.

A small increase in the fidelity of social learning will transform cultural habits from being short lived to virtually immortal. (page 151)

Mathematical models suggest that the accumulation of culture depends crucially on minimizing the loss of traits, which means increasing the fidelity of copying.

Laland and colleagues found that copying works best when it is strategic (undertaken in circumstances where copying is most likely to produce benefits), accurate, and cost-effective. Laland argues that the human brain’s ability to copy in this manner was gradually increased through natural selection. For example, because visual acuity is a key determinant of copying ability, and copying ability improves fitness, our brains developed high visual acuity.

Another example of a capability that was selected for is “theory of mind.”

Effective copying might plausibly favor the evolution of a theory of mind and enhanced abilities for perspective-taking, with their associated neural underpinnings. Such abilities would allow the observer to comprehend the goals and procedures of the demonstrator more effectively, as well as allowing “teachers” to understand the state of mind ability of their “pupils.” (page 129)

Indeed, the phenomenon of teaching is strikingly more prevalent among humans than among animals. Laland argues that the evolution of teaching ability and language were spurred by culture. Among our prehistoric forebears there was

… plenty to teach, because our hominin ancestors subsisted on a broad omnivorous diet and were reliant on a large number of extractive foraging and tool-using skills. This period in human history was the dawn of cumulative culture, when our ancestors first began manufacturing stone tools… cumulative culture would help make teaching widely adaptive.

Perhaps, then, language originally evolved to enhance the efficiency and scope of teaching. (page 185)

He argues that this is more plausible than other accounts of language acquisition, such as the view that it emerged to help coordinate hunting. Thus, his anthropological interpretation is that the accumulation of knowledge helped stimulate teaching, which the value of teaching helped to stimulate the development of language, and that teaching and language helped to stimulate the development of a theory of mind.

In the end, Laland writes

… this entire book is one long advocacy for the significance of evolutionary feedback that encompasses a cultural drive mechanism initiated by natural selection that favored accurate and efficient copying. That selective feedback propelled the evolution of cognition in some primate lineages, and ultimately was responsible for the awesome computational power of the human brain. (page 208)

Apart from its interpretation of human cultural-genetic evolution, I found that Darwin’s Unfinished Symphony raised other issues that interested me.

• 1. I thought that Laland weaved together mathematical models, simulation exercises, experiments, and observations in a way that was much more persuasive than most social science. I recommend that economists read the book in order to stimulate thinking on how to improve our research methods.
• 2. The importance of fidelity in copying suggests that human progress should be related to the copying methods available. For example, the invention of writing would help to preserve some knowledge that otherwise might be lost, and this might be a factor in the advent of agriculture. The printing press would produce another leap in copying ability, and this might be a factor in the industrial revolution. Finally, the development of computers and communications technology should produce another similar surge in human progress.
• 3. While the advantages to an individual of copying are clear, the social benefits and costs may differ at times. There will be times when people copy behavior that seems to work in the short run but is harmful in the long run. There will be times when for each individual it appears to be risky and expensive to engage in asocial learning, but society as a whole needs more innovation in order to solve difficult problems, such as a deep recession.