This print conversation accompanies a This View of Life Podcast episode.
This series of conversations is centered on entrepreneurship but places it in the context of positive cultural change of all sorts and at any time or place in history. Up to now, it might seem that we have set the stage broadly to include economic and political history worldwide. This conversation with Peter Turchin sets the stage still more broadly to include our emergence as a species and the last ten thousand years of human history, stretching back to when the largest human societies were mere tribes of a few thousand people.
Peter is uniquely qualified to provide this broadest of views. He is Professor of Ecology and Evolutionary Biology at the University of Connecticut with joint appointments in Anthropology and Mathematics. As a biologist, he began his career studying the population dynamics of nonhuman species such as rodents and bark beetles. Then he began to train the same skills on the study of human history, a field that he christened Cliodynamics, which combines Clio, the Greek muse of history, with dynamics, the study of why things change over time.
Even more recently, Peter has worked with colleagues to create a global history databank called Seshat, named after the Egyptian goddess of wisdom, knowledge and writing. His books include War and Peace and War: The Rise and Fall of Empires, Historical Dynamics: Why States Rise and Fall, Ultrasociety: How Ten Thousand Years of War Made Humans the Greatest Cooperators on Earth, and Ages of Discord: A Structural-Demographic Analysis of American History. The Seshat project has resulted in an edited book titled Seshat History of the Axial Age.
David Sloan Wilson: Greetings Peter! In the spirit of full disclosure, you are past Vice President and currently serve on the Board of the Evolution Institute, which administers some of your support for the Seshat project. Is there anything you would like to add to my introduction?
Peter Turchin: Thank you, David, for starting this conversation. As to your question, I have just accepted the position of Group Leader at Complexity Science Hub–Vienna. For the next five years, I will be splitting my time—Spring Semesters teaching at the University of Connecticut and Fall Semesters leading the group on Social Complexity and Collapse at CSH in Vienna.
DSW: Congratulations! I’d like to begin our conversation by establishing some basic facts about genetic and cultural evolution. The relevance for the theme of the Third Way will soon become clear. Let’s begin with the distinction between “natural” and “artificial” selection. This distinction is relatively clear-cut for the study of genetic evolution in non-human species. In both cases, organisms vary in their measurable traits (their phenotypes) and these differences make a difference in terms of survival and reproduction (fitness). For natural selection, the fitness differences are imposed by the environment inhabited by the organisms, such as cold climates selecting for the thick fur of polar bears. For artificial selection, the fitness differences are imposed by humans, such as a poultry farmer selecting for egg productivity in hens. Artificial selection was already a common practice when Darwin wrote Origin of Species, so he could use it as an analogy when proposing his theory of natural selection.
However, the distinction between natural and artificial selection is not always clear-cut. In sexual selection, the traits of one sex (typically males) are selected not by the environment but by the preferences of members of the other sex (typically females). Is that “natural” or “artificial”? Group-living species select each other’s social traits, such as cooperativeness, a process called “social selection” and “self-domestication”.
When it comes to human genetic and cultural evolution, the distinction between “natural” and “artificial” selection becomes even more blurred. In a recent book titled The Goodness Paradox: The Strange Relationship Between Virtue and Violence in Human Evolution, Richard Wrangham attributes the genetic evolution of our species to self-domestication. Wrangham’s thesis for human genetic evolution dovetails with your thesis for human cultural evolution in your books, which we will be discussing in detail. Given all of this, is there any useful distinction to be made between “natural” and “artificial” selection in our species?
PT: The truth is I don’t find this distinction particularly useful when we talk about humans. Yes, domesticated plants and animals, for example, have evolved substantially as a result of their association with humans. But so did humans. One well-known example is the evolution of lactase persistence, resulting from the need for human adults to digest milk. But there is much more. Humans from populations that have cultivated wheat for thousands of years evolved better ability to detoxify this food resource (see my blog post The Dark Side of Cultural Evolution). Populations that had not benefited from such evolution, such as Polynesians, experienced catastrophic health effects when exposed to wheat-based foods. Capacity to process foods, of which fire was a very important component, enabled us to dispense with big guts, such as those characterizing our nearest relatives (chimpanzee and gorilla). So, to me the key concept is coevolution. Humans co-evolved with other organisms, and it is not clear who domesticated whom. Our biology coevolved with technology. Most generally, the key process has been culture-gene coevolution, a theory initially developed by Robert Boyd and Peter Richerson (also called Dual Inheritance Theory).
DSW: Thank you! Now let’s bring in Multilevel Selection (MLS) theory, which shows that natural and artificial selection can operate on single-species social groups and multi-species communities, in addition to individuals and their genes. Eusocial insect colonies are an example of natural group-level selection. Most of the traits of eusocial insects (ants, bees, wasps, and termites) evolve by virtue of fitness differences among colonies, not fitness differences among individuals within colonies. When selection within colonies occurs, it tends to result in antisocial behaviors that would be regarded as “cheating” in human terms. Poultry breeding practices are an example of artificial group-level selection. The best way to increase the egg productivity of hens is by selecting groups on the basis of their collective egg-laying ability. Selecting the most productive egg-layers within groups favors antisocial behaviors—hens laying their eggs at the expense of other hens—as surely as within-colony selection favors cheating in social insects.
As we know and will soon relate to our audience, MLS theory has proven to be transformative for the study of human genetic and cultural evolution. We are an ultra-cooperative species because group-level selection has been a very strong evolutionary force; first at the scale of small groups during our genetic evolution and then at the scale of increasingly larger groups during our cultural evolution, leading to the mega-societies of today. Is there anything that you would like to add to my account before I connect all of this to the theme of the Third Way?
PT: Well, as you know, you are preaching to the choir! Ever since switching from population ecology to cultural evolution, I have found the CMLS (cultural multi-level selection) theory to be an incredibly useful conceptual framework to guide my theoretical and empirical research. Its use is particularly relevant to the evolution of large-scale complex societies, which is the primary focus of my research.
The key here is cooperation. Our species is the champion of cooperative ability in the animal kingdom, at least if you measure it by the scale at which human societies cooperate. As I wrote in Ultrasociety, which you were kind to mention at the beginning, for tens of millions of years the champions of super-cooperation were the ants. We have surpassed the scale at which they cooperate (millions of individuals) only two or three millennia ago. A number of human societies number in hundreds of millions, and we are capable of such feats of international cooperation that placed the International Space Station in the orbit, the most expensive structure ever built by humans. Of course, our cooperation, especially on a large scale, is still quite fragile, as witnessed by our inability to stop wars. But it is still impressive, with all the caveats.
Why is this important? Because cooperation is key for well-functioning groups, organizations, and whole societies. Understanding how it evolved, and how we can cooperate better at all levels of organization is one of the major and still unresolved questions in science. It affects many aspects of our life, including economic production, which, as I expect, this conversation will eventually go to—but I’ll let you get there first.
DSW: Right! But I’d like to challenge your claim that our species is the champion of cooperative ability in the animal kingdom. That award goes to multicellular organisms as societies of cells! A single human being is a society of trillions of cells that functions with a precision that puts a Swiss watch to shame. This is thanks only to a process of between-organism selection, similar to the between-colony selection that results in the high cooperativity of the ultrasocial insects and the between-group selection that results in human cooperativity at the scales that we find it. For me, a major insight of MLS theory is that every entity that we call an organism, such as multi-cellular organisms or single-celled creatures, did not evolve by small mutational steps from other organisms, but as societies of lower-level units that evolved to be so cooperative that we bestowed a new name upon them. This is also why we feel impelled to call ultrasocial insect colonies superorganisms and also to employ the same metaphor in the description of human societies, from Aristotle to Hobbes. What MLS theory tells us is that the concept of society as an organism (when it is a unit of selection) is no longer metaphorical. It is a scientific fact. Even more amazing, multicellular organisms are not just a society of cells with genes that are mostly identical to each other, but also an ecosystem of microbes that number in the thousands of species and trillions of individuals and also function in the economy of the whole. Of course, it is unnecessary to award a prize for cooperative ability, but I think it is important to appreciate the conceptual similarity of all major evolutionary transitions, stretching back to the origin of life.
PT: We are largely in agreement, but let me qualify it on two counts. The first one is terminlogical. The now-standard definition of “ultrasociality” is large-scale cooperation among genetically unrelated individuals. The cells in our bodies, of course, are nearly 100% genetically identical (apart from some somatic mutation). Workers in social insect colonies are also highly related, with typical coefficients of relatedness between 0.5 and 0.75. For this reason, I prefer to use the more traditional term “eusociality” for social insects. The second qualification, with which you will probably agree, is that the boundary between “organisms” and “superorganisms” is not black-and-white. Putting it another way, organisms come in a variety of forms, some better integrated, others less so. A human body is a pretty well-integrated organism, but even we are vulnerable to insurrections of constituent elements (cells), known as cancer. Human societies are typically well-less integrated, with a lot of dysfunction and potential for collapse. But the main point, that it is selection between organisms/groups/superorganism/societies that make them more integrated and functional—in the long run.
DSW: We agree on the main point and on the fuzzy boundary between “organism” and “superorganism”, but I want to dwell more on the issue of genetic relatedness because it will be important for what follows. What you said about the cells in our bodies ignores the burgeoning literature on microbiomes, a term that wasn’t even coined until the turn of the twenty-first century. Now we know that the simple story of individuals as groups of nearly identical genes is complicated by the fact that individuals are also ecosystems of thousands of species and trillions of individuals, which interact with the genes of the organism in an intricate fashion. Can selection take place at the level of the microbiome, in addition to the level of the genotype? That depends upon whether the microbiomes vary among individual organisms and how much of the variation is heritable.
It turns out that heritable variation does exist at the level of microbiomes. Some laboratory experiments that I conducted with my Ph.D. student William Swenson, which were published at the same time that the word “microbiome” was being coined, are instructive in this regard. We created microbial ecosystems in the laboratory and selected them on the basis of their measurable properties, just as if they were individuals. The ecosystems were initiated by millions of individuals from a single well-mixed source, so initial variation caused by sampling error was vanishingly small. Nevertheless, the ecosystems became different from each other over time and there was a response to selection on this variation, which is proof of heritability.
What caused the heritable variation among the ecosystems? Not genetic relatedness, but complex interactions within each ecosystem that magnified initial variation caused by sampling error—what complex systems theorists call sensitive dependence on initial conditions. Against this background, we can begin to see that the importance accorded to genetic relatedness is misplaced. We treat genetic relatedness as important because we use it as a proxy for heritable variation. We think that phenotypic variation among groups will be proportional to genetic relatedness among groups, but this expectation is based on models that assume additive genetic interactions as a simplifying assumption. Once we focus on what really matters—heritable phenotype variation at various levels, genetic relatedness becomes only one of the multiple factors and there is no reason to single it out in our terminology.
Why is this important for the Third Way series of conversations? Because everything I have said about microbial ecosystems also goes for cultural evolution in human groups. Imagine colonizing a number of islands with a 1000 people per island, chosen at random from a single pool of people. Initial genetic, cultural, and phenotypic variation among islands based on sampling error would be very small. But it is highly likely that as the individuals on each island start interacting with each other, chance differences among islands will emerge that will compound over time, precisely as with my laboratory microbial ecosystems, causing divergent genetic evolution over the longer term.
This is why cultural variation among human groups exists at all scales, including nations of hundreds of millions of people. It’s not a matter of genetic vs. cultural evolution. It’s a matter of complex systems vs. simple systems as units of selection. Our entire intuition that gives genetic relatedness a special status, compared to the many other factors that influence multilevel selection, needs to be re-examined.
PT: I agree. To me, what’s important is heritable information, but how it is encoded—as DNA, RNA, knowledge in our brains, or electronic texts—is secondary. This was actually the point made by George Price when he discussed the significance of what is now known as the Price equation.
DSW: Returning to human cultural multilevel selection (CLMS) the first thing everyone needs to know is that the last 10,000 years of human history has not been a smooth increase in the scale of society but a tug of war between levels of selection, with the lower levels sometimes prevailing. As you have been pointing out since your book War and Peace and War, CMLS explains the collapse of empires in addition to the rise of empires, something that might well describe our current moment in history. Please elaborate on this theme for our audience, including what you call the Z-shape curve of human history over the last 10,000 years.
PT: A key theoretical insight that we gained studying models of multilevel selection is that competition between groups (including whole societies) promotes cooperation, while competition within groups/societies destroys it. This allows us to think about both the rise of large-scale complex societies and its inverse, their breakdown, using the same conceptual framework. Historical research by myself and colleagues shows that one of the most important factors explaining imperial collapse is what we call elite overproduction (by elites I simply mean the small proportion of the population that concentrates social power in their hands). This is what happens when there are too many elite aspirants competing for a limited number of power positions. I explain why all complex societies periodically experience elite overproduction in the book you mentioned, War and Peace and War (and, focusing on the United States, in Ages of Discord).
Some competition is normal and beneficial, but when the intensity of competition reaches the level when a large majority of aspirants have to become losers, its effects are entirely negative. The normal functioning of human societies requires a high degree of cooperation both among the elites and between elites and the rest of the population. Intra-elite competition corrodes cooperation among the elites and leads to political polarization, the formation of rival factions, and internecine conflicts that often lead to bloody civil wars and sometimes social collapse.
Obviously, I am simplifying a very complex set of factors that together bring about state breakdown. But the main idea here is that cooperation plays a very important role in both the rise and demise of large-scale societies.
Let’s talk about the Z-curve in the context of your next question.
DSW: Right! One quantum leap in the scale of human society was the advent of the major religions during the period that historians call the Axial Age. Please describe how CMLS goes beyond traditional historical scholarship on the Axial Age, especially with the publication of Seshat History of the Axial Age.
PT: Although human history in the last 10,000 years is sometimes portrayed as a smooth increase in the scale of societies, growth of technological knowledge, and increased economic productivity, as we have just discussed, there were also periodic breakdowns and back-sliding along all those dimensions. On top of these cycles, we discern another macro-historical pattern that my colleagues and I have called “the Z-curve of human egalitarianism.” When taking the long view of human evolution, from very unequal societies of our great ape ancestors, characterized by strict social hierarchies, to the societies of today, which at least emphasize equal rights (even if often failing to live up to such declarations), we see that there were two great turns, together making up the “Z”.
Foraging societies, in which human beings lived until roughly 10,000 years ago, are quite egalitarian. But the first centralized societies, chiefdoms and archaic states, which arose after humans started to settle down in agricultural villages, were extremely unequal, even despotic. The chiefs and early kings wielded enormous power over commoners. These early complex societies had huge differentials of power and wealth, rulers that set themselves up as gods, and human sacrifice, often on a gigantic scale. And then something happened. Starting 2.5 thousand years ago first a few, and then increasingly more world regions turned away from such extreme forms of inequality. Naturally, they never returned to the days of hunter-gatherers (even our modern democratic societies are much more unequal than a typical foraging band). But the new trend was towards more prosocial rulers, who gradually abandoned their godhood pretensions and started to care about the well-being of their subjects.
The historians call this turn, which started in the Middle East, Eastern Mediterranean, North India, and China during the first millennium BC, the Axial (or Pivotal) Age. Can cultural evolution help us make sense of this Axial Age?
What happened was that escalating competition between the states, taking the form of more intense large-scale warfare, put an enormous selection pressure on them to become more cooperative. Because inequality corrodes cooperation, those states and empires that abandoned the worst forms of it, survived and even thrived under the new circumstances. As a result, new social norms and institutions that made societies less unequal and despotic spread.
I develop this theoretical argument, explicitly based on CMLS, at greater length in Ultrasociety. But we can ask, could this theory be tested empirically? It turns out that the historical and archaeological record allows us to do so. Seshat: Global History Databank, an international and interdisciplinary project that my colleagues and I started in 2011, have now collected an enormous amount of data that has allowed us to thoroughly test this and other theories of cultural evolution. We have summarized much historical information, relevant to the Axial turn of the Z-curve in recently published Seshat History of the Axial Age.
DSW: Indeed! Do I understand correctly that the shift toward egalitarianism occurred independently in different geographical regions? That it was a case of parallel cultural evolution caused by similar conditions (between-group conflict at a large scale requiring within-group cooperation), rather than a single cultural event that spread worldwide? This would make it unnecessary to posit a historical connection between early Christianity and Buddhism, for example. Please elaborate just a bit, because this bears upon how a CMLS perspective adds to the extensive traditional historical scholarship on the topic.
PT: Not completely independently. What happened was that the spread of iron weapons and especially horse-riding through the Great Eurasian Steppe, starting around 1000 BCE, resulted in a sharp increase in the intensity of warfare, both within the Steppe and in the belt of agrarian regions south of the Steppe. This is why agrarian societies from China to Anatolia (modern Turkey) were subjected to much more intense inter-societal competition nearly simultaneously (within a few hundred years, depending on how fast new forms of warfare reached these regions). Different regions responded to this selection in different ways: Monotheism in the Near East, the idea of Karmic Retribution in North India, and Confucianism in China, to give a few examples. What was similar was that intensification of inter-societal competition everywhere resulted in the spread of more prosocial ideologies.
DSW: I want to make sure that I have this right, for both myself and our readers. All of these societies were facing the same selection pressure (intensity of warfare) but were adapting culturally in different ways (Monotheism, Karmic Retribution, Confucianism, etc.), like different ways of skinning the proverbial cat. And whatever took place in the Americas would follow a very different timetable but would be somewhat predictable based on the scale and intensity of intergroup competition. Right?
PT: Yes, assuming that we have correctly identified the general principles governing social evolution. So if the Americas were cut out in 1491 and moved to a different Earth, I would expect that they would continue inventing the ever more destructive military technologies. The increasing pressures of warfare would eventually result in the same evolution that we saw in the Old World, including the greater social scale of states and empires, and the spread of ideologies that reduce the worst forms of inequality.
DSW: Great! Now I’d like to engage your help in resolving a major paradox about human cultural evolution that is central to the thesis of the Third Way. On one hand, human cultural evolution has a huge unpredictable component. Life consists of many inadvertent social experiments, a few that hang together and most that fall apart. The result is societies that work (at least to a degree) without having been designed by anyone. They work without anyone knowing how they work.
On the other hand, human cultural evolution is suffused with intentions. Nearly every single person is trying to do something with goals in mind, either as individuals or in various groupings within the culture. More often than not, those groupings are socially constructed, which has a huge impact on the balance between levels of selection. Selection couldn’t take place at the scale of a large society if that society didn’t have a name, symbols, and practices associated with it, and so on. That’s why “niche construction” is such a key term in the study of human cultural evolution.
So, how do we reconcile the intentional component of human cultural evolution with the unpredictable component?
PT: I am not sure that the distinction between “intentional” and “unpredictable” is a good one. To give a historical example, when the Assembly of the Notables failed to endorse the tax proposals put by Louis XVI of France to them in 1787, they acted intentionally. But surely their aim was not to trigger the French Revolution in which most of them lost their heads? The problem is that human brains are not great at figuring out the consequences of their interventions into complex systems with nonlinear feedbacks (like our societies). Thus, intentional actions may (and often do) result in unintended consequences.
In my view, human intentionality is not a huge issue for evolutionary approaches to understanding how our societies change. The main mechanism, to me, is the selection by consequences. Groups and whole societies try different ways to achieve their goals. Those ways that lead to bad unintended consequences are abandoned, while approaches that work are adopted and spread. And this is very similar to natural selection.
DSW: I agree and think that what we have covered so far affirms the thesis of the Third Way for human cultural evolution writ large. When you say that unregulated competition within a society disrupts cooperation, that’s a generalized version of “laissez-faire doesn’t work”. When you say that intentions frequently have unintended consequences, that’s a generalized version of “centralized planning doesn’t work”. And when you say that it all boils down to selection by consequences, that lands us in the ballpark of the Third Way.
But more needs to be discussed about the role of intentions in structuring selection by consequences. You just said that “groups and whole societies try different ways to achieve their goals.” Those goals are their targets of selection and their different ways are the variation oriented around the targets. You rightly stress that the variation has a big unintended component in addition to an intended component, but either way, outcomes are selected on the basis of whether they take the system closer to the goal.
Here are some examples to make the discussion less abstract. In Ultrasociety, you quote the pre-Axial Age ruler Tiglath Pileser I, whose goal was to pillage and plunder (“The heavy yoke of my empire I imposed upon them.”). You also quote the Axial Age ruler Ashoka the Great, whose goal was to create an administration that works “for the welfare, happiness, and benefit of the people in the country.” In both cases, the actions of the rulers probably had unintended consequences, but everything they tried to do or selected on the basis of their consequences was oriented toward their respective goals—their target of selection. It is extremely unlikely that Tiglath would create a fair society or Ashoka would create a brutal despotic society by accident. Their intentions shaped both the variation and selection of parts of the cultural evolutionary process.
Here is another example based on my own work. While I was writing my book Darwin’s Cathedral, I made a special study of the origin of Calvinism in the city of Geneva during the Protestant Reformation. All of the Protestant reformers were trying to tailor the Christian religion to various polities. I was impressed by how much Calvin’s catechism and ecclesiastical ordinances were tailored to the details of Genevan life. Nevertheless, Calvinism, as it unfolded, was not Calvin’s conscious invention, because his efforts had unintended consequences and collided with the efforts of others. The fact that the reform movement in Geneva hung together while similar reform movements in cities such as Bergen fell apart was largely a matter of chance. Given that it did hang together, Geneva became known as “The City on the Hill” and was widely emulated by other reformers, who didn’t necessarily understand the active ingredients of what they were copying. In this and many other examples that could be cited, everything you say about unintended consequences, selection by consequences, and semi-blind copying is true, but intentions still play an essential role in defining the target of selection and variation oriented around the target. Agreed?
PT: Yes!
DSW: Now I’d like to apply some of these ideas to current affairs, based on your great book Ages of Discord. You show that America has experienced two peaks of relative well-being, during the first half of the 19th century and middle of the 20th century, separated by valleys of discord, during the turn of the 19th century and the present. As you are careful to say, cycles of history are not like mechanical pendulums. They are based on decisions that people made, no matter how contested or uncertain in their consequences. What were the decisions that pulled America out of its first Age of Discord? What decisions are required to pull America out of its current Age of Discord?
PT: Exiting from our first Age of Discord was a slow process that occurred during the first third of the twentieth century—with reforms that started in the Progressive Era and then were clinched during the New Deal. During most of the Nineteenth Century worker wages lagged behind economic growth. A very vivid indicator of the declining well-being for the majority of the population was shrinking average stature (height) of native-born Americans, as well as declining life expectancy. This factor (declining well-being) created an enormous degree of instability which reached a peak around the 1920s. In fact, it was so bad that the US was in a “revolutionary situation”. The forward-looking segments of the political elites realized that unless they solve the problem of declining well-being the whole system would collapse. It was not easy, because many segments of the elites had to give up their privileges, but eventually, the reformist elites prevailed and saved the capitalism from itself, so to speak.
DSW: Right! That’s why one biography of Franklin Delano Roosevelt is titled Traitor to his Class.
Now I’d like to discuss the speed of cultural evolution during the long stretch of human history. Cultural evolution might be fast compared to genetic evolution, but it can still require decades and centuries, which is unacceptably slow as far as modern policy formulation and implementation is concerned. Why does cultural evolution take as long as it does and is there a way to make it go faster? Please give some examples so that the conversation does not become too abstract.
PT: I think we should use the approach of adaptive management, which was developed in ecology and the environmental sciences. For example, let’s choose a concrete problem to solve. Over the past four decades, the median worker wage scaled by GDP per capita has declined to historically low levels. This means that the fruits of economic growth our society has experienced have not made the well-being of the majority of the population better (they primarily went to the upper 1 percent). This happened in the US and several Western European countries
How do we reverse this trend? First, we need experimentation from different countries and regions within them. They try different approaches. For example, the state of Washington has mandated a substantial increase in the minimum wage. Are there negative consequences, e.g. increased unemployment rate, or business moving away and taking jobs with them? If yes, what additional measures can be taken to correct them? Meanwhile, other countries/regions try different approaches. The goal is bringing median worker wage scaled by GDP back on track so that large swaths of the population experience better wellbeing. As different regions are more or less successful in achieving this goal, their approaches are copied by other regions. Some negative consequences may take years to develop, so this would be a slow process, but much faster than centuries needed for unmanaged cultural evolution.
DSW: That sounds like the Third Way to me! Thanks for helping me set the stage so broadly and for your great work.
