Collective Learning 

Collective Learning

The term collective learning is used here for an idea that appears under different names in different disciplines (Christian 2004, 2010). Near synonyms include culture or social learning (Richerson and Boyd 2005). Although their meanings overlap, there are subtle but important differences between these terms. Collective learning is defined here as the ability of a species to share information so efficiently and so precisely that learning takes place not just at the level of the individual, but also at the level of the community and the species.

The human species has crossed a fundamental tipping point in communicative efficiency, so that information can keep accumulating without any apparent limits at the level of the whole species. Already by the eighteenth century, the Scottish philosopher Adam Ferguson had grasped the importance of our capacity for collective learning: “In other classes of animals, the individual advances from infancy to age or maturity; and he attains, in the compass of a single life, to all the perfection his nature can reach: but, in the human kind, the species has a progress as well as the individual; they build in every subsequent age on foundations formerly laid” (Ferguson 1767, sec. 1).

Why Is Collective Learning Important?

Properly understood, collective learning is an idea of profound importance because it helps explain what makes our species, Homo sapiens, unique in the history of the biosphere. Collective learning explains why we are the only species in almost 4 billion years to have a history of long-term change. It explains the distinctive nature of change in human history as the idea of natural selection explains the distinctive forms of change we see in biological history. The idea of collective learning helps explain the remarkable and terrifying power we wield today, and why to many scholars human activities seem to threaten humanity and perhaps much of the biosphere. As the Dutch climatologist Paul Crutzen has argued, we now have become so powerful that, for better or worse, we are the dominant force for change in the biosphere, a species capable of transforming the climate, the oceans, the rivers, and the landscapes of an entire planet. That is why he argues that we have entered a new geological era: the “Anthropocene,” or the era dominated by human beings (Crutzen 2002; Steffen, Crutzen, and McNeill 2008).

Like natural selection, the idea of collective learning may seem simple, but it, too, contains important and subtle nuances.

First, the adjective collective is important. There are limits to what an isolated individual can learn, whether the individual is a human or a member of any other intelligent species. Human brains are indeed larger than those of our closest relatives, chimpanzees. But the roughly threefold difference in brain capacity is not enough to account for the much greater differences between the cumulative, diverse, and highly changeable historical trajectory of Homo sapiens and the relatively stable historical trajectories of all other species. Nor have humans gotten any brainier. We have no evidence that individuals today can store more information in their brains than our Paleolithic ancestors could, even if we may be better at accessing the vast stores of information that exist outside our individual brains. What distinguishes us from all other species is that we can share information rapidly, efficiently, and precisely, creating a large and growing stock of information that we share collectively. In principle, this stock of information can grow without limit. As the Canadian psychologist Merlin Donald puts it, “The key to understanding the human intellect is not so much the design of the individual brain as the synergy of many brains” (Donald 2001, xiii).

The second important nuance is contained in the notion of a “threshold” or “tipping point” in communicative efficiency. In a limited sense, many species are capable of collective learning. They have languages and can share information, so they can be said to have “cultures.” Primatologists know that different communities of primates vary slightly in their technologies and behaviors. Some chimp communities, for example, use sticks to extract termites from termite mounds, in a practice known to primatologists as “termiting,” and it seems clear that young chimps learn these culturally specific behaviors from their elders (Goodall 1990). So it makes sense to talk of different communities having different cultures and to presume the existence among many primates of some form of social learning.

But the notion of social learning blurs a critical distinction. Some information does indeed circulate within the collective memory of primate communities and other intelligent species, but it circulates so inefficiently, so slowly, and with so many leaks that gains are eventually canceled by losses. That is why the behaviors of other intelligent species do not appear to change in fundamental ways on scales of centuries or millennia. We have no evidence that chimp technologies have improved significantly over time. Nor is there any evidence of sustained accumulation of cultural or technological information in any species apart from ourselves. Indeed, if there had been such a species, its existence surely would show up in evidence that the species had expanded its range and transformed its environment.

In what follows, the phrase social learning refers to cultural sharing in general, and the word culture refers to the products of that sharing. The phrase collective learning is confined to social learning that operates so powerfully that learning begins to accumulate without clear limits. Defined in this way, collective learning is unique to our species and should be considered as a defining feature of Homo sapiens.

In short, collective learning is social learning that has crossed a tipping point beyond which culture does not merely exist; it evolves, changes, and gains increasing power. The tipping point can be found where new and more powerful forms of communication emerge, allowing such efficient sharing of information that more information accumulates in the collective memory than is lost through misunderstanding, forgetfulness, “leakiness,” or simple chaos. The difference between either side of the threshold may seem small, a matter, perhaps, of minor rearrangements in the brain. Perhaps, as the US linguist Noam Chomsky suggests, it is grammar that explains the critical increment in efficiency. Or, as the US anthropologist Terrence Deacon has argued, it may have been the ability to use language symbolically (Deacon 1997). However we explain it, the crossing of this threshold in communicative efficiency was an event of profound significance in the history of our planet. It is what physicists might describe as a “phase change”: a small change in some parameter that proves transformative. Imagine water flowing into a tub and out through a drain. For a time, the water level will remain steady, but if the flow increases, there will come a tipping point when water begins to flow in faster than it flows out. Suddenly, the water level will start to rise, and it will keep rising, without clear limits, as long as the high flow is maintained. Where previously there was stasis, now there is change. The difference, though small, is transformative.

There are powerful reasons for thinking that our species may be the first in the history of the biosphere to have crossed this important threshold. Once a species begins accumulating cultural information, there is no limit to the process, so in principle, information can accumulate until such a species acquires powers that are dangerous both to itself and to its environment. Eventually, such a species will start transforming its home planet. This is why the notion of collective learning may have much to tell us about the Anthropocene era.

Species of such power should show up in the archaeological record, as our species certainly will, on scales of millions or even hundreds of millions of years (Zalasiewicz 2009). Yet we have no evidence that such a species has ever existed before us. Furthermore, as we move back a few hundred million years in paleontological time, the size of the largest brains diminishes and the likelihood of such a species having existed fades to nothing. So there are good reasons to think that we are the first species capable of collective learning in the 4-billion-year history of Earth.

The existence of a threshold also suggests that at any one time there is likely to be only one such species on the planet. This is because collective learning unleashes mechanisms of change so much faster than those of natural selection that they will close off all opportunities for other species to evolve down similar pathways. This is apparent in today’s world, in which human numbers are rapidly increasing, while our closest relatives, some of the most intelligent species on the planet, are close to extinction. The ecological principle of “competitive exclusion” explains why two nearly identical species can never share the same niche; in a zero-sum competition, tiny differences ensure that one species will drive out the other. So it explains why, even if Neanderthals crossed the same linguistic threshold as our ancestors, and at about the same time, only one of these species was likely to survive, particularly as both species may have been expanding the niches they occupied. Our propensity to keep widening our niche may also explain why we seem to have driven so many other species to extinction, beginning with our closest relatives.

The idea of collective learning helps explain why we are unique on scales of billions of years, and why humans alone have a history of sustained, long-term change. It also explains the technological precocity that has created the world of the Anthropocene and that may be threatening the sustainability of the biosphere as a whole.

Collective Learning and the Trajectory of Human History

This discussion may seem to imply that collective learning has generated a steady trickle of cultural and technological change throughout human history. But of course that is not quite right. The pace and nature of change has varied enormously in different eras of human history and in different environments. On occasion, after wars, or natural or ecological or epidemiological disasters, there also have been periods of regression, when information was lost faster than it was found, even if the long trend has undoubtedly been toward sustained and accelerating cultural accumulation.

So how does collective learning work in detail? How has it shaped human history in different eras and different environments? Why are cultural and technological changes so fast today; and why were they so slow in the Paleolithic era? As such questions suggest, the idea of collective learning can generate some rich historical research agendas. The following sections sketch some of the more important principles that explain why collective learning operates in different ways and with differing effects in different environments.

Feedback Cycles and Accelerating Change

Perhaps the most important general principle about the workings of collective learning is that it generates many positive feedback loops. Collective learning feeds on itself. An obvious example is when new technologies, such as writing, printing, or the Internet, improve the efficiency of information sharing and storage, so as to increase the power of collective learning in general. Multiple feedback loops explain why the long trends in human history, whether of population growth, control of biospheric resources, or cultural accumulation, have accelerated over the course of human history, culminating in the frenetic changes of today. To understand how these feedback mechanisms work, it may help to begin by describing how collective learning worked in the relatively simple communities of our Paleolithic ancestors.

Collective Learning in Paleolithic Societies

The archaeological record shows that, by modern standards, change was glacially slow in the Paleolithic era. Most anthropologists agree that our species appeared between 100,000 and 200,000 years ago (Scarre 2005). If collective learning is what makes our species so distinctive, it makes sense to equate the origin of our species with the beginning of collective learning. But identifying the earliest evidence of collective learning is difficult. Human communities were small, and if technologies were accumulating, they did so slowly—so slowly that change may be almost undetectable from the available archaeological evidence. Signs of growing technological diversity might be one indication that local cultural accumulation was driving different communities along divergent cultural pathways, but detecting such diversity is difficult. If collective learning depends, however, on improved forms of communication, then it may be worth looking for evidence of new, perhaps symbolic, forms of communication. That makes early evidence of symbolic activities, such as the use of ocher (an iron ore used as a pigment, presumably to paint bodies or objects), highly significant (McBrearty and Brooks 2000). The US anthropologists Sally McBrearty and Alison Brooks offer tentative evidence both for technological innovation such as hafting (or adding a handle to a tool or weapon) and symbolic activity (such as the use of ocher) from African sites that are 200,000 years old. By 100,000 years ago, evidence of both kinds is more common and more reliable. Sites such as Blombos cave in South Africa, which dates to almost 100,000 years ago, make it all but certain that by then collective learning was under way and generating considerable cultural and technological creativity.

But of course, cultural accumulation would have been slow in the small and relatively homogenous communities of the Paleolithic. In foraging communities, ideas were exchanged by small groups with broadly similar experiences. Their size and homogeneity limited possibilities for interesting cultural synergies. This may explain the slow pace of change in the Paleolithic era. Nevertheless, even in this period change was extremely rapid compared to the biological realm, where change is driven by genetic rather than cultural evolution. Only in comparison with today’s world does technological change appear slow in the Paleolithic era of human history. The piecemeal migrations that took modern humans to all continents apart from Antarctica by 15,000 years ago depended on constant innovation, as communities learned how to exploit new environments and new plant and animal species and to deal with extreme climatic conditions, such as those of northern and eastern Siberia. (See figure 1 below.)

Click to view larger

Figure 1. Behavioral Innovations of the Middle Stone Age in Africa

Source: McBrearty and Brooks (2000, 530).

*Note: “Notational pieces (incised)” refers to carved pieces of bone, stone, and red ocher.

Figure 1 shows the evolution of various implements and other innovations throughout the Middle Stone Age in African history, from the development of blades (c. 160,000 years ago) to the relatively recent appearance of images (c. 20,000 years ago).

Agriculture and Collective Learning

Agriculture, which appeared about 11,000 years ago, transformed the workings of collective learning and generated powerful new synergies. Agriculture consisted of a suite of technologies that diverted more of the products of photosynthesis toward our own species by altering environments in order to eliminate species we cannot use (“pests” or “weeds”) and to favor the growth of species we can use (“domesticates”). As a result, humans began to commandeer more and more of the energy and resources of the biosphere, and human populations (and those of their domesticates) began to increase. (See figure 2 below.)

The new synergies were powerful. Larger communities meant more people sharing more information. The impact of greater numbers was not just additive; it was exponential. The mathematics is simple. In a group of three people, three distinct links are possible, but in a group of four, six links are possible, in a group of ten, forty five links are possible, and in general, the number of possible links in a group of n people is proportional not to n but to n × (n − 1)/2. (With large numbers, this is close to half of n2.) More people does not just mean more information sharing, but lots more sharing. Population increase alone thus multiplied the potential synergy of collective learning in agricultural societies.

Agricultural communities were also more diverse than foraging communities because surpluses began to support nonfarming groups, encouraging a growing division of both labor and knowledge. Specialization distilled and concentrated expert knowledge in different sectors of society, filing it within distinct groups. This process significantly increased the total stock of available information. Markets also encouraged exchanges of goods and information between scattered communities. In all these ways, agriculture created new synergies and new feedback mechanisms within expanding networks of collective learning.

Click to view larger

Figure 2. Human Population Growth since 100,000 Years before Present (in millions)

Source: Adapted by Berkshire Publishing Group from Christian (2004), Maps of Time, page 143, plus interpolation.

*Note: The human population has increased from 6.3 billion to 7 billion since the publication of the original figure in 2004.

There has been a sharp spike in the global population since around 11,000 years ago, when the spread of agriculture diverted more of the products of photosynthesis toward our own species. The use of agriculture alters environments in order to eliminate species we cannot use (so-called weeds) and favors the growth of species we can use (e.g., wheat or tomatoes). The population continues to rise, passing 7 billion people in early 2012.

For all these reasons, it should be no surprise to find that the pace of technological change accelerated after the appearance of agriculture. Accelerating technological change increased human control over the resources of the biosphere, which encouraged further population growth, creating a new feedback loop of extraordinary power. As it spread, agriculture generated larger and more diverse communities, which stimulated technological change, which allowed the emergence of even larger and more diverse communities over several thousand years. The impact of these feedback loops is apparent from any attempt to calculate human population growth over thousands of years.

Increasingly diverse communities and a division of labor and knowledge also made for more complex social structures and ensured that different groups would have differential access to information. As a general rule, those best connected within networks of information exchange were those with the greatest power. They were, in the familiar cliché, “better connected.” Traders, city dwellers, and elite groups tended to be better connected than peasants and that meant they had better access to the information filed away within particular specialist groups. This emerging hierarchy of connectedness and access to information may help explain the development of hierarchies of power, prestige, and wealth within agrarian societies.

Networks of collective learning share important properties with networks in general. Since the middle of the twentieth century, mathematicians, economists, computer programmers, biochemists, and specialists in many other fields have realized that networks are ubiquitous, and they share common features, whether they consist of reactions within a living cell or servers on the Internet. Networks are much more than the sum of their parts; they are active forces, and their topologies and evolutionary rules shape how they work.

Networks have two main components: points and links. Mathematicians call the points nodes and the links edges, while they commonly refer to entire networks as graphs. One of the most important properties of networks is that variations in the connectedness of different nodes may affect the efficiency with which the entire network is connected. If most nodes are connected only to close neighbors, there are many “degrees of separation,” so that getting from one random node to another can take a long time. (This is surely a plausible description of a world of small peasant communities.) But add a few random long-distance connections, perhaps a few itinerant peddlers or migratory workers, and then add markets, cities, and rulers, and suddenly everything changes. Now the best-connected nodes create information peaks from which information can move more efficiently to other parts of the network. Then, as the US physicist Mark Newman and the Australian physicist Duncan Watts have shown, the network seems to shrink as the number of steps between any two nodes is suddenly reduced.

Let’s say we have a population of 1,000 people with 10 friends each and no “random” friends. That is, everyone’s friends are drawn only from a strictly defined social circle. Then the average degree of separation is 50; in other words, on average it will take 50 hops to get from one randomly selected person to another. But if we now say that 25 percent of everyone’s friends are random, that is, drawn from outside their normal social circle, then the average degree of separation drops dramatically to 3.6.

(Beinhocker 2007, 146)

These general principles suggest that the diverse and hierarchical structures of complex societies must have greatly enhanced the synergies of collective learning.

Collective Learning in Agrarian Civilizations

Agrarian civilizations magnified the power and scope of collective learning in many ways. Cities developed trading links reaching over hundreds or even thousands of miles. Eventually, those links extended beyond the borders of particular regions until, by 4,000 years ago, Eurasian networks of exchange began to span the entire continent, sharing technologies such as horse riding and bronze making and goods such as silk and jade (Christian 2000). As links of various kinds developed over larger and larger distances, the number and diversity of information exchanges increased until, 800 years ago, there emerged in the Mongol Empire, a polity reaching from Korea to the Mediterranean.

Similar changes also occurred in other parts of the world, but at different speeds. In Australia, there were no agricultural communities when European colonizers arrived in the eighteenth century. Nevertheless, there was plenty of cultural and technological change, and change was accelerating in recent millennia in ways that are reminiscent of the “affluent foragers” that lived in southwest Asia in the Fertile Crescent, just before the appearance of agriculture. Populations increased in favored regions, some groups became more sedentary, and new technologies appeared.

In the Americas, agriculture appeared several thousand years later than in the core agrarian regions of Afro-Eurasia, but it generated similar changes. Despite the lack of any clear links with Afro-Eurasia, farming communities, cities, states, writing systems, monumental architecture, and extensive networks of exchange appeared. But with smaller populations and fewer metropolitan centers than in Afro-Eurasia, networks of exchange in the Americas were smaller and less diverse, and collective learning operated less powerfully. For example, there is little evidence of significant links between the major agrarian civilizations of Mesoamerica and the Andes. When the Americas and Afro-Eurasia came into contact after 1492, these differences mattered. The technologies of metallurgy and gunpowder making and horse riding, which had not evolved in the more limited networks of the Americas, gave European invaders significant military advantages. Even more important were the epidemiological networks through which diseases, like information, had been exchanged in Afro-Eurasia over large areas. Afro-European colonizers brought to the Americas hardened immune systems, while Americans, whose exchange networks had exposed them to fewer diseases, succumbed in millions to Afro-Eurasian diseases, such as smallpox.

So it really did matter if collective learning operated within large, populous, diverse, and dynamic networks or within networks that were smaller and less varied. These differences had a profound impact on the pace and nature of change in different regions and different eras.

Collective Learning and the Modern World

The linking of local networks into a single global system of information exchange in recent centuries has created larger, more diverse, and more efficient networks of exchange than ever before in human history. Their evolution may help explain the extraordinary acceleration in the pace of change in the modern era, as crops, livestock, and commodities began to travel across the entire world, along with new technologies such as the trilogy of gunpowder, paper, and the compass, inherited by European societies from China. For a time, European societies, which sat at the center of the first global networks of exchange, benefited disproportionately from these global flows of information and ideas. Since 1800, railways, steamships, the telegraph, the telephone, the radio, automobiles, air travel, and the Internet have magnified the efficiency and power of modern networks of collective learning by many orders of magnitude.

Today, highly efficient global networks of collective learning are generating a tsunami of new information that threatens to overwhelm us while it accelerates change in general. The astonishing power of modern networks helps explain the rapid increase in the ecological power of our species during the last two centuries and the fact that our species has sharply increased its control over the resources and the energy of the biosphere, particularly since learning how to exploit the energy of fossil fuels. Not surprisingly, as our species has commandeered more resources, more energy, and more ecological space, other species have suffered.

Collective Learning and the Future

The idea of collective learning can help us make sense of the large trajectories of human history, with their many accelerating trends. Indeed, the idea of collective learning is so powerful that it has something of the power of a paradigm, in the sense made famous by the US historian of science Thomas Kuhn (1970). That is to say, it has the potential to help coordinate the research of scholars working in many different humanities disciplines.

Can the idea of collective learning also help us think more clearly about the future? If the ideas described above are correct, they suggest that our astonishing technological precocity arises from our very nature as a species. Something like today’s Anthropocene era was predictable as soon as the first humans crossed the threshold in communicative efficiency that switched on the engine of collective learning. Although many different factors may have shaped the timing, pace, and geography of change, the general—and accelerating—trend of collective learning can be traced all the way back to our species’ origins. As a species, we cannot help being creative and innovative, even if some of our innovations, such as the increasing use of fossil fuels or the creation of nuclear weapons, may pose grave threats to our future. The challenges we face today are no accident of history.

This may seem a pessimistic conclusion, suggesting that the dangers of today’s world are inherent in our nature as a species and that we are, therefore, doomed to self-destruction. But the notion of collective learning also points to more optimistic conclusions. If the dangers and challenges of today’s world are products of our capacity to learn collectively, so too is our ability to find and implement appropriate solutions. Rather than winding back the technological clock, we will need to use our creativity in more disciplined ways in the future, to develop technologies that are more sustainable, to build political and social relationships that enhance global collaboration and collective learning, and to redefine our understanding of a good life and a good society in ways that allow a more durable relationship with the biosphere as a whole. These are all tasks that will depend on exploiting to the full the powerful synergies we see in today’s interconnected world, in which 7 billion individuals are linked, more or less in real time, within a single global network of collective learning.

Our collective intellectual power as a species is surely increasing as fast as the scale of the military, ecological, and resource problems that we face. Far from disowning our remarkable creativity, we need to direct it in new ways. This conclusion suggests the profound importance of education, of communications, of science, of effective collaboration—in short, of all the skills humans will need to fully exploit our remarkable capacity for collective learning