Thomas Homer-Dixon

Writing

April 25th, 2007 —

Conflict in a Nonlinear World

Complex Adaptation at the Intersection of Energy, Climate, and Security

Ingar Moen Memorial Lecture, Science and Technology Symposium, Defence Research and Development Canada

It is an enormous privilege to have the opportunity to present the first Ingar Moen Memorial Lecture. I understand that Dr. Moen was the initiator of this annual symposium and was widely regarded as a brilliant, creative, and unorthodox thinker. I will, I hope, offer some ideas today that he would think in keeping with the spirit of this event.

In the abstract prepared for this year’s symposium, I noted the following sentence: “More and more often, solutions to complex human conflict require complex solutions—solutions involving diverse organizations such as police forces, first responders, other government departments, non-government organizations (NGO/charities) and militaries. As a result, the politics of these operations can be Byzantine, the logistics overwhelming, and the moral and ethical considerations dizzying in their implications.”

I think this comment is deeply perceptive. It reflects an understanding of what theorists of complex adaptive systems call “the law of requisite variety”—the proposition that there is a positive correlation between the complexity of a problem, the complexity of actors generating solutions to the problem, and the complexity of the solutions themselves. According to this law, a successful adaptive system must have a repertoire of behaviors at least as wide as the range of behaviors expressed by its surrounding environment. To widen its repertoire of behaviors, in turn, an adaptive system must increase its internal complexity.

Although the premises of complexity theory appear to inform this symposium, elsewhere a conventional mechanistic ontology—and a concomitant “management” approach to policy—underpin thinking about war, peace, and international security. This ontology assumes that conflict systems have relatively easily discernible boundaries, that the behavior of a system is an additive consequence of the behavior of its parts, that effect is proportional to cause, that it is possible to discriminate among multiple causes in terms of their causal power, and that the “gold standard” of explanation involves the identification of a single, necessary and sufficient cause of a given phenomenon.

These assumptions are often invalid. Conflict systems are often fundamentally “complex” in that they are characterized by feedbacks, causal openness, causal synergy (interactivity), disproportionality of cause and effect (i.e., nonlinear behavior), and contingency.

A Complex-Systems Ontology

Let me unpack these concepts.

How do we know a complex system when we see one? What are its key features? First and most obviously, a complex system is composed of a multiplicity of things; it is made up of a large number of entities, components, or parts. Systems with more components are generally more complex than those with fewer. But multiplicity is not sufficient: if it were, a complex system would merely be complicated. In complex systems the multiplicity of parts enables a wide range of potential behaviors, as I indicated above. Machines like windup clocks or car engines may be extremely complicated—they may have thousands of parts—but all their parts work together to produce a system with a relatively narrow and predictable range of behaviors.

The second feature common to complex systems is the dense web of causal connections among their components; in other words, their components have so many links to each other that these components affect each other in many ways. In general, the more causal connections among a system’s components, the greater its complexity. A particularly important result of all this dense connectivity is causal feedback, in which a change in one component affects others in a way that eventually loops back to affect the original component.

Feedback can be positive or negative. Positive feedbacks reinforce or amplify the initial change and are inherently unstable: they create self -reinforcing spirals of behavior. Negative feedbacks, on the other hand, help maintain the stability or equilibrium of a system by counteracting the initial change.

The third feature of complex systems is their openness to outside environments: they are not self-contained, but are affected, sometimes profoundly, by outside events. As a result, it is often hard to locate a complex system’s boundary—that is, the point where the system ends and the outside world starts. Where, for instance, do we draw the boundary between a human society and its encompassing ecosystem of other living things, like the soils in which it grows its food and the forests that help recycle water through its environment?

Complex systems also normally show a high degree of synergy among their components – a fourth common feature. Synergy means, in everyday language, that “the whole is more than the sum of its parts.” Likewise, when we talk about complex systems, synergy means that the combined effect of changes in two or more of a system’s components differs from the sum of their individual effects. In the social sciences, this phenomenon is commonly referred to as causal interactivity.

Fifth, complex systems exhibit nonlinear behavior. Nonlinearity is a notoriously difficult concept that even complexity specialists have difficulty explaining, but in the simplest terms it means that a change in a system can produce an effect that is not proportional to its size: small changes can produce large effects, and large changes can produce small effects—or even no effects at all.

Nonlinear behavior can happen as disturbances or changes in the system, each one relatively small by itself, accumulate. Outwardly, everything seems to be normal: the system does not generate any surprises. At some point, though, the behavior of the whole system suddenly shifts to a radically new mode. This kind of behavior is often called a threshold effect, because the shift occurs when a critical threshold—usually unseen and often unexpected—is crossed.

Conflict systems often exhibit nonlinear behavior. They might evolve slowly over time, with no more than incremental changes in their key components and variables, and then suddenly exhibit a sharp shift in behavior—perhaps the outbreak of a war, rebellion, or genocide—as they cross a critical threshold.

The behavior of a complex system with all these characteristics is highly contingent—a sixth common feature. How the system behaves at any given time, and how it evolves over time, depends on a host of factors, large and small, knowable and unknowable. The further we try to predict into the future, the more bewildering the task of predicting the system’s route becomes.

How might these concepts help us better understand contemporary human conflict and the challenges our societies face in coping with this conflict? I’ll address this question by briefly reviewing two emerging security concerns that fall at the interface between nature and society: first, climate change as a threat to international peace and security and, second, energy scarcity as a deep cause of societal collapse.

Climate Change and Security

Does climate change threaten international peace and security? The British government thinks so. As this month’s head of the United Nations Security Council, the United Kingdom convened a debate on the matter just over two weeks ago. One in four UN member countries joined the discussion—a record for a thematic debate.

Countries rich and poor, large and small, and from all continents—Japan, Bangladesh, Ghana, Mexico, much of Europe and, most poignantly, a large number of small island states endangered by rising seas—recognized the security implications of climate change. Some other developing countries like India, Brazil and Cuba and most of the biggest producers of fossil fuels and carbon dioxide—including Russia, China, and Qatar—either questioned the very idea of such a link or argued that the Security Council is not the right place to talk about it.

But these skeptics are wrong. Evidence is fast accumulating that, within our children’s lifetimes, severe droughts, storms and heat waves caused by climate change could rip apart societies from one side of the planet to the other. Climate stress may well represent a challenge to international security just as dangerous—and more intractable—than the arms race between the United States and the Soviet Union during the Cold War or the proliferation of nuclear weapons among rogue states today.

The US Congress and senior American military leaders are taking heed: Legislation under consideration in both the US Senate and the House calls for the director of national intelligence to report on the geopolitical implications of climate change. And two weeks ago a blue-ribbon panel of 11 retired generals and admirals warned that climate change is already a “threat multiplier” in the world’s fragile regions, “exacerbating conditions that lead to failed states—the breeding grounds for extremism and terrorism.”

Addressing the question of scientific uncertainty about climate change, General Gordon R. Sullivan, a former Army Chief of Staff who is now retired, wrote, “Speaking as a soldier, we never have 100-percent certainty. If you wait until you have 100-percent certainty, something bad is going to happen on the battlefield.”

In the future, that battlefield is likely to be staggeringly complex. Climate change will help produce the kind of military challenges that are difficult for today’s conventional forces to handle: violence in the form of insurgencies, guerilla attacks, ethnic clashes, gang warfare, and terrorism that’s diffuse, chronic, and subnational.

In the 1990s, a research team I led at the University of Toronto examined links between various forms of environmental stress in poor countries — cropland degradation, deforestation and scarcity of fresh water, for example — and violent conflict. In places as diverse as the Philippines, Pakistan, Haiti and South Africa, we found that severe environmental stress multiplied the impact of a society’s existing vulnerabilities, including its ethnic cleavages and skewed distribution of land, wealth, and power.

Rural folk who depend on the local environment for their day-to-day livelihood become poorer, while powerful elites take control of—and extract exorbitant profits from—increasingly valuable land, forests and water. As these resources in the countryside dwindle, people sometimes join local rebellions against landowners and government officials. In mountainous areas of the Philippines, for instance, chronic land degradation has exacerbated poverty and helped drive peasants into the arms of the Communist New People’s Army insurgency.

Other times, people migrate in large numbers to regions where resources seem more plentiful, only to fight with the people already there. Or they migrate to urban slums, where unemployed young men can be primed to join criminal gangs or radical political groups.

Climate change will have similar effects, if nations fail to aggressively limit carbon dioxide emissions and develop technologies and institutions that allow people to cope with a warmer planet. The recent report of Working Group II of the United Nations Intergovernmental Panel on Climate Change identifies multiple pathways through which warming will hurt poor people in the Third World and hinder economic development there more generally. Large swaths of land in subtropical latitudes—zones inhabited by billions of people—will experience more drought, more coastal damage from storms, higher mortality from heat waves, worse outbreaks of agricultural pests and an increased burden of infectious disease.

The potential impact on food output is a particular concern: in semiarid regions where water is already scarce and cropland overused, climate change could devastate agriculture. (There is evidence that warming’s effect on crops and pastureland is a deep cause of the Darfur crisis.) Many cereal crops in tropical zones are already near their limits of heat tolerance, and even a couple degrees of warming could lead to much lower yields.

By weakening rural economies, boosting unemployment and dislocating people’s lives, global warming will increase the frustrations and anger of hundreds of millions of people in vulnerable countries. Especially in Africa, but also in some parts of Asia and Latin America, climate change will undermine already frail governments—and make challenges from violent groups more likely—by reducing revenues, overwhelming bureaucracies and revealing how incapable these governments are of helping their citizens.

At this intersection of climate change and national and global security, we see many of the characteristics of complex systems behavior I spoke of earlier. Climate induced scarcity of water resources will encourage rent-seeking behavior by elites that worsens water scarcity—a positive feedback. Conflict in the social systems in question will be exacerbated precisely because they are open to influence from the larger climate system. Climate change will have its greatest impact by synergistically interacting with societies’ existing social and ecological vulnerabilities. And these synergistic combinations of pressures will tend to have nonlinear consequences through multiple and highly contingent causal pathways. They will accumulate with little obvious effect for long periods of time and then produce—perhaps because of the influence of a local and seemingly insignificant factor of leadership, clan rivalry, or land ownership—a sudden outbreak of violence.

Energy, Complexity, and Societal Collapse

A complex-systems ontology also helps us better understand the links between energy availability, complexity, and societal collapse—the second topic I will use to illustrate my thesis today.

What causes societies to collapse, and are our modern societies at risk of collapse themselves?

In the last few years, these questions have been on people’s minds. In a bestselling book, the American evolutionary biologist Jared Diamond described how past societies collapsed and why today’s societies might share the same fate. Martin Rees, President of the Royal Society of London, has also argued that humankind will confront an unprecedented crisis this century. James Lovelock, the originator of the Gaia hypothesis, has recently declared that damage to Earth’s environment is now so severe that we’ve already passed the point of no return. Calamity is inevitable.

Much of this writing is weakly argued, which makes it easy for our societies’ policymakers, elites and “experts” of various stripes to dismiss it with a patronizing wave of the hand. Jared Diamond, for instance, singles out environmental causes of collapse. He doesn’t fully acknowledge that the greatest risks arise from the simultaneous convergence of multiple stresses—including non-environmental ones. Also, he relies heavily on evidence of collapse from ancient history, without really showing why modern societies aren’t vastly more adaptive and resilient—and far less likely to collapse. Most importantly, he doesn’t really say much about what we should do if some kind of collapse occurs.

All the same I think the deeper intuition of these writers is correct: real trouble does lie ahead.

When a society collapses, it rapidly loses complexity. Its internal organization and institutions, laws, and technologies become dramatically simpler, while its inhabitants’ range of social roles and potential behaviors is sharply narrowed. Many people suffer, because without complex institutions, technologies, and social roles, societies can’t keep large populations living well. After collapse, people consume far less, move around far less, communicate far less, and die far sooner. Simple maxims—such as “might is right”—guide conduct.

At one level, it seems preposterous to say that something like this could happen to our modern societies. So much around us seems so real, consequential, and permanent. We can see with our own eyes skyscrapers in our cities, an endless infrastructure of highways, airports, and shopping malls, and imposing buildings that house private and public institutions like banks, corporations, courts, and bureaucracies. What imaginable forces could humble such wealthy and smart societies?

For a theory of societal collapse better than that proposed by, for instance, Jared Diamond, we have to bring together research from the fields of sociology, anthropology, and ecology. The work of Jack Goldstone, a sociologist at George Mason University and a leading authority on revolution, is a good place to start. Goldstone has shown that societies are far more likely to break down when they’re overloaded by converging stresses—say, rapid population growth, scarcity of key resources, and a financial crisis. “Massive state breakdown is likely to occur,” he writes, “only when there are simultaneously high levels of distress and conflict at several levels of society.”

Converging stresses won’t cause collapse by themselves, though. Something also has to limit societies’ ability to cope, and here the research of the American anthropologist Joseph Tainter is helpful. After studying the histories of many societies, ancient and modern, Tainter has concluded that they generally respond to stress by boosting the complexity of their institutions and technologies. A society dealing with a prolonged drought, for example, might build elaborate irrigation systems so it uses water more efficiently on its farms, and it might create another layer of bureaucracy to make sure everyone follows water-sharing rules. In the short and medium terms, this greater complexity often produces big benefits—like more food—and most people are better off.

But Tainter has also found that greater complexity doesn’t produce benefits forever, because it’s costly. The cost is paid in the currency of energy: energy is needed to build an irrigation system and keep it from falling apart, just as it’s needed to create and maintain an irrigation bureaucracy. “Not only is energy flow required to maintain a sociopolitical system,” Tainter writes, “but the amount of energy must be sufficient for the complexity of that system.”

Lastly, Tainter has determined that investments in complexity eventually produce what economists call “diminishing marginal returns.” This simply means that at some point an additional unit of energy spent on complexity gives us less benefit than the immediately previous unit. Why? Because societies almost always try first those solutions that give the biggest return for the least cost, leaving for later costlier and less effective solutions.

In time, the benefits of greater complexity fall to zero and can even become negative. As an expanding portion of a society’s wealth is sucked into further boosting complexity, its reserves to deal with unexpected contingencies fall, making it more susceptible to sudden, severe shocks from the outside.

Tainter’s theory has gaps: it doesn’t fully explain why societies become more complex even when it doesn’t do them any good; and, most importantly, it doesn’t really tell us why rising complexity’s consequences are sometimes so bad.

So for the final pieces of the puzzle, we have to turn to the work of Buzz Holling, a renowned Canadian ecologist. Since the early 1970s, Holling’s research has attracted wide attention, and he has become something of a guru for an astonishing number of people studying complex adaptive systems.

According to Holling and his colleagues, any living system—from forest ecologies to modern economies—naturally tends to become more complex, internally connected, and efficient over time, regardless of whether it needs complexity to solve its problems. Eventually it becomes so well adapted to a specific range of circumstances—and so well organized as an efficient and productive system—that when a shock pushes it outside that range, it can’t cope. And the system’s high connectedness helps any shock travel farther and faster across the system as a whole. Overall, then, the system becomes more rigid and brittle—in a word, less resilient.

Taken together, what do the theories of Goldstone, Tainter, and Holling tell us about our situation today? Is some form of collapse becoming more likely? Disturbingly, the answer appears to be yes.

In the last half-century, largely because of the enormous growth and relentless integration of the world’s economy, humankind and the natural environment it exploits have evolved into a single “socio-ecological” system that encompasses the planet. This system has become steadily more connected and economically efficient. Partly as a result, a financial crisis, a terrorist attack, or a disease outbreak can now have almost instantaneous destabilizing effects from one side of the world to the other. The system has also developed increasingly severe internal pressures—because of, among other things, gaps in wealth between rich and poor people, worsening environmental stresses like climate change, and the diffusion of technologies for mass violence away from governments to small groups of people (including terrorists).

Managing these pressures demands steadily more complex institutions and technologies and, in turn, steadily higher inputs of high-quality energy. Yet our efforts to manage them aren’t working very well. In Tainter’s terms, greater complexity seems to be producing diminishing returns—it’s not solving our problems. The wealth gaps between us are steadily widening, we’re having little success reducing our output of carbon dioxide, and we see the horrible effects of the diffusion of technologies of violence everyday in places like Baghdad.

At the same time, largely by coincidence, the world is entering a critical transition from an era of abundant and cheap high-quality energy sources (mainly conventional oil and natural gas) to one of scarcer and more expensive, lower-quality energy sources (like solar and wind power and non-conventional oil from places like the Alberta tar sands).

And it’s this contradiction that’s at the crux of our predicament today: our global system is becoming steadily more complex, yet the abundant, cheap, high-quality energy we need to cope with this complexity will soon be less and less available.

Societies without access to enough energy to sustain rising complexity and to manage worsening internal stresses risk the kind of overload Goldstone identifies. They’re more likely to break down—to succumb to economic crisis and political disorder—when they can’t cope with sudden, severe shocks. It’s impossible to say what such breakdown might look like in the future, or where or when it could start. It could be global in its scope—if, for instance, terrorists simultaneously detonated nuclear weapons in two or three of the world’s main financial centers. It’s more likely, though, to proceed in stages from weak peripheral areas—from poor countries already afflicted by strife and dysfunctional governments—towards the global system’s most powerful and complex centers, including North America and Europe.

And here’s the key point: a complex-systems ontology tells us that just because everything looks relatively calm on a society’s surface, we shouldn’t assume that everything is fine underneath. The research of Goldstone, Tainter, Holling and other scholars tells us that incremental, long-term, and largely invisible shifts in a society’s complexity and interconnectedness, in the quality of the resources it depends on, and in its relations with its natural environment can induce huge social stress while they sap social resilience. The society becomes progressively more vulnerable to sudden, sharp bursts of change—threshold events—that can be enormously disruptive.

Co-evolution and a New Paradigm for Problem Solving

What can we do to cope with these enormous challenges? In truth, a great deal. I don’t have time today to review in detail my prescriptions for how we should respond, but I do want to put before you a new way of thinking about the task ahead.

First of all, as we develop our energy, climate, and security policies, we need to reconceptualize the relationship between human society and its larger context of natural systems. As illustrated in Figure 1, our energy, climate, and security policies are embedded in a human society that often responds nonlinearly to both endogenously and exogenously generated pressures. Society in turn is embedded in—and co-evolves with—natural systems that are themselves often nonlinear.

Figure 1: Co-evolution of Society and Nature

Figure 1: Co-evolution of Society and Nature

The key idea here is co-evolution. Matter, energy, and information flow both ways between society and nature and profoundly affect the behavior of both. Most particularly, human societies no longer simply respond to the exigencies of the natural world. They actively shape that world on a macro scale. We’re now so large in our numbers and so powerful with our technologies that we’ve become a planetary force. We have the raw power to alter the basic characteristics of Earth’s biosphere—a life-supporting layer no thicker, proportionately, than the skin of an apple.

Take, for instance, how we changed the planet’s landscape. When we build houses, office towers, and highways, and when we till our fields, excavate gravel, and erect dams, the 6.5 billion of us move countless millions of tons of dirt and rock—an amount that is now about ten times larger than the total moved through the natural action of wind and water around Earth. Partly as a result, we’ve transformed and often degraded about half of the world’s ice-free land. Row-crop agriculture, cities with their suburban sprawl, and industrial zones alone cover about 15 percent of Earth’s ice-free land—an area larger than the United States, Canada, and Mexico combined.

But perhaps nothing better shows how we dominate the biosphere than our use of Earth’s plant energy. In a famous paper published in 1986, the ecologist Peter Vitousek of Stanford University and his colleagues estimated that each year humans and their livestock consume directly (in the form of food, fuel, fiber, or timber) about 4 percent of the energy that land plants trap and store through photosynthesis. Four percent may not seem like much, but then they calculated that in the ecosystems we dominate—like cropland, pastureland, and cleared forest land—we manage or destroy an additional 27 percent. If we add the energy that land plants would have stored if their ecosystems had been left in their natural state, then, Vitousek and his colleagues estimate, we use, waste, or disrupt nearly 40 percent of all the energy that Earth’s land plants could potentially store. The scientists conclude, “An equivalent concentration of resources into one species and its satellites has probably not occurred since land plants first diversified [some four hundred million years ago].”

These perturbations of the natural systems around us—including our simplification of these systems by regulating, overusing, or damaging them—increase the probability that they’ll behave in nonlinear ways, by flipping to new equilibria perhaps not optimal for human wellbeing. Such events may include ecosystem collapses and sudden reorganizations of the air and ocean currents that drive Earth’s climate. Greater likelihood of harmful nonlinear behavior in natural systems in turn boosts the likelihood of harmful nonlinear behavior in human societies—including conflict and even catastrophic collapse. Thus do our globalized human society and Earth’s natural systems now co-evolve.

I have argued recently that social nonlinearities—what I term social “breakdown” in the vernacular—can be productive. Under the right circumstances, they can create both the motivation and opportunity for deep change in entrenched social relations, institutions, and technologies. The central question for our societies is then this: How can we exploit potentially productive nonlinearities in our societies to help us better cope with or adapt to destructive nonlinearities in nature?

The answer to this question isn’t yet clear. But, if there is an answer, I expect it will involve another model of co-evolution—this time the co-evolution of two paradigms of problem solving that we have hitherto regarded as mutually exclusive. I define the key characteristics of these two paradigms in Table 1.

Table 1: Two Problem-Solving Paradigms

Conventional Complex Adaptive
Defining ontology Mechanistic Complex
Social organization Centralized/hierarchical Decentralized/distributed
Competence / Knowledge High, technocratic, explicit Mixed, experiential, tacit
Scale of testing Small number of large tests with high consequence of failure Abundant small scale, safe-fail experimentation
Sources of legitimacy/power Policy communities, management elites Civil society, democratic action, markets
Social location Top Bottom and middle
Goal Optimization of expected utility (according to explicit, well-defined preferences) Satisficing of multiple, often conflicting, and sometimes incommensurable values

The conventional model of problem solving still dominates the elite management culture of Western societies, including our security and defence establishments. To the extent that the complex-adaptive alternative is even recognized within these elites, it is assumed to produce radically suboptimal results. The complex-adaptive alternative is recognized—and often adopted—within much of civil society, but these groups reciprocate the contempt of conventional management elites. Most importantly, both sides assume that the two approaches to problem solving are antithetical—that they cannot be combined or synthesized in any way.

This assumption is, I have come to believe, deeply misguided. As humankind confronts this century the gravest challenges it has ever faced, we must learn how to make these two problem-solving paradigms work together. Our technocratic, top-down management procedures will better learn what approaches to a particular problem have the highest likelihood of succeeding if they incorporate—and perhaps co-evolve with—distributed social networks in which tens of thousands of agents are exploring the landscape of potential solutions. And these networks will explore the solution landscape vastly faster if are willing to draw from conventional management institutions the best scientific and technical knowledge available.

Canadian and human security—broadly defined—urgently requires a radical new pragmatism. It’s too late for the ideological grandstanding that passes for political discourse in our societies much of the time, and, more profoundly, it’s too late for stubborn adherence to the rigid ontology that informs our conventional approach to problem solving. If our children and their children are to have a life worth living, we need to break through these barriers. We need to get as many heads working on our energy, climate, and security problems as possible. We need to initiate an explosion of experimentation, learn from rapidly our failures, and communicate rapidly our successes across broad social networks. And we need to get ready for a future of enormous volatility—and enormous possibility.

Print Friendly