DOZENS OF KILOMETERS beneath the Colosseum, the mighty forces of Earth’s geology are at work. The African tectonic plate, a thick slab of the planet’s crust stretching from the Mediterranean to Antarctica, is driving northward. Long ago, it smashed into the even larger Eurasian plate. This monumental collision crumpled and tore the edges of the two plates along a front line extending the length of the Mediterranean basin. The collision also buckled the crust upward to form mountain ranges like the Alps, Apennines, and Carpathians. Under Italy, the crust has been pushed downward, where it melts into magma and occasionally bursts back to the surface in spectacular volcanoes.1
All this churning beneath the Italian peninsula has left its mark on the civilizations that have come and gone on the land above. It has sometimes brought disaster, as in 79 CE, when Mount Vesuvius’s eruption buried Pompeii. But it has also created a land of fertility, geographic variety, and great abundance. Volcanism forged the stone so characteristic of Rome’s magnificent buildings: subterranean magma heated water deep underground, which then dissolved limestone lying in layers of rock above. As the superheated water bubbled to the surface, it cooled, and its burden of dissolved calcium carbonate precipitated out to form the vast beds of travertine that are still mined in places like Tivoli today.
So it’s a bit ironic that the very seismic activity that generated much of the rock that Romans used to build the Colosseum has also harmed the building—repeatedly and sometimes grievously. In the past two thousand years, more than a dozen significant earthquakes have struck Rome. Experts believe that temblors seriously damaged the Colosseum in 443 CE, 508, 801, 847, 1349, 1703, and 1812. The 847 quake probably destroyed the columns around the upper floor of the amphitheater that supported its roof, and the 1349 quake toppled much of the southern wall that forms the building’s external ring.2 In addition to these disasters, the Colosseum was plundered for its stone for a thousand years, often with the pope’s authorization and even encouragement. The once-grand amphitheater was reduced to a ruin. Restoration began only when Napoleon briefly controlled the city in the early nineteenth century.
If the builders of the Colosseum had known what we know now about earthquakes, they might not have constructed it on its current site. The southern part of the building sits on soft alluvial material that amplifies ground motion, which explains why, over the centuries, that part has sustained far more earthquake damage. Still, while we know much more about earthquakes than the Romans did, we’re little better at predicting when they’ll occur. Taken together, Earth’s geophysical processes are a complex system, and the earthquakes that it produces are awesome examples of “threshold events.”
Deep under Earth’s surface, tectonic plates grind into each other a centimeter or two a year, and enormous pressure builds up incrementally and invisibly. At some point, the pressure crosses a critical threshold, something snaps, and an earthquake occurs. But how and when it occurs—whether, for instance, the pressure is released through a long series of micro-quakes or a Colosseum-shattering disaster— depends on many subtle and often unknowable aspects of local geology: the type and porosity of the rock, the extent and direction of the rock’s fractures and fissures, the amount of water in these spaces, and even the types of compounds dissolved in this water.3
Earthquakes aren’t just geophysical events of special interest to those of us who live on major fault lines—whether in Italy or many other places such as Japan, Indonesia, California, and British Columbia. They are also instances of a more general kind of system breakdown—a kind that’s important to us today, as we try to understand the risk of social breakdown. The population, energy, environmental, climate, and economic stresses affecting our world are just like tectonic stresses: they’re deep, invisible, yet immensely powerful; they’re building slowly; and they can release their force suddenly without warning.
The analogy between earthquakes and social breakdown is surprisingly apt. For example, earthquakes occur when two adjacent tectonic plates run into each other. Sometimes such collisions aren’t catastrophic, because one plate slips beneath or alongside the other without too much trouble. Other times, though, the plates lock together along the fault line between them. Then the pressure building between the plates mounts and is eventually released in one great burst. Similarly, at the human level, our institutions and political and economic systems can lock up or become rigid, which prevents the release of stress and keeps societies from adapting to new circumstances. For instance, racial segregation remained widespread in the United States even after the U.S. Supreme Court declared state segregation laws illegal in the 1950s. Eventually in such situations, frustrations can erupt in events like the U.S. civil rights movement in the 1960s, the recent riots by black and Muslim immigrant communities in French cities, or the popular revolutions in Eastern Europe in 1989—revolutions that were born of people’s deep frustrations and that quickly brought the collapse of the Soviet east bloc.
Seismologists have recently learned that major earthquakes often relieve stress in one region while boosting it in others, so, over time, earthquakes often come in chains, as the first triggers a second, which triggers another and another. The final earthquake in the chain might occur a very long way from the first—in fact, the December 2004 earthquake off Indonesia that caused the devastating tsunami in the Indian Ocean set off smaller quakes as far away as Alaska.4
Seismologists call this phenomenon “stress triggering,” and we can see something similar in social systems: a disturbance in one part of an economy or society can dramatically—and often unexpectedly— increase stress in other parts, leading to cascades of economic or social change. When Hungary’s new reformist parliament tore a hole in the Iron Curtain by opening the country’s borders with Austria in August 1989, tens of thousands of East Germans saw their chance to escape to the West if they could reach Hungary. This precipitated a remarkable chain of events: within months, all border points between Eastern and Western Europe had been opened, the Berlin Wall had been ripped down, and one east bloc regime after another—from Czechoslovakia to Romania—had crumbled.
But there’s one similarity between earthquakes and social breakdown that’s particularly key for us here. Just as seismologists can’t forecast a geological earthquake’s precise timing or character, social scientists can’t forecast the precise timing or character of a major social earthquake. It might happen tomorrow, two decades from now, or never. Yet seismologists can still say useful things about the varying risk of earthquakes from one region to another or from one time to another.5 In the same way, we can say something useful about the risk of sociopolitical earthquakes at different times and places. Understanding these risks gives us a chance to make sensible, even life-saving judgments about which futures are plausible and which are wholly unlikely. The ability to make such judgments is a vital attribute of what I call the prospective mind.
One feature of earthquakes is an especially intriguing tool for prediction—and a potentially powerful one too, because it could help us prepare in advance. About half of all earthquakes are preceded by one or more foreshocks—small tremors that happen up to three days before the quake. “Foreshocks don’t occur before all earthquakes,” notes a leading seismologist, “but they occur commonly enough that [they suggest] a larger event might be coming.”6 There’s a problem, however. As a prediction tool they generate far too many “false positives”—too many false alarms that an earthquake is on its way.7 Still, foreshocks are fascinating, because if seismologists could better understand them, they might be able to warn people to move to safer ground, evacuate vulnerable buildings, or store extra supplies against the devastation that could follow.
Our social systems often exhibit foreshocks too. The Eastern European revolutions of 1989, for instance, were preceded by the rise of the independent trade union in Poland known as Solidarity and by Soviet President Mikhail Gorbachev’s policies of “glasnost” and “perestroika.” In this light, in the early twenty-first century, diverse events like the Iraq war, the 9 /11 attacks, the 2005 urban riots in France, and hurricane Katrina might be seen as foreshocks of a coming global breakdown.
The science of earthquakes, then, can help us understand sharp and sudden change in types of complex systems that aren’t geological— including societies. Future energy scarcity, especially, could shake our world from top to bottom, by making it far harder for us to cope with the other stresses converging on our societies. Peak oil could aggravate the impact of the world’s population imbalances by weakening economic growth and worsening unemployment in poor megacities and in immigrant ghettos in rich countries.
But the analogy between earthquakes and social breakdown has its limits. It focuses our attention on a single dominant cause of breakdown, for example, because an earthquake is caused mainly by one factor—a buildup of pressure along a fault line. In my travels to investigate the problems we face around the world, an experience in California helped me see that if we’re really going to understand social breakdown, we need to grasp what happens when diverse stresses combine—when, in other words, there are multiple causes of breakdown.
In the sun of a late-November afternoon, I was standing on a ridge overlooking the city of San Bernardino, a hundred kilometers east of Los Angeles. It was Thanksgiving Day 2003, just six months after my visit to the Roman Forum.
An American flag flapped in the breeze. It was tied to a length of charred pipe that someone had jammed into a mound of twisted debris. Around this lonely symbol of pride and perseverance were the blackened remains of a refrigerator and a hot-water tank; a naked, ash-covered brick chimney; and the concrete foundations of the house that once stood there. Beyond this perimeter was a forest of dead pine trees, their burned, needleless branches reaching up toward the sky as if appealing for salvation.
Just a month before, wildfires had ripped across Southern California—from the suburbs of San Diego northward three hundred kilometers to Simi Valley. Fires had ravaged the San Bernardino Mountains, including the ridge where I was standing. Hundreds of houses had been destroyed, both in these mountains and within the city’s residential zones nestled below.
Southern Californians live with the constant danger of fire on the outskirts of their cities. After summer’s withering heat, forest and bush are tinder dry, and fierce Santa Ana winds can turn any spark into a conflagration. But firefighters had never seen anything in their lives like this inferno. Flames as high as thirty-story buildings—extending across fronts dozens of kilometers long—raced over hillsides, up valleys, and through suburban tracts, devouring everything in their path. By the time the fires had been put out, over three thousand homes had been destroyed, twenty-four people had died, and the total cost ran into billions of dollars.
Nothing conveyed the pain better than the naked chimneys—eerie sentinels left behind amid the wrecked houses on both sides of me. They must have been grand, these houses perched on the mountain’s edge, boasting splendid views of the city from their decks and patios. At each site, I could see that families had come back to collect a few artifacts of everyday life, carefully plucking memorabilia from the ashes—a pot, a ceramic bowl, the melted remains of a prized chandelier, a vacant photo frame—and putting them to one side in little clusters, to be picked up later.
And then, looking across the city, its dead-straight streets converging to a vanishing point in the distance, I saw an immense brown cloud moving in from the west, as if someone were pulling a dirty curtain across a stage. The afternoon winds were whipping across the charred hillsides, lifting the newly exposed soil high into the air, and churning it into a dust storm.
As I observed this scene, I thought about another Californian conflagration some six hundred kilometers to the northwest and almost a century before—the great fire in San Francisco in 1906 that I described in this book’s opening pages. The two episodes were quite similar in their staggering economic cost.8 And they were similar in another way too, because both, I realized, were examples of negative synergy.
Synergy happens when people, things, or events combine to produce a larger impact than they would if each acted separately. We’ve all heard about the phenomenon; business consultants constantly babble about exciting synergies achieved when people work together on a new project. We tend to assume that synergy is always positive and beneficial, but it can just as easily be negative and harmful. For instance, a person might be able to knock back a few drinks and still drive quite competently under normal weather conditions; the same person, without the drinks, might be able to navigate a car through a sudden snowstorm. But if the person combines the drinks and snowstorm, the result might easily be fatal. Similarly, in the aftermath of the 1906 San Francisco earthquake, sparks and fuel combined to cause an explosion of fires across the city. Either the sparks or the fuel alone wouldn’t have been enough—but together they produced calamity.
Policy makers, social scientists, and commentators almost always overlook the potentially ferocious power of negative synergy, perhaps because they don’t fully understand its implications.9 Yet its implications were clear on the ridge overlooking San Bernardino. A month earlier, when the wildfires were burning, the media had given them lurid coverage, and everyone reacted as if they’d burst from nowhere. But in reality, they were just the latest chapter in a story of converging stresses that had been ignored till then and that has received little attention since.
Three developments combined to cause the conflagrations, any two of which probably wouldn’t have been enough by themselves. First, in previous decades suburban building had expanded into the bushlands and forested zones across California, including the San Bernardino Mountains above Los Angeles. Many people had built their homes in those attractive edge zones where cities meet nature. Second, in 2003, several years of harsh drought—including, in 2001–02, the driest year in more than a century—had desiccated Southern California, sucking the moisture out of the soil and weakening trees and other vegetation. And third, an infestation of bark beetles had exploited the weakened trees and laid waste to vast stretches of pine forest. Today, experts estimate that 90 percent of Southern California pine forests will eventually die, including the entire San Bernardino National Forest.10 Bark beetle infestations are in fact chewing their way through pine forests from Alaska and British Columbia to Arizona. Some experts think these infestations are an early sign of human-induced climate change, because warmer temperatures have extended beetles’ life spans and geographic range.11
I could see the infestation’s grim results. Farther back from the ridge—beyond the point where the fire had finally died out—needles from dead and dying pines were still piled in thick layers across driveways, gardens, and sometimes right up to the foundations of unburned houses. In some places, this layer of explosively inflammable duff was a third of a meter deep. I noticed that the ridge, trees, and blanket of needles were all bone-dry. A single spark would produce another inferno. The people still residing in the houses on that hillside were living inside a giant tinderbox.12
The October 2003 fires were a dramatic breakdown of the human/nature system in Southern California. Two severe, simultaneous, and causally related stresses—drought and a beetle infestation—synergistically combined with ill-advised patterns of residential construction to create a nightmare of wrecked homes, devastated landscapes, and collapsing property values. One might think that such an outcome would encourage residents to question the wisdom of living in such a spot.
Smoke from the October 2003 wildfires in southern California, as seen from space
Yet everywhere I looked new telephone poles had been rigged with power lines and lots were being cleared of debris for reconstruction. What defiance! But then it occurred to me that unlike the Romans when they built the Colosseum, the residents of the ridge now knew for sure that they were constructing their buildings in a danger zone. When, I wondered, does defiance become denial?
When I talk about the breakdown of complex systems like the one along the San Bernardino ridge what, exactly, do I mean? We typically use the word “breakdown” when something’s regular function is disrupted or its movement stops—when we talk about our car’s breakdown, for instance. We also use it to refer to a sudden change in normal affairs. Here I’ll use the word a bit differently—to refer to a rapid loss of complexity. The breakdown of a system—whether an ecosystem, economy, or an imperial government like Rome’s—simplifies its internal organization and reduces its range of potential behaviors.13
Think, for example, of what happens when someone experiences a psychological breakdown. Their activities and goals become much simpler. Life becomes a matter of survival, of satisfying basic needs on a day-to-day basis. The same is true when ecological, technological, and social systems break down. The San Bernardino fires simplified the region’s ecology—wiping out many of its plants, animals, and ecosystems—just as it simplified the human society that lived there (after all, most people left). And during the blackout across eastern North America in August 2003, we lost the communication and transport technologies that make possible our high-velocity interactions and our multiple professional, commercial, and social roles. Almost all business meetings, financial transactions, and shopping excursions abruptly stopped. Our range of options was radically reduced and our goals radically simplified—first and foremost, we wanted to get home. And if we were already home, we wanted to secure basic necessities—water, food, and light.14
People often use the words “breakdown” and “collapse” synonymously. But in my view, although both breakdown and collapse produce a radical simplification of a system, they differ in their long-term consequences. Breakdown may be serious, but it’s not catastrophic. Something can be salvaged after breakdown occurs and perhaps rebuilt better than before. Collapse, on the other hand, is far more harmful: the damage endures— it may even be permanent—and there’s far less knowledge, wealth, or information left behind to use in a process of renewal.
What factors make a society more likely to experience breakdown or even collapse? Historians and social scientists have endlessly debated this question, and they may never reach a conclusion.15 All the same, a large body of research now points to the rough answer I suggested in chapter 1: a society is more likely to experience breakdown when it’s hit by many severe stresses simultaneously, when these stresses combine in ways that magnify their synergistic impact (just as happened in the San Bernardino mountains), and when this impact propagates rapidly through a large number of links among people, groups, organizations, and technologies.
When a society has to confront a bunch of critical problems at the same time, it can’t easily focus its resources on one and then move to the others. Unfortunately, the key stresses that I’ve pinpointed in this book, and that are affecting our world right now, are all getting worse at the same time. Some of them, like the demographic imbalance between rich and poor countries and environmental damage in poor countries, are already severe. Others, like energy scarcity and climate change, while perhaps not yet severe, will probably pass critical thresholds before too long. They may not be bad now, but in the future they likely will be.
At the heart of my argument is the idea of overload. A society overloaded with stresses breaks down.16 Whether a society is overloaded depends not only on the nature of the stresses it encounters but also on whether it can manage or adapt to them. Societies vary a lot in their ability to cope with stresses. At one end of the spectrum, Western capitalist democracies are probably the most adaptive societies in history. During the past century alone, they’ve spent unimaginable quantities of blood and treasure in two world wars, recovered from a deep economic depression, absorbed untold numbers of immigrants, and won a hideously expensive, half-century political competition and arms race with the Soviet Union. They emerged from the century not just intact but wealthier and more powerful than ever.17 At the other end of the spectrum, we find societies— including many in sub-Saharan Africa and some in Asia and Latin America—that have much lower ability to manage or adapt, because of poverty, environmental damage, low education levels, chronic internal violence, and weak and corrupt governments. A few, like Somalia and Haiti, have completely succumbed to the stresses pounding away at them: these countries may still exist as a matter of international law, but they have, for all intents and purposes, ceased to exist as coherent societies.18 In general, the greater the number and severity of stresses affecting a society—and the more they combine synergistically—the greater the chance of social breakdown. But two other factors also affect the impact of stresses on a society. And it was an earlier experience in California that helped me see why these two factors are critically important.
I was driving a brand-new rental car in the fast lane of a divided highway heading south from San Francisco. For once on California’s choked roads, the traffic was moving quickly, and I was speeding along at about 130 kilometers per hour. Less than ten meters behind me, an SUV hugged my bumper, followed close behind by a string of other cars. To my right, between the highway’s shoulder and me, were two lanes of bumper-to-bumper vehicles.
I had pulled into the fast lane to pass a semitrailer, and I was now almost halfway down the rig’s length to my right. But I suddenly realized my car’s engine had gone dead.
I pumped the accelerator. The engine didn’t show a hint of life, and the pedal seemed to have gone to mush. In an instant, I realized I was in a potentially fatal situation. As I shed speed, the SUV closed in behind me; I could see the vehicle’s nose drop as the driver snubbed the brakes. All escape routes to my right appeared blocked. And, without engine power to drive the car’s power steering, my wheel had gone leaden.
I don’t really recall exactly what happened next. I remember realizing that I mustn’t touch my brakes because my momentum was precious—I needed it to get off the road. In any case, there was a tailgater hard on the rig’s heels, which meant I couldn’t get behind the truck. So I use my remaining momentum to overtake the semitrailer and squeak into the small space in front of him. The truck’s horn growled behind me, and all I could see in my rearview mirror was his grille. From there, even though I was losing speed quickly, I made the jump to the right lane and then to the narrow shoulder.
Coasting to a stop, I found that there was barely enough room to get my car off the road. Traffic roared by so close that I couldn’t open the driver’s-side door. I flipped on the emergency flashers, struggled out the passenger’s door, and tumbled down the highway’s steep embankment, my heart racing.
Four hours later, after I’d made many calls to the rental agency and California Highway Patrol, a flatbed truck arrived to take the wretched car away. It all made for a good story when I finally arrived—in a different car and many hours late—at my destination, a conference on Monterey Bay. I wouldn’t have thought much more about the incident if it hadn’t been for something that happened at the conference itself.
It was May 2001, and the conference was an annual event of the Northern California World Affairs Council. About one thousand delegates had converged on the magnificent Asilomar conference center to discuss the dramatic twists and turns of international affairs in the early years of the new millennium. I’d been invited to join a plenary panel on globalization, and I recall that the large hall was packed. The audience wasn’t to be disappointed, because the panelists’ viewpoints differed sharply, and fur was soon flying. Toward the end of the session, one of my fellow panelists, a gentleman who’d made a fortune in Silicon Valley during the dot-com boom, said something that especially caught my attention.
“The more connectivity in our world, the better,” he declared.
In my experience, absolutist, unqualified statements like this one are almost always wrong. And when they’re converted into government and corporate policy dogma—as has happened over the past couple of decades to the general idea that greater connectivity among people, technologies, economies, and societies is always better—we’re in real trouble. There is, of course, no doubt that our world’s soaring connectivity and the ever-greater speed with which material, energy, and information move along our world’s connections can often lead to good things. But sometimes they don’t; in fact, sometimes they multiply a society’s stress. Whether or not this happens depends on the specific characteristics of a society’s connectivity and on the specific characteristics of the stresses the society encounters. On these matters, the devil is in the details.
Together, connectivity and speed are, in fact, the first of the two “multipliers” I mentioned in chapter 1 (the other being the escalating power of small groups to destroy things and people). These multipliers combine with stresses to make breakdown more likely and, when it happens, more disruptive.19
To better understand how and why connectivity and speed work this way, we’ll need to take a brief detour into recent research on the implications of connectivity in complex systems. To begin, it’s useful to think of all complex systems—including the myriad economic, political, social, and technological systems in our own societies—as networks. They consist of sets of nodes and the links connecting those nodes. Nodes can be machines, people, groups, organizations, and even whole countries, while links can be anything that carries material, energy, or information between these nodes. The U.S. economy, for instance, is made up of nodes like corporations, factories, business associations, banks, labor unions, and urban agglomerations; it’s also made up of links among these nodes in the form of fiber-optic cables, electrical grids, gas pipelines, highways, and rail lines.
As societies modernize and become richer, their networks become more complex, interconnected, and faster: they add more nodes, they increase the number of links among the nodes, and they boost the speed at which stuff moves from node to node along the links. The new connections run horizontally as they link similar entities together within a single level of social organization—for instance, people are linked with people, cities with cities, and countries with countries. They also run upward and downward across the levels—for example, a person becomes linked to a city, which is in turn linked to a country, which is finally linked to a vast array of international organizations and institutions.20
Researchers have learned a great deal about the benefits and costs of greater connectivity and speed.21 Greater connectivity often boosts economic productivity by creating larger markets that allow companies to reap the benefits of economies of scale and that encourage people and companies to specialize. Also, when people are better connected, they can usefully combine their diverse ideas, skills, and resources. Greater speed means more goods and services can be produced, transported, sold, and bought in a given period of time. Together, greater connectivity and speed often make economies and societies more resilient to shock because they can respond faster and draw from their larger networks a wider range of skills, resources, capital, and goods and services.
But that’s not the end of the story. At the Asilomar conference, the Silicon Valley entrepreneur’s bold claim struck me as particularly dubious, as I had just experienced the downside of hyper-connectivity and speed on the California highway. We often hear of multi-vehicle pileups caused partly by people following each other too closely and too fast. My car and its surrounding cohort of vehicles—all roaring down the highway together—weren’t physically linked to each other, but we were still connected through information flows (via our eyes and ears) and our mutual vulnerability and interdependence. Our high speed and dangerous nearness tightened this connectivity, so we had almost no room for error, accident, or mechanical failure. In other words, we had little slack or buffering capacity to compensate for surprises. Together we’d become what experts call a “tightly coupled” system.22 When something went wrong at this system’s heart—when my car’s engine failed, for example—the consequences could easily have been catastrophic for many people as well as me.
Tight coupling at high speed contributes to multi-car accidents.
So the first cost of greater connectivity is that damage or a shock in one part of the system—the failure of a machine (like my car engine), the release of a computer virus, or a local financial crisis—can cascade farther and faster to other parts of the system. This is especially true when the nodes in the network, or the elements in the system, are packed so closely together that the links among them are very short—that is, when they’re tightly coupled. Then problems with one node or element can ramify outward before anyone can intervene. Such domino effects happen not only in multi-car pileups but also in telephone, air-traffic, and financial systems—and, as we saw in the 2003 blackout in North America’s electrical grid.
Our world’s tight connectivity also promotes the rapid spread of disease. In fact, we are now seeing a negative synergy between the massive size of the human population and its internal connectivity that helps new diseases—like HIV/AIDS, severe acute respiratory syndrome (SARS), and perhaps soon avian influenza—develop and propagate around the planet faster than ever before. Collectively, humankind now makes up one of the largest bodies of genetically identical biomasses on Earth: all of us, taken together, weigh nearly a third of a billion tons. Combined with our proximity in enormous cities, and our constant travel back and forth across the globe, we’re now a rich environment— just like a huge Petri dish brimming with nutrients—for the emergence and spread of disease.23
Connectivity harbors other risks too. As we create more links among the nodes of our technological and social networks, these networks sometimes developed unexpected patterns of connections that make breakdown more likely. They can, for instance, develop harmful feedback loops—what people commonly call vicious circles—that reinforce instabilities and even lead to collapse. A stock market crash or financial panic is such a vicious circle, because selling drives down prices, begetting fear in the market and more selling, which lowers prices even more.24
Also, the new links we create can connect previously separate systems or system components, so failures or accidents that would before have been isolated from each other now combine in unexpected and harmful ways.25 This is the best explanation for the terrible Three Mile Island and Chernobyl nuclear-reactor accidents and the Challenger and Columbia space-shuttle disasters. In each case, while the designers and managers may have understood the system’s bits and pieces, they didn’t fully understand what could happen when all the bits and pieces combined—in other words, they didn’t really understand the system as a whole. The systems became so complex and interconnected that they exhibited emergent properties—the whole became more than the sum of its parts. So those in charge didn’t anticipate all possible combinations of component failures or possible negative synergies of combined failures, and tragedy was the result.26
These were the roots, too, of the 2003 blackout: the deregulation of the North American power grid in the 1990s caused long-distance electricity sales to skyrocket, which stimulated a surge in connectivity between regional electricity production and distribution systems that had previously been isolated from each other. At the time of the blackout, the new integrated system included six thousand power plants run by three thousand utilities overseen by 142 regional control rooms. The intricate rules developed decades earlier to manage a grid in which most power was generated reasonably close to its consumers were suddenly obsolete.27 Now technicians had to have, as one expert said at the time, the reflexes of “a good combat pilot managing an aircraft that has been badly damaged.” On the day of the blackout, even those reflexes weren’t good enough.28
Although researchers have mainly focused on technological systems, this kind of synergistic failure is just as likely in highly connected social systems. In fact, the dividing line between technological and social systems is never distinct, and almost all the systems we rely on—from electrical grids to banks to governments—are intricate combinations of machines, people, and organizations. In a world of ever-increasing connectivity and speed, unanticipated interactions among previously separate systems happen more often, as do unanticipated combinations of failures within systems. And the likelihood rises that some of these combinations will cause catastrophe.
Greater connectivity harbors another peril too: it can increase our vulnerability to terrorism. To see why, we need to delve just a bit deeper into the new science of networks.
In recent years, scientists have discovered that there are two main types of network. Specialists call them “random networks” and “scale-free networks,” for technical reasons we needn’t go into here. For our purposes, the critical thing is that these two types have different patterns of links connecting their nodes.29
A random network looks like the U.S. interstate highway system, in which the nodes are cities and towns, and the links are highways. In such networks, most nodes have a moderate numbers of links to other nodes, while a few nodes have a small number of links, and a few others have a large number of links. No node has a very large number of links to other nodes. A scale-free network, on the other hand, looks a lot like the U.S. air-traffic network in the 1990s.30 Most nodes have a very small number of links, while a very few, called hubs, have a huge number of links to other nodes.31 Although researchers long assumed that most networks were like the interstate highway system, recent study shows that a surprising number of the world’s networks—both natural and human-made—are more like the air-traffic system. These scale-free networks include most ecosystems, the World Wide Web, large electrical grids, petroleum distribution systems, and modern food-processing and supply networks.32
Such differences have huge implications for the resilience of the network. In a random network the loss of a small number of nodes can cause the overall network to become incoherent—that is, to break into disconnected subnetworks. In a scale-free network, such an event usually won’t disrupt the overall network because most nodes don’t have many links. But there’s a big caveat to this general principle: if a scale-free network loses a hub, it can be disastrous, because many other nodes depend on that hub. An ecosystem, for example, will have a certain number of “keystone species”—species that provide vital services, like pollination, to a wide range of other species—and these keystone species are essentially hubs in the ecosystem’s larger network of species. If enough of these hubs are lost, the ecosystem can collapse.33
The world’s networks can be divided into two general categories.
Scale-free networks are particularly vulnerable to intentional attack: if someone wants to wreck the whole network, he simply needs to identify and destroy some of its hubs. And here we see how our world’s increasing connectivity really matters. Scientists have found that as a scale-free network like the Internet or our food-distribution system grows—as it adds more nodes—the new nodes tend to hook up with already highly connected hubs. New e-mail users on the Internet tend to connect themselves to well-established server computers, and new farms tend to sell their food to large and already dominant wholesale food distributors. Also, as already existing nodes create more links among themselves, they’re more likely to attach themselves to the biggest and most connected hubs. We can see this most clearly on the Web: dominant Web pages—including massive hubs like amazon.com, google.com, or nytimes.com—tend to receive the large majority of all new links from other pages. So in a scale-free network, greater connectivity tends to make already dominant hubs even more dominant. In doing so, it can actually make the network more vulnerable to attacks directed against those hubs.
Once a hub is damaged, intentionally or otherwise, its problems can spread quickly far and wide. For example, cities are hubs in the global network of human population, and they can play a powerful role spreading new diseases. The virus that causes SARS first appeared in early 2003 in Guangdong in southern China—a relatively poor, densely populated region where some people live in close contact with animals that are sometimes diseased. From there, wealthy and mobile residents and visitors carried it to Hong Kong. Then other travelers carried it on planes to cities around the world, including Toronto.
The way SARS spread shows what could happen with tomorrow’s new pathogens. A key danger is that a highly contagious and virulent new disease—something like SARS, except worse—will work its way into the densely packed squatter settlements of the megacities in poor countries. In the vast slums of São Paulo, Delhi, Dacca, Calcutta, Lagos, and Mexico City, health-care facilities are rudimentary. Respiratory and intestinal infections are already endemic, in part because of appalling urban air and water pollution, so selective quarantine is impossible. And members of local elites, who regularly travel around the world, often live nearby. Once a tenacious pathogen is established, poor urban zones could become what specialists call “epidemiological pumps”—that is, they would be hubs that are a permanent disease reservoir and continually pump the pathogen back into the planet’s human population.34
Almost certainly, malicious individuals and small groups, including terrorists, are starting to understand how to exploit our interconnected and high-velocity networks to multiply their disruptive power.35
Sometimes we see the results of this multiplication effect in a very personal way—especially when we have to deal with viruses on our home or office computers: a single kid hacking away in his basement can cause disarray in computer systems around the world. Yet this kind of thing is a minor nuisance compared with the chaos that could follow a major attack on the scale-free networks that provide us with the goods and services we need to survive—especially our energy, information, and food.36 Our energy systems, which encompass everything from networks of gas pipelines to the electricity grid, have countless hubs such as oil refineries, tank farms, and electrical substations, many of which are easily accessible. When energy demand is at peak, as it often is in air-conditioned North America in mid-summer, the electricity grid especially is extremely tightly coupled. An attack against key hubs— such as a number of substations—could have a devastating impact.37 Our food-supply system, too, has countless hubs, including huge factory farms and food-processing plants, with many connections to other nodes and then to all of us who eat the food. Unprotected grain silos dot the countryside, and trains made up of railway cars filled with grain often sit for long periods on railway sidings. Attackers could easily break into these silos and grain cars and lace the grain with contaminants—which would then diffuse through the food system.
But we shouldn’t exaggerate the risks. Terrorists have to be clever to exploit the vulnerabilities of our networks. They have to attack the right hubs in the right networks at the right times, or the damage will remain isolated and the overall network will be resilient. We can also protect ourselves by introducing ways of tracing attacks, as we do already by using batch numbers on drugs and food products; these numbers help authorities find items that have been contaminated and identify people who had access to them. Another protective factor is a network’s redundancy—that is, its ability to offload the functions once served by damaged hubs onto undamaged hubs. But strategies like improving tracing and redundancy aren’t foolproof, and the incentives for terrorists are large: if they succeed in attacking our complex networks, they can spark cascading failures causing immense hardship.38 In fact, according to Langdon Winner, a theorist of politics and technology, the first rule of modern terrorism might be “Find the critical but nonredundant parts of the system and sabotage … them according to your purposes.” Referring to the great Prussian military theorist Carl von Clausewitz, Winner concludes that the science of complexity awaits its own Clausewitz “to make the full range of possibilities clear.”39
We also create extraordinarily attractive targets for attack when we concentrate high-value assets—people included—in geographically small locations. When we build larger factories, we generally lower cost per unit of output, and when we concentrate expensive equipment and highly skilled people in one place, we can access and use them more easily and efficiently. Bringing them all together also creates those valuable synergies among people, things, and ideas that are an important source of our wealth. This is one reason we build places like the World Trade Center.
But terrorists are learning they can cause a huge amount of damage in a single strike against such a target. On September 11, 2001, a complex of buildings that took seven years to build was destroyed in an hour and a half, killing almost three thousand people, obliterating fifteen million square feet of office space, and exacting upward of $20 billion in direct costs. A major telephone switching office was destroyed and another heavily damaged, while some key transit lines were buried under rubble. The attack also immobilized the Bank of New York, one of two banks that provide clearance services for Wall Street’s fixed-income transactions, and an institution vital to the market for U.S. Treasury securities.40
Yet, despite the horrific damage to the area’s infrastructure, the immediate operation of the financial system, and the New York economy, the attack did not cause catastrophic failures in U.S. financial, economic, communication or other networks. The World Trade Center may have been an attractive target, but it was not—as it turned out—a critical, nonredundant hub.
The attack nevertheless had a crucial effect on another kind of network: a tightly coupled and sometimes unstable psychological and emotional network. We’re all nodes in this particular network, and the links among us consist of Internet connections, satellite signals, fiberoptic cables, talk radio, and twenty-four-hour seven-days-a-week news television and radio. In the minutes following the attack, information about it flashed across this network. We then sat in front of our televisions for hours on end; we plugged phone lines checking on friends and relatives; and we sent each other millions upon millions of e-mail messages—so many, in fact, that the Internet was noticeably slower for days afterward.
Along these links, from TV and radio stations to their audiences, and especially from person to person through the Internet, flowed raw emotion—grief, anger, horror, disbelief, fear, and hatred. It was as if we were all wired into one immense, convulsing, and reverberating neural network. Thanks to new communication methods such as the Internet and twenty-four-hour-news television, we’ve created a psychological network that acts like a huge megaphone, vastly amplifying the emotional impact of terrorism. As a result, the ultimate, indirect effects of a terrorist act can now be incomparably greater than its direct effects.
So, the biggest impact of the September 11 attacks wasn’t the direct disruption of America’s financial, economic, communication, or transportation networks—physical things, all. Rather, by working through the psychological network we’ve created among ourselves—a network that extends around the planet—the attacks’ biggest impact was their shock to our subjective feelings of security and safety. Such shocks don’t remain subjective: they soon have huge, real-world consequences. Among other things, when we’re scared, insecure, and grief-stricken, we don’t buy a lot of stuff. We aren’t ebullient consumers. Instead we behave cautiously, and we save more. Consumer demand drops, corporate investment falls, and economic growth slows.
In the end, thanks to the multiplier effect provided by our highly connected information and emotional networks, the 9/11 terrorists may have achieved an economic impact far greater than they ever dreamed possible. The total cost of lost economic growth and decreased equity value around the world ultimately exceeded $1 trillion—and that total doesn’t even include the increased spending on security measures and the later Afghanistan and Iraq wars.41 Since the cost of the attack on the World Trade Center to Al Qaeda was probably only a few hundred thousand dollars, the terrorists multiplied their impact well over a million-fold.
We could see the same amplification effect in the American public’s response to the September and October 2001 anthrax attacks. Only five people died and fewer than two dozen were infected, probably as a result of no more than ten anthrax-laced letters mailed through the postal system. Yet the 9/11 attacks had so frightened the media and public that they reacted almost hysterically to each piece of news about anthrax. Again, we can be sure that terrorists around the world took note.
“If there’s another major attack, people will leave the city in droves.”
Andrew, a colleague of mine in New York City, was sitting in his office in a building not far from Grand Central. It was October 2001, and I’d phoned him from Canada to discuss some business. But our conversation quickly turned to the city’s fevered mood. After the attack on the World Trade Center and the string of anthrax letters, New York’s normally thick-skinned inhabitants seemed near their breaking point.
Of course, another attack did not occur, so we’ll never know just how close New York’s citizens came to leaving the city en masse. But Andrew clearly had the idea that social breakdowns are like earthquakes: stress accumulates over time, then some kind of external trigger releases the stress to produce a sudden reorganization of the system. The psychological pressure on the people of New York, Andrew implied, had reached just such a threshold.
As stress accumulates in a system, there’s a dangerous buildup of something akin to what physicists call “potential energy.” The tectonic plates grinding into each other along the coast of California or under the Mediterranean basin store enormous potential energy in the ground. In the same way, prior to the 2003 California wildfires, drought and infestation combined to cause a buildup of highly inflammable material in residential areas—material, again, with huge potential energy. The accumulated energy may eventually be released in a single mammoth event—as happened in the earthquake that devastated San Francisco. Or it may be released in a series of smaller, rapid-fire shocks. Such staccato shocks—any one of which could be managed on its own—can overload a system and causes its breakdown, just the way a boxer sends his opponent reeling with a series of sharp blows.
This kind of thing is actually quite familiar to most of us, although probably not personally. Most of us know of someone who, through bad planning, carelessness, or simply sheer misfortune, has confronted a string of life crises in rapid succession. They might have lost a job, suffered a severe illness, and had their marriage fall apart—all around the same time. Perhaps these events were causally linked (the loss of the job might have precipitated the illness and the marriage problems), or perhaps they were entirely independent. In either case, the result has been catastrophic: although the person might have coped with two of these events, the combination of three pushed him or her over a threshold. Such unlucky people can end up on the street, without social support and without a home.
Some social scientists propose something similar to explain large social crises, including history’s great revolutions in England in the mid-seventeenth century, in France in the late eighteenth century, and Russia in the early twentieth century. The scholars argue that revolutions happen when inflexible societies experience multiple shocks—or body blows—at many levels simultaneously or in quick succession. As Jack Goldstone, one of the world’s leading theorists of revolution writes, “Massive state breakdown is likely to occur only when there are simultaneously high levels of distress and conflict at several levels of society—in the state, among elites, and in the populace.”42 In revolutions, he writes, “there is a crisis of national government, but there are also crises of local government. There are conflicts with the state, but also regional conflicts and even conflicts within families. There are elite rebellions, but also a variety of rural and urban popular movements.”43
Accumulated stress can also, over time, make a system less supple or resilient. In other words, just like a stick that has dried out and become brittle, the system becomes more likely to break if exposed to too much outside pressure or a sudden sharp shock. Ecosystem collapse often occurs this way: when pollution, overharvesting, or some other long-term stress severely damages a fishery, forest, or grassland, the ecosystem can lose much of its resilience and become more vulnerable to breakdown.44 The overfishing of the great cod stocks off eastern North America in the early 1990s, for instance, weakened the stocks so much that they may not have been able to cope with otherwise normal fluctuations in ocean salinity and temperature.45
In history, we can find many examples of civilizations that have been pushed over the edge to collapse by the combination of multiple stresses and weakened resilience. A good example is classic Mayan civilization, which flourished in and around Mexico’s Yucatán peninsula from the third to the eighth centuries. In this case, a key long-term stress was an expanding population that became too large for the region’s resource base, given the limited productivity of Mayan agriculture. Once peasant farmers had fully exploited the fertile valleys, they moved up the nearby hillsides, cutting down the forests to get wood for fuel and land for farming. But the hillsides’ soil was thin, acidic, and soon depleted of nutrients. And because it was no longer stabilized by the root systems of forests and vegetation, it washed down into the valley bottoms, plugging drainage and irrigation systems and lowering crop yields.
The scarcity of good cropland and the loss of food supplies brought about constant warfare among the Mayan kingdoms. Rulers and political elites focused on short-term gains: enriching themselves, extracting wealth from peasants, fighting each other, and building monuments to their glory. Then another potent stress was added to this mix: between 750 and 800 CE, Central America was hit by the most severe and prolonged drought of the millennium. But by this time overpopulation, cropland and energy scarcity, and chronic warfare had so debilitated many of the kingdoms that they had little reserve capacity to cope. Some kingdoms imploded, and eventually the region’s population declined by over 90 percent.46
Rome is another example of a civilization that succumbed to multiple stresses and weakened resilience. Some scholars argue, however, that such a conclusion about the fall of Rome is fundamentally flawed because the western empire never really collapsed—it merely transformed itself into something else.47 Much Roman culture and some Roman institutions sustained themselves (via, among other things, the Holy Roman Empire) through the Middle Ages and even up to the modern era. Also, these scholars continue, we can’t really fix the exact peak of Roman power, wealth, and accomplishment, so we can’t establish the precise start of its decline. And if getting the timing right is hard, so is specifying the geographical location of Roman decline: during periods when parts of the empire were clearly in trouble, other parts were often expanding and prospering.48
Yet if collapse involves a severe loss of complexity, it seems that in some ways Rome did indeed collapse. Recent archaeological excavations show that, particularly in the west, social complexity ebbed, as did urban populations (as we saw in chapter 3), large-scale administration of people and territory, and interregional travel and trade.49 People often reverted to simpler technologies: for instance, they stopped using wick lamps and returned to the open-saucer lamps common in Paleolithic times.50
By modern standards of social change, though, this transformation didn’t happen fast. We can’t point to an episode or moment when the western empire experienced a sudden, sharp failure in structure, organization, and function. If we measure the process by the decline in the empire’s total territory—a reasonable gauge of its subsiding administrative capacity—the process unfolded over nearly a century.51 As was true in the waning days of many ancient polities, the western Roman empire’s fall was a slow-motion crisis (which is why many scholars prefer to call it a decline rather than collapse).52
But then, by modern standards, almost everything happened far more slowly in Roman times. The very slowness of Rome’s crisis underscores a lesson for us today. If social breakdown occurs now—whether it encompasses one country, many countries, or the whole world, and whether its consequences are moderate or severe—we can be sure it won’t unfold at the same leisurely pace as seventeen hundred years ago. The underlying mechanisms may be the same—a combination of accumulated stresses, weakened resilience, and multiple shocks. But today our global social, technological, and ecological systems are so tightly linked together, and they now operate at such velocity, that the duration of any future breakdown or collapse is likely to be dramatically compressed.
If we use the contraction of an empire’s total territory as a crude indicator of the empire’s decline, we find that ancient and early-modern empires usually (though not always) declined over centuries. Empires as diverse as the Islamic caliphate, the Ming dynasty, and the Mongol empire took about two centuries to vanish, and Byzantium declined over a thousand years. In striking contrast, almost all modern empires seemed to disappear virtually overnight: the Spanish, Ottoman, French, British, and Soviet empires all disintegrated within decades. Sometimes, as in the case of the Soviet empire, they vanished in a few years.53
We don’t usually mourn the disappearance of empires.54 Most people assume it’s a good thing when colonizing societies vanish, for the freedom of the subjugated peoples if not for their prosperity. Still, the evidence of the quickening tempo of imperial breakdown hints at a critical conclusion, and one particularly germane to our own time: as human societies’ connectivity and speed increase, social breakdown, when it does happen, generally happens faster.
Modern empires have collapsed more quickly than early-modern empires.
Our societies’ rising connectivity and speed have a final disturbing effect. In the past, cascading failures usually occurred within single systems—like electrical grids or banking systems—but now these failures are more likely to jump system boundaries. If, for example, terrorists use a new genetically engineered organism to contaminate a Western country’s food supply, the disruption will spread in a flash beyond the food system to our larger economic and political systems. Because today’s communication technologies vastly multiply our emotional reaction to shocking events—something we saw in full force in the wake of the 9/11 attacks and anthrax letters—this kind of terrorism could easily cause a financial panic and even civil disorder, despite the fact that the threat posed to any one person from the attack might be very small. The scale, connectivity, and speed of our modern food-supply system could also spread a new bio-terror organism far and wide before authorities have figured out what’s going on, increasing the chances that it would jump from the food system to affect ecological systems in nature.
The 2003 SARS outbreak may have been a foreshock of this kind of boundary-crossing phenomenon. The virus swept around the world in weeks and sparked instantaneous social and economic emergencies in places as diverse as Hong Kong, Singapore, Vietnam, and Toronto—the almost immediate cost to Asian economies, largely because of a sudden drop off in air travel, was between $11 and $18 billion.55 With SARS we were lucky because the virus, contrary to popular perception, wasn’t very contagious or virulent; next time we may not be so lucky.
If our societies are already brittle because accumulating stresses have eroded their resilience over time, what starts as a local and seemingly manageable breakdown could jump boundaries and quickly spread around the globe, and might even trigger a collapse of global economic and political order. Such an outcome would be a tangible example of what, in chapter 1, I called “synchronous failure”—an event caused by multiple, simultaneous, and synergistic stresses that together generate multiple, simultaneous, and synergistic failures. We can’t see the future, of course, so we can’t possibly know whether such a thing might occur and, if so, what exactly it would look like. In fact, the whole notion might seem pretty outlandish. But then again, we’ve never before lived in a world like the one we’ve created today—in which we can disrupt the planet’s most fundamental natural processes, carry a new disease to distant continents in days, and move terabytes of information across the planet in a second—and in which half a dozen of us, with the right materials, could destroy an entire city.
“It all looks beautifully obvious—in the rear mirror,” wrote the novelist and social philosopher Arthur Koestler. “But there are situations where [one] needs great imaginative power, combined with disrespect for the traditional current of thought, to discover the obvious.”56