5

COOPERATION WHEN IT COUNTS

In 1906, British physiologist and pharmacologist Sir Henry Dale made a most unusual discovery. He had recently started experimenting with the liquefied extract of the human posterior lobe of the pituitary gland, the kidney-bean-sized organ at the base of the brain. Dale suspected that the gland’s hormonal secretions played an important role in the onset of labor in pregnancy, so on a hunch he injected the extract into the uterus of a pregnant cat.

Dale’s hypothesis proved true and then some: His feline subject delivered her litter of kittens in record time. Appropriately, he named the drug using a combination of Greek words: “fast” (oxy) and “birth” (tocos). Sir Henry Dale had discovered the hormone oxytocin.

When word got out about Dale’s discovery, a whole industry of labor-inducing drugs was born. Slaughterhouses began to harvest the postpituitary extracts of cattle and sell them to physicians under the name Pituitrin. When adverse reactions were reported from Pituitrin, pharmaceutical companies took over and started developing chemically purer formulas. By 1953, chemist Vincent du Vigneaud, riding a wave of acclaim after successfully synthesizing penicillin, used his knowledge to fully isolate and synthesize Dale’s “fast birth” discovery: The chemistry of birth would now be available in IV drip form. By the 1960s, physicians were administering oxytocin to millions of pregnant women under a pharmaceutical name familiar to anyone who sets foot in a hospital delivery room today—Pitocin.

For many decades, it seemed that oxytocin’s main claim to fame would be its role in bringing babies into the world. More recent findings, however, reveal this hormone has a much bigger and more nuanced role to play in human nature. Oxytocin is among the most important neurochemicals involved in our experience of two linked human traits that are essential for our survival: trust and cooperation.

As we will see throughout this chapter, resilience is predicated on trust in a system, allowing potential adversaries to move seamlessly into cooperative mode—and quickly, during the moments when it counts the most.

33 LIBERTY STREET

On Friday, September 12, 2008, the treasury secretary of the United States, Henry “Hank” Paulson, and a team of officials from the Federal Reserve called thirty or so of Wall Street’s most powerful executives to the Fed’s offices in Lower Manhattan for a secret evening meeting. Limousines, town cars, and taxis pulled up to the giant slab of white stone at 33 Liberty Street just before six o’clock. Some of the lower-level bankers got out in front of the building, making their way through the onslaught of press. Savvier senior executives opted for the building’s underground garage, to avoid the scrum of reporters.

There was good reason for their low profiles: All at once, and with a frightening speed and apparent randomness, seemingly unassailable emblems of the U.S. financial industry were being felled like giant sequoia in the forest. Just one week earlier, Washington had seized control of the failing mortgage giants Fannie Mae and Freddie Mac. Months prior to that, government officials had brokered a deal to save Bear Stearns, agreeing to put up $30 billion to help complete an acquisition of the failing bank. Bankers in the business for twenty and thirty years described the sequence of events leading up to the September 12 weekend as “scary,” “terrifying,” and “extraordinary.” Alan Greenspan, the former chairman of the U.S. Federal Reserve, called the financial crisis the worst of his career “by far.” Suddenly, even monolithic institutions like Merrill Lynch—the populist brokerage outfit with branches in almost every city in America—seemed dangerously close to collapse.

None more so than Lehman Brothers, the storied firm that had suffered a slow but steady decline over the course of the past year. As we discussed previously, Lehman was heavily exposed to troubled real estate investments—caught like so many others in the subprime mortgage debacle described in chapter 1, and though the firm’s leaders had tried to raise fresh capital, they had made little headway. By Monday, September 8, Richard Fuld, Lehman’s CEO, received the proverbial nail in the coffin: A possible rescue plan that Lehman had been cobbling together with a Korean bank had collapsed. When the market caught wind of the news, Lehman’s shares plummeted 41 percent—this after falling 90 percent from a peak at the beginning of 2008.

By Friday, September 12, Paulson and his team—including then head of the New York Federal Reserve Tim Geithner and Federal Reserve chairman Ben Bernanke—were trying to navigate the waves of panic roiling the marketplace. Despite their previous bailouts over the course of the year, the federal government wanted all the bankers to understand that they were indeed willing to let Lehman Brothers fail.

At 6:00 p.m., the group of twenty or so financial titans—including John Mack, chief executive of Morgan Stanley; John Thain, head of Merrill Lynch; Vikram Pandit, chief executive of Citigroup; Lloyd Blankfein, of Goldman; and Jamie Dimon, head of JPMorgan Chase—were all settled in and seated around the table. Paulson declared his position to the players: It was time to reestablish moral hazard. The bankers would have to clean up their own mess. With no political will for another bailout, the only way to save Lehman Brothers—and the stability of the market, if not global capitalism itself—would be through the collective effort of the bankers sitting around the table. “You have a responsibility to the marketplace,” Paulson told them.

A show of such large-scale cooperation was not unheard of. Most of the bankers around the table remembered an eerily similar message delivered ten years earlier by Geithner’s predecessor, William McDonough. In 1998, the lauded hedge fund Long-Term Capital Management had collapsed, threatening to send cascading defaults throughout the markets. At McDonough’s urging, a group of financial titans, including some of the same names and faces that met at 33 Liberty Street, ultimately agreed to join forces to bail out the firm, saving the entire financial system from destructive turbulence in the markets. Over the September 12 weekend, however, the answer would not be nearly so straightforward.

To begin with, no one was entirely certain of the government’s resolve. Although Paulson and Geithner laid out their case adamantly, there was still some speculation that the Friday night session might really be a big game of poker. Were they bluffing? Surely, if push came to shove, Paulson and his team would blink and the threat of market discipline would give way to a quick infusion of cash. After all, they had once offered up the same claims of austerity for Bear Stearns before ultimately caving in. When all the posturing and grandstanding over moral hazard was over, wouldn’t Paulson realize that Lehman Brothers was simply too big to fail?

Not only was there a sense of distrust in the government’s position, there was also a sense of distrust in the current state of the market system. CEOs like John Mack of Morgan Stanley and Lloyd Blankfein of Goldman Sachs, who, only a month earlier, were considered immune to the toxicity of the subprime mortgage debacle, had found themselves vulnerable. The global financial markets were beginning to look like a line of precariously placed robust-yet-fragile dominos, one default pushing down another in a chain of interlinked interests and overleveraged debt packages, failures followed by bad faith followed by more failures. On Friday, September 12, the bankers did not fully understand how Lehman Brothers’ toxic debt fit into the greater financial banking ecosystem. Even more frightening, no derivative formula or whiz-kid risk assessment manager could tell them how their own banks were doing in the midst of such dramatic and turbulent change. The players sitting around the table at 33 Liberty Street didn’t just distrust the word of the government; they were having a proverbial crisis of faith in the whole system of global capitalism. Were the gods of the market dead?

By 7:00 p.m. on Friday, the government officials were laying out a few potential courses of action for the bank CEOs. One involved an orderly dismantling of Lehman that would allow the firm to die through bankruptcy. For purposes of simplicity, we’ll call this option Let Lehman Fail.

Paulson and the regulators then offered an alternative: What if the Wall Street firms joined forces to take on Lehman’s most toxic assets? The failing bank was a bit like a sports car that sat still in the driveway: It impressed from a distance but one peek under the hood revealed a lemon of a motor. What if they all worked together to make Lehman’s motor look as good as the rest of its body? This option would call for the banks to form a team, pool their resources, and, ultimately, sweeten the deal for an eventual Lehman buyer. We’ll call this option Teamwork Saves the Day.

This latter option would not come cheap. Though the costs would be shared among the players, the effort would require underwriting $85 billion of Lehman’s most toxic assets—a pretty penny even in the best of times, never mind amidst the gathering financial crisis. More important, it would demand deep cooperation from the players—executives who, under normal circumstances, were fierce rivals. Could the bankers in the room at 33 Liberty muster the strategic presence of mind to join forces to save Lehman Brothers—and themselves? In a crisis of faith, would they be able to make a choice that served not their own best interests but the interests of the whole system?

The government emphasized the urgency of the situation. Virtually all of the institutions in the room were connected either directly or indirectly to Lehman’s toxic assets; if Lehman went down, the hairball of its interconnected liabilities would ensnare the market as a whole. Timing was also of the essence: A decision needed to be made in a matter of days. As the sun started setting, Paulson’s gravelly voice hung in the air: “Everybody is exposed.”

But lingering doubts and questions remained: Both Bank of America and Barclays had expressed interest in acquiring Lehman Brothers at a fire sale price, and both of them had representatives present in the room at 33 Liberty Street. That left a bad taste in the mouths of the other CEOs. First, they were being asked to pony up a lot of cash for someone else’s lemon of a motor and then, to add insult to injury, when the whole crisis was over, they would be handing over a fixed-up sports car to Bank of America or, even worse, Barclays: a direct competitor that was not even American.

By 8:00 p.m., the initial meeting was coming to a close. Paulson and his team had presented the options. The Wall Street executives had asked their questions. And now all of them had the night to think over their respective positions.

At 9:00 a.m. on Saturday morning, a most dangerous game of “Will We Save Lehman Brothers?” would officially begin.

•     •     •

It’s worth taking a step back to look at the structure of the meeting at 33 Liberty Street described thus far. Structures create their own behavior, so it’s no surprise that a meeting called in secret, convened under unclear rules, and rooted in a lack of transparency about each one of the individual banks’ books would breed suspicion and distrust.

What’s more, with the prospect of looming carnage, there was still the lingering possibility that one or more of the banks could really clean up on the deal. In the midst of losers, the bankers thought, there might also be some winners. But only if they played their cards right.

With such a start to the weekend, Teamwork Saves the Day was the underdog option. And yet the urgency and magnitude of the crisis had everyone’s attention. If ever there was a day for inspired leadership to guide the emotions in the room, Saturday September 13 was it.

•     •     •

On Saturday morning, the CEOs and their closest advisers were ferried back to 33 Liberty Street in their town cars. Once again upstairs in the meeting room, the bankers broke up into groups to discuss the options placed before them, Let Lehman Fail and Teamwork Saves the Day. One Teamwork scenario group discussed the possibility of having every bank borrow from the Federal Reserve under the emergency lending provision it had recently started to offer. With their newly borrowed cash, the banks would be able to buy up Lehman’s toxic debt, preventing it from bankruptcy and protecting themselves from exposure to the company’s failing assets.

Another Teamwork group focused on helping Barclays or Bank of America buy Lehman’s good assets, placing the toxic debts in a bad bank stuffed to the brim with $85 billion in investments gone wrong. Once that had been accomplished, the Teamwork group proposed that all the other Wall Street firms would inject sufficient capital into the bad bank to keep it afloat and unwind the debts over time.

With two possible cooperative solutions on the table and the memory of the Long-Term Capital Management lingering in everyone’s minds, there were several possibilities for Teamwork to come together and save Lehman. Yet with every step forward toward cooperation, the level of distrust mounted. Would Paulson cave in? Would they even have the liquidity in their own coffers to hand over money to another firm? And would the system reward them with stability if they took such a chance? Perhaps most important, would this be the end of the crisis? Would the government secure them against possible failures in the future?

As the day progressed, grumbles of resentment started to drown out the expressions of fear and confusion. Several CEOs openly questioned why they should bear the cost of Lehman’s problems when others who also faced exposure—institutional investors, hedge funds, and foreign investors—were not being asked to do the same. Even though they hadn’t been the only proverbial binge drinkers at the bar, the bankers felt they had been left with the tab for the whole industry’s decade-long bender. And if they paid up this time, wouldn’t they just get suckered into paying next time too? Morgan Stanley’s CEO, John Mack, expressed this suspicion to the room: “If we’re going to do this deal, where does it end?”

With little faith that the government would protect them and even less faith that the system would reward them for choosing the collective good over their individual firms, the bankers were left in a stalemate on Saturday afternoon. In a market system in which big bets were typically backed by reams of data, it was the soft value of trust that would win—or lose—the day. But trust cannot just be turned on and off like a light switch.

Or can it?

Paul Zak, founding director of the Center for Neuroeconomics Studies at Claremont Graduate University, is one of several scientists starting to investigate the neurobiology of trust. His most recent research offers proof that our brains actually do have a finely honed on/off switch for the soft value of trust. The electricity powering this switch is none other than oxytocin—the neurotransmitter first discovered by Sir Henry Dale some hundred years earlier. Zak and his research colleagues demonstrated the role oxytocin plays in negotiations with an oft-used economics experiment called the Trust Game.

The game always involves two people, the subject and the stranger—someone the subject has never met. It also involves the diva of all lust and trust interactions: money.

In a control version of the game, the subject of the experiment is asked to withdraw money from a computerized bank account and then give some of it to the stranger (without any face-to-face contact). The subject has been informed that on receipt of the funds, the stranger will reciprocate, returning either the same amount or an even greater amount of money, at a later date.

In Zak’s version of the experiment, however, a research team (headed up by Zak and Ernst Fehr of the University of Zurich) took 194 male students and instructed them to inhale a dose of oxytocin formulated as a nasal spray. The researchers compared their behavior with subjects who inhaled a placebo. Both the control and the experiment groups were then given ten dollars and told that they could share some, none, or all of it with a second person. If they did offer any portion to the second person, the amount dispersed would be tripled. The second person would then be asked to give as much back to the original subject as he felt inclined to give.

If the subjects trusted their recipient to be fair and give back half of the final total, both would gain. Without trust, however, their more obvious choice would be to hold on tight to the ten dollars and not even attempt to engage in the potential deal. The experiment shared some striking similarities to the bankers’ predicament: If they all put money in to save Lehman, they would have a shot at collective success: market stability. But they needed to trust that they would be rewarded for this choice, by both the system and the government.

The experiment results were striking: Subjects who received the oxytocin nasal spray gave about 17 percent more money to the strangers than did the placebo-sprayed control group. Even more remarkably, twice as many oxytocin-treated subjects exhibited maximal trust by choosing to give their entire cash amount to the second person.

The team concluded that “oxytocin specifically affects an individual’s willingness to accept social risks arising through interpersonal interactions.” Their experiment became so well known that it led New Scientist journalist A. C. Grayling to offer a suggestion: “Instead of pumping billions into the money markets, governments might more cheaply have sprayed oxytocin up the noses of bankers and speculators.”

•     •     •

By 5:00 p.m. on Saturday, the group at 33 Liberty Street was at an impasse. The Bank of America representative made it clear to Paulson and his team that they were pulling out of the running to buy Lehman Brothers, choosing instead to work out a deal with John Thain’s Merrill Lynch. This left only Barclays as a possible buyer, but they refused to move forward unless the other players agreed to buy up Lehman’s toxic debt. Mistrust was breeding mistrust and emotions in the room were running high.

Not surprisingly, Zak and his colleagues’ groundbreaking research on the nature of trust also tells us something about the neurochemistry of distrust. He noted the release of a testosterone derivative in men during moments of distrust and stress. This derivative is called dihydrotestosterone (DHT), and it’s no small force to counteract: Elevated levels have been proven to boost the desire for confrontation—often physical—during intensely anxious social situations.

So aside from shooting hormones up the noses of the most powerful men on Wall Street, what could Paulson and his team have done to better establish an atmosphere of trust? If systems create their own behavior, why was Paulson’s particular negotiating structure producing such distrust? In essence, how do we work together to solve a vital issue when it is in our collective, but not necessarily our individual, best interest?

One of the greatest challenges to successful cooperation was the complex role that Paulson played. As a former head of Goldman Sachs, Paulson’s role as a neutral representative from the government was compromised. To the bankers at 33 Liberty Street, he was still “one of our kind,” a member of the investment banking tribe. When push came to shove in the negotiating process, the bankers never really believed that one of their own would allow Lehman, much less the whole system, to fail. Such a belief was not unfounded: Paulson had blinked and caved when he had decided to save Bear Stearns only months before. Surely he would do the same for Lehman. After all, Paulson had been a banker for a lot longer than he had been a regulator. And when his stint in government was over, he might just become a banker again.

•     •     •

The bankers returned for another marathon session of talks on Sunday morning, and the CEOs acknowledged intractable disagreements over Teamwork Saves the Day. When the Barclays representatives turned to Paulson for taxpayer money to help them buy Lehman, he refused. A few hours later, Barclays backed out altogether as a potential buyer. The game was over.

Jim Wilkinson, chief of staff to Treasury Secretary Paulson, spoke to the room as the bankers started getting ready to leave: “This would be extremely interesting from an analytical perspective if it wasn’t happening to us.”

Briefcases snapped, suits and coats buttoned and zipped: 33 Liberty Street was closing up shop for good. A silence fell over the room as the players wandered out, stunned by a collective sense of dread. One by one, they brushed off the first signs of dust from the fall of the House of Lehman and attempted to focus their fears on how Lehman’s collapse would affect their own precarious state of affairs.

When the CEOs left the meeting on Sunday evening, they did not know exactly how the markets would respond to their decision, but they all understood the underlying game play of the weekend: There would be no consensus for collective action.

Over the weekend of September 12, Lehman Brothers filed the largest bankruptcy in American history. Thousands of employees arrived for work on Monday wheeling small suitcases and toting gym bags to empty out their offices. Meanwhile, Merrill Lynch, famously catering to the middle class with its “thundering herd” of brokers, made a frightened dash into the arms of Bank of America for safety.

On Monday September 15, Wall Street had its worst day in seven years. The Dow Jones Industrial Average lost more than 500 points, its steepest point drop since reopening after the September 11 attacks. In one single day, $700 billion disappeared into thin air, leaving retirement plans, government pension funds, and investment portfolios empty. Shares of Morgan Stanley and Goldman Sachs took a nosedive. Over the following days, both companies reconfigured themselves as commercial banks, leaving behind their storied roles as securities firms.

In just one weekend, the decades-long winning streak involving some of Wall Street’s greatest gamblers came to a crashing end. On September 15, it was time to walk away from the table and face the harsh light of day. It was every man for himself.

Of course, there were seemingly countless impediments to the success of Teamwork Saves the Day during that ill-fated weekend. We mentioned the air of secrecy and suspicion that hung over the meeting. Not only did that breed mistrust, it also made it impossible for the bankers to recruit other members of their network for support. The lack of transparency did not just apply to their own balance sheets, it was across the entire financial market system. Bad information led to poorly informed decision making and a great deal of uncertainty in the room. It was unclear how effective Paulson was in his role as referee. The bankers’ decades of training in negotiating within competitive markets discouraged whole-systems thinking, much like to fishermen using maximum sustainable yield as their default mode rather than ecosystem-based fishery management. And, as if all of the aforementioned were not enough, the terms around the table implied that there might, in fact, be winners among the losers, lending the entire negotiation process the tenor of a zero-sum game.

You would be forgiven for thinking that, considering all of the above, the bankers could not have reached any other conclusion. And yet, the lens of game theory suggests otherwise. The key to how that could be lies in one simple, four-line computer code that has single-handedly changed the entire framework of science’s attitude toward cooperation. Its name could not have been more appropriate: Tit for Tat.

TIT FOR TAT

Ever since the age of Darwin, scientists had been puzzling over one seemingly simple but impenetrable question: If living things evolved through competition, how did cooperation ever emerge? In the late 1970s, long before Paulson’s tense meetings at 33 Liberty Street, a political scientist at the University of Michigan named Robert Axelrod set himself the task of exploring this vexing question. His research started with the standard paradigm used to measure the evolution of cooperative behavior: a non-zero-sum game called the Prisoner’s Dilemma.

The rules of the game sound more like an outtake from The Godfather than a mathematical puzzle:

Imagine two burglars, partners in crime, are being interrogated separately for suspected involvement in a jewel heist. Although convinced of their guilt, police do not have sufficient evidence to convict them of the most serious offense but do have enough to convict both prisoners of a lesser charge—let’s say six months in prison for loitering outside the jewelry store.

Police, of course, want to convict the criminals of the most serious charges with the longest sentences possible, so they use a classic divide and conquer strategy: In separate rooms, and out of earshot of each other, they put pressure on each man to confess and implicate the other. They make the following offer: “Look, we’ve got videotape of you hanging around outside the jewelry store. You’re already going to jail for six months, minimum. But if you confess to the robbery, and your partner doesn’t, we’ll let you off scot-free and use your testimony to put that other guy away for ten years. Of course, if you don’t confess and your partner does, you’ll go to prison for ten years and he’ll walk away free as a bird. And if you both confess, we’ll put you both away for five years.”

Because the dilemma itself can be complicated to explain, it is often rendered in a matrix that looks something like this:

	Prisoner B Stays Silent	Prisoner B Betrays
Prisoner A Stays Silent	Each serves 6 months	Prisoner A: serves 10 years; Prisoner B: goes free
Prisoner A Betrays	Prisoner A: goes free; Prisoner B: serves 10 years	Each serves 5 years

If you put yourself in the prisoners’ shoes, you can see where the dilemma part comes in: While it’s in both of your interests to cooperate in keeping your mouths shut (following the old Mafia rule of omertà) and get a minimal sentence, it’s otherwise in your personal interest to defect and rat out your accomplice. The situation is made worse by the fact that you can’t communicate or coordinate with your partner. Think how easy it would be if you could just call him up on his mobile phone and make a plan to get you both off with a minimal sentence.

Yet it is this very restriction that allows the Prisoner’s Dilemma to mimic our ancient evolutionary landscape. Our mammalian ancestors, after all, had to cooperate long before we evolved the language to ask for it. The game roughly parallels the position of animals, unable to make any promises of a payment and yet fully engaged in cooperation: “You scratch my back, I’ll scratch yours.”

In short, while the details of the Prisoner’s Dilemma sound awfully specific, it functions as an abstract formulation of common situations in which there are optimal payoffs for everyone if they can agree on mutual cooperation (win/win) but strong incentives for each of the two players individually to defect (win/lose).

When Axelrod first became interested in the Prisoner’s Dilemma in the 1970s, research into cooperation was limited primarily to science that focused on cooperation within a gene pool. Biologist William Hamilton wowed the world with an elegant equation to explain cooperation that worked in service of gene replication, known as kin selection. Axelrod’s interest, however, was in an altruism, or cooperation, that was strategic rather than genetic. What would ever make one cooperate if it wasn’t an effort to propagate genetic code—either one’s own genes or the genes of a family member? Evolutionary biologists, beginning with the brilliant polymath Robert Trivers, began referring to a cooperation based on strategy as “reciprocal altruism.”

If the concept of reciprocal altruism strikes you as technical or vague, consider situations in your own life. Most people have, at one time or another, encountered a neighbor or coworker who abided by a principle of “gimme gimme gimme” without ever giving anything back. It doesn’t take long for the sensible person to “defect” and cut ties with such an egoist. More difficult, however, is the person in your life who cooperates some of the time and then defects at others. It is harder to stop interacting with such part-time meanies: Remember the Queen Bee in junior high who deigned to speak with you only once in a blue moon, or the annoying friend who paid for dinner only one time out of three. What about your colleague, the guy who goes out drinking every night and then calls in “sick” for half of your group meetings? Do you cooperate again and again, or do you finally defect? And after a defection, when should we forgive and start cooperating again?

In a one-round match of Prisoner’s Dilemma, where the players will never meet again, it often makes the most sense to defect immediately, because that is where the payoff is potentially the biggest. Two strangers hailing the same cab have no chance of seeing each other again, so there is little incentive to be accommodating. Most of us, however, live in a world of iterated rounds of the Prisoner’s Dilemma, or rounds that involve the same players in the same game time after time after time. Anyone who has ever been married or negotiated living quarters with a partner or roommate intuitively knows the cycle of cooperation and defection that characterizes the iterated rounds. Who took out the garbage last night? Who’s going to take it out tonight?

Our forebears existed in small tribes—negotiating over and over again with the same few people—so we have finely evolved mechanisms for strategizing for the longest arm of time. Common sense tells us to cooperate when we know there will be reciprocity at some point farther down the road. Unless, that is, cooperation comes at a cost too high to bear.

All of this brings us back around to our bankers at 33 Liberty Street. For they, too, were playing an iterated round of Prisoner’s Dilemma when they struggled over whether or not to save Lehman. All of them had negotiated with one another in the past, and, considering the trajectories of their careers, they would be sure to do so again in the future. Given this, which move would prove the most strategic for their complex dilemma?

In the late 1970s, Axelrod decided to explore this very question using what was then a truly novel strategy: He would hold a computer tournament. He simplified the whole scenario of the Prisoner’s Dilemma so that it could be programmed in code and played in simulated rounds over and over again. Then he invited various experts in the Prisoner’s Dilemma to submit strategies, expressed as computer code, and played them against one another—a kind of battle royal of cooperation and defection.

In each battle, two players—computer programs—faced off against each other for two hundred rounds. In each of these rounds, the players were given two choices: cooperate or defect (C or D). Each made a choice without knowing what the other would do. Defection, D, yielded a higher payoff of points than cooperation, C. But if both players chose to defect, they would both do worse than if both had chosen to cooperate.

Axelrod initially received fourteen different submissions for the contest, incorporating a dizzying array of strategies. One, called Massive Retaliation, initially cooperated, but at the first sign of defection, continuously defected until the end of the game. Another, called Tester, tried occasionally to defect, but at the first sign of retaliation would retreat back to a cooperative mode, effectively trying to sneak one by its opponent.

When Axelrod tested all of these strategies against one another, he was surprised at the winner. So surprised, in fact, that he put out a call for a second set of submissions. This time, he received sixty-three entries, including the same winner from the first round. Again, Axelrod programmed all sixty-three entries in a round robin tournament, and again, almost to his disbelief, the same four lines of code won the entire tournament. It was, by far, the simplest strategy, a rule set that even your two-year-old child can follow: Tit for Tat.

Submitted by eminent mathematician and game theorist Anatol Rapoport, Tit for Tat had an astoundingly straightforward way of navigating the game. It cooperated on the first round—making it a “nice” strategy—but then it simply repeated the move of its opponent for the rest of the game. If the opponent was a nice guy, Tit for Tat responded to every cooperation with a mutual cooperation. This meant that any nice strategy would find a safe harbor of love and harmony with Tit for Tat.

By contrast, in the presence of a defector, Tit for Tat would give as good as it got. A defection would be immediately punished with a defection. If the defector did it again, Tit for Tat would do so as well. But Tit for Tat was also forgiving—it didn’t hold a grudge. As soon as its opponent stopped defecting, it would as well.

Axelrod discovered that over the long run, Tit for Tat does better than any other strategy. But an even bigger surprise was that Tit for Tat was not the only nice strategy to emerge successful. The eight top-ranking entries were nice or primarily cooperative and forgiving. Out of all the winning programs, none of them were the first to defect. Tit for Tat never did better than its immediate opponent, but it was able to cooperate with itself and with other nice strategies. Conversely, the defector strategies tended to cannibalize each other, leading to fewer gains. The meanies were also unable to take advantage of Tit for Tat over time because Tit for Tat retaliated. When this winning strategy represented a large enough proportion of the population, other nice strategies could also effectively settle in and, ultimately, thrive.

In a subsequent iteration of the tournament, Axelrod allowed the programs in the competition to make copies of themselves at the end of each round, reproducing in proportion to how well they had done. He then ran the simulation over and over, providing a rough proxy for evolution. Axelrod found that, as long as there was a minimally sufficient number of nice strategies that could meet and reproduce, these strategies could grow and thrive, even in a world dominated by defectors.

Over the long haul, Tit for Tat showed how cooperation could evolve as a stable strategy in an otherwise competitive evolutionary framework. When we take Tit for Tat out of the realm of game theory and into a real-world context, however, the story changes slightly.

Axelrod’s computer tournaments focused exclusively on the logic of two abstract dispassionate actors who execute their moves flawlessly. But the real world is nothing like this. People make mistakes—they sometimes cooperate when they mean to defect, and vice versa. And this messiness has important consequences for how the game is played.

For example, when Axelrod studied the results from the first round of his tournament, he discovered that Tit for Tat might have been beaten by a slightly more forgiving strategy, called Tit for Two Tats. As the name implies, this strategy retaliates after only two back-to-back defections by its opponent. It “turns the other cheek,” at least initially, before responding.

When Tit for Two Tats competed against another nice strategy like Tit for Tat, it fared marginally better. Yet in another iteration of the computer tournament, in which hostile strategies were also included, Tit for Two Tats was a major loser, finishing twenty-fourth. More aggressive strategies exploited its niceness—defecting, but never enough to trigger retaliation.

Yet, intriguingly, recent research suggests that Tit for Two Tats might actually be a much more successful strategy in the real-world realm of human affairs. To test this, Harvard psychologist David Rand and his colleagues created a modified version of the game. In it, human subjects played a series of repeated Prisoner’s Dilemmas, with a twist: In each turn, players had a one-in-eight chance that their move would be changed to the opposite of what they intended—a cooperation would turn into a defection, and vice versa. This condition was added to account for the fact that people aren’t perfect—they sometimes make execution errors.

In this error-prone, “noisy” condition, players who followed at Tit for Two Tats actually earned the highest payoff, because they were able to look past the occasional, unintended defection. Players following a stricter Tit for Tat strategy, on the other hand, ended their cooperation too early, and lost out on a beneficial relationship. “In an uncertain world,” says Rand, “it can be advantageous to be slow to anger and fast to forgive.”

The real-world also varies from Axelrod’s simulations in other ways. Human beings are social primates, and like all social primates, our ideas about when to cooperate and when to defect have been strongly influenced by the groups on which we depend for survival. Thinking in social rather than individual terms will often lead us to make counterintuitive choices that are as much about the impact on other members of our group as about the benefits we ourselves receive (and the ones we suspect others are receiving).

Primatologist Frans de Waal, together with his research partner Sarah Brosnan, investigated cooperation and perceptions of inequity among capuchin monkeys. In one of their experiments, two monkeys were given either a grape or a piece of cucumber for completing a simple task. If both monkeys were given the same reward, de Waal and Brosnan didn’t observe any problems. They confirmed that the grapes were far preferred—de Waal noted that all primates enjoy a sugar rush—but even if both monkeys received a cucumber, they seemed to have no problem repeatedly performing the given task.

The experiment started to get interesting, however, when monkeys received different rewards. The one who got the lesser reward—a cucumber instead of a grape—would start to hesitate, eventually putting up a fight by either not eating the cucumber or not performing the task.

De Waal concluded that this is an irrational response: “If profit-maximizing is what life (and economics) is about, one should always take what one can get. Monkeys will always accept and eat a piece of cucumber whenever we give it to them, but apparently not when their partner is getting a better deal.”

Humans have this same instinct—referred to as inequity aversion—but economists, game theorists, and mathematicians have historically counted on us to care more about the pursuit of self-interest over fairness. Yet when it comes to the distribution of scarce resources, there is little evidence that humans actually do act purely in accordance with their own self-interest. Even when we have a fresh, tasty cucumber in our hands, humans have evolved the tendency to start fixating on the asshole with the grape.

All of this is to say that, despite the dictates of Adam Smith or Milton Friedman, the bankers at 33 Liberty Street were still using their social and emotional brains. Amidst discussion surrounding systemic risk and toxic debts, every one of the suits around the table was also assessing who would get stuck with the cucumber and who would walk away with the grape.

In addition, the bankers also brought a whole set of cognitive biases that help to inform how all of us, as social creatures, make decisions. Each CEO had worked directly with a team of hundreds, perhaps thousands, of bankers and had met many more in the course of his career. Yet this is just a small fraction of the whole—in the global world of finance, there are many thousands of players; even a CEO in an alpha position such as Mack, Blankfein, or Thain couldn’t possibly know every single banker in the field. These CEOs—like the rest of us—use heuristics (mental shorthand) to identify colleagues in different firms, akin to different tribes: “You know Bob; he’s with Morgan Stanley.” These tribes in turn have their own values, cultures, myths, risk tolerances, and styles, which help bind them together and reinforce distinctions for their members. And tribal thinking can assert itself with a vengeance.

As the bankers considered bailing out Lehman, for example, they were threatened by the thought of handing over their scarce resources to help Barclays sweep in and score with a strategic and profitable move. Barclays registered in mental shorthand not just as “threatening” but also as “British.” In cognitive terms, when negotiating over possible money coming in from the American taxpayers, Barclays was an outside group—a foreign tribe. The Wall Street Journal noted that the bankers “were loath to provide support when a rival like Barclays might still buy Lehman.” Later they reported, “By Sunday morning, the U.K.’s Barclays looked like the sole potential buyer. That further minimized the chances of a government bailout: If the Bush administration wouldn’t help to fund a Wall Street solution, aiding a foreign buyer was even less likely.”

Is it possible that oxytocin—our love and trust elixir—is the very same hormone underlying such in-group and out-group biases? Scientists like Carsten K. W. de Dreu, a psychologist at the University of Amsterdam, believe so. He gave Dutch students a standard moral dilemma: Do you save five people in the path of a train by throwing one bystander over the tracks? He left the five people nameless while giving the sacrificial victim either a Dutch or a Muslim name. Dutch subjects given a whiff of the oxytocin before beginning the experiment were far more likely to push Muhammad over the tracks than Maarten.

A famous study from 1954 suggests similar hormones at work: Psychologist Muzafer Sherif and his colleagues gathered twelve-year-old boys at a summer camp, divided them up into two groups, and instructed them to compete against each other. The boys quickly started to antagonize members of the other group, doing things like holding their noses in disgust upon seeing them pass by—behavior known as out-group derogation. Meanwhile, within the groups, cohesiveness increased, and the groups almost immediately became both more hierarchical and more internally cooperative.

This kind of in-group, out-group tribalism has a powerful logic. Game theorists Steve Rytina and David L. Morgan mathematically modeled this bright line between “Nice Nellies” and “Meanies” by dividing an imaginary society into two groups, the Blues and the Reds. They watched what happened when members of each tribe followed a variant called Discriminatory Tit for Tat (DTFT), which is just like Tit for Tat except when dealing with someone of a different-color group. A Red will always defect with a Blue, and vice versa.

When two Reds engage in negotiation for the first time, each cooperates. When two Blues with no history negotiate, both cooperate. But what happens when a Red and a Blue interact? Defection is automatic because “you can’t trust those guys.” Think of the Sharks and Jets; the Montagues and the Capulets; the Hatfields and the McCoys. Unfortunately, we understand this story all too well.

Rytina and Morgan demonstrated that the DTFT game play is not only stable, it’s all but intractable. In the early rounds, an individual who tries to play regular, color-blind Tit for Tat is worse off than one who defects with the out-group. Why? When a Red and Blue interact for the first time, even if the Blue contemplates cooperating (as in regular Tit for Tat), the Red player will almost certainly be playing DTFT and will defect. That means that the Blue player—only trying to start off nice—will get the sucker payoff and lose points.

This does not mean that DTFT is a more successful strategy. Tit for Tat still leads in terms of long-term resilience. The problem with DTFT is in its insidious stability. Once it is entrenched, it punishes individual efforts to attempt cooperation across enemy lines.

So what is the takeaway of such results? As in any complex system, it’s a mixed bag. The bad news is that we have evolved cognitive biases that lead us to see our playing field in terms of “us” and “them.” The good news is that, although this is a stable game play, it has been mathematically proven to be less successful (over the longest time periods) than the more generous strategies of Tit for Tat.

But considering the rational and emotional rewards we get from aligning ourselves with a tribe (DTFT style), how do we trigger cooperation between opposing groups?

Fortunately, there is a surprising amount of flexibility within delineations of in-groups and out-groups. As the famous sociobiologist E. O. Wilson has written, “Altruism is characterized by strong emotion and protean allegiance. Human beings are consistent in their codes of honor but endlessly fickle with reference to whom the codes apply.” Note the use of the word “fickle.” It’s our very plasticity that gives us our evolutionary advantage. Yes, we love “our” people. But our minds have evolved malleable markers for the definition of just who “our” people are. Wilson goes on to write, “The important distinction is . . . between the in-group and the out-group, but the precise location of the dividing line is shifted back and forth with ease.”

The trick, then, is to get people to expand their definition of “us.” But how?

ENLARGING THE TRIBE

Arthur Aron and Elaine Aron are husband and wife as well as research partners and writing collaborators at Stony Brook University. They have made it their life’s work to study the ways in which we form intimate ties with those in our inner circle as well as with members from an out-group. Over the last decade, they have been bringing together diverse pairs of complete strangers—black and white, Latino and Asian, black and Latino—to participate in an unusual experiment.

The Arons asked each of these couples to come together for a series of four intimate, hour-long sessions. During their first session, each member of the pair was asked to share his or her answers to a list of questions, everything from, “Would you like to be famous? In what way?” to “If you could change anything about the way you were raised, what would it be?” During the second session, the pairs competed against other pairs in timed games like charades, word play, and logic puzzles. Then, in the third, they were guided through a series of intimate conversations with questions about their personal lives and their feelings of affiliation toward their ethnic groups.

In the last hour-long session, they did a blindfolded trust walk, taking turns navigating a maze wearing the blindfold or serving as the guide.

Although the activities conjure up a weekend at seventh-grade summer camp, Dr. Arthur Aron argues that these four hours create a relationship that is as close as any in a person’s life. In essence, they are saying that these relationships, when successful, are just as effective as any at releasing the hormone oxytocin. Once the “cuddle chemical” is triggered—whether through the structure of the Arons’ exercises or a whiff of Zak’s nasal spray—trust and cooperation naturally follow.

In fact, research from the Arons’ tests showed that the four-hour sessions almost immediately lowered subjects’ score on a variety of prejudice measures. Stress hormone tests, conducted on the subjects’ saliva, showed significantly reduced anxiety for both members of the pair when they encountered a social interaction with a member of their partner’s ethnic group.

Psychologists suggest that these kinds of powerful bonds—creating new definitions for in-group and out-group—occur through our own highly evolved process of empathetic mimicry. We are rarely conscious of it but we, as humans, are constantly copying the facial expressions, manners of speech, postures, and body language of those around us.

All of these copying behaviors, through a rather elegant neural feedback loop, allow us to actually experience the emotions associated with the particular sort of behavior we are mimicking. This helps to explain why it is so excruciating to watch someone else cry: Your neural feedback is allowing you to feel an analogous sensation. It makes human tears—costly in energy for the body to produce—a bona fide evolutionary bargain when it comes to the empathic and cooperative payoff.

It’s no coincidence that the Arons take their couples on a walk, the quintessential exercise for conflict resolution. In 1982, when arms negotiators from both the United States and the Soviet Union were looking for a way to draft an arms reduction proposal, they met in Switzerland to meander on that now famous walk in the woods. Face-to-face interactions often bring conflict to the forefront of the conversation, whereas walking side by side allows space for a more considered dialogue. What if we could take not just a few pairs, but hundreds, and then thousands of people on a trust walk together? Could a walking path help to enlarge the tribe in the midst of one of the most fraught tribal conflicts on the planet?

•     •     •

William Ury, a world-renowned expert on conflict negotiation, was committed to finding a way to bring religious storytelling into the sphere of political discussion in Middle East conflicts. Despite his stature in the field of conflict resolution—Ury is cofounder of Harvard’s Program on Negotiation—most of his colleagues counseled him to let go of religion as any kind of negotiating tool. But throughout the Oslo talks in 2000, Ury was dismayed to see that the political process failed to address the two aspects of the conflict that he deemed essential: an acknowledgment of the identity issues wrapped up in the distribution of the land and the practical, economic conditions of the Palestinians (in decline since the 1990s).

“Things were so stuck that you really needed out-of-the-box thinking. The question was and is: What is the game-changing move?”

An anthropologist by training, Ury looked to some of his work in other cultures for answers. He realized that the conflict needed a way to call upon the community he described as the “third side.”

“Spending time with groups like the Bushmen in the Kalahari, I noticed how they used the surrounding community as a third side. When they approach conflict, everyone gets in a circle and they all have a say. They play an enormously healing role in which they create a container. Even a destructive conflict can be gradually transformed when it is contained inside community.” In other words, when everyone is a member of the same tribe, the oxytocin rush of pitting us against them disappears.

Ury was looking for a set of people to act as a buffering force. But who was the third side in the Middle East?

“Every culture has an origin story,” Ury told us, “and the origin story in the Middle East engages almost all of us on some level. Four thousand years ago a man and his family walked across the Middle East. The world has never been the same since.”

That man, of course, was Abraham.

Ury started referring to him as our symbolic third side.

“What if you took the story of Abraham—the symbol of hospitality—and used it as the antidote to terrorism, a vaccine of sorts against intolerance?” By calling upon the story of Abraham, Ury could unite the three groups of faith and, at the extreme fringes, nullify the faiths in relation to one another. All three faiths defined under one tribe—People of the Book—had a shared history, shared values, and a shared sense of God that conferred respect on everyone.

“In most negotiations I’ve been in, there is something so simple that is required. I would call it respect. It’s the cheapest concession you can give as a negotiator—it doesn’t cost anything—but it’s amazing how little we use it. Respect is an interesting element because it’s a positive-sum element. If I give you respect, I don’t have less respect for myself. In fact, you’re likely to give me more. That is what Abraham actually stands for: respect and the positive value of the other human being.”

But it wasn’t enough for people simply to remember or recite the story of Abraham, Ury realized. For the symbolic third side to really play a role in the conflict, people would need to experience the story. And the best way to do that would be to go for a walk.

“Walking has power; it’s what made us human. A path that followed in the footsteps of Abraham—crossing ten different countries—has the power to transform hostility to hospitality, terrorism to tourism, all in the name of Abraham.”

Ury decided to trace the journey of Abraham through the Middle East and use it to create a cultural, ecotourist walking path. He named it the Abraham Path Initiative.

The route Ury and his colleagues originally followed started in Abraham’s birthplace in Urfa, Turkey, then wound through Harran, Turkey—the place where many sources suggest Abraham heard the call from God. It continued onto Syria, down from Jordan to Jerusalem, and then into the West Bank before ending in Hebron, or al-Khalil, described in the book of Genesis as Abraham’s burial place. Ury describes this trek—more than six hundred miles—as a journey from “womb to tomb.”

Just as the Arons surmised in their experiments on intimacy, the simple act of walking provides an opportunity for participants to engage in small acts of trust building. These interactions build opportunities for hospitality: the cooperative contract between guest and host based on mutual respect.

“In the West,” Ury told us with his characteristic wry chuckle, “we perceive the conflict through a frame of HOSTILITY. But you take HOSTILITY and add PITA bread, and you get HOSPITALITY.”

And that is exactly what we discovered when we donned our hiking boots and experienced the Abraham Path for ourselves. Our five days of walking through the shepherds’ hills of Palestine took in but a small section of William Ury’s game-changing vision, but they put us—and our fellow walkers—in touch with dozens and dozens of new faces as well as frame-changing experiences.

Not only were we served fresh pita bread drizzled with olive oil—made daily in the caked mud ovens at the back of the homes—we also encountered an almost dizzying array of responses toward religion, the role of women in society, and the identity of the modern Palestinian. We were welcomed in villages where the men refused to shake the hand of a woman, much less sit in a room with her. But we also sat and drank sweet tea in villages where the women functioned as powerful matriarchs—outspoken and well educated—while the men sat back in respect. The one constant was the contract of hospitality—respect—spoken in the name of Abraham.

When we walked up the steep hills of Awarta, five miles southeast of Nablus, we were even pulled to participate in a town wedding without anyone ever batting an eye. There, in a meeting house in the center of town, hundreds of traditionally dressed women from the village were dancing and singing with abandon while a young bride and groom sat elevated on chairs, watching them. The singing—referred to in Arabic as al-mardudeh—was a raucous call and response between the lead singer, known as the badda’a (the talented one) and the hundreds and hundreds of women all packed into the tiny room. Before we had time to fully enter the room, we were placed next to the bride and the groom on the elevated seats overlooking the crowd of celebrants.

“You are special guests here in the village,” a woman wet with sweat told us in between claps and chants. “We must celebrate you!”

Not only invited to a complete stranger’s wedding but deemed a guest of honor: It is a gesture of such generous hosting that it is almost inconceivable in most Western countries. Frédéric Masson, the French-born coordinator for the Palestinian arm of the Abraham Path, explains that this extraordinary culture of hospitality is being threatened by the lack of employment and social structure in the occupied territories.

“With the current situation, the villagers are isolated from the world. They are so eager to share their experiences but they are disconnected and that leads to a great despair, especially for a culture that values hospitality. When you attended their wedding, you were honoring them and their expression of hospitality.”

This leads to the second essential aspect of peacemaking that Ury is eager to address: the idea of mutual prosperity, another non-zero-sum element. The tourism along the path brings in income for the villagers: selling food, teaching classes, offering shelter. Their culture will be valued in name but also through economic improvements, conferring even more respect on the people offering up their hospitality.

Over the last five years, thousands of people have started to walk the path through Syria, Jordan, Israel, and Palestine. And people unable to journey to the Middle East are organizing walks in their own cities and communities. Much like the layered garden of Willie Smits, the Abraham Path attempts to address both immediate economic needs of the community as well as its needs for respect and trust over the longest arm of time.

“Today the path is like an acorn,” he told us. “Tomorrow it may look like an oak tree.”

His words remind us that though Abraham serves as the symbolic third side, we, the global community, are the ones who will walk the walk.

Now think back to our bankers at 33 Liberty Street. As we have already established, the secrecy of their meeting impeded any efforts to enlarge their tribe. The nature of the negotiations created strong incentives for side dealing, as the participants engaged in two very different kinds of behavior. It is hard to think in zero-sum and non-zero-sum terms simultaneously. Worst of all, the bankers—even with Paulson in the room—represented only a tiny fraction of the real stakeholders in the crisis. What might have changed had Ury’s third side—ordinary citizens, other kinds of companies, outside economists, or activists—been present to enlarge the tribe and seek out a constructive whole-systems solution?

Despite all the positive evidence from neuroscience and game theory, and the demonstrated power of strategies like enlarging the tribe and employing the third side in complex negotiations, one might still doubt that the bankers at 33 Liberty Street could ever overcome the seemingly insurmountable odds against them and find a path to successful cooperation.

And yet, less than two years after their fateful meeting, in the middle of an even more profound crisis, with a deadline similarly measured in hours, and with life and death hanging in the balance, a group of total strangers showed just how such a cooperative effort could emerge. Their story—of Mission 4636—holds lasting lessons for eliciting cooperation everywhere.

MISSION 4636

The devastating earthquake near Port-au-Prince, Haiti, on January 10, 2010, wasn’t the most powerful earthquake ever to hit the poorest nation in the Western Hemisphere, but it was certainly the most catastrophic. As the quake ripped through the ground below, it triggered the collapse of nested socioeconomic, political, racial, and physical fragilities above. In mere moments, it generated statistics of the sort one normally hears tallied after a war: 316,000 Haitians were killed, and another 300,000 were injured. A million more were made homeless; in all, three million were affected. Physical infrastructure was leveled, entire departments of the government were obliterated, the citizenry was traumatized, as were countless international aid organizations who were already operating in the country. There was never any doubt that this would be a humanitarian crisis of the first order.

And yet, as terrible as the days’ events were, the earthquake might have been an even greater disaster had it happened just a few decades earlier. For on January 10, 2010, the instantaneous news of the disaster also unlocked a uniquely collaborative, global response that could not have existed before the age of the Internet—one that harnessed a global community of volunteers, technologists, first responders, and Haitian citizens both at home and abroad.

What’s more, this remarkable collaboration unfolded in approximately the same amount of time as the meetings held over that now-famous weekend at 33 Liberty Street.

Our story starts 1,700 miles away from Haiti in the snowy winter weather of Somerville, a suburb of Boston. Rosalind Sewell—known as Roz—had just returned to the East Coast to resume her first year of graduate school at the Tufts Fletcher School of Law and Diplomacy when she heard the news of the quake. As a former Fulbright fellow in Morocco, Sewell was already interested in emerging forms of social media and technology around the globe. That’s why, when she checked her email the week of the tenth, one message in particular caught her eye.

“The email said, ‘Hey, if you’re in town, we’re sitting around in my apartment mapping social media as a response to the earthquake.’”

The man behind the email invitation was a PhD candidate at the Fletcher School named Patrick Meier. Sewell recognized him as the leader of a group of Tufts students particularly interested in crisis mapping, a new field of disaster analysis that used a combination of satellite imagery (now widely available online from organizations like Google) and data sent in through SMS texts, Twitter, and Facebook to map out the impacts of an unfolding crisis and coordinate response.

For Sewell, the email promised an opportunity to learn more about crisis mapping and the community that Meier had cultivated at Tufts. When she showed up at Meier’s apartment that evening, she expected to see a group of Tufts students calmly reading Twitter and Facebook and then uploading the coordinate information to an online map. What she encountered, instead, was a full-blown operations center, the nerve center for information sharing and incident reporting on the Haitian earthquake.

“My family is army, so I’ve seen the Army Op Center with the thousand computer screens and the big monitors. When I walked into Patrick’s living room, it had that exact feel. A couple of people were running around on telephones, putting information in, really focused. Someone else was trying to train new people. The whole apartment had this really electrifying energy that sucked me in.”

She arrived that night planning to map for a couple hours and then return home. Instead, she stayed inextricably linked to the crisis mapping effort—both in person and online—around the clock for the next three weeks. And Roz Sewell was not alone. Thousands of people all over the globe—from Istanbul to Geneva to Washington, D.C.—collaborated in response to the earthquake, making it the largest crowd-sourced crisis mapping effort ever undertaken.

Mission 4636 actually had its roots halfway around the world in response to another crisis: a terrifying wave of postelection violence that had buffeted Kenya in 2007. In the wake of the election, ethnic violence swept through the countryside—a cycle of horrible defections among groups: Discriminatory Tit for Tat, gone terribly wrong.

In the midst of the crisis, Kenyan residents had virtually no reliable real-time information on the violence outside of the cities. Ory Okolloh, a Kenyan-born Harvard-trained lawyer, who had returned to Kenya to observe the elections, started to collect and geotag all the incidents of violence on her own personal blog, but the reports were coming in too quickly. Much like financial regulators before 2008, Kenyans needed a better mapping tool, one that would dynamically change to reflect real-time information on violent outbreaks. But they also needed a system they could access themselves, and easily. She reached out to the blogosphere with a challenge to help her create such a tool. Days later, volunteer software developers answered her call by quickly prototyping a web platform that allowed Kenyans to contribute anonymous reports of riots, rapes, deaths, and displacements of people using just their cell phones. These crowdsourced crisis reports could be mapped on a single display, and citizens could then access them not only to see what had happened and where, but to see larger trends over time—how the violence had spread through the community and where it might go next. The loose network of technologists and human rights activists called the tool they had built Ushahidi, which means “to bear witness” in Swahili.

In the immediate circumstances of the election, not enough people were aware of the program to be able to make it truly usable during the crisis, so the tool served as more of a prototype, a vision for what might be. One thing, however, was clear to the developers right away: The platform could be used by a community in any situation that was lacking good, real-time systemic information—not just a violence outbreak, but things like public health epidemics and ecological disaster. They decided to package it for a wider variety of contexts.

Over the last few years, the Ushahidi platform has been used by communities to track anti-immigrant violence in South Africa, map violence in eastern Congo, report pharmacy stockouts in several East African countries, and to monitor elections across the world. The organizational goal of Ushahidi was to make the platform easy to use and free of charge to anyone who wanted to bring greater transparency to an issue. Like many software developers, the organization that created Ushahidi never intended to get involved in any of the actual deployments after their initial launch in Kenya.

On January 10, all of that changed.

Patrick Meier, the man behind the email sent to Roz Sewell’s account, is the director of crisis mapping at Ushahidi and the cofounder of the International Network of Crisis Mappers. He was watching CNN in Boston a few hours after the earthquake struck, shocked by the severity of the damage. Meier called Ushahidi’s main technical director and told him that he really wanted Ushahidi to do something beyond just disseminating the platform. Considering the extremity of the disaster and the impoverishment of the country, Meier felt that the quake was set to be a tipping point—a systemic flip—for the whole country.

Within an hour of getting Meier’s call, Ushahidi’s technical director worked with its main programmer to customize the platform for the Haitian crisis. Whereas the Ushahidi platform was used in Kenya solely as a mapping visualization tool, in the Haitian disaster, it was hoped it could also be used to help coordinate responses, connecting all of the stakeholders in the crisis: first responders, NGOs, the government, and most important, Haitians themselves. This was a fundamental shift in the way the platform functioned: It went from being simply a mapping tool to a tool for intervention.

“Once I made the call to set up the platform, it was clear that there was no way Ushahidi’s small, primarily tech team was going to be able to do the near real-time crisis mapping,” Meier told us. “After Kenya, this is the only other time we’ve ever done a deployment.”

The first requirement was to track down and sort all the information streams coming out of Haiti, including tweets, Facebook updates, people finders on various sites all across the web, and any text messages that were managing to make their way out of the country via mobile phones.

Meier knew that monitoring all of this information would require the constant attention of a whole team of people. But that was only the beginning. After tracking all the pieces of information, they would need to be processed into different categories and then geotagged, mapped onto a dynamic digital map of Greater Port-au-Prince with accurate coordinates. Considering the fact that neither Meier nor any of the other immediate Ushahidi team was familiar with the streets, back alleys, and local haunts of Port-au-Prince, the task of finding the right coordinates seemed next to impossible.

To aid in the seemingly insurmountable effort, Meier called upon Ushahidi’s latent network, sourcing potential volunteers through Twitter, Tufts University, college campus networks, and the social circles of a generation of digitally savvy development professionals. The Ushahidi team was also part of a large and informally linked patchwork of likeminded, technologically innovative humanitarian NGOs, technologists, and researchers that had been building steadily for years, many of which, at various points, would play important collaborative roles in the Haitian crisis response.

The SOS email that Roz Sewell received went out to the thousands of other people on the Tufts University email list. As an affiliate of the Fletcher program, Meier had already given some presentations on Ushahidi at the school, galvanizing a small group of friends and acquaintances engaged in the effort. After his initial email, the core group of a dozen volunteers showed up, accompanied by a few new faces like Sewell. Meier trained his initial group on the basics of tracking and tagging, emphasizing the urgency of the effort while giving them ideas and tools for finding more people to help. He told them that the work was remote: As long as they signed into the group Skype chats and logins, they could process and map information from anywhere in the world.

And then he left them alone.

After the initial handful of people showed up to help in Meier’s living room, the numbers of volunteers grew quickly. A few days later, twenty people were squeezing onto the couches and floor space. Then thirty, forty, and, finally, when more than fifty people were coming to help, the group arranged to take over a study space in a Tufts University building. By the following week, a training session arranged for Tufts undergraduates was overflowing with more than eighty students and the op center was given an official name: Ushahidi-Haiti @ Tufts.

Soon Meier was receiving email notifications that training sessions and ancillary crisis mapping situation rooms were popping up all over the world. A group of Tufts alumni emailed Meier and told him that they were replicating the crisis mapping op room down in D.C. with login information passed on from another volunteer. An exchange student from the Fletcher School studying abroad at the Graduate Institute in Geneva emailed Meier and the group and told them that she was covering Geneva with crisis mapping outposts. Another Fletcher student still in Portland for the holiday break started organizing students to do real-time mapping from the West Coast. Then the seeds of organization blew to cities all over the world, creating small start-up communities of crisis mappers in London, Montreal, Providence, Istanbul, and other locations across the globe. Anywhere there was at least one person with the training to log in to the Skype chat and the platform, there was a crisis mapping node.

Just like the previously discussed, malevolent AQAP networks, these pro-social crisis mappers were not bound together via traditionally strong command-and-control structures, but by informal social connections and a shared set of values. Over time, hundreds of crisis mappers were contributing to the effort—a fully fledged swarm. Meier had not met a single one of them. In some cases, he was completely unaware they existed until the entire effort was over.

“I told people, ‘We’re going to let this be emergent,’” Meier explained. “There are so many things that need to happen every single hour and so many things that need to keep evolving in such a short amount of time. I have to just let it flourish and deal with what happens when it starts getting inefficient.”

The open nature of the platform—both the code that powers Ushahidi and the collaborative nature of the mapping—meant that people could easily be recruited to perform discrete, useful tasks with a minimum of formal authority. That helped the platform, and the recruitment process, to spread.

“The whole story of the volunteers is one of the most amazing things that I’ve ever been a part of,” Meier reflected. “Within about five or six weeks, people trained more than three hundred individuals on how to use Ushahidi’s platform, and I didn’t organize a single one of the sessions.”

In the earliest hours of the crisis, this ad hoc global crisis mapping community was pulling its data primarily from news reports, Twitter feeds, and Facebook updates. But to truly transform from a mapping tool to a full-blown intervention, Ushahidi would need first to directly connect with Haitians in need, and then in turn connect them with the first responders (who could look at the map and determine where they could be the most helpful). While the crisis mappers were in the midst of recruiting and training volunteers to assist in the deployment, another initiative, led by a constellation of technology-led NGOs including FrontlineSMS, InSTEDD, Energy for Opportunity, and the Thompson Reuters Foundation, was under way to figure out how to enable Haitians on the ground to add in their own situation reports to the map.

On January 10, minutes after the earthquake hit, Josh Nesbit, a self-described health-care junkie in his midtwenties who worked for FrontlineSMS (now MedicMobile), made a call to the State Department. Within an hour, he was strategizing with media-savvy staffers there about using radio and the telcos to reach the Haitians directly.

“I basically knew that in a country like Haiti where only about one percent of the population is online—even less than that after the quake—land lines are pretty much nonexistent,” Nesbit told us. “On the other hand, forty percent of the population owns a cell phone, and if you look at people who have access to those phones, the number is probably much higher, closer to seventy-five or eighty percent.”

Nesbit and his contacts at the State Department knew that the success of the effort would hinge on the ability of the mapping platform to directly connect with Haitians on the ground. They all agreed that securing an SMS channel was a first priority.

SMS—or short message service—is really just a technical term for a text message. What Nesbit and the State Department wanted, however, was a specific channel for receiving incident reports from earthquake victims. Much as cities often set aside phone numbers for specific functions—in New York, people know they can dial 311 to reach the offices of the city government—Nesbit wanted to get all the Haitian telcos to agree on an earthquake emergency text number.

From his apartment in Washington, D.C., he cataloged all the nongovernmental organizations he had worked with over the past six months and pinged them all, looking for what he referred to as an “SMS hub.” He spent the rest of the night connected, circling round and round phone calls, tweets, and texts with the team at State and his own personal network of global health contacts. In the flurry of communication, Nesbit rushed out a quick tweet to his entire network of followers:

Reaching out to @FrontlineSMS users in #Haiti with hopes of establishing local SMS gateway for http://haiti.ushahidi.com.

Almost immediately, Nesbit received contact from one of his Twitter followers in Cameroon in West Africa, telling him to reach out to Jean-Marc Castera, Digicel’s IP manager in Haiti. Within minutes, Nesbit was Skyping with Castera, making a plan for creating a short code.

“Now, if anyone ever asks me why I ‘waste my time tweeting,’” Nesbit told us, “I have a pretty good response for them.”

Finding Jean-Marc Castera was like hitting the jackpot for Mission 4636. Instead of approaching Nesbit’s request with suspicion or reservations, Castera immediately agreed to get in touch with the CEO of Digicel, Haiti’s largest mobile phone telco. What he came back with was better than anyone could have imagined. He proposed that the entire disaster relief mission co-opt the SMS code 4636. Under normal circumstances, 4636 was a simple service providing weather information through an auto reply. Now it could function as an SMS hub for incoming disaster-related messages from Haitians in distress.

Turning that technical possibility into a reality, however, would prove a daunting, on-the-ground challenge, and here the effort saw new, largely unsung heroes enter the story: Eric Rasmussen and Nicholás di Tada from an organization called InSTEDD, a nonprofit organization dedicated to innovating new responses to diseases, disasters, and emergencies.

A year before the earthquake, the Thomson Reuters Foundation, based in the United Kingdom, had commissioned InSTEDD to develop the Emergency Information Service, a web- and mobile-based platform that could be used by journalists to help them, and survivors, better communicate in a disaster. Within hours of the Haitian earthquake, the foundation asked InSTEDD to fly to Haiti and set up operations. Sixty hours after the first tremor, Rasmussen and di Tada and several foundation staffers landed and set up operations at the east runway of Port au Prince’s international airfield. It was one of the first organizations of any kind on the ground after the quake.

From their base, the InSTEDD team was able to do several critically important things: First, they built and connected the underlying infrastructure for receiving 4636 text messages, working with both of the dominant telecoms in Haiti (Digicel and another company, Comcel) to ensure that 90 percent of Haitian mobile users could send in texts to the code. Next, they worked with Ushahidi to ensure that they and other groups could get access to the messages being sent in to the system. Again, the latent network of connections paid off handsomely: At the time of the earthquake, InSTEDD’s Vice President of Engineering, Ed Jezierski, was on site with Ushahidi in Nairobi and was able to ensure smooth integration between the two platforms.

Then, InSTEDD and Thomson Reuters Foundation teams traveled to radio stations all over Port-au-Prince and got the emergency short code announced over the airwaves. For all of its technical novelty, Mission 4636 would have been irrelevant without the old-school platform of radio: Everyone listening to the radio heard the message about 4636 and they helped to spread the word to friends and neighbors. Haitians hurt, trapped, or in need of food and water could text their message and geo-coordinates to 4636 and the information would be instantly transmitted and integrated into the dynamic map by crisis mappers.

Two days after the quake, most of the technical platforms were in place, the recruitment process for the crisis mapping volunteers was gaining momentum, and a short code had been secured. There was still one enormous obstacle to success, however: Almost all of the victims’ messages would be written in Haitian Kreyol—and almost all of the crisis mappers and first responders spoke English. Somehow Mission 4636 needed to find a group of people fluent in both Haitian Kreyol and English who could serve as translators. In the spirit of emergent cooperation, a most unlikely candidate stepped forward to take on the job.

A year earlier, while Nesbit was working for his own health-care start-up, Frontline SMS, in Malawi, he connected with a computer linguist named Robert Munro. After leaving Malawi, Munro returned to Stanford University to continue his doctorate research investigating ways to process large volumes of multilingual SMS messages. Twenty-four hours after the quake, Munro was taking a train from Palo Alto to his home in San Francisco. He checked his email en route and saw a new message from Nesbit.

“Josh asked me if I could adapt the SMS classification work that I’d been doing with him in Malawi for Haiti.”

By the time Munro arrived back at his apartment, he was already logged in on several email chains and Skype chats. He would spend every waking hour on Mission 4636 for the next ten weeks.

At first, he was deeply involved in the technical discussion with Ushahidi about how best to process the text messages that would filter in through 4636.

“As we scrambled to launch the technical platform,” Munro told us, “we realized that we didn’t have anyone to process the messages. I speak some Sierra Leonean Krio, and some French, but not enough that I was comfortable translating messages for many weeks. None of us had really slept more than an hour or so in the past three days, and it felt like everything I was coordinating was about to fail.”

Munro realized that he would have to take charge not only of finding translators but also of training and guiding them throughout the effort. Considering his lack of experience in management, this was a daunting realization, but the collective energy and momentum of the mission made it inconceivable to give up. He forged ahead, looking for ways to connect with Haitian Kreyol speakers.

Considering the minute-by-minute contact the 4636 team had with one another via Skype chats, one imagines that they all knew each other well, or that they were, at the very least, in the same time zone. The reality was just the opposite: Many of the individual players had never met in person, and many of the new recruits were communicating via the Skype chats from all over the world. While Munro tackled the task of finding Kreyol speakers also fluent in English, Nesbit was helping new partners plug into the platform.

“The ability to explain the system and the partners involved ended up being hugely important,” Nesbit told us. It is no coincidence that none of the technology from Frontline SMS—Nesbit’s organization—was in use with Mission 4636. “Having an independent connector with no skin in the game fostered trust much faster.”

Back in his apartment in San Francisco, Munro finally found a solution for contacting potential translators by using Facebook. The Haitian diaspora had started to cluster around several groups on the site, and many of them expressed a strong desire to offer their services. Munro was soon coordinating clusters of volunteering groups all over the world including Union Haiti, a group of Haitian expats in Montreal, and members of the Service Employees International Union (SEIU) in the United States, who worked tirelessly in long shifts to translate the texts. As a group familiar with the streets and signposts of Port-au-Prince, their tactical knowledge proved invaluable to the effort.

Within mere days after the quake, the entire system was in place and functioning. Radio broadcasts all over Haiti announced 4636 as an emergency short code. Haitians—many in desperate need—used their cell phones to send text messages into the 4636 SMS hub. These messages were written in Haitian Kreyol with street names and addresses that would make sense only to native residents of Port-au-Prince. The information was then instantly transmitted to Haitian Kreyol speakers all over the world who quickly translated the messages into English while using their on-the-ground knowledge of the city to find exact geographic coordinates on the map. This English information was then instantaneously sent to crisis mappers working around the world, who took each one of the translated messages and geotagged them on the real-time map. This entire process, spanning the entire globe and thousands of volunteers recruited from different organizations, universities, and companies, took approximately ten minutes in the earliest stages of the effort and less than two minutes toward the end.

The last and final goal—to connect these mapping efforts to the community of first responders—came when Ushahidi was linked to UNDAC (Disaster Assessment and Coordination), which directed search and rescue teams to the addresses with the greatest need. Within twelve hours, the Ushahidi map had become an important resource for many first responders from the United Nations, the U.S. Southern Command (SOUTHCOM), and the Marine Corps.

Together, the 4636 team, along with thousands of strangers from all over the globe, had built a pioneering digital disaster response system, largely from scratch—without a single organization or person in charge.

A testimonial on Mission 4636’s blog speaks to the effectiveness of the commitment exhibited by Meier and others:

“I cannot overemphasize to you what the work of the Ushahidi/Haiti has provided. It is saving lives every day,” Craig Clark, open source intelligence analyst for the Marines, wrote in during the earliest stages of phase one. “I wish I had time to document to you every example, but there are too many and our operation is moving too fast. . . . The Marine Corps is using your project every second of the day to get aid and assistance to the people that need it most. . . . Keep up the good work! You are making the biggest difference of anything I have seen out there in the open source world.”

Several days into the effort, Mission 4636 started shifting from phase one to phase two. According to Nesbit, “There weren’t a lot of people to be pulled out of rubble anymore. We moved to more general needs like food and water requests and medical services.”

The U.S. Coast Guard, along with Southern Command, became the primary response team during this second phase, helping to connect the victims with food, water, and health services. And an increasing number of organizations were using the maps to plan and coordinate relief efforts.

The Kreyol-speaking translators came together online to share maps and information, bringing members of the Haitian diaspora deeper into the information loop. Eventually, the job of translating the messages was turned over to Haitians still residing in Haiti. Through the collaborative support of CrowdFlower, a private company that develops technical platforms for crowdsourcing, and Samasource, a nonprofit specializing in microwork digital projects, the translation tasks were successfully passed off to paid workers.

There are no official statistics about how many lives were saved using Mission 4636, but CrowdFlower reports that the average response time to translate, map/geocode, and categorize a message never exceeded two minutes. In total, they processed more than 100,000 SMS messages and, at peak volume, more than 5,000 SMS messages were processed in one hour.

By the time Mission 4636 was finished, a whole ecosystem of collaboration—coders and software architects, volunteer mappers, Haitian American translators, Haitian citizens, NGOs, first responders from the Marines, the Red Cross, the State Department, and the UN all worked together on it—most without once having met one another in person.

TRIBES, NETWORKS, AND TEAMS

What made a collaborative effort like Mission 4636 work against such long odds? A big part of the story rests in the structure of the social network that tied together the volunteers.

Social networks are classically described by being comprised of weak and strong ties. Your relationship with a close friend or family member, for example, is usually a strong tie—rooted in shared experiences, a lot of trust, a deep sense of reciprocity, and a great deal of interaction. Because birds of a feather flock together (a phenomenon network researchers call homophily), you’re likely to have more in common with your strong ties, especially those you’ve chosen rather than inherited. These are your people, your proverbial tribe. A weak tie, on the other hand, is someone you might know only distantly—a business acquaintance, or someone you know distantly through a strong tie, or a friend of a friend.

In a classic 1973 study, Stanford sociologist Mark Granovetter interviewed dozens of people to find out what role their social networks had played in their landing a new job. He discovered that most had found their jobs through a weak tie—not a close friend, but an acquaintance. Ever since, the power of weak ties has been lauded in sociology and network theory; weak ties have been found to be important for all different kinds of social mobility and innovation diffusion. “Because weak ties connect people across very different neighborhoods of a social network, they are critical for quickly finding new information—like where a new job opening might be—that was unavailable to you and your immediate connections,” says Sinan Aral, an information economist at New York University. These weak connections form bridges over holes in the social fabric, where people are less well connected. “This can be critical, particularly in the case of an emergency, for propagating information across a network.”

But strong ties have a vital place too. “Having a team built on strong ties—close connections—is essential for the intense, collaborative work of creating or synthesizing new material, especially complex material,” says Aral.

In fact, Aral’s research suggests that Granovetter’s insight about the strength of weak ties may be incomplete. For all of the connectivity that weak ties provide, most people still get the majority of their novel information from their strong ties, not their weak ones. That’s because the frequency and intensity of the interactions between you and your tribe overwhelms the relatively infrequent and narrower exchanges you have with your acquaintances. “Let’s say you talk to your best friend three times a week and an old golfing buddy once a year. Even if only a tiny portion of what your best friend tells you is new, in the end, that’s still, proportionally, where most of your new information is going to be coming from,” says Aral.

This is part of a larger phenomenon called the diversity/bandwidth trade-off. As we expand the diversity of social connections we have, the bandwidth we can commit to each of those connections becomes more limited, and the information that comes across them gets weaker and narrower. And that in turn makes weak ties suitable for certain kinds of work and strong ties suitable for others.

“The most powerful constellation seems to be to have small, diverse teams of strongly tied collaborators, who each have a large and diverse weak-tie network—the best of both worlds,” adds Aral.

We see the interplay of strong and weak ties throughout the story of 4636. A platform like Twitter, for example, makes it incredibly easy to maintain weak ties—to passively follow people you don’t know well, and through them, to get to people you might need in an emergency. And that’s exactly how Josh Nesbit used it to find Jean-Marc Castera at Digicel. That was a linchpin event—one of many—and had it not happened, the project could not have succeeded as it did.

On the other hand, the volunteers weren’t entirely random groups of strangers meeting for the first time. At the center of the project were small teams made up of people who knew and trusted one another already—small teams dominated by strong ties, like the tribe of engineers building the software platform itself.

The real power of the 4636 project came in the way it enabled weakly tied volunteers to rapidly turn into strongly tied, committed collaborators. Though only a small percentage of the collaborators had met in person, the informal network provided a baseline of modest trust. Just as William Ury hopes to tap into the narrative of Abraham to remind the three religious tribes of their shared culture and values, the open nature of the 4636 platform—both the code that powers Ushahidi and the collaborative nature of the mapping—enlarged the tribe of disaster relief responders from officials at international bureaucratic organizations to a swarm of volunteers all over the world. The 4636 project was able to spread—as evidenced by Meier’s description of the growth of the volunteer network—because the platform allowed people to be recruited in to perform discrete, useful tasks, with a minimum of formal authority.

For Ushahidi-Haiti @ Tufts, this was greatly aided by the culture of graduate school itself—a collaborative environment filled with people with lots of unstructured time who could make face-to-face commitments with that time fairly easily.

The 4636 project also had leaders who modeled desired behavior in the network—commitment to a shared goal above all else—and didn’t show any sign of defecting from the common purpose. At the highest-stakes moments, players like Meier and Munro never gave up. They forged ahead with difficult tasks and offered encouragement for others to do so as well. The InSTEDD team worked tirelessly, and often thanklessly, behind the scenes for the good of all. Josh Nesbit set his own self-interest aside by taking his technology out of the mix, instilling trust by presenting himself as a neutral arbiter and connector.

Finally, there was a sense that the platform was showing results—feedback loops that motivated the team members and increased their sense of agency. The more they did, the more valuable it became; the more valuable it became, the more they wanted to do. By the time a volunteer like Sewell was on her third or fourth day of mapping, she felt not just needed but essential to the relief effort.

Yet we must be careful not to overglamorize Mission 4636. While it mostly succeeded spectacularly as a collaborative effort (which is why we explore it here), there are also some important places where it failed.

First, because the entire system was being built in real time, there was no process for deep integration with systems that drive traditional large-scale disaster response. Responders arriving in Port au Prince in the days following the quake faced a near-complete information gap: Many of the most basic datasets that describe a country’s infrastructure, its roads, hospital networks, schools, and water systems, were on computers now buried in the rubble. Many of the people who “owned” that data, whether in Haitian government, the UN, or various NGOs, were missing or dead. The country’s Ministry of Education, for example, was obliterated, and not a single list of schools—where one might house the homeless—survived. Finding and reconstructing these datasets obsessed the formal response organizations in the first hours after the quake. In this sense, there were important mismatches between what the formal institutions felt they needed, what the crisis mappers were providing, and how they were providing it. Paradoxically, even as they were missing the “usual” data on which to proceed on the ground, the volume and format of 4636 data gave some of the traditional players a sense of information overload—they were neither prepared for it, nor had protocols for how to act on it.

There was also a question about how crisis-mapping information was authenticated. How accurate were the reports? People in a crisis have strong incentives to sound alarm bells—with limited resources, how could the mappers and responders validate that the needs were real? This was why, on the ground, the Ushahidi maps were frequently used less to respond to singular events and more to establish emergent “centers of gravity”—trends and clusters in the data that showed where needs were emerging or still unmet.

Finally, there was a simple matter of scale. The new and nimble technology players were mostly tiny, and a few were barely organizations at all. Few had the infrastructure and proper resources needed to deliver replicable results—some could barely cover the cost of pizza to feed their volunteers. Weeks after the Haitian earthquake, an even larger quake hit Pakistan. Yet nothing like the collaborative 4636 model emerged. Why? Partly it was simple fatigue, and the continuing deep involvement of many of the best players in Haiti. The remoteness of the location from North America was also a factor. And there was a deeper issue: Haiti had represented a significant opportunity for many of these new, smaller organizations to demonstrate their value to their funders. Yet the process of sorting out who had contributed what, and how important each element was, was messy. Some organizations’ roles were overemphasized, and others underemphasized, leading to some hurt feelings behind the scenes. Especially in such fluid, collaborative contexts, proper credit and proper rewards matter all the more.

Yet for all of these concerns, there can be no doubt that the 4636 effort was a success. For the first time, a global community of technical volunteers, affected citizens and global diaspora communities made material contributions to a major disaster response. Many of the failures were unavoidable by-products of the effort’s novelty and will be addressed with improvements and the development of new protocols for next time.

The lessons are not just for disaster relief. Think back one last time to our bankers around the table at 33 Liberty Street. In a high-stakes weekend humming with the undercurrents of “us” against “them,” what might have happened if the structural elements in place had insured space for such a 4636-style, inclusive, collaborative, and innovative “third side” response? As we’ll see in the next chapter, recruiting a more diverse array of players—enlarging the tribe through both technological platforms and collaborative processes—not only might have changed the moral vector of the discussions, it could have opened up entirely new vistas in the modes of thinking around the table.