HEY!”
She looks up.
Jeremy starts to run. “Hey!”
He picks up steam, steps into the bicycle lane on the edge of the street to avoid a half dozen colleagues walking, spilling out from under two shared umbrellas.
The woman slips around to the driver’s seat, hops in. The car starts to peel away. It’s something bland, Jeremy thinks, a blue-gray Hyundai. He’s in the street now, fully unleashing healthy, practiced legs, decent lungs, and DNA that made him a sufficiently capable track athlete to win a Rhodes. Not sufficiently capable to catch a sedan, accelerating. He recognizes the woman, right? Same person from the night before?
He hears the horn. From behind. Another car approaching. It swerves, splashing rainwater onto Jeremy’s jeans.
It dawns on him he might want to get a ride from the car passing him, try to chase the sedan. Instead, he finds himself yelling: “Watch it, asshole!”
Brake lights go on in the car, a BMW. The driver slides down his window, then thinks better of it. Takes off. Jeremy yells: “The prom queen called. She wants her low-end Beamer back!”
Ten minutes later, damp, furious, Jeremy descends into a packed BART station to head to Berkeley. Discovers train delays. Thirty minutes due to some refuse on the tracks near City Center in Oakland.
Jeremy forces his way onto a bench in the tunnel and pulls out his iPad. He looks at the map. Red, red, red. He opens a new window and clicks on a link that will let him delve further into the variables that, allegedly, have prompted the computer to predict war. Chief among those variables: changes in conflict rhetoric, language that presages war.
While the computer whirs, calling up the data, Jeremy marvels at this particular capability—the one that allows him to track the language of the world. It is, to Jeremy, one of the most powerful tools afforded by the Internet. It is the equivalent of giving the world a blood test, taking its temperature, assessing its mood. Or, rather, it will become that. Eventually. For now, the Internet is remarkable at capturing what everyone is saying, and even organizing that data—by region, topic, media (Twitter versus blog versus newspaper), communicator (politician versus CEO versus activist).
It was amazing for Jeremy to watch the Arab spring and the protests in Russia, organized around Twitter feeds, and social networks, spontaneous calls to action in which the language elicits, organizes and stokes conflict. An amazing nearly one-to-one relationship between words and thoughts and action.
More broadly, Jeremy thinks that this development of mining and sifting the world’s conflict rhetoric could help answer an age-old philosophical question about the relationship between language, thought and action.
Philosophers and linguists have for millennia debated the relationship. To what extent are the words we choose insights into what we think—not what we want to communicate but what we really think? On one hand, of course, it is very easy to lie about what we’re thinking, making words fundamentally untrustworthy. Clearly, Hitler did intend to invade Russia, despite his protestations otherwise, and George W. Bush did not believe Iraq had weapons of mass destruction. Words, not true thoughts. To that end, Plato and others who engaged in the most powerful early analysis of communication thought of language in terms of rhetoric: not what do we think but what do we want to communicate? Or, put another way, what do we want other people to think we think, or what do we want to persuade them to think?
But where Jeremy thinks the Internet is so powerful is in the way it creates such a huge sample size of language that it betrays what we, the human race, think. All the linguistic data, unprecedented insights into the human psyche, a global inkblot test, a linguistic prism into our collective subconscious, where we’ve been emotionally and where we’re going, a digital augur, or, as Jeremy sometimes prefers, the Freud machine.
Part of a larger embrace of Big Data, aimed at predicting the stock market, of course, but also forecasting political tastes, weather, consumers’ susceptibility to pricing changes and their shopping desires.
Just before Jeremy’s fall from grace, there was a party held in his honor at a two-tone mansion off Fillmore. It was hosted by one of the handful of Silicon Valley socialites with that knack for convening, at a moment’s notice, fifty interesting and smart people, along with journalists. The evite heralded: “Make PC, Not War.” Computers, not conflict.
Partygoers came mostly for the interesting cocktails, using pomegranate juice or fresh leaves picked from such-and-such garden. But also to connect to, or stay connected to, Jeremy, just in case he became the next Zuckerberg. His own vertical, his little satellite world, opportunists, but also, given the highly politicized nature of the quest, ideologues, academics, curious government officials domestic and international and, of course, venture capitalists with military backgrounds—a staple in Silicon Valley going back to the creation of the region by the military brats who started Hewlett-Packard.
What they didn’t know, or some may have only sensed, was that Jeremy soon would be not the next Zuckerberg but, rather, the next Hindenburg. A week before the party, the Pentagon had told him the conflict algorithm had failed. He was waiting for Andrea to call to tell him when he could get on an airplane to Iraq, or Afghanistan, to check the results for himself, do the field research. He’d already packed a bag so, per Andrea and the Pentagon’s admonitions, he could pick up and climb on a plane at any moment.
Meantime, he pacified himself with figs and goat cheese and high-potency frozen vodka.
Drunk, in the shadows, he overheard the half dozen late-night stragglers descend into an argument about whether Big Data could be used to predict individual behaviors, like the likelihood of suicide. For instance, could someone’s communications or movement patterns—as measured by a mobile phone—be a predictor of whether the person is getting depressed or uncommunicative, suggesting eventual suicidal ideations?
Could it predict whether someone is becoming hostile, or even homicidal?
Jeremy heard one of them say: “I have a prediction. Sometime soon, Stillwater will let loose with a hostile outburst.”
Another said: “You don’t even need a calculator to tell you that.”
Lots of hushed laughter.
Jeremy, unseen, receded to the basement, to a guest room. He called Emily, who told him to calm down. He hung up and slammed the iPhone down on the bed. I’ll fucking show them.
He passed out that night, furious, with an idea, the seeds of “Program Princip,” the algorithm named for the assassin of Archduke Franz Ferdinand. Jeremy’s idea was to try to go deeper than merely understanding when a conflict might start, to understanding who might spark it. Not the obvious person—like an aggressive and politically motivated President Bush—but someone lesser known. Could Jeremy take all the increasingly public information about the connections between and among people and figure out who belongs at the center of a conflict?
Another brainstorm, a terrific one, marketable, at least. But personal too. At its heart, Jeremy wondered if he might be able to use the program to figure out who was lying to him—Andrea, Harry, Evan. Who, even then, was undermining him? Who is Jeremy’s enemy? Or who is his most central enemy?
It never dawned on Jeremy that night that the very problem might be Jeremy.
In the present, his computer beeps. It is returning the data. Unfolding before him, the report on where the rhetoric of the world has changed so much that the computer now projects Armageddon. T minus three days. Fighting off fatigue, he takes in the screen, a menu at the top. It reads: “five entities.”
Jeremy purses his lips. There has been a material change in the language of conflict in five different categories. The categories materialize:
North Korea (related), + 14 percent
Moscow (related), + 9 percent
Mexico (related), + 46 percent
Congo (related), + 6 percent
Fertile Crescent (related), − 12 percent
He taps his right toe on the ground, nervous energy. With an index finger, he points at the word “Mexico.” His finger is so close it nearly touches the screen. Forty-six percent—a veritable explosion in conflict language coming from Mexican authorities or from other figureheads around the world relating to Mexico. Something’s going on to the south.
He calls up another window with the map. It’s seething red, pulsing around the globe. He runs his mouse over Mexico. A pop-up box shows: 112,336,538, 8.8 percent. The population and its annual growth.
He looks on the bottom right of the map, at the countdown clock. 55:19:27. Hours, minutes and seconds.
Jeremy startles at the buzz in his pocket. Private number. He picks it up from the desk. “I’m on the Do Not Call Registry and have retained a legion of rabid plaintiff attorneys to enforce it with threat of lawsuit and execution.”
He hears nothing on the other end of the phone. Then a sound in the background, a horn or loudspeaker, maybe a hint of breathing from a caller. Then the line goes dead.
He recalls the list of regions with changing rhetoric: North Korea, Moscow and Congo experiencing rising tensions. Might mean something. The Fertile Crescent, the Middle East, experiencing dulling tensions. That’s at least one variable he can toss out.
He slips the phone into his pocket. On his tablet, in the window with the Rhetoric statistics, he swipes the line about Mexico, and discovers a place consumed by the language of war.