The hope of this book is that we can catch a glimpse of an epochal change in our understanding that makes sense of an array of phenomena that are puzzling on their own, and more so because we’ve so easily come to take them as normal. It is a change in how we think things happen, which means it is a change in how we think the future arises from the present, the next from the now.
And that changes everything.
Chapter One
The Evolution of Prediction
When I was a lad, you could beat the local weather forecaster’s batting average simply by predicting that tomorrow was going to be like today. Or so my fifth-grade science teacher told us.
Now, in a series of photos on the NASA website, you can graphically see one reason the app on your phone predicts the weather with scary accuracy ten days out.1 In 2005, satellite images showed the amount of water vapor—an important predictor of hurricanes—as colored blocks each representing fifty kilometers of land. In the images from 2015, the blocks are about one-tenth the size, providing far greater detail. As an article on the site puts it, “Imagine going from video game figures made of large chunky blocks to detailed human characters that visibly show beads of sweat on their forehead.”2
You’d think with these new predictive superpowers we’d be able to predict earthquakes by now.3 After all, these are planet-scale events caused by the collision of massive tectonic plates that move maybe a couple of inches a year. If you slowed footage of a head-on car crash down to that pace as part of a forensic investigation, you’d be desperately pressing the “Go Faster!” button on the playback device. But earthquakes can be cataclysmic because a tiny triggering event—perhaps the slowly crashing plates vaporize a stone stuck between them, or other uncountable incidents of stress finally reach the tipping point—can set off a cascade of energy that buckles the ground and tumbles towers. As our predictions in multiple fields get better by being able to include more data and by tracking more interrelationships, we’re coming to realize that even systems ruled by relatively simple physical laws can be so complex that they are subject to these cascades and to other sorts of causal weirdness.
We are thus simultaneously getting better at prediction and are being brought to recognize just how profoundly unpredictable our world is. We are less patient when predicted departure times are delayed but also more on edge about cataclysms, attacks, and our collective future. That is at least seemingly a paradox. But it is no mere curiosity, for, as we will see, how we predict shows us how we think the future happens and thus how the world works.
In this chapter we’ll explore the “before” of prediction. In the next chapter we’ll look at the “after,” or at least the “now” and the “emerging.” We’ll see that our enhanced predictive powers are based on our new technology’s ability to grapple with a world so detailed, so densely interconnected, and so variable that its complexity overwhelms our understanding.
Prediction’s Sweet Spot
Think about the times you have uttered a prediction. Very likely you didn’t just state what you think will happen and then clam up. You probably hedged it to some degree. “I bet the marriage of these two celebrities won’t last a year,” you might have said. Or, “If this candidate makes it through the debate without saying, ‘Good jobs at good wages,’ I’ll eat my hat.” The hedge words that acknowledge a lack of complete certainty—“I bet,” “If so”—help flag a statement as a prediction.
Predictions live in a sweet spot between surprise and certainty. That’s why “It looks like it’ll rain on Friday” is a clearer example of a prediction than “The sun will come up tomorrow”: tomorrow’s sunrise is too certain. On the other hand, who will be invited to your newborn’s eightieth birthday party is not subject to prediction because it’s too uncertain. A prediction is a statement that attempts to say what’s going to happen when there’s room for belief but also for serious doubt and disagreement.
Still, not every such statement is a prediction. If I ask you why you picked a particular lottery number and you say, “I just have a hunch,” your pick isn’t a prediction, or at least it isn’t a good example of one. In the clearest cases, we expect there to be some grounds for a prediction. For a prediction of rain, those grounds might be statistical (“It usually rains a lot in April”), might be based on a model that uses scientific laws (“A mass of moist warm air is going to collide with a cold front, causing precipitation”), or might come from the far more complex data models used by machine learning that we’ll discuss in the next chapter.4
These two characteristics of predictions—we have a degree of certainty, but not too much, and we have grounds for that certainty—makes them an essential clue to understanding our ideas about how things happen. If you predict that it might be rainy tomorrow because tonight you saw dark clouds on the horizon, you’re probably speaking not only from experience but because you also think there are scientific theories that link tonight’s clouds with tomorrow’s rain—something about the steadiness of the movement of clouds through the atmosphere, the composition of clouds as water vapor that can condense, and so forth. If the number 66 bus doesn’t arrive at 8:17 as the printed schedule says it should, you may suppose that the traffic was bad or one of the earlier buses broke down, or you might just accept the bus system as not very reliable because of the complexity of life in a city.
In fact, predictions are such a particular type of speech that simply having them in a culture’s linguistic toolkit reveals much about how that culture thinks about the world: What causes change, how regular is it, and what is the human role in its occurrence?
Let’s look briefly at three early cultures that understood the world in ways that kept them from making what we would today recognize as predictions.
* * *
For three thousand years, the Egyptians held to a cyclical view that year after year proved itself to be true: the seasons came and went, life in the farms and villages remained basically the same, and the idea of progress was as foreign as soft-serve ice cream. The Egyptians didn’t even bother to give sequential numbers to their years: each new pharaoh reset the clock to year one and increased it by one when the taxes were levied every two years.5 A cyclical culture that remains unchanged for millennia is not a culture of predictions.
The ancient Hebrews had a more linear view of time, but still didn’t make predictions the way we think of them. They had a promise from God that someday they would be returned to the Promised Land and the world would be redeemed, but a promise is not a prediction. That promise began a grand linear narrative for them—very different from the Egyptian cyclical sense of time—but the completion of their journey was conditional upon the Hebrews’ fulfilling their side of their Covenant with God. That’s why the words of the prophets are generally too intended to influence behavior to count as predictions: if our people continue in these wrongheaded ways, then we will face deprivation and punishment, but if we follow the word of God, then we will be blessed and our people’s story will continue.
For the ancient Greeks, the situation was different depending on whether they were looking up or down. Looking up, they saw the same wheeling stars that the Egyptians did, and believed in their regularity just as firmly. But here on Earth, there was no telling what would happen. As a culture of sailors and traders living in tumultuous times, the contrast between the predictable, orderly heavens and the randomness of life on Earth was for the Greeks a fundamental fact of life.6 After all, one morning the playwright Aeschylus woke up not knowing that an eagle would that day drop a tortoise on his head, killing him—out of the blue, quite literally.7
Life for mortals was so irregular because it was controlled by so many intersecting forces. The Fates determined your life span and some of the broad-brush themes, such as whether your marriage was going to be happy. The gods could not undo the Fates’ decrees, but they could change a mortal’s life in just about any other way they felt like at the moment. Then there were the ancient spirits, or daimons, who
intervened in individual lives in untamed ways. So it was quite a mix of superhuman forces that determined the turning points in one’s life, with only limited human control and predictability.8
True, the Oracle at Delphi had divine access to the future—or perhaps “was a priestess in a cave who became disoriented by volcanic fumes and babbled incoherently,” as an article in Scientific American claims—but her pronouncements were so cryptic that sometimes they couldn’t be applied until it was too late.9 Just ask King Oedipus, who knew the words the Oracle had proclaimed but still ended up saying “Mom???” to his wife before plunging a knife into his eyes.
That’s why the Greeks talked about the future not as what lay ahead of them but as what was behind them, according to the late Harvard professor of Greek, Bernard Knox. The future, he explained, was for the Greeks primordially unknowable—as invisible as what is going on behind us.10 Nor were the Greeks alone in this. Some maintain that the ancient Israelites talked about the future in the same terms for the same reason.11 Likewise, a scholar of African religion writes of “African peoples” who consider time to be a present that “moves ‘backward’ rather than ‘forward.’ ”12
When the future is so unknowable that we think of it as perpetually behind us, predictions are no more possible than are prayers for societies of atheists or limericks in languages that have no rhymes.13
* * *
The point of this mini history is not that those three early cultures were too dumb to come up with the idea of prediction as we understand it. Rather, it is to make it clear that our idea of prediction requires a delicate balance of ideas about how the world works. In turn, predicting reveals how the world works in very particular ways.
For example, let’s jump ahead to the origins of modern weather forecasting. In 1900, a Norwegian scientist named Vilhelm Bjerknes thought that we could understand the dynamics of global weather using just seven variables—the three spatial dimensions and the air’s pressure, temperature, density, and water content—and laws derived from the work of Isaac Newton.14 At last we had a model that explained how weather happens, enabling us to make predictions based on the laws of physics, rather than extrapolating from observations of the weather’s prior behavior.
Over time, scientists refined Bjerknes’s crude model and eventually ran it on some of the very earliest computers, making weather predictions generally good enough to tell you if you should take an umbrella to work, but still distressingly unreliable, especially for forecasts more than a couple of days ahead. Still, for all of its deficiencies, Bjerknes’s approach had at last fully brought the weather into the realm of modern predictability.
If you predict tomorrow’s weather by choosing seven factors governed by Newton’s laws, you’re revealing the world as a place that is orderly, rule-based, and knowable, at least at the large scale. If you foretell the weather for the coming season by reading bird entrails, you are revealing the world as the sort of place where what happens depends on hidden connections of meaning.15 If you can tell that a storm is coming by sticking your finger into the wind, you reveal the world as consisting of systems so tightly interwoven that a single factor can tell the story of the whole. If you predict the weather by doing a statistical analysis of years of prior data, you reveal the world as governed by laws that may be unknown or too hard to apply because of the many factors but that tend toward repetitive patterns. And, to switch examples, if you use A/B testing to decide which version of an ad will generate the most clicks, you reveal at least the online world as a place where the causes can be so minuscule and contextual that the old ways of predicting don’t work and don’t much matter.
The story of prediction is therefore also the story of how we have understood how what happens happens.
To see how that story is about to hit an inflection point, we have to go back to the sweet spot that allowed the emergence of what we now think of as prediction. That sweet spot has a name: Sir Isaac Newton.
Tick. Tock.
Newtonian physics gave us a universe ripe for prediction. It’s governed by rules. Those rules are knowable. They are the same for all things and throughout the universe. And they apply to a universe the state of which we can know well enough to make predictions but not so well that we are omniscient gods who don’t need to make predictions.
Crucially, Newton’s simple, knowable laws are sufficient to explain how things happen. They don’t need gods, fates, or spirits to intervene. They don’t require us to stipulate that, as acorns become oaks, things change in order to achieve their essence, as the Greeks believed. Newtonian predictions reveal the universe as a self-contained clockwork that can be explained purely on its own terms.
Think of the innards of an old windup watch or grandfather clock, each gear meshing so perfectly with the others that the only noise the system makes is the tick of the mechanism that releases the clock’s pent-up energy all at once, moving the second hand one notch. The mechanism follows rules: this gear turns that one, and that one turns the other one, one tooth at a time: simple rules with simple results. The clockwork is perfectly knowable: you can open up the clock and trace how it works. Its next state depends entirely on its current state and thus is predictable: if the time is 12:01, in sixty seconds the minute hand is going to point to the 2, not the 57. If you can’t predict exactly where a clock’s hands will be in one hour, you’re looking at a clock that may not be right even twice a day. In fact, you might check that you’re not staring at your clothes dryer.
Clocks thus became a standard way of understanding and expressing how things happen: perfectly regular changes that occur in small increments, following simple, knowable rules that would be entirely sufficient to explain the change … if we had perfect knowledge. But, of course, we don’t. That’s what opens the space for prediction.
In 1814 Pierre-Simon, Marquis de Laplace, drew an inevitable, and unsettling, conclusion from Newton’s work. He imagined a godlike intellect—frequently referred to as “Laplace’s demon”—who could know the position of every item in the universe at any one instant, all the forces at play, and the Newtonian laws governing them. (This demon will pop up more than once in this book.) “For such an intellect,” Laplace wrote, “nothing would be uncertain and the future just like the past would be present before its eyes.”16 This great intellect could apply Newton’s laws to any single moment in history and deduce the entire future of the universe, as well as its entire past—“postdicting” (so to speak) as well as predicting.
Laplace was sometimes called “the Newton of France” because of his work extending the use of Newton’s gravitational laws, including to explain phenomena that seemed not to fit, such as minor perturbations of the orbits of Jupiter and Saturn.17 But in the previous passage, Laplace draws the conclusion that made Newton personally uncomfortable with the clock metaphor: once the universe is set in motion, it doesn’t need God’s help to keep it going. The idea that God might have no post-Creation role in the universe was unthinkable for a person as deeply religious as Newton. But he thought he saw a way to save a place for Him. Newton’s theory of universal gravitation—every body attracts every other—meant that the combined gravitational pull of all the objects in the universe would slowly draw the planets out of their beautifully elliptical orbits. Newton hypothesized that God might therefore have to lob the occasional comet on exactly the right path for its gravitational pull to tug the heavenly bodies back into their perfect ellipses.18
Laplace, on the other hand, was a forthright atheist who felt no need to make work for God, or even to defend free will, which he viewed as “the expression of our ignorance of the true causes” of our actions.19 Because we humans can’t know everything about the universe at any one moment, prediction is for us a matter of probability, as the title of Laplace’s book—A Philosophical Essay on Probabilities—makes clear. But since everything that happens is completely and perfectly determined by prior causes, it is perfectly predictable to an all-knowing demon.
The clockwork metaphor stuck with our culture. It felt natural because clockworks were exquisitely built, serving as representations of the magnificence of God’s own handiwork. And clockworks are completely understandable within themselves, like a system of logic, making them an appropriate metaphor for the newly dawning age of reason.
Over one hundred years ago, Albert Einstein dealt the death blow to the clockwork metaphor among physicists; not only isn’t the universe like a steadily ticking clock, time itself isn’t a simple sequence of synchronized ticks and tocks. Then digital clocks undid the metaphor for the general public. Yet in ordinary life we still assume that everything that happens—every tock that follows a tick—arises in a knowable and determinate way from the present state of the mechanism. We still make predictions using laws that let us “skip ahead” to see the future—for example, predicting how far a plane will have flown in four hours as easily as predicting the distance it will have traversed in two hours.
Beneath these predictions is a confidence that not only are the rules pretty simple, but simple rules create a predictable world. We can think of this as the first level of predictability.
But from the beginning, Newton knew that there’s a second level where simple rules get complicated quickly.
A Summer of French Aristocrats
In 1676, Isaac Newton modestly wrote to his rival Robert Hooke, “If I have seen further it is by standing on the shoulders of Giants.”20 Yet in the preface to his masterpiece, the Principia, there is only one giant whom Newton acknowledges: “the most acute and universally learned Mr. Edmund Halley,” who not only edited the book but was responsible for nagging Newton into publishing it.21 Halley even helped to finance it after the Royal Society had blown too much of its budget on a history of fish, thereby avoiding what would have been perhaps the worst trade-off in history. Newton appreciated his friend’s efforts: he sometimes referred to the Principia as “Halley’s book.”22
Everyday Chaos Page 3