A Safe Prediction: There Are No Easy Answers

The Galileo spacecraft cruised the solar system for six years before it got to Jupiter--a distance of 2.3 billion miles. Yet its aim on arrival was picture perfect.

This is the kind of spectacular success that leads people to believe science is good at predicting just about anything: the next hurricane, the next cancer victim, the next stock market crash, the global climate 20 or 200 years from now.

Scientists often play into their popular image as soothsayers when they use the phrase “the theory predicts.” What they don’t explain is that “predicts” in this sense means predicting the present, not the future--what is happening, not necessarily what will.

Einstein’s theory of gravity, for example, predicted that light from a distant star should bend as it dipped into the gravitational field of a massive object such as our sun. He wasn’t predicting an event that might happen next week. Since time began, starlight has been getting bent out of shape as it skims by massive stars. Before Einstein’s prediction, however, no one thought to look for it.

Einstein’s theory also predicted that space could bend so much it could knot itself into black holes. He didn’t mean he thought black holes would suddenly appear at some time in the future. He meant they were there for the seeing, if only we could figure out where to look and how.

Examples abound. When Hertz heard about Maxwell’s theory that light was an undulation of electromagnetic energy in space, he “predicted” the existence of waves much longer than the eye could see and “discovered” radio.

*

A scientific prediction is less like a weather forecast than a train of thought. If x, then y. If curved space, then black holes. The better the theory, the easier it is to find whatever the theory “predicts” should be there.

Galileo got to its target not by predicting the future, but by following well-known patterns to their logical conclusions. Objects in motion follow well understood paths as they coast and fall and loop around bodies in space. If you know the patterns, it becomes a matter of mere calculation to get Galileo to Jupiter on target.

Unfortunately, nature has conspired to make even pattern perception unreliable under many common conditions. For example, no amount of understanding of the behavior patterns of atoms allows one to predict just where an atomic particle will be at a certain time. The best you can get is a probability. And if you do pin down the particle’s position precisely, you can’t say anything very precise about its velocity.

Encoded into physics as the Heisenberg uncertainty principle, this weird fuzziness of the subatomic realm seems to put a natural limit on what we can know.

But you don’t need to resort to Heisenberg to see that scientific knowledge has limits. Even such a simple prediction as where a particular drop of water will fall as it crashes over Niagara Falls is beyond our capabilities. It simply requires too much information.

*

For much the same reason, an experienced baseball player can predict the trajectory of a batted ball better than a physicist, even though, as physicist Frank Oppenheimer pointed out, the physics of batted balls is “ludicrously simple” and described by “one of the most highly perfected mathematical theories in physics.”

So perhaps it is fitting that the hottest field in physics today is known as complexity theory--in essence, a theory of unpredictability. It’s a strange kind of field because it covers everything from economic systems to human consciousness, from the formation of galaxies to the behavior of clouds, from the study of the Earth’s core to the evolution of stars.

But the core idea is simple: Take any simple thing--a water droplet, a star, the firing of a single neuron. Each, on its own, might be perfectly predictable. But put a bunch of them together, and you’ve got clouds, galaxies, the mind--all unpredictable phenomena.

*

The reason is that in complex systems, every part influences what all the other parts do--making the web of causes and consequences much too tight to untangle.

Curiously, certainty can creep back into prediction when the numbers get huge. Crowds are more predictable than individual people. You can’t predict whether a single coin will land heads or tails, but you can predict pretty accurately that a million coins will come up half heads, half tails.

Either way, the idea of science as society’s crystal ball is unrealistic--especially among scientists themselves. Or as geophysicist Raymond Jeanloz at UC Berkeley pointed out, “Not everything is describable or predictable by simple formulas. At some point, you almost go back to the farmer’s knowledge that it’s impossible to predict the weather.”