Wednesday, January 06, 2010

Can Physics Predict the Future?

You and I each make predictions about the future continuously throughout our day. As we drive to work, we predict that the path we take will look much like it did yesterday. When we practice a sport, we begin to learn how hard we need to kick or hit a ball to make it go where we want. And soon we understand the forces of gravity, friction, and our own muscle, enough to score a goal or block a line drive. Each of us is constantly predicting the future. And sometimes we're right.

So if physics is a science meant to break down the world that we live in - picking apart it's smallest particles, figuring out it's hidden mechanisms, and working out the equations that describe it - then shouldn't physics be able to predict the future?

This is a question, or an assumption, that many people have about physics. And I'll tell you right now that it's wrong.

Physics can, to some extent, tell us what to expect from the world. Just like you and I can reasonably assume that because our path to work or school has been mostly the same every day for the past however many days, weeks, and years, it will be pretty much the same tomorrow. Just like we can learn to play a sport and begin to control the outcome of the game. But the goal of physics simply isn't to predict the future.

The goal of physics is to understand the world as it is now; here in the present. Physicists want to know why things are the way they are, why the world works the way it does. Using that knowledge, it is often possible to learn more about the world around you than you originally knew. We can't see black holes, but by learning about the world around us, and coming up with equations to describe it, we can find other ways to look for black holes. We know that matter around a black hole will behave in a certain way, so we look for that behavior and aha! A black hole. This kind of discovery requires that we uncover the rules and equations that describe the world; that make it run. Some scientists are even searching for a single equation, a "theory of everything," from which all other equations could be derived. Could such an equation predict the future?


K.C. Cole, a tremendous science writer for the LA Times and author of an incredible new book on the life of Frank Oppenheimer, wrote a short post over at the NPR Blog 13.7: Cosmos and Culture, about the notion of a "theory of everything." She makes the point that although scientists may (heavy on the 'may') find a single equation from which all other equations and governing limits of the universe can be derived, it will still not allow us to predict the future. That was never the intention of figuring out physical laws.

If a fortune teller told you that tomorrow you would go to work and the road you drove on would bend in the same direction that it always has, that the building you work in would be in the same place, and that the sun would rise and set at about the same time it did yesterday - well, you'd be pretty disappointed with that fortune teller. Even if she was right.

You'd want to know about something much more complex and less predictable. Lets say you chose, on your way to work or school, to forget about your destination and allowed yourself to be pushed and pulled by the stimuli in your world - turning a corner because a gust of wind pushed you that way; crossing a street because the person next to you did; pausing for a moment because you smelled something good - well, then that same fortune teller would probably fail to tell you anything that actually happened. Even you might not be able to predict where you'd end up. The possibilities would overwhelm both of you. And so do possibilities overwhelm physics' ability to predict what will happen next.

One of Isaac Newton's three laws states that an object in motion tends to stay in motion unless acted upon by a force. Seems pretty straight forward. Consider the above scenario where you decide to let go of the force of will that moves you toward work or school each day. Could you predict where you would go? You could be pretty sure that you would not start to float upward and land on top of the Empire State Building (but you might take an elevator there). You will obey the laws of physics, and yet the many ways in which those laws will act upon you would overwhelm your ability to predict your fate.

Now, as a human, you have more control over where you go, so Cole uses the example of a drop of water in a water fall. As that drop of water reaches the crest of the fall, if we could pause and try to predict it's path and where it would land, we would be utterly overwhelmed. We'd have to incorporate the motion of every other drop in the river, the force of gravity and wind, the force of the rocks below. The equations would grow to fill volumes, they would swamp a supercomputer, and we'd be lost. We can assume that the drop of water will obey the laws of physics - but those laws will act upon that drop of water in so many ways that we will lose all hope of figuring out where it ends up before it ends up there.

Systems like this one are too complex to predict. And unfortunately, nearly everything we really want to know about the future is more complicated than predicting what your trip to work will be like. Even then, you never know what will happen. Who will you bump into? What sounds will you hear as you walk? What emotion might overtake you when you see something beautiful or ugly? Will you run into the person you are meant to marry? We can't say for sure until it's already happened. And the goal of physics is to best describe the current situation we are in. What does the universe look like right now? Why do things behave the way they do? Those are questions that physics hopes to answer.

That's not to say that we can't predict how things will go at all, or that we can't use those predictions to our advantage. We can't predict the path of a single electron for very long, but we can still utilize electricity. We can still count on our computers to work (most of the time). And as science progresses, we can get a clearer and clearer idea of what is happening when these things work the way they do. In essence, there are fundamental rules and regulations for the universe - like the force of gravity - that help us understand the way things are. That is, gravity helps us understand why we don't float away. But they can't predict the future.

"But wait!" you exclaim (I knew you would); aren't things like Moore's Law an example of scientists predicting the future? Moore's Law shows that about every two years, computer processor speeds double. So it predicts that in four years, processor speeds will be four times as fast, no? Isn't this a prediction of the future? Well, yes and no. Maybe now we're getting to the point where we need to define what we mean by predicting the future.

Is predicting the future seeing something that can't possibly be determined by looking at cause and effect? Fortune tellers promise to reveal things like the date you will get married, how many children you will have, or how much money you will make. Of course, to get specific about these things they'd need to know everything from your genetic history to what the weather will be like for the next thirty years. If, say, you believed that a fortune teller instead was given this information without having to know all those things; that he or she had some sort of phone line running into the future, well, then, I can't say I'd know what to say to that. So lets say that "predicting the future" means using cause and effect to determine the outcome of something.

So with Moore's Law, scientists are, in a way, predicting the future. Moore's Law doesn't tell us who will come out with the faster processor first, or what they'll eat for breakfast the day they announce it. And Moore's Law would fail if all the engineers in the world stopped working. It's not a law that is set in stone. Instead, it's more like predicting that the lights in your house will turn on assuming your bill is paid and the wiring is in good shape. It's likely, but not set in stone. Though we can't control every cause and effect in the world, we can use what we know of how the world works to reasonably predict what those causes will lead do; what things will be like tomorrow, or ten years from now. We obviously couldn't do much with our lives if we didn't understand cause and effect well enough to have a good idea of how many things would go. We predict that big buildings won't spontaneously switch places (although you could insert a quantum argument that it's not impossible), we predict that machines we build will run the way they did yesterday, and we predict that the sun will rise. And because of science, we know why the sun appears to rise. We'd know enough about the current state of the world to deal with situations that we can't predict.

Now I'm not even mentioning quantum effects - because once again, this is more a qualitative discussion about the objective of physics, rather than a discussion of the nitty gritty details that prevent us from estimating what will happen next.

So just because we can't predict the future doesn't mean we have no idea what will happen. What is more important to understand is that the purpose of physics is not to predict the future, but to describe the system we live in now.

That said, there was an interesting article in Wired today about the work of New Zealand-based physicist Sean Gourley, who gave a talk at TED discussing how he and a group he was working with thought they'd come up with a model to predict insurgencies in war. My immediate reaction to this is that there would far too many variables, not the least of which is human decision making, that determine what happens in a war. To think that we could come up with an equation to model these events seems to simplify an incredibly complex situation. It would be like a single equation to show the path of a water droplet down a water fall. But Gourley and his team thought they had found it, and last month they had a paper on the cover of Nature called “Common ecology quantifies human insurgency.” Nature is the top science journal in the world, so approval of this research gives it some major support.

Wired contributor Katie Drummond has some major issues with this. Namely, that the model Gourley and his team used was based on information that came from the media. So their map of insurgencies only includes those that were reported, and only those that included fatalities. Unfortunately, it's sort of widely known that the media coverage of insurgencies isn't always accurate. Partly because of how close the media can cover the entire war, and how much the military is willing to share. This is only part of Drummond's objection and I encourage you to read her piece.

The article seriously calls into question whether or not Gourley's model is worth anything. But the objections are mainly to his methods: what he uses for input data to construct the model and assumptions he makes about what defines an insurgency (does it only count if the insurgent forces start it, as opposed to the counter-insurgents?). With more accurate data and some discussion with army personnel to define the parameters, could we make a model to predict the patterns of war? I'm sure it's been attempted and is part of military strategy, but who knows how many complicating factors you would run into when trying to make a universal equation to describe all of these conflicts.

That's not to say there haven't been studies of systems that depend on human behavior: things like traffic patterns and the best way to board airplanes come up rather frequently. Albeit, those could definitely be viewed as much simpler systems, with much more concrete definitions and traceable data. It would certainly be wonderful if we could use physics to learn something about wars, with the ultimate objective of ending them faster and with fewer casualties.

But beware of those peddling snake oil: no one has a phone to the future.


  1. I don't agree with you. I've had plenty of experiences to prove this theory wrong.

  2. Mr.Anonymous where is your explanation? I would like to see it. I remain unbiased in evaluating the problem. I just want to see your side of the argument in contrast with this article.

  3. I liked how you described in detail on the reason why it can not predict the future, i am using quantum mechanics for school, i am writing a short story about scientific fiction to predict the future :) keep up the good work

  4. No particle in the Observable Universe(except Blackhole) can disobey the laws of motion of Physics. Even a 'thought' in the Human Brain occurs due to pre-existing thought and surrounding environment directly affects the thinking. Prediction of Physics (that too precisely) is possible. Even the motion of a tree leaf can be known after a week. No doubt that the equation predicting future would be a bit complex, it would give us most accurate result. Let's say for an instance that a man is travelling on his bike(consider Ideal isolated situation) at a constant speed v. We can accurately determine his position at time t by a simple equation d=vt. The equation above is also predicting future but at a smaller scale. I strongly believe that there is an equation that governs the world. Mathematics is a powerful tool. If I am not wrong, the equation has already been discovered...

  5. you need to change the title of this article.

    can physics predict the future? maybe. everyone with half a brain and access to google knows that this issue can only be decided by looking at the potentially random nature of the universe below the atomic level.

    the argument you use does not match the title. the title you use should be "can we predict the future". because all this article says is " once there is more than a few variables it gets too complicated to predict". the only reason this is true is due to lack of data and processing power, as in that the limiting factor is our own technology. this is not related to the nature of physics, only what we are/will be able to achieve.