Sunday, October 10, 2010

Forecasting the future

The French scientist Pierre-Simon Laplace, when asked by Napoleon why he didn't mention God in his work on celestial mechanics, famously responded that "I had no need of that hypothesis."

But Laplace, one of many scientists to deal with Newton's deterministic view of reality, is best known for inventing the idea of a being that has complete, perfect knowledge of the speeds and velocities of every particle in the universe. With such knowledge, he could in principle predict the future of the entire universe, exactly and forever.

Setting aside quantum mechanics and its brand of uncertainty and sticking to classical physics, Laplace's idea stuck around for a long time. I had young friends who remained wowed by the concept. But I, no scientist, still smelled a rat. And sure enough, along came chaos theory.

Edward Lorenz stumbled onto the concept while using computer programs to study weather forecasting. It turns out that weather - and indeed nearly all of what you can measure of the workings of the natural world - is extremely dependent on initial conditions. If you could gather weather data from around the world with 100 times greater accuracy, you probably would only extend your ability to forecast weather for a day or two.

Newton's laws are indeed deterministic, but in practice nobody actually living in this universe could muster the precision to predict the future. Maybe Laplace's "being" could, if it knew each data point to an infinite number of decimal places, perform such a miracle. But, of course, that would make it a "God."

In this case, anyway, it looks like Laplace needed that hypothesis after all.

No comments:

Post a Comment