Intermediate Form

A New Kind of Science

Previous Entry | Home | Next Entry


Yesterday, Stephen Wolfram came to campus to discuss/promote/sell his book, A New Kind of Science. In it, he makes some interesting points. I think that his idea of using CS principles as a tool to understand natural science is interesting, as is his idea that the natural world may be subject to some of the same undecidability that computer programs are.

(Undecidability in a nutshell: There is no general way to figure out what an arbitrary program will do except to run it.)

Unfortunately, Wolfram's work has problems. An important one is that he doesn't give much of a reason to believe that what he thinks is right. He switches freely from speaking about models of the universe to speaking about the nature of the universe, making it hard to determine what he believes it is possible to know about the universe itself.

His main principle is the use of cellular automata, although he also uses other systems. (A cellular automaton uses some number of neighboring cells to choose the color of a cell in the next generation.) He seems to like to study cellular automata by enumerating all possible configurations. This is an acceptable method when used with the simple examples he gives in his book, which have 2 colors and 3 input cells, for a total of 64 possible configurations, discounting mirror images and color-swaps.

But when we expand to 5 input cells, there are 1,073,741,824 possible automata. This is a large number, but perhaps still feasible. With 7 cells, we get


possible automata. This is far too many to look at all of them. Wolfram didn't mention any method of study other than enumerating all of them.

I also question the value of a theory that claims that most problems are undecidable. Modern natural science at least tries to come up with theories with some predictive power. Wolfram doesn't do that, making his new kind of science less useful than the traditional kind, even if he did stumble onto something fundamental about how the universe works.

(I actually like the idea of the universe being some sort of computer simulation. That's the answer to the uncaused cause problem that most appeals to me. But I'm a geek, after all.)

Overall, I consider Wolfram's New Kind of Science to be quite similar to a Segway. Both are somewhat interesting, but not nearly as important as their inventors claim.

- Tom | permalink | changelog | Last updated: 2003-03-08 11:57

Previous Entry | Home | Next Entry


Commenting has been suspended due to spam.