Saturday, July 26, 2008

Defining Theoretical Success

I'm clearly never going to understand string theory, but as it has no falsifiable results, I just don't see the point. Or rather, I'm unimpressed by elegance and consistency, because if you take a fixed set of phenomenon, people are smart enough to create an elegant, consistent set of mathematics to describe it. But only if it predicts something new, integrates (simplifies), or solves a puzzle. I don't know how many times I've read an economic paper that has 'rich' results because it can accommodate everything in some very complex way. If you think that math is good in of itself, this is really great stuff, but math has too many 'degrees of freedom'. Sure there are examples of neat math seeming irrelevant and then becoming really useful, but those are exceptions (Riemann spaces and General relativity). Neat math is not useful in itself, especially today when there are tens of thousands of theorems published every year.

Leonard Susskind, one of the founding fathers of string theory, is interviewed here. He is asked, "Can you cite any published results that support the main contentions of string theory?" He answers, "Yes! The existence of gravity, the existence of particles."

That's what I would call an 'in sample' result.

No comments:

Post a Comment