Wednesday, August 14, 2002

Doug Turnball has taken my idea of "satisficing" (which was perhaps at the heart of this recent conflict, as it formed the basis of my initial defense against Steven Den Beste's critique) and has explained how it relates to the scientific method beautifully. He discusses the apparent "perversity" of scientists attempting to disprove hypotheses instead of proving them, how often in science there are few enough hypotheses that proof of one hypothesis can serve to disprove others (which is rarely the case in analysis; there's always a metric assload of things a human being or collection of them can do, whereas particles are usually somewhat more predictable), and why the problem of satisficing doesn't necessarily lead to some sort of PoMo hell where nothing is ever remotely provable- a topic that I hadn't touched on.

I've heard variations on it before, but I really like his explanation of what science actually is:

Science, by its own admission, is not an attempt to arrive at Truth. It’s an attempt to arrive at a valid explanation which has predictive power. That is, I want to come up with a theory that explains existing data and will allow me to predict the result of experiments in the future. A scientific theory only gains acceptance if it works, in this sense. Any alternative hypothesis which equally fit the data is, by definition, equivalent to the accepted theory. They produce the same predictions, and hence are interchangeable. So even if there were infinite alternate hypotheses, they’d all be equally good explanations of the world, rather than being equally bad.
That latter bit about infinite possible hypotheses is, sadly, not really something that you can get away with in analysis- after all, in the end people only do one thing at a time. It's still a useful response to the postmodern critique, though, because it's inherently positive- if you have a hypothesis that hasn't been disproven, then it's just as good as any other and you're entitled to it. (At least until it is disproven.)

In any case, I enjoy "philosophy of science" stuff, and Doug's post is an excellent example of science-positive writing in that field.

No comments:

Post a Comment