but Cosma says I’ve got another think coming! He’s blogging the Ockham’s razor conference I mentioned in the previous post, and starts out today’s entry with the following bombshell:
The theme of the morning was that Ockham’s razor is not a useful principle because the truth is inherently simple, or because the truth is apt to be simple, or because simplicity is more plausible, or even because simple models predict better. Rather, the Razor helps us get to the truth faster than if we multiplied complexities without necessity, even when the truth is actually rather complex.
I have always thought of the utility of parsimony as derving from a tendency of true things to be simple. But am I fooling myself? I tend to think that mathematical truths are apt to be simple — for instance, that when I have truly understood a difficult mathematical argument I see that the main idea is simple and elegant, while the visible complications are somehow inessential. But you could argue that this is just prejudice on my part, and I denigrate the complicated part as inessential just because it is complicated.
And certainly I don’t think the truth about big biological or social systems is apt to be simple. In fact, because I know people are prejudiced to believe in simple explanations, I find myself leaning against them; the fact that a simple explanation is widely believed by people I trust is less compelling as evidence than it would be, if the explanation in question were prickly or technical or otherwise unpleasant to believe.