Information Theory and Natural Selection

Abstract

Recently philosophers of biology have debated over the status of evolutionary theory; is it interpreted as realistic or anti-realistic? I discuss this issue by using Fisher's formula of natural selection as an example. In generalizing evolutionary phenomena, we need only partial information. So I compare the formula of natural selection with information theory, which gives us a reasonable tool of inference on the basis of partial information, and show that the formula of natural selection may be regarded as a part of information theory. I conclude that there is the mixture of realistic and anti-realistic factors in the formula of natural selection, and that the aim of evolutionary theory is rational prediction or explanation of phenomena on the basis of partial information, not complete description of it on the basis of full information.