Inferring Physical Probabilities from Symmetries

We humans are surprisingly good at figuring out the values of the physical probability distributions attached to a system from a rather small amount of information about the physical symmetries of the system. A simple example is our inference that the probability that a toss of a fair coin lands heads is one half. More interesting and useful inferences of this sort can be found in statistical physics and evolutionary biology.

In these papers, I ask what method we use to make these inferences and why the method works so well. There is a positive and a negative part to the project. The negative part argues against the suggestion that we use a principle of indifference or some other ignorance-driven principle to make the inferences. The positive part attempts to give the correct account of the inferences.

Published Work

Work In Progress

Unpublished Paper