Intelligent Epistemology

The paper begins with MU—the claim that consistent inference is possible—and asks what any conclusion commits itself to when it is said to follow from its grounds. Equivalent descriptions must agree, while every answer-relevant dependence must be carried by the stated problem.

From that discipline, called Epistemic Zero, it derives four quantitative forms across their stated domains: probability, relative entropy, maximum-entropy projection and KL updating. It then applies the same method to classical problems in epistemology, scientific reasoning, social knowledge and artificial intelligence.