Looking from the inside and from the outside

Abstract

One often sees a sharp distinction in mathematics between descriptions from the outside and from the inside. Think of defining a set in the plane through an algebraic equation, or dynamically as the closure of the orbit of some point under iterations of a given mapping. In logic one sees this dichotomy in the descriptions of sets of tautologies through semantics and proofs. Logic provides several tools for making outer descriptions of mathematical objects. This paper concerns a slightly complicated mixture of themes related to inner descriptions and formal proofs. We use the notion of feasibility to embed mathematical structures into spaces of logical formulas, from which we can obtain new structures through proofs. We present new geometries on finitely generated groups through proofs, and new structure on the rational numbers (or other fields) which is susceptible to dynamical processes, such as the action of $SL(2,Z)$ by projective transformations. We consider the topological notion of {\em Serre fibrations}. This entails more difficulties of formalization, but basic points arise already for {\em torus bundles}, which present exponential distortion through cycling in a nicely geometric way. One of our goals is to bring out mathematical structure related to cuts and cut elimination. Our geometry on groups through proofs is far from the word metric precisely because of the cut rule. We want to explore the idea that in general the existence of short proofs with cuts should be related to internal symmetry or dynamical processes in the underlying mathematical objects. We also want to bring ordinary mathematical proofs closer to proof theory. In this regard the topological example is attractive for presenting realistic difficulties.