Fixed-point proportion of geometric iterated Galois groups
Jorge Fariña-Asategui, Santiago Radi
TL;DR
The paper resolves the problem of determining the fixed-point proportion of geometric iterated Galois groups, proving that only Chebyshev-type polynomials yield positive $\mathrm{FPP}$; it develops a unified framework using mixing self-similar groups, martingale methods, and ergodic theory, and extends the analysis to rational functions via exceptional sets and Thurston orbifolds. The main technical innovation is a commutator trick that, combined with a pseudomixing property, shows that mixing groups have $\mathrm{FPP}=0$ except in Euclidean orbifold cases. A complete classification is obtained for polynomials, linking positive $\mathrm{FPP}$ to Euclidean orbifolds of type $(2,2,\infty)$ and, over separably closed fields, twisted Chebyshev polynomials. These results unify and extend prior partial findings and have direct implications for the distribution of periodic points over finite fields and related Arakelov-type dynamical problems.
Abstract
In 1980, Odoni initiated the study of the fixed-point proportion of iterated Galois groups of polynomials motivated by prime density problems in arithmetic dynamics. The main goal of the present paper is to completely settle the longstanding open problem of computing the fixed-point proportion of geometric iterated Galois groups of polynomials. Indeed, we confirm the well-known conjecture that Chebyshev polynomials are the only complex polynomials whose geometric iterated Galois groups have positive fixed-point proportion. Our proof relies on methods from group theory, ergodic theory, martingale theory and complex dynamics. This result has direct applications to the proportion of periodic points of polynomials over finite fields. The general framework developed in this paper applies more generally to rational functions over arbitrary fields and generalizes, via a unified approach, previous partial results, which have all been proved with very different methods.
