Rigidity of Newton dynamics (Q2172226)

This well-written paper contains a detailed study of the dynamics of Newton maps of polynomials (or, to use the terminology of the paper, of polynomial Newton maps) and in particular of their rigidity properties. This is particularly interesting because polynomial Newton maps are rational maps, and the rigidity theory of rational maps is in general not as well developed as the rigidity theory of polynomials. This paper deals with two kinds of rigidity: dynamical rigidity and parametric rigidity. Roughly speaking, dynamical rigidity concerns the possibility of distinguishing different points by means of their orbits with respect to a suitable defined symbolic dynamics, while parametric rigidity concerns instead the possibility of distinguishing (up to quasiconformal conjugation) two polynomial Newton maps by means of a suitably defined combinatorical structure of their Julia and Fatou sets. One of the main technical advancement contained in this paper consists precisely in setting up the construction of suitable symbolic dynamics and combinatorics. This is done by introducing the notion of complex box mapping, which is a far-reaching generalization of the notion of polynomial-like maps obtained by extending ideas due to Lyubich, Kozlovski, van Strien and others. A \textit{complex box mapping} is a holomorphic map \(F\colon U\to V\) between two open sets \(U\subset V\subset\widehat{\mathbb{C}}\) with the following properties: \begin{itemize} \item[1.] \(F\) has finitely many critical points; \item[2.] \(V\) is the union of finitely many open Jordan disks with disjoint closures, while \(U\) is the union of finitely or infinitely many open Jordan disks; \item[3.] every connected component \(W\) of \(V\) either is a connected component of \(U\) or \(W\cap U\) is a union of Jordan disks compactly contained in \(W\) and with pairwise disjoint closures; \item[4.] for every connected component \(Y\) of \(U\) the image \(F(Y)\) is a connected component of \(V\) and the restriction \(F|_Y\colon Y\to F(Y)\) is a proper map. \end{itemize} Using the iterated inverse images of \(V\) under \(F\), the authors are able to construct a nested sequence of puzzles, and thus a symbolic dynamics associated to this sequence of puzzles, and a notion of renormalizable dynamics. A point \(z\in U\) is \textit{non-escaping} if \(F^n(z)\in U\) for all \(n\in\mathbb{N}\); the \textit{fiber} of a non-escaping point \(z\) is the set of points in \(U\) having the same symbolic dynamics as \(z\) with respect to the given sequence of puzzles. Finally, roughly speaking, a complex box mapping has a \textit{renormalizable restriction} if there exists a puzzle piece containing a critical point so that the restriction to this puzzle piece is a polynomial-like map with connected Julia set. The authors are then able to prove the following rigidity theorem (Theorem C) for a complex box mapping \(F\). Given an arbitrary non-escaping point \(z\), at least one of the following cases occurs: \begin{itemize} \item[(T)] the fiber of \(z\) is trivial, that is no other point has the same symbolic dynamics as \(z\); \item[(R)] \(z\) belongs, possibly after applying a finite iterate, to the filled Julia set of a renormalizable restriction; \item[(CB)] the orbit of \(z\) converges to the boundary of the domain of definition of \(F\); \item[(NE)] \(z\) eventually maps to a periodic connected component that maps surjectively onto itself under some iterate. \end{itemize} The authors, then, find inside any polynomial Newton map a dynamical structure well represented by a complex box mapping (with special properties allowing to exclude some of the possibilities given by Theorem C) and use this to prove the following dynamical rigidity theorem (Theorem A) for any Newton map \(N_p\) of a polynomial \(p\) of degree at least two. Given an arbitrary \(z\in\widehat{\mathbb{C}}\), at least one of the following cases occurs: \begin{itemize} \item[(B)] \(z\) belongs to the basin of attraction of a root of~\(p\); \item[(T)] the fiber of \(z\) is trivial, that is no other point has the same symbolic dynamics as \(z\); \item[(R)] \(z\) belongs, possibly after applying a finite iterate, to the filled Julia set of renormalizable dynamics (a polynomial-like restriction of \(N_p\) with connected Julia set). \end{itemize} This theorem has many consequences. For instance, the authors use it to prove that the boundary of the basin of a root for a polynomial Newton map is always locally connected (this was also recently proved, in a different way, by \textit{X. Wang} et al. [``Dynamics of Newton maps'', Ergodic Theory Dyn. Syst. 43, No. 3, 1035--1080 (2023; \url{doi:10.1017/etds.2021.168})]). Furthermore, they are also able to prove that a large class of polynomial Newton maps have locally connected Julia sets; this holds, for instance, for Newton maps of cubic polynomials without Siegel disks. Finally, the authors turn to the question of parametric rigidity. To do so, they associate to each polynomial Newton map \(N_p\) a graph called \textit{Newton graph} (very much involved also in the construction of the puzzles associated to \(N_p\)), that roughly speaking represents how the basins of the roots are connected to each other. They then say that two polynomial Newton maps are \textit{combinatorially equivalent} if their Newton graphs coincide. It turns out that two polynomial Newton maps that are quasiconfomally conjugated in a neighbourhood of their Julia set are combinatorially equivalent. Conversely, the authors are able to prove the following parametric rigidity theorem (Theorem B). If two polynomial Newton maps are combinatorially equivalent then they are quasiconformally conjugated in a neighbourhood of the Julia set provided that \begin{itemize} \item[(1)] either they are both non-renormalizable; or, \item[(2)] they are both renormalizable essentially in the same way (more formally: there is a bijection between their domains of renormalization that respects hybrid equivalence between the little Julia sets as well as their combinatorial position). \end{itemize}