Reinhardt domains determined by their endomorphisms

Rafael B. Andrist, Włodzimierz Zwonek · 2026

Read Reinhardt domains determined by their endomorphisms by Rafael B. Andrist, Włodzimierz Zwonek (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-19364.txt.

Introduction

Andrist and Zwonek prove that pseudoconvex Reinhardt domains in 2D with isomorphic semigroups of holomorphic endomorphisms are equivalent, either biholomorphically or anti-biholomorphically. They also prove the same for every Stein manifold that folds back to a copy of the punctured complex line. It is fixed, up to the same two kinds of match, by its semigroup of holomorphic endomorphisms.

Catalog record

CREATORRafael B. Andrist, Włodzimierz Zwonek
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50601
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.19364

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Reinhardt domains determined by their endomorphisms Rafael B. Andrist, Włodzimierz Zwonek Abstract . We show that pseudoconvex Reinhardt domains in dimension two with isomorphic semigroups of holomorphic endomorphisms are biholomorphically or anti-biholomorphically equivalent. Moreover, we show that every Stein manifold that retracts to a properly embedded copy of the punctured complex line, is determined (up to biholomorphic or anti-biholomorphic equivalence) by its semigroup of holomorphic endomorphisms. 2020 Mathematics Subject Classification 32Q02, 32E10, 20M20 1. Introduction It is a classical result that Stein manifolds are determined by their rings of holomorphic functions – up to (anti-)biholomorphic equivalence which goes back to a theorem of Bers in the case of one complex variable. Instead of functions, we consider the question whether the holomorphic endomorphisms, i.e. the holomorphic self-mappings, determine a complex manifold. The holomorphic endomorphisms of a complex manifold X consist at least of all the constant mappings and the identity, and thus form a monoid which we denote by \enmo(X) . The following lemma is due to Schreier [ 13 ] : Lemma 1.1 . Let X and Y be complex manifolds. Let \Phi\colon\enmo(X)\rightarrow\enmo(Y) be an isomorphism of semigroups. Then there exists a bijective mapping \varphi\colon X\rightarrow Y such that \Phi(f)=\varphi\circ f\circ\varphi^{-1}\quad\forall\,f\in\enmo(X) A variation of this Lemma with slightly different conditions can be found in the work of Buzzard and Merenkov [ 4 ] . Instead of an abstract isomorphism of semigroups, we can therefore study the bijection \varphi\colon X\to Y between complex manifolds and try to establish that \varphi is (anti-)biholomorphic. The Lemma motivates the following definition: Definition 1.2 (see Andrist [ 2 ] *Definition 2.3) . A (set-theoretic) mapping \varphi\colon X\to Y between complex manifolds is called a conjugating mapping if it is bijective and if it induces a homomorphism \Phi\colon\enmo(X)\to\enmo(Y) of the endomorphism semigroups by conjugating \enmo(X)\ni f\mapsto\varphi\circ f\circ\varphi^{-1}\in\enmo(Y) . If, in addition, \Phi is an isomorphism, then \varphi is called an iso-conjugating mapping . In the holomorphic category, the question of determining a complex manifold by its semigroup of endomorphisms has first been studied by Eremenko [ 5 ] who established this result for Riemann surfaces that admit bounded holomorphic functions. His method crucially relies on first showing the continuity of \varphi and on the theory about the solutions to the Schröder equation. The case of the complex line \mathbb{C} had been studied before by Hinkkanen [ 7 ] who investigated the possible automorphisms of \enmo(\mathbb{C}) , i.e. he showed that an iso-conjugation \varphi\colon\mathbb{C}\to\mathbb{C} is (anti-)biholomorphic. Merenkov [ 12 ] extended Eremenko’s methods and result to the case of bounded domains in \mathbb{C}^{n} . Using a result of Buzzard and Hubbard [ 3 ] about certain Fatou–Bieberbach domains that form a subbasis of the topology of \mathbb{C}^{n} , Merenkov and Buzzard [ 4 ] extended the result later also to \mathbb{C}^{n} . The first author generalized the result for \mathbb{C}^{n} to any Stein manifold that contains a properly embedded complex line \mathbb{C} [ 2 ] by a new method that does not rely on establishing continuity a priori. The analogous result also holds for affine varieties over any algebraically closed field, as shown by Kraft and the first author [ 1 ] . In this article, we first study the one-dimensional case that had resisted for the longest time, the punctured complex line \mathbb{C}_{\ast} : Proposition 1.3 . Let \varphi\colon\mathbb{C}_{\ast}\to\mathbb{C}_{\ast} be a conjugating mapping. Then \varphi is biholomorphic or anti-biholomorphic. For the applications, it is important that we do not require \varphi to be iso-conjugating in this proposition, but just conjugating. Next, we cover the case of two-dimensional pseudoconvex Reinhardt domains: Theorem 1.4 . Let D_{1},D_{2}\subset\mathbb{C}^{2} be pseudoconvex Reinhardt domains with an isomorphism \Phi\colon\enmo(D_{1})\to\enmo(D_{2}) of semi-groups. Then there exists a biholomorphic or anti-biholomorphic \psi\colon D_{1}\to D_{2} such that \Phi(f)=\psi\circ f\circ\psi^{-1} . For unbounded domains, we can’t drop the requirement of pseudoconvexity; we provide a concrete counterexample in Theorem 7.1 . In higher dimensions, we can prove the following general result: Theorem 1.5 . Let X be a Stein manifold and let i\colon\mathbb{C}_{\ast}\to X be a proper holomorphic embedding such that i(\mathbb{C}_{\ast}) is a topological retract of X . Let \Phi\colon\enmo(X)\to\enmo(Y) be an isomorphism of semi-groups where Y is a complex space. Then there exists a biholomorphic or anti-biholomorphic \varphi\colon X\to Y such that \Phi(f)=\varphi\circ f\circ\varphi^{-1} . In particular, this also implies that if \enmo(Y)\cong\enmo(\mathbb{C}_{\ast}) , then Y is (anti-)biholomorphic to \mathbb{C}_{\ast} . 2. Conjugating mappings for \mathbb{C}_{\ast} Lemma 2.1 . Let \varphi\colon\mathbb{C}_{\ast}\to\mathbb{C}_{\ast} be a conjugating mapping with \varphi(1)=1 . Then \varphi is multiplicative, i.e. it is actually a group isomorphism of \mathbb{C}_{\ast} . Proof. The result is deduced by looking solely at the automorphism group of \mathbb{C}_{*} , consisting of the mappings z\mapsto a\cdot z^{\pm 1},a\in\mathbb{C}_{\ast} . Therefore, with w=\varphi^{-1}(z) , we can write: \varphi(aw)=A(a)\cdot\left(\varphi(w)\right)^{(\pm 1)_{a}} Since \varphi(1)=1 , it follows that in fact A(a)=\varphi(a) . The cases (\pm 1)_{a}=-1 can be excluded as follows: The elements of order 2 in \aut(\mathbb{C}_{\ast}) are z\mapsto-z and z\mapsto a/z for any a\in\mathbb{C}_{\ast} . Hence it follows that \varphi(aw)=\varphi(a)\cdot\varphi(w),\;\text{ for all }a\in\mathbb{C}\setminus\{0,-1\} We only need to exclude that \varphi(-w)=\varphi(-1)/\varphi(w) . For this we consider z\mapsto iz whose second iterate is z\mapsto-z . Since \varphi(iw)=\varphi(i)\cdot\varphi(w) , we can conclude that \varphi(-w)=\varphi(-1)\cdot\varphi(w) . It also follows \varphi(-1)=-1 and \varphi(i)=\pm i . ∎ Lemma 2.2 . Under the same assumptions, \varphi conjugates z\mapsto\exp(z) to z\mapsto\exp(cz) or to z\mapsto\exp(c/z) for some c\in\mathbb{C}_{\ast} . Proof. Let p_{n}(z):=z^{n} . Multiplicativity of \varphi yields \varphi\circ p_{n}\circ\varphi^{-1}=p_{n} . Every holomorphic endomorphism of \mathbb{C}_{\ast} is of the form z\mapsto z^{m}\cdot\exp\left(\sum_{k=-\infty}^{\infty}c_{k}z^{k}\right),\quad m\in\mathbb{Z} Therefore, z\mapsto\exp(z) is necessarily conjugated by \varphi to the form above. Now, consider: \varphi\circ p_{n}\circ\exp\circ\varphi^{-1}=p_{n}\circ\varphi\circ\exp\circ\varphi^{-1} The right-hand side can be written explicitly as: z^{nm}\cdot\exp\left(\sum_{k=-\infty}^{\infty}nc_{k}z^{k}\right) Using the multiplicativity of \varphi , the left-hand side can be written explicitly as: \displaystyle\varphi\circ p_{n}\circ\exp\circ\varphi^{-1}(z) \displaystyle=\varphi(\exp(n\varphi^{-1}(z))) \displaystyle=\varphi(\exp(\varphi^{-1}(\varphi(n)\cdot z))) \displaystyle=(\varphi(n)\cdot z)^{m}\cdot\exp\left(\sum_{k=-\infty}^{\infty}c_{k}(\varphi(n)\cdot z)^{k}\right) Equating both sides, we now obtain: \displaystyle m \displaystyle=0 \displaystyle c_{k}\neq 0,k\neq 0\Longrightarrow\varphi(n^{k}) \displaystyle=n \displaystyle c_{0}\neq 0\Longrightarrow 1 \displaystyle=n\mod 2\pi i\mathbb{Z} Since the mapping can’t be constant, there exists exactly one k\in\mathbb{Z} with c_{k}\neq 0 , and, therefore, \exp is conjugated to z\mapsto\exp(c_{k}z^{k}),\quad k\in\mathbb{Z} Let \zeta\in\mathbb{C}_{\ast}\setminus\{1\} be a primitive k -th root of unity and let q(z):=\zeta z . Such q is conjugated to \varphi(\zeta)z where \varphi(\zeta) is also a k -th root of unity. We observe that \exp\neq\exp\circ q , but their conjugates are equal. Hence, k=\pm 1 . ∎ Corollary 2.3 . If we assume (after applying z\mapsto 1/z if necessary) that \varphi conjugates z\mapsto\exp(z) to z\mapsto\exp(cz) and if we set \varphi(0):=0 , then \varphi is additive. Proof. By multiplicativity of \varphi we have: \varphi(\exp(a+b))=\varphi(\exp(a)\cdot\exp(b))=\varphi(\exp(a))\cdot\varphi(\exp(b)) Applying Lemma 2.2 to both sides, we obtain: \exp(c\varphi(a+b)))=\exp(c\varphi(a))\cdot\exp(c\varphi(b))=\exp(c(\varphi(a)+\varphi(b))) This implies that \varphi is additive modulo \frac{2\pi i}{c}\mathbb{Z} . Combining this with multiplicativity, it is easy to see that \varphi is fully additive: \varphi(a+b)=\varphi(\varepsilon\cdot(a/\varepsilon+b/\varepsilon)) The l.h.s. equals \varphi(a)+\varphi(b)+\frac{2\pi i}{c}k\quad\text{ for some }k\in\mathbb{Z}, and the r.h.s. equals \varphi(a)+\varphi(b)+\varphi(\varepsilon)\frac{2\pi i}{c}\ell\quad\text{ for some }\ell\in\mathbb{Z}. Choosing \varepsilon s.t. \varphi(\varepsilon) is irrational, requires k=\ell=0 . ∎ Proposition 2.4 . Such a mapping \varphi is continuous, and hence holomorphic or anti-holomorphic. Proof. Assume to get a contradiction that \varphi is not continuous at 1 . Then there exists a sequence (a_{n})_{n}\subset\mathbb{C}_{\ast} such that \lim\limits_{n\to\infty}a_{n}=1 but \lim\limits_{n\to\infty}\varphi(a_{n})\neq 1 . Since \varphi is additive and multiplicative, it equals the identity (if necessary, after

Read the complete text