On the (Dis)connection Between Growth and Primitive Periodic Points

Adi Glücksam, Shira Tanny · 2025

Read On the (Dis)connection Between Growth and Primitive Periodic Points by Adi Glücksam, Shira Tanny (2025) 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-2503-20034.txt.

Introduction

Adi Glücksam and Shira Tanny answer a question in complex dynamics in twenty twenty-five. In nineteen seventy-two, Cornalba and Shiffman showed that zeros of a certain holomorphic function can grow in number arbitrarily fast. Glücksam and Tanny extend that finding to dynamics. For holomorphic functions of vanishing order in two or more variables, the number of isolated primitive periodic points can also grow arbitrarily fast. The result settles a recent question posed by Buhovsky, Polterovich, Shelukhin, and Stojisavljević.

Catalog record

CREATORAdi Glücksam, Shira Tanny
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-20034

Browse the full library · About the Tower of Babel

Full text

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

On the (Dis)connection Between Growth and Primitive Periodic Points Adi Glücksam, Shira Tanny Abstract In 1972, Cornalba and Shiffman showed that the number of zeros of an order zero holomorphic function in two or more variables can grow arbitrarily fast. We generalize this finding to the setting of complex dynamics, establishing that the number of isolated primitive periodic points of an order zero holomorphic function in two or more variables can grow arbitrarily fast as well. This answers a recent question posed by L. Buhovsky, I. Polterovich, L. Polterovich, E. Shelukhin and V. Stojisavljević. 1 Introduction Bézout theorems seek to explore the relation between the growth of a holomorphic function and the size of its zeros set. The classical Bézout theorem states that the number of zeros of n polynomials over n variables is bounded by the product of their degrees. A natural question is whether this theorem extends to general holomorphic functions. Here, the notion of degree is replaced by the notion of growth rate, measured by the asymptotic behavior of its maximum modulus function , M_{f}(r):=\max_{\left|{z}\right|\leq r}\left|{f(z)}\right|. Jensen’s formula gives a full description of this relation in complex dimension 1. Formally, assuming that f(0)\neq 0 , Jensen’s formula implies that for every a>1 there exists C=C(f,a)>0 such that \#\left\{{\left|{z}\right|\leq r:f(z)=0}\right\}\leq C\log M_{f}(a\cdot r)\qquad\text{for all}\qquad r>0. In [ 2 ] Cornalba and Shifmann showed this phenomenon does not extend to higher dimensions. They constructed a holomorphic function F:\mathbb{C}^{2}\rightarrow\mathbb{C}^{2} satisfying \log M_{F}(r)=O\left({\log^{2}r}\right) , whose zeros grow arbitrarily fast. Here, and everywhere else in the paper, O\left({\log^{2}r}\right) means that there exists a constant 0<c for which \log M_{F}(r)\leq c\log^{2}(r) . In a recent paper, [ 1 ] , L. Buhovsky, I. Polterovich, L. Polterovich, E. Shelukhin and V. Stojisavljević proposed a new way to obtain a transcendental analog to Bézout theorem, replacing the traditional zero count with the coarse zero count. Roughly speaking, the coarse count of zeros of a holomorphic function in \mathbb{C}^{n} is the number of connected components on which it is small, that contain a zero. They proved that this coarse count can be bounded in terms of the maximum modulus function, thus establishing an analog to Bézout theorem in the setup of coarse count. This version of Bézout theorem can be imported into the world of complex dynamics. It implies that the coarse count of p -periodic points of holomorphic functions is also bounded by a function of their growth rate. Moreover, the Cornalba-Shifmann example (shifted by the identity map) shows that the honest count of fixed points does not obey such a bound. In [ 1 ] they asked whether the construction carried out by Cornalba-Shifmann can be generalized to construct an order zero holomorphic function with arbitrarily many primitive periodic points of higher periods. Definition 1.1 Let f:\mathbb{C}^{n}\rightarrow\mathbb{C}^{n} be a holomorphic function. • We say z\in\mathbb{C}^{n} is p - primitive periodic point for f (or p -PPP for short) if f^{\circ j}(z)\neq z whenever 1\leq j\leq p-1 , but f^{\circ p}(z)=z . • Denote by \nu_{p}(f,r) the number of p -PPP lying in a ball of radius r centered at the origin, namely \nu_{p}(f,r):=\#\left({B(0,r)\cap\left\{{z,z\text{ is a p-PPP for }f}\right\}}\right). Question 1.2 ( [ 1 , Question 1.12] ) Does there exist a transcendental entire map, f , of order 0 (i.e., the modulus M_{f}(r) grows slower than e^{r^{\varepsilon}} for every \varepsilon ) for which \nu_{p}(f,r) grows arbitrarily fast in p and r ? For a fixed period, p , a simple modification of the Cornalba–Shiffman example produces a holomorphic function, f , of order zero for which \nu_{p}(f,r) grows arbitrarily fast in r ; see Example 3.4 below. However, producing a holomorphic function with many primitive periodic points of varying prescribed periods is more complicated. Our main result states that this is indeed possible. Theorem 1.3 For every sequence of periods, \left\{{p_{n}}\right\}_{n=1}^{\infty} , and for every rate, \left\{{m_{n}}\right\}_{n=1}^{\infty} , there exists a holomorphic function F:\mathbb{C}^{2}\rightarrow\mathbb{C}^{2} satisfying \nu_{p_{n}}(F,2^{n})\geq m_{n}\quad\text{and}\quad\log M_{F}(r)=O\left({\log^{2}(r)}\right). The constant in O(\log^{2}(r)) is independent of the sequences. Unlike the example by Cornalba and Shiffman, the holomorphic functions constructed in Theorem 1.3 are not completely explicit. In particular, it is not clear how to estimate or even bound from above or below their coarse zero count. The main idea of the proof is to utilize the simple example constructed in Example 3.4 with different fixed periods, and then use a ‘dispatcher’ function to choose between different periods. The main tool in the construction is a theorem of Hörmander (see Theorem 3.3 below) which guarantees the existence of solutions to non-homogeneous \bar{\partial} -equations with certain integral bounds. This theorem allows one to construct entire functions with distinct behaviours in different regions. The use of Hörmander’s theorem requires an underlying subharmonic function which is large in the area between distinct regions, but very negative where we want a good approximation. Section 2 contains the construction of subharmonic functions that will be used in the application of Hörmander’s theorem in Section 3 . In Section 3 we construct the required holomorphic functions and prove Theorem 1.3 . 2 Preliminary Results on Subharmonic functions We commence the discussion with the following lemma whose role is to modify a given subharmonic function assigning -\infty to prescribed points. Lemma 2.1 (The Puncture Lemma) Let u be a continuous subharmonic function and, for N\in\mathbb{N} or N=\infty , let \left\{{B_{k}}\right\}_{k=1}^{N} be a sequence of pairwise disjoint disks, B_{k}=B(z_{k},r_{k}) . Assume that for every k , \underset{z\in B_{k}}{\inf}\;\Delta u>0 . Then there exists a subharmonic function v satisfying 1. u(z)=v(z) whenever z\not\in\bigcup_{{k}={1}}^{N}B_{k} . 2. For every \delta\in\left({0,1}\right) , and k\in\left\{{1,2,\cdots,N}\right\} \underset{z\in B(z_{k},\delta\cdot r_{k})}{\max}\;v(z)\leq\underset{z\in B_{k}}{\max}\;u(z)-c\cdot r_{k}^{2}\cdot\underset{z\in B_{k}}{\inf}\;\Delta u\log\left({\frac{1}{\delta}}\right), where c>0 is some uniform constant. See Figure 1 for an illustration of the image of the subharmonic function v . Figure 1 : An illustration of the the image of the subharmonic function created by Lemma 2.1 While the proof can be found in [ 4 , Section 3.1] , where \mathcal{K}_{0}(t)=\log(t) and the function v is any function, we include it here for the reader’s convenience. Proof. Define the function v(z)=\begin{cases}\left({P_{\mathbb{D}}u_{k}}\right)\left({\frac{z-z_{k}}{r_{k}}}\right)+A_{k}\log\left({\frac{\left|{z-z_{k}}\right|}{r_{k}}}\right),&z\in B_{k}\\ u(z),&\text{otherwise}\end{cases} where u_{k}(z)=u\left({r_{k}\cdot z+z_{k}}\right) (and therefore u(B_{k})=u_{k}(\mathbb{D}) ), the constants A_{k}>0 will be chosen momentarily, and P_{\mathbb{D}}f is Poisson integral of a function f defined by P_{\mathbb{D}}f(re^{i\theta}):=\frac{1}{2\pi}\int_{-\pi}^{\pi}P_{r}(\theta-t)f(e^{it})\ dt,\quad\text{for}\quad P_{r}(\theta):=\frac{1-r^{2}}{1-2r\cos\theta+r^{2}}. Note that v is continuous, since the Poisson integral of any continuous function agrees with it on \partial\mathbb{D} . We need to choose the constants A_{k} such that sub-harmonicity is preserved. We will use the following claim: Claim 2.2 ( [ 3 , p.24] ) Let \Omega\subseteq\mathbb{C} be a domain, and let \Omega_{1}\Subset\Omega be a domain, so that \partial\Omega_{1} is an orientable smooth curve. Every function f which is continuous on \Omega and subharmonic on \Omega_{1}\cup\left({\Omega\setminus\overline{\Omega_{1}}}\right) , is subharmonic on \Omega if on \partial\Omega_{1} it satisfies \frac{\partial f}{\partial n_{1}}\leq\frac{\partial f}{\partial n_{2}}, where n_{1} is the outer normal to \Omega_{1} along \partial\Omega_{1} and n_{2} is the outer normal to \Omega\setminus\Omega_{1} along \partial\Omega_{1} . We will use this claim with \Omega:=\mathbb{C} , and \Omega_{1}:=B_{k} . In this case, we need to show that along \partial B_{k} \frac{\partial v}{\partial n_{1}}\leq\frac{\partial v}{\partial n_{2}}=\frac{\partial u}{\partial n}, and the latter is known. To calculate \frac{\partial v}{\partial n_{1}} we will use Poisson-Jenssen’s formula (see e.g. [ 7 , Theorem 4.5.1] ),

Read the complete text