Revisiting cyclic elements in growth spaces

2025

Read Revisiting cyclic elements in growth spaces (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-2503-23987.txt.

Introduction

Linus Bergqvist, Adem Limani, and Bartosz Malman revisit which functions can seed orbits that span a space of disk functions in a 2025 paper. For the shift operator, a cyclic seed's shifts and scalings must reach every function in the space. Growth spaces weigh how fast functions may swell near the rim, and finite Nevanlinna type bounds their wildness.

The team gives sharp answers for tame Dini weights and for the border zone where a log integral just fails to settle. Their proofs build the needed seeds by hand more than past work did. In doing so they simplify and stretch a classic result by Korenblum and Roberts and a fresh theorem by three other authors. The paper tightens the map of when single seeds suffice.

Catalog record

YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-23987

Browse the full library · About the Tower of Babel

Full text

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

Revisiting cyclic elements in growth spaces Abstract . We revisit the problem of characterizing cyclic elements for the shift operator in a broad class of radial growth spaces of holomorphic functions on the unit disk, focusing on functions of finite Nevanlinna characteristic. We provide results in the range of Dini regular weights, and in the regime of logarithmic integral divergence. Our proofs are largely constructive, enabling us to simplify and extend a classical result by Korenblum and Roberts, and a recent Theorem due to El-Fallah, Kellay, and Seip. † † authors: Linus Bergqvist, Adem Limani & Bartosz Malman August 24, 2026 1. Introduction 1.1. Cyclic Nevanlinna functions in growth spaces Let W:(0,1]\to(0,1] be a continuous non-decreasing weight (positive function) with \lim_{t\to 0+}W(t)=0 . We denote by A^{p}(W) the space of holomorphic functions f in the unit-disc \mathbb{D} equipped with the metric \left\lVert f\right\rVert_{A^{p}(W)}=\left(\int_{\mathbb{D}}\left\lvert f(z)\right\rvert^{p}W(1-|z|)dA(z)\right)^{\min(1,1/p)}<\infty. Since the weight W is radial, it is well-known that the polynomials form a dense subset in A^{p}(W) (for instance, see Proposition 3.1 in [ 1 ] for a neat proof). We shall also consider the weighted growth space A^{\infty}(W) consisting of holomorphic functions f in \mathbb{D} satisfying \lim_{\left\lvert z\right\rvert\to 1-}W(1-|z|)\left\lvert f(z)\right\rvert=0. Equipped with the norm \left\lVert f\right\rVert_{A^{\infty}(W)}:=\sup_{z\in\mathbb{D}}W(1-|z|)\left\lvert f(z)\right\rvert<\infty, it becomes a separable Banach space, containing the polynomials as a dense subset. Let \mathcal{N} denote the Nevanlinna class, which consists of holomorphic functions in \mathbb{D} having finite Nevanlinna characteristic: \sup_{0<r<1}\int_{\partial\mathbb{D}}\log(1+\left\lvert f(r\zeta)\right\rvert)dm(\zeta)<\infty, where dm denotes the unit-normalized Lebesgue measure on the unit-circle \partial\mathbb{D} . Given a Nevanlinna class function f on \mathbb{D} , we consider the classical problem of when the set \{f(z)z^{n}:\,n=0,1,2,\dots\},\qquad z\in\mathbb{D} forms a dense linear span in A^{p}(W) . Such functions f are said to be cyclic in A^{p}(W) (with respect to the shift operator M_{z}f(z)=zf(z) ). Questions of this type originate back to the work of Keldysh in [ 10 ] and to Beurling [ 2 ] . Since the topologies in A^{p}(W) induce uniform convergence on compact subsets of \mathbb{D} , cyclic functions f in A^{p}(W) can certainly not have any zeros in \mathbb{D} . The classical Nevanlinna representation allows one to express any zero-free function f\in\mathcal{N} as f(z)=\exp\left(\int_{\partial\mathbb{D}}\frac{\zeta+z}{\zeta-z}d\mu_{f}(\zeta)\right),\qquad z\in\mathbb{D}, where \mu_{f} is a finite real-valued Borel measure on \partial\mathbb{D} . In fact, a more refined Lebesgue and Jordan decomposition, in conjunction with standard properties of Poisson kernels, implies that d\mu_{f}=\log|f|dm+d\nu_{f}-d\sigma_{f} where \nu_{f},\sigma_{f} are mutually singular positive finite Borel measures on \partial\mathbb{D} , both singular wrt dm . We shall often refer to \sigma_{f} as the associated negative singular part of f (instead of \mu_{f} ). This gives the refined Inner-outer factorization of f , defined by f(z)=\mathcal{O}_{f}(z)\Theta_{\sigma_{f}}(z)/\Theta_{\nu_{f}}(z),\qquad z\in\mathbb{D}, where \mathcal{O}_{f} denotes the so-called outer factor of f , and \Theta_{\sigma_{f}},\Theta_{\nu_{f}} are singular inner functions. For a detailed treatment of Nevanlinna factorization and Hardy spaces, we refer the reader to the excellent book [ 7 ] . In what follows, we shall solely restrict our attention to continuous non-decreasing weights W , which satisfy the following additional weak regularity condition: (1) \log\frac{1}{W(t/2)}\leq C\log\frac{1}{W(t)},\qquad\text{for some}\,\,\,C>1. From now and onward, we shall refer to such weights as good weights . We will use the notation A\lesssim B to indicate that A\leq cB for some absolute constant c>0 . When both A\lesssim B and B\lesssim A hold, we simply write A\asymp B . Occasionally, absolute constants may appear when carrying out estimates, but the reader should note that these constants may vary from line to line. 1.2. Dini-regular weights In this section, we shall restrict our attention to weights W which tend to zero in a slightly slower fashion. More precisely, we shall assume that there exists a constant C>0 , such that (2) \int_{0}^{x}\log\frac{1}{W(t)}dt\leq Cx\log\frac{1}{W(x)},\qquad 0<x<1. In this regime, it turns out that there are zero-free holomorphic self-maps f on \mathbb{D} which are not cyclic in A^{p}(W) . Results of this kind were initially proved by Korenblum in [ 11 ] and independently by Roberts in [ 14 ] . For a certain range of weights W , their results assert that the cyclicity of f in A^{p}(W) is entirely contingent upon whether the associated Nevanlinna measure \mu_{f} assigns any mass to certain exceptional W -sets on \partial\mathbb{D} . Below, we clarify these points. Throughout, we let \kappa_{W} be the associated gauge-function with respect to W , defined by \kappa_{W}(t)=t\log\frac{1}{W(t)},\qquad 0<t<1. A compact set K\subset\partial\mathbb{D} of Lebesgue measure zero is said to have finite \kappa_{W} -entropy if \sum_{k}\kappa_{W}(\ell_{k})<\infty. where (\ell_{k})_{k} are the lengths of the connected components (I_{k})_{k} of \partial\mathbb{D}\setminus K . When W(t)=t^{\alpha} some \alpha>0 , such sets are typically referred to as Beurling-Carleson sets, and they play a crucial role in function theory. For instance, they precisely classify all zero sets on \partial\mathbb{D} of holomorphic functions in \mathbb{D} which are smooth up to \partial\mathbb{D} (see [ 15 ] ). We remark that the condition ( 1 ) is equivalent to the doubling property of the gauge function \kappa_{W}(t/2)\asymp\kappa_{W}(t) , while ( 2 ) is typically referred to as \kappa_{W} being Dini-regular . Our main intention is to prove the following generalization of the Korenblum-Roberts Theorem. Theorem 1.1 . Let 0<p\leq\infty and W be a good weight which satisfies the condition ( 2 ). Then a function f\in A^{p}(W)\cap\mathcal{N} is cyclic in A^{p}(W) if \sigma_{f}(K)=0 for all sets K\subset\partial\mathbb{D} of finite \kappa_{W} -entropy. The above theorem was initially proved by Korenblum and Roberts in the classical setting of the Bergman spaces A^{p}(W) , corresponding to weights of the form W(t)=t^{\alpha} . They also showed that the above condition on \sigma_{f} is not only sufficient, but also necessary. For a wider range of weights, the same conclusion was also recently confirmed in [ 12 ] , indicating that ?THM? is sharp. Our proof of ?THM? is carried in the following steps. First, we simply reduce the problem to cyclicity of f to the associated singular inner factor \Theta_{\sigma_{f}} . Secondly, we shall utilize a Roberts-type decomposition adapted to the corresponding weight W , allowing us to decompose singular measures. In the last step, our approach substantially deviates from Korenblums proof and from Roberts, where the former involves an implicit linear programming argument (see [ 11 ] ), while the latter invokes a quantitative version of Carleson’s Corona Theorem (see [ 14 ] ). Instead, we shall carry out a fairly explicit construction of bounded holomorphic functions (h_{n})_{n} in \mathbb{D} , such that \Theta_{\sigma_{f}}h_{n}-1 have small A^{p}(W) -norms. 1.3. Logarithmic integral divergence We now restrict our attention to weights which tend to zero rapidly. That is, we assume that \log W is not integrable: (3) \int_{0}^{1}\log W(t)dt=-\infty. Note that the above condition is slightly stronger than ( 2 ), and is equivalent to the assertion that the associated gauge-function \kappa_{W}(t) is not Dini-continuous. In this setting, our second result provides a new proof of Theorem 1.1 from [ 6 ] on cyclic singular inner functions in A^{p}(W) , due to El-Fallah, Kellay, and Seip. We shall re frame it in the following general form. Theorem 1.2 . Let 0<p\leq\infty and W be a good weight which satisfies the condition ( 3 ). Then any f\in A^{p}(W)\cap\mathcal{N} with no zeros in \mathbb{D} is cyclic in A^{p}(W) . A version of this Theorem was initially proved in

Read the complete text