A critical majorant for the Khinchin-Ostrowski property
Bartosz Malman · 2025
Read A critical majorant for the Khinchin-Ostrowski property by Bartosz Malman (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-2506-06911.txt.
Introduction
Malman proves a sharper version of a classic result on when boundary data force a function to vanish. A set on the unit circle has the Khinchin-Ostrowski property if test sequences that fade on it must fade everywhere. Earlier theorems by Kegejan and Khrushchev set terms for when this holds.
One step in Khrushchev's proof needed the growth bound to be log-integrable. That need blocked some generalized Beurling-Carleson settings from the result. Malman drops that assumption with a new domain shape. The domain cuts out areas bounded by hyperbolic arcs between gaps of the set. Harmonic measure estimates then go through without the extra condition.
Catalog record
| CREATOR | Bartosz Malman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-06911
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A critical majorant for the Khinchin-Ostrowski property Thanks: Part of this work has been done during a visit of the author at the Department of Mathematics and Statistics, University of Reading, funded by the University of Reading and the London Mathematical Society. He gratefully acknowledges the financial support. The author is also partially supported by Vetenskapsrådet (VR2024-03959). Bartosz Malman Abstract . In this short note we prove an optimal version of a classical result. Given a majorant determining a growth restriction on functions in the unit disk 𝔻 \mathbb{D} , we say that a set E E on the unit circle 𝕋 \mathbb{T} is a uniqueness set , or has the Khinchin-Ostrowski property , with respect to the majorant, if any sequence of analytic polynomials satisfying the growth restriction which converges in an appropriate sense to 0 0 on E E , in fact is forced to converge to 0 0 in 𝔻 \mathbb{D} also. Theorems proved by Kegejan and Khrushchev state that if E E has positive Lebesgue measure and satisfies a generalized Beurling-Carleson condition, then for an appropriate majorant the Khinchin-Ostrowski property is satisfied. A technical point in Khrushchev’s proof is the estimation of the harmonic measure in a Privalov-type domain which requires logarithmic integrability of the majorant. This forbids the application of his result to certain types of generalized Beurling-Carleson conditions. Here, we dispose of the integrability assumption on the majorant. To do so, we use a Joukowski-Privalov domain which is obtained by removing from the unit disk the areas enclosed by hyperbolic geodesics between the endpoints of intervals complementary to E E . For this type of domain the method of Khrushchev applies, but the harmonic measure may be estimated more accurately by simple explicit formulas for conformal mappings. As a consequence, we find the critical majorant at which Beurling-Carleson type conditions stop determining the Khinchin-Ostrowski property of a set, and above which the containment of intervals is the only relevant characteristic. We discuss also weighted versions of the Khinchin-Ostrowski property, and apply our result to establish the remarkable precision of a one-sided spectral decay condition which detects the local logarithmic integrability of a function. 1. Introduction 1.1. Main result Let E be a closed subset of the unit circle \mathbb{T}:=\{z\in\mathbb{C}:|z|=1\} . If (p_{n})_{n} is a sequence of analytic polynomials in a single complex variable, and p_{n}\to 0 on E in some appropriate sense, then by analyticity of the polynomials involved we may expect for the sequence (p_{n})_{n} to collapse to 0 also in the unit disk \mathbb{D}:=\{z\in\mathbb{C}:|z|<1\} . In the literature, one often refers to such results as Khinchin-Ostrowski properties , after the classical result on pointwise limits of sequences in the Nevanlinna class of the disk (see [ HJ12 , p. 278] ). To prove a rigorous statement of this type, one needs a growth restriction on (p_{n})_{n} in the unit disk, and E needs to be sufficiently massive to make the null-convergence of (p_{n})_{n} on E influence the sequence sufficiently strongly inside the disk. The purpose of this note is to reconcile the related results of Khrushchev from [ Khr78 ] and the author from [ Mal23a ] , and force the two works to meet in a remarkably sharp statement of the kind just described. The article [ Mal23a ] can be interpreted as a study of the Khinchin-Ostrowski property, and related matters, in the more general context of weights on \mathbb{T} instead of its subsets E . Let us state precisely what we mean by the weighted context, and how it specializes to the classical setting. Definition 1.1 ( Khinchin-Ostrowski property ) . Let \lambda(x) be an increasing positive function (a majorant) defined for x\in[1,\infty) , and let w be a non-negative Borel function on \mathbb{T} . We say that the pair (\lambda,w) satisfies the Khinchin-Ostrowski property, if for every sequence of analytic polynomials (p_{n})_{n} satisfying the two conditions (i) |p_{n}(z)|\leq C\lambda\Big(\frac{1}{1-|z|}\Big),z\in\mathbb{D} , for some C>0 , (ii) \lim_{n\to\infty}\int_{\mathbb{T}}|p(z)|w(z)|dz|=0 , we also have that (p_{n})_{n} collapses to 0 inside the disk: \lim_{n\to\infty}p_{n}(z)=0,\quad z\in\mathbb{D}. Setting w=1_{E} , the indicator function of a set E , the above definition reduces the the classical setting (the convergence p_{n}\to 0 on E being interpreted as convergence in L^{1}(E) ). In that case, we shall write (\lambda,E) instead of (\lambda,1_{E}) . The unweighted context is our main concern, but our results are both inspired by and have implications in the weighted context (see the theorems in Section 1.4 ). We will unveil a special role for the Khinchin-Ostrowski property of the following family of exponential majorants ( c>0 ): (ExpGrowth) \lambda\bigg(\frac{1}{1-|z|}\bigg)=\exp\bigg(\frac{c}{1-|z|}\bigg),\quad z\in\mathbb{D}. A simple consequence of the results in [ Mal23a ] (and, in the unweighted context w=1_{E} , of the results of Khrushchev in [ Khr78 ] ) is that if the integral \int_{I}\log w(z)\,|dz| diverges over every interval I\subset\mathbb{T} , and \lambda is as in ( ExpGrowth ), then the pair (\lambda,w) does not satisfy the Khinchin-Ostrowski property. A fortiori, the property is not satisfied if \lambda in ( ExpGrowth ) is replaced by a larger majorant. For instance, under the two conditions in ?THM? we can still ensure that p_{n}\to 1 on \mathbb{D} . Reversing the roles of the disk and the circle, in fact (p_{n})_{n} can be made to converge in L^{1}(w) to any given function in that space, while simultaneously converging to 0 in \mathbb{D} . Note that if w=1_{E} , then the divergence of the above logarithmic integral is clearly equivalent to E containing no intervals (if E is not closed then we interpret this statement as |I\setminus E|>0 for every interval I ). So such a set rather spectacularly fails to satisfy the classical Khinchin-Ostrowski property when paired with a majorant in ( ExpGrowth ). Conversely, if \log w is integrable over some interval (or if E contains an interval), then the Khinchin-Ostrowski property will hold even if the majorant in ( ExpGrowth ) is replaced by a much larger one (see [ Mal23a ] for details, and much more general weighted results). Our goal here is to understand exactly for which majorants the Khinchin-Ostrowski property can be described in terms of intervals alone, as in the results mentioned above. According to what was said there, relevant for this question are the majorants which are smaller than the ones in ( ExpGrowth ). Therefore, let us introduce largest majorants which are strictly smaller than the ones in ( ExpGrowth ). Consider (1.1) \lambda_{h}\bigg(\frac{1}{1-|z|}\bigg):=\exp\bigg(\frac{h(1-|z|)}{1-|z|}\bigg),\quad z\in\mathbb{D}, where h is a non-negative continuous function satisfying h(0)=0 . In order to avoid trivialities, we assume that h decays slow enough to have \lim_{x\to 0+}h(x)/x=\infty. We will also work under the following rather mild regularity assumptions (Reg) h(x)\text{ is increasing, and }h(x)/x\text{ is decreasing, in }x but we will impose no further conditions on h . In particular, h(x) may tend to 0 arbitrarily slowly as x decreases to 0 . Note that if h is a non-negative function satisfying \lim_{x\to 0+}h(x)=h(0)=0 but not ( Reg ), then there exists a continuous function \widetilde{h} pointwise larger than h for which ( Reg ) is satisfied, and \widetilde{h}(0)=0 . For instance, we may take \widetilde{h} to be the least concave majorant of h . Establishing