Hörmander’s multiplier theorem on H p H^{p} -spaces in the rational Dunkl setting

Jacek Dziubański, Agnieszka Hejna-Łyżwa · 2026

Read Hörmander’s multiplier theorem on H p H^{p} -spaces in the rational Dunkl setting by Jacek Dziubański, Agnieszka Hejna-Łyżwa (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-2603-20555.txt.

Introduction

Dziubański and Hejna-Łyżwa prove a Hörmander multiplier theorem for Hardy spaces in the rational Dunkl setting. The work sits on Euclidean space with a root system and a weight. A bounded function that meets the classic Hörmander smoothness terms defines the multiplier. They show this operator extends to a bounded map on Dunkl Hardy spaces. Those spaces are defined with a Littlewood-Paley square function. The proof uses atomic and molecule descriptions of the spaces. The range covers exponents p from zero up to one.

Catalog record

CREATORJacek Dziubański, Agnieszka Hejna-Łyżwa
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE91026
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.20555

Browse the full library · About the Tower of Babel

Full text

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

Hörmander’s multiplier theorem on H p H^{p} -spaces in the rational Dunkl setting Jacek Dziubański, Agnieszka Hejna-Łyżwa Abstract . On ℝ N \mathbb{R}^{N} equipped with a normalized root system ℛ \mathcal{R} and a multiplicity function k ≥ 0 k\geq 0 , let d ​ w ​ ( 𝐱 ) = Π α ∈ ℛ ​ | ⟨ 𝐱 , α ⟩ | k ⁡ ( α ) ​ d ​ 𝐱 dw(\mathbf{x})=\Pi_{\alpha\in\mathcal{R}}|\langle\mathbf{x},\alpha\rangle|^{k(\alpha)}\,d\mathbf{x} , 𝐍 = N + ∑ α ∈ ℛ k ⁡ ( α ) \mathbf{N}=N+\sum_{\alpha\in\mathcal{R}}k(\alpha) denote the associated measure and the homogeneous dimension of the system ( ℛ , k ) (\mathcal{R},k) respectively. Let ℱ \mathcal{F} stand for the Dunkl transform. For 0 < p ≤ 1 0<p\leq 1 , let m m be a bounded function on ℝ N \mathbb{R}^{N} , which satisfies the classical Hörmander’s condition with smoothness s > 𝐍 / p s>\mathbf{N}/p . We show that the multiplier operator 𝒯 m ​ f = ℱ − 1 ​ ( m ​ ℱ ​ f ) \mathcal{T}_{m}f=\mathcal{F}^{-1}(m\mathcal{F}f) , initially defined on H Dunkl p ∩ L 2 ​ ( d ​ w ) H^{p}_{\mathrm{Dunkl}}\cap L^{2}(dw) , has a unique extension to a bounded operator in H Dunkl p H^{p}_{\mathrm{Dunkl}} , where the space H Dunkl p H^{p}_{\mathrm{Dunkl}} is defined by means of a Littlewood-Paley square function. To prove the theorem, we use special atomic and molecule characterizations of H Dunkl p H^{p}_{\mathrm{Dunkl}} . 2020 Mathematics Subject Classification primary: 42B30, secondary: 42B35, 33C52, 42B10, 35K08 1. Introduction In classical harmonic analysis, Hörmander’s multiplier theorem [ 19 ] reduces the problem of operator boundedness on the spaces L^{p}(\mathbb{R}^{N}) and H^{p}(\mathbb{R}^{N}) to controlling the behavior of the symbol on dyadic scales. By replacing strict pointwise bounds on derivatives with a Sobolev-type integral condition, this theorem establishes the smoothness threshold s>N/p-N/2 required for the multiplier (see e.g., [ 6 ] , [ 29 ] ). This result serves as a standard tool in operator theory, finding direct applications in, among others, proving the boundedness of singular integral operators, investigating the properties of Bochner-Riesz means, and analyzing solutions to partial differential equations (e.g., the wave and Schrödinger equations). The aim of the present paper is to investigate an analogous sufficient condition for multiplier operators on the Hardy spaces H^{p}_{\mathrm{Dunkl}} in the rational Dunkl setting. On the Euclidean space \mathbb{R}^{N} equipped with a root system \mathcal{R} and a multiplicity function k\geq 0 , let dw(\mathbf{x})=\Pi_{\alpha\in\mathcal{R}}|\langle\mathbf{x},\alpha\rangle|^{k(\alpha)}\,d\mathbf{x} be the associated measure and \mathbf{N}=N+\sum_{\alpha\in\mathcal{R}}k(\alpha) the homogeneous dimension of the system (\mathcal{R},k) , respectively. Let E(\mathbf{x},\mathbf{y}) , \mathbf{x},\mathbf{y}\in\mathbb{C}^{N} , be the associated Dunkl kernel. The Dunkl transform \mathcal{F}f(\xi)={\mathbf{c}}_{k}^{-1}\int_{\mathbb{R}^{N}}E(-i\xi,\mathbf{x})f(\mathbf{x})\,dw(\mathbf{x}), originally defined on L^{1}(dw) , is an isometry on L^{2}(dw) and preserves the Schwartz class of functions \mathcal{S}(\mathbb{R}^{N}) . Moreover, its inverse is given by \mathcal{F}^{-1}f(\mathbf{x})=\mathcal{F}f(-\mathbf{x}) . Let \Delta_{k}=\sum_{j=1}^{N}T_{j} be the Laplace-Dunkl operator, where T_{j} are the Dunkl operators. The operator \Delta_{k} generates the Dunkl-heat semigroup \{H_{t}\}_{t>0} , of contractions on L^{p}(dw) , 1\leq p\leq\infty , strongly continuous and holomorphic for 1\leq p<\infty . For a reasonable function f on \mathbb{R}^{N} , (for example for f\in L^{2}(dw) ) we define the square function Sf(x) : (1.1) Sf(\mathbf{x})=\Big(\int_{0}^{\infty}\int_{\|\mathbf{x}-\mathbf{y}\|<t}|t^{2}\Delta_{k}H_{t^{2}}f(\mathbf{y})|^{2}\frac{dw(\mathbf{y})}{w(B(\mathbf{x},t))}\frac{dt}{t}\Big)^{1/2}. Following [ 18 ] and [ 11 ] , for 0<p\leq 1 , let (1.2) \mathbb{H}^{p}_{\mathrm{Dunkl}}=\{f\in L^{2}(dw):Sf\in L^{p}(dw)\}. We equip the space with the quasi-norm \|f\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}}=\|Sf\|_{L^{p}(dw)} . Then \|f-g\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}}^{p} is a metric in \mathbb{H}^{p}_{\mathrm{Dunkl}} . Definition 1.1 . The Hardy space H^{p}_{\mathrm{Dunkl}} is defined as the completion of the space \mathbb{H}^{p}_{\mathrm{Dunkl}} in the quasi-norm \|\cdot\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}} , where, for a Cauchy sequence (f_{n})_{n\in\mathbb{N}} in \mathbb{H}^{p}_{\rm Dunkl} , its equivalence class \big[(f_{n})_{n\in\mathbb{N}}\big] has the quasi-norm \|\big[(f_{n})_{n\in\mathbb{N}}\big]\|_{H^{p}_{\rm Dunkl}}=\lim_{n\to\infty}\|f_{n}\|_{\mathbb{H}^{p}_{\rm Dunkl}} . Let us note that if (f_{n})_{n\in\mathbb{N}} is a Cauchy sequence in the quasi-norm \|\cdot\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}} , then the limit \lim_{n\to\infty}\|f_{n}\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}} exists, since for all n,m\in\mathbb{N} we have \Big|\|f_{n}\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}}-\|f_{m}\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}}\Big|\leq\|f_{n}-f_{m}\|_{\mathbb{H}^{p}_{\mathrm{Dunkl}}}. We will show that every element of H^{p}_{\mathrm{Dunkl}} is identified with a unique tempered distribution f\in\mathcal{S}^{\prime}(\mathbb{R}^{N}) , where \mathcal{S}(\mathbb{R}^{N}) denotes the class of Schwartz functions on \mathbb{R}^{N} . For s>0 , denote by W_{2}^{s}(\mathbb{R}^{N}) the classical Sobolev space, that is, W_{2}^{s}(\mathbb{R}^{N})=\Big\{f\in L^{2}(dx):\int_{\mathbb{R}^{N}}|\widehat{f}(\xi)|^{2}(1+|\xi|)^{2s}\,d\xi=\|f\|_{W_{2}^{s}}^{2}<\infty\Big\}. We are now in a position to state our main result. Theorem 1.2 . Let 0<p\leq 1 and let \psi\in C^{\infty}_{c}(\mathbb{R}^{N}) be a non-zero radial function such that \text{supp }\psi\subset\mathbb{R}^{N}\setminus\{0\} . Assume that m is a bounded function on \mathbb{R}^{N} which satisfies the Hörmander condition (1.3) M_{0}:=\sup_{t>0}\|\psi(\cdot)m(t\cdot)\|_{W^{s}_{2}}<\infty for certain s>{\mathbf{N}}/p , then the multiplier operator \mathcal{T}_{m}f=\mathcal{F}^{-1}(m(\cdot)\mathcal{F}f(\cdot)) originally defined on \mathbb{H}^{p}_{\mathrm{Dunkl}} , has a unique extension to a bounded operator on H^{p}_{\mathrm{Dunkl}} . Moreover, \|\mathcal{T}_{m}f\|_{H^{p}_{\mathrm{Dunkl}}}\leq C_{p,s}M_{0}\|f\|_{H^{p}_{\mathrm{Dunkl}}}. The theorem extends the results of the authors of [ 14 ] , where Dunkl multipliers on Lebesgue spaces L^{p}(dw) , 1\leq p<\infty , and the Hardy space H^{1}_{\rm Dunkl} were studied. In the case of the space H^{1}_{\rm Dunkl} , the multiplier theorem was proved using a characterization of this space by relevant Riesz transforms. We refer the reader to [ 21 ] , in which the Dunkl-transform multiplier operators on Lebesgue spaces L^{p}_{r}(dw) , 1<p\leq 2 , of radial functions were considered. Let us also remark that for the Hörmander’s multiplier theorem on the classical Hardy spaces one needs smoothness s>N/p-N/2 for the multiplier m , where N is the Euclidean dimension which, in this case, coincides with the homogeneous dimension. It should be emphasized that Theorem 1.2 applies to general, non-radial multipliers. The requirement of a higher smoothness order s>\mathbf{N}/p in our proof, compared to the classical setting, is a consequence of a fundamental obstacle in Dunkl theory: the lack of general L^{p}(dw) -boundedness for Dunkl translations and convolutions. Whether the Dunkl translation operators are bounded on L^{p}(dw) for p\neq 2 remains a well-known open problem. Currently, affirmative results are restricted either to product root systems or to the subspace of radial functions. To overcome this structural difficulty, our strategy avoids general L^{p} estimates. Instead, we exploit two crucial features of the Dunkl setting. First, we rely on the fact that Dunkl translations and convolutions are well-behaved bounded operators on L^{2}(dw) . Second, we make essential use of the support theorem, which states that if a function f is supported in a ball B(0,r) , then its Dunkl translation \tau_{\mathbf{x}}f is supported in the orbit set \mathcal{O}(B(\mathbf{x},r)) (see Theorem 2.2 ). This result was previously known only for radial functions [ 23 ] , but in [ 14 ] it was generalized to non-radial functions, which serves as one of our main tools (see also [ 2 ] , [ 30 ] ). Furthermore, recent results from [ 17 ] , [ 20 ] , and [ 12 ] indicate that this orbital support condition cannot be essentially improved. By combining the L^{2} -theory with the support estimate, we are able to effectively employ atomic decompositions and Hardy space techniques to obtain the desired bounds. One of the main elements of the proof of Theorem 1.2 is the usage of characterizations of the space H^{p}_{\mathrm{Dunkl}} by means of special atoms and molecules. In the case of Hardy spaces associated with semigroups satisfying Gaussian bounds, or more generally Davies-Gaffney estimates, such characterizations were considered in [ 18 ] , [ 11 ] . Since in the estimates for the Dunkl-heat semigroups kernels two distances occur, namely the Euclidean distance and an orbit distance, and for the reader

Read the complete text