Reflections in L 2 ​ ( 𝕋 ) L^{2}(\mathbb{T})

Esteban Andruchow Β· 2025

Read Reflections in L 2 ​ ( 𝕋 ) L^{2}(\mathbb{T}) by Esteban Andruchow (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-2504-11600.txt.

Introduction

Esteban Andruchow studies reflection operators acting on square-integrable functions on the unit circle. For each point inside the unit disc, a MΓΆbius-type map defines a composition operator that sends a function to its composite with the map. Squaring that operator gives the identity, so it acts as a reflection, though not a symmetric one.

Two related symmetric reflections are built by weighting the composition with kernel functions. One of them is the unitary part of the original operator's polar decomposition. Andruchow characterizes the eigenspaces of all three operators for their two possible eigenvalues. He then studies how those eigenspaces sit relative to each other as the disc point changes, including their overlaps and perpendicular parts. The work sorts out the geometry of a natural family of reflections in a classical function space.

Catalog record

CREATOREsteban Andruchow
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-11600

Browse the full library Β· About the Tower of Babel

Full text

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

Reflections in L 2 ​ ( 𝕋 ) L^{2}(\mathbb{T}) Esteban Andruchow Abstract Let 𝔻 = { z ∈ β„‚ : | z | < 1 } \mathbb{D}=\{z\in\mathbb{C}:|z|<1\} and 𝕋 = { z ∈ β„‚ : | z | = 1 } \mathbb{T}=\{z\in\mathbb{C}:|z|=1\} . For a ∈ 𝔻 a\in\mathbb{D} , consider Ο† a ​ ( z ) = a βˆ’ z 1 βˆ’ a Β― ​ z \varphi_{a}(z)=\frac{a-z}{1-\bar{a}z} and C a C_{a} the composition operator in L 2 ​ ( 𝕋 ) L^{2}(\mathbb{T}) induced by Ο† a \varphi_{a} : C a ​ f = f ∘ Ο† a . C_{a}f=f\circ\varphi_{a}. Clearly C a C_{a} satisties C a 2 = I C_{a}^{2}=I , i.e., is a non-selfadjoint reflection. We also consider the following symmetries (selfadjoint reflections) related to C a C_{a} : R a = M | k a | β€– k a β€– 2 ​ C a ​ and ​ W a = M k a β€– k a β€– 2 ​ C a , R_{a}=M_{\frac{|k_{a}|}{\|k_{a}\|_{2}}}C_{a}\ \hbox{ and }\ W_{a}=M_{\frac{k_{a}}{\|k_{a}\|_{2}}}C_{a}, where k a ​ ( z ) = 1 1 βˆ’ a Β― ​ z k_{a}(z)=\frac{1}{1-\bar{a}z} is the Szego kernel. The symmetry R a R_{a} is the unitary part in the polar decomposition of C a C_{a} . We characterize the eigenspaces N ⁑ ( T a Β± I ) N(T_{a}\pm I) for T a = C a , R a T_{a}=C_{a},R_{a} or W a W_{a} , and study their relative positions when one changes the parameter a a , e.g., N ⁑ ( T a Β± I ) ∩ N ⁑ ( T b Β± I ) N(T_{a}\pm I)\cap N(T_{b}\pm I) , N ⁑ ( T a Β± I ) ∩ N ​ ( T b Β± I ) βŸ‚ N(T_{a}\pm I)\cap N(T_{b}\pm I)^{\perp} , N ​ ( T a Β± I ) βŸ‚ ∩ N ⁑ ( T b Β± I ) N(T_{a}\pm I)^{\perp}\cap N(T_{b}\pm I) , etc., for a β‰  b ∈ 𝔻 a\neq b\in\mathbb{D} . 2020 MSC: 47A05, 47B33, 47B91 Keywords: Reflections, Symmetries, Disk automorphisms, Grassmann geometry. 1 Introduction Let {\cal H}=L^{2}(\mathbb{T}) with normalized Lebesgue measure. For each a\in\mathbb{D} let \varphi_{a}(z):=\displaystyle{\frac{a-z}{1-\bar{a}z}}. Note that \varphi_{a}\circ\varphi_{a}(z)=z and |\varphi_{a}(z)|=1 for z\in\mathbb{T} , and therefore the composition operator C_{a}:{\cal H}\to{\cal H},\ C_{a}f=f\circ\varphi_{a} is well defined, and is a reflection in {\cal H} : C_{a}^{2}=I . A reflection C is a symmetry if additionally C^{*}=C . Clearly C_{a} is a symmetry only if a=0 . The adjoint of C_{a} is easily computed: C_{a}^{*}=(1-|a|^{2})M_{\frac{1}{|1-\bar{a}z|^{2}}}C_{a} , where M_{g} denotes the multiplication operator (by g ). There are two symmetries closely relaed to C_{a} . First the unitary part R_{a} in the polar decomposition C_{a}=R_{a}(C_{a}^{*}C_{a})^{1/2} of C_{a} (in [ 9 ] , G. Corach, H. Porta and L. Recht noted this fact, that the unitary part of a reflection is in fact a symmetry). Next the operator W_{a} given by W_{a}=M_{\psi_{a}}C_{a} , where \psi_{a} is the normalized Szego kernel \psi_{a}=\frac{(1-|a|^{2})^{1/2}}{1-\bar{a}z} . The object of this paper is the study of eigenspaces of these symmetries, and their relative position when one changes the parameter a\in\mathbb{D} . By relative position of two subspaces {\cal S} , {\cal T} , we mean the computation of the intersections {\cal S}\cap{\cal T} , {\cal S}^{\perp}\cap{\cal T}^{\perp} , {\cal S}\cap{\cal T}^{\perp} and {\cal S}^{\perp}\cap{\cal T} . Of particular interest to us are the last two, for their geometric significance: if these two intersections have the same dimention, it means that there exists a geodesic of the Grassmann manifold of {\cal H} that joins them; if they are trivial, the geodesic is unique. When the four intersections are trivial, the subspaces {\cal S} , {\cal T} are said to be in generic position. This paper is a continuation of [ 4 ] , where composition operators with symbols \varphi_{a} where considered in the Hardy space of the disc. Several results otained there will be useful here. The contents of the paper are the following. In Section 2 the eigenspaces of C_{a} , R_{a} and W_{a} are characterized. In this characterization, an important role is played by the unique fixed point \omega_{a} of \varphi_{a} inside \mathbb{D} . In Section 3 we make some brief remarks on the maps a\mapsto C_{a} and a\mapsto R_{a} . In Section 4 we obtain partial results on the relative position of the eigenspaces of these reflections induced by different a,b\in\mathbb{D} . The main result is Theorem 4.6, which states that for a\neq b N(C_{a}-I)\cap N(C_{b}-I)=\langle 1\rangle\ \hbox{ and }\ N(C_{a}+I)\cap N(C_{b}+I)=\{0\}. In Section 5 we consider the special case of the subspaces N(C_{a}-I) and N(C_{a}+I) , where a complete study can be done. In Section 6, using the conjugation properties R_{a}C_{a}R_{a}=C_{a}^{*},\ R_{a}C_{0}R_{a}=R_{\Omega_{a}},\ W_{a}C_{a}W_{a}=M_{\frac{k_{a}^{2}}{\|k_{a}\|_{2}^{2}}}C_{a}=M_{\frac{k_{a}}{|k_{a}|}}C_{a}^{*}\ \hbox{ and }\ W_{a}C_{0}W_{a}=W_{\Omega_{a}} we can compute further intersections between the eigenspaces. Here \Omega_{a}=\frac{2a}{1+|a|^{2}} is characterized by \omega_{\Omega_{a}}=a , i.e., a is the unique fixed point of \varphi_{\Omega_{a}} in \mathbb{D} . In Section 7 we examine the consequence of these intersections for the geometry of the Grassmann manifold of {\cal H} : they determine wether given pairs of subspaces can be joined by a geodesic of this manifold. Aknowledgement: this work was completed with the support of PICT 2019 04060 (ANPCyT), Argentina. 2 Preliminaries and notations If {\cal S} is a closed subspace of {\cal H} , P_{\cal S} will denote the orthogonal projection onto {\cal S} . For a bounded operator T , |T| will denote the modulus of T , i.e., |T|=(T^{*}T)^{1/2} ; N(T) and R(T) will denote the nullspace and range of T , respectively. Note that C_{a}^{*}C_{a}=(1-|a|^{2})M_{\frac{1}{|1-\bar{a}z|^{2}}}\ \hbox{ and }\ |C_{a}|=(C_{a}^{*}C_{a})^{1/2}=(1-|a|^{2})^{1/2}M_{\frac{1}{|1-\bar{a}z|}}. (1) In [ 9 ] , Corach, Porta and Recht, made the following remarkable observation: Remark 2.1 . Let C be a reflection in {\cal H} , and let C=R|C| be its polar decomposition. Then the unitary operator R is in fact a symmetry. Let us denote by R_{a} the symmetry obtained in the polar decomposition of C_{a} . After routine computations one obtains that R_{a}=C_{a}|C_{a}|^{-1}=\frac{1}{(1-|a|^{2})^{1/2}}C_{a}M_{|1-\bar{a}z|}=(1-|a|^{2})^{1/2}M_{\frac{1}{|1-\bar{a}z|}}C_{a}. (2) Note that if C^{2}=I and C=R|C| is the polar decomposition, then C^{*}=|C|R , so that RCR=C^{*} . In particular, for C=C_{a} we have R_{a}C_{a}R_{a}=C_{a}^{*}. (3) There is another symmetry in {\cal H} associated with \varphi_{a} , known in the literature (at least when acting in the (invariant) Hardy subspace H^{2}(\mathbb{T})\subset{\cal H} , [ 10 ] ): W_{a}=M_{\psi_{a}}C_{a}=(1-|a|^{2})^{1/2}M_{\frac{1}{1-\bar{a}z}}C_{a}. (4) Note the similarity between R_{a} in ( 2 ) and W_{a} in ( 4 ). We shall say that an element f\in{\cal H} is even if f(z)=f(-z)\ p.p. in \mathbb{T} . Equivalently, if the Fourier coefficients \hat{f}(n) of f are 0 for n odd. We shall denote by {\cal E} the closed subspace of even elements. Similarly, {\cal O} will denote the subspace of odd elements of {\cal H} ( g\in{\cal H} such that g(-z)=-g(z)\ p.p. in \mathbb{T} or \hat{g}(n)=0 for n even). Note that for a=0 , C_{0}=R_{0} , and N(C_{0}-I)={\cal E} and N(C_{0}+I)={\cal O} . Finally, let {\cal H}^{+} be the Hardy space, as a subset of {\cal H} , {\cal H}^{+}:=H^{2}(\mathbb{T})=\{f\in{\cal H}:\hat{f}(n)=0\hbox{ for }n<0\} , identified with the Hardy space of the disk H^{2}(\mathbb{D})=\{g(z)=\sum_{k=0}^{\infty}a_{k}z^{k}:\sum_{k=0}^{\infty}|a_{k}|^{2}<\infty\}. A reflection T is completely described by its two eigenspaces N(T-I) and N(T+1) . The object of this study is the characterization of the eigenspaces N(T-I) and N(T+I)

Read the complete text