Rearrangement Invariant Orthogonal Sums in Kreĭn Spaces. II

Michael A. Dritschel, Alejandra Maestripieri, James Rovnyak · 2025

Read Rearrangement Invariant Orthogonal Sums in Kreĭn Spaces. II by Michael A. Dritschel, Alejandra Maestripieri, James Rovnyak (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-2509-15428.txt.

Introduction

Michael Dritschel, Alejandra Maestripieri, and James Rovnyak continue a study of sums inside Krein spaces. A Krein space is like a standard inner product space, but some lengths can be negative. Part one of their work dealt with sums of well behaved parts of such a space. Here they turn to a wider class of parts, called quasi-pseudo-regular.

Such a part splits into a well behaved piece and a neutral piece, where all lengths are zero. The trio give new ways to tell when a part has this form. They study endless sums of these parts with a flexible sum notion based on ranges of operators. The paper is dense and aimed at experts in operator theory.

Catalog record

CREATORMichael A. Dritschel, Alejandra Maestripieri, James Rovnyak
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE93 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-15428

Browse the full library · About the Tower of Babel

Full text

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

Rearrangement Invariant Orthogonal Sums in Kreĭn Spaces. II Thanks: Alejandra Maestripieri was partially supported by the Air Force Office of Scientific Research (USA) grant FA9550-24-1-0433. Michael A. Dritschel, Alejandra Maestripieri, James Rovnyak Abstract . Part I of the paper considered infinite orthogonal sums of regular subspaces in a Kreĭn space (that is, of subspaces which are themselves Kreĭn spaces). How precisely these sums should be defined and conditions for when such a sum is itself regular were examined. These included, for example, a boundedness condition for the sum of the corresponding orthogonal projections. The same problem is addressed here for (quasi-)pseudo-regular subspaces. Such subspaces happen to be the orthogonal direct sum of a regular space and an isotropic, or neutral, subspace. Alternate characterizations of such subspaces are given, along with various examples and counter-examples. The infinite orthogonal sum of these is then considered. 1991 Mathematics Subject Classification 47B50, 47A05, 46C20 1. Introduction and background A fundamental result in the study of Hilbert space geometry is that orthogonal sums of Hilbert spaces are themselves Hilbert spaces. This works whether the number of spaces involved is finite or infinite. Kreĭn spaces are indefinite analogues of Hilbert spaces, and so it is natural to ask which properties of Hilbert spaces carry over to these. Beginning in 1979, Brian McEnnis made a detailed study of this topic in a series of papers considering shift operators on Kreĭn spaces [ 10 , 11 , 12 ] . Part I of the current series [ 14 ] contains an elaboration and expansion of McEnnis’ work. This, an earlier work of the third author [ 13 ] , and McEnnis’ papers were primarily concerned with regular spaces , which are the natural analogue in Kreĭn spaces of closed subspaces of Hilbert spaces, in that they are themselves Kreĭn spaces and orthogonal sums are automatically direct. Part II will examine orthogonal sums of pseudo-regular and quasi-pseudo-regular or qpr spaces (formally defined in Definition 2.1 ), which generalize regular spaces in that they are the orthogonal direct sum of regular spaces and neutral subspaces (where all inner products are 0 ). A natural question is how the orthogonal sum of such subspaces should be defined, and what conditions are needed to guarantee that orthogonal sums exist. Here, even for finitely many orthogonal spaces, the sum need not exist (as a space of the same type), or it may not be direct, along with other possible difficulties. Theorem 1.5 originates in Part I of this paper [ 14 ] and highlights the status regarding orthogonal sums of regular spaces. There, infinite sums of regular spaces are well-defined in terms of rearrangement invariance of infinite direct sums of vectors from these spaces, or equivalently, by means of a boundedness condition of related projection operators. While it is perhaps too much to ask that such definitive results hold for pseudo-regular and quasi-pseudo-regular spaces, analogues are nevertheless available — see Theorems 5.1 and 6.1 , as well as Corollary 6.3 . The first of these theorems is a strengthening of Theorem 4.3 in [ 13 ] , the main ideas of which were discovered by the first two authors and communicated to the third. Quasi-pseudo-regular (or qpr) spaces are new, being an outgrowth of work of Aurelian Gheondea on pseudo-regular spaces [ 5 , 6 ] . The latter spaces are assumed to be closed, while this is relaxed in the quasi-pseudo-regular setting to the assumption that they are operator ranges. Generalizations of Gheondea’s results form the bulk of Section 2. A deeper understanding of infinite sums of operator ranges is required when considering orthogonal sums of quasi-pseudo-regular spaces. Section 3 takes this up, expanding on fundamental work of Fillmore and Williams [ 4 ] . Section 4 deals with nets of projections in Kreĭn spaces, playing a key part in understanding orthogonal sums of pseudo-regular spaces. The last two sections put all of this together in theorems on infinite orthogonal sums of pseudo-regular and quasi-pseudo-regular spaces. Now for a brief summary of the needed background in Kreĭn space theory. Recall that a Kreĭn space is a complex inner product space ({\mathcal{H}},{\left<\cdot,\cdot\right>}_{{\mathcal{H}}}) which admits a fundamental decomposition {\mathcal{H}}={\mathcal{H}}_{+}\oplus{\mathcal{H}}_{-} such that {\mathcal{H}}_{+} is a Hilbert space and {\mathcal{H}}_{-} is the anti-space of a Hilbert space (that is, a Hilbert space with the inner product replaced by its negative). Every Kreĭn space {\mathcal{H}} has a unique strong topology which is induced by any such decomposition, though in general the decomposition is not unique. A subspace {\mathcal{M}} is called regular if it is closed and is itself a Kreĭn space in the inner product of {\mathcal{H}} . In analogy with Hilbert spaces, the range of a selfadjoint projection operator on {\mathcal{H}} is a regular subspace, and every regular subspace is the range of a unique selfadjoint projection. Here selfadjoint means with respect to the indefinite inner product. Definition 1.1 . Let be {\mathcal{M}}_{1} and {\mathcal{M}}_{2} be ( not necessarily closed ) subspaces of a Kreĭn space {\mathcal{H}} . (1) {\mathcal{M}}_{1} and {\mathcal{M}}_{2} are orthogonal if for any f\in{\mathcal{M}}_{1} , g\in{\mathcal{M}}_{2} , \left<f,g\right>=0 . ( Unlike in Hilbert spaces, this does not necessarily imply that {\mathcal{M}}_{1}\cap{\mathcal{M}}_{2}=\{0\} . ) (2) The sum of orthogonal subspaces {\mathcal{M}}_{1} and {\mathcal{M}}_{2} is written as {\mathcal{M}}_{1}\oplus{\mathcal{M}}_{2} and called the orthogonal sum . (3) If {\mathcal{M}}_{1}\cap{\mathcal{M}}_{2}=\{0\} , the sum of {\mathcal{M}}_{1} and {\mathcal{M}}_{2} is written as {\mathcal{M}}_{1}\,{\overset{\smash{\mathchoice{{{\hbox{\scalebox{1.8}{$\displaystyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\textstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptscriptstyle.$}}}}}}}{+}}\,{\mathcal{M}}_{2} and is called the direct sum . (4) If {\mathcal{M}}_{1} and {\mathcal{M}}_{2} are orthogonal and {\mathcal{M}}_{1}\cap{\mathcal{M}}_{2}=\{0\} , the sum of {\mathcal{M}}_{1} and {\mathcal{M}}_{2} is written as {\mathcal{M}}_{1}\,{\overset{\smash{\mathchoice{{{\hbox{\scalebox{1.8}{$\displaystyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\textstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptscriptstyle.$}}}}}}}{\oplus}}\,{\mathcal{M}}_{2} and is called the orthogonal direct sum . Definitions (2)–(4) extend obviously to any finite number of spaces. The notation is at odds with that used in Part I [ 14 ] , but since the goal here is to consider orthogonal sums which are not necessarily direct, the above notation is better suited to the task. Finite orthogonal sums of closed spaces in a Hilbert space are direct and closed, and likewise finite orthogonal sums of regular subspaces are direct and regular. Theorem 1.2 . Let {\mathcal{M}}_{1},\dots,{\mathcal{M}}_{n} be pairwise orthogonal regular subspaces of a Kreĭn space {\mathcal{H}} . (1) The span {\mathcal{M}} of {\mathcal{M}}_{1},\dots,{\mathcal{M}}_{n} is regular, and {\mathcal{M}} is the orthogonal direct sum of the subspaces: {\mathcal{M}}={\mathcal{M}}_{1}\,\,{\overset{\smash{\mathchoice{{{\hbox{\scalebox{1.8}{$\displaystyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\textstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptscriptstyle.$}}}}}}}{\oplus}}\,\cdots\,{\overset{\smash{\mathchoice{{{\hbox{\scalebox{1.8}{$\displaystyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\textstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptstyle.$}}}}}{{{\hbox{\scalebox{1.8}{$\scriptscriptstyle.$}}}}}}}{\oplus}}\,{\mathcal{M}}_{n}. The selfadjoint projection P onto {\mathcal{M}} is given by P=E_{1}+\cdots+E_{n} , where E_{1},\dots,E_{n} are the selfadjoint projections onto {\mathcal{M}}_{1},\dots,{\mathcal{M}}_{n} . (2) Let \widetilde{\mathcal{M}}={\mathcal{M}}_{1}\times\dots\times{\mathcal{M}}_{n} be the Cartesian product of the subspaces viewed as Kreĭn spaces in the inner product of {\mathcal{H}} . Then \widetilde{\mathcal{M}} is a Kreĭn space, and the mapping U\colon(f_{1},\dots,f_{n})\to f_{1}+\cdots+f_{n} from \widetilde{\mathcal{M}} onto {\mathcal{M}} is a Kreĭn space isomorphism. As indicated earlier, the generalization of Theorem 1.2 to infinite families is standard for Hilbert spaces. This is not the case for Kreĭn spaces. While the orthogonal sum of regular subspaces is direct, it may fail to be regular, or it may be regular but not every vector in the closed span is representable as a sum of vectors from the individual subspaces [ 10 , Example 7.3] , [ 14 , Example 6.2] . Likewise, the notion of an infinite Cartesian product is problematic for Kreĭn spaces. In what follows, \mathrm{span}_{\lambda}\,{\mathcal{M}}_{\lambda} stands for the vector space of finite sums

Read the complete text