On preservers of strong Birkhoff-James orthogonality between C ∗ C^{*} -algebras
Bojan Kuzma, Srdjan Stefanović, Ryotaro Tanaka · 2025
Read On preservers of strong Birkhoff-James orthogonality between C ∗ C^{*} -algebras by Bojan Kuzma, Srdjan Stefanović, Ryotaro Tanaka (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-2510-24308.txt.
Introduction
Bojan Kuzma and two co-authors study maps between C-star algebras that keep a strong form of a geometric relation. The relation is called strong Birkhoff-James orthogonality. It extends the plain idea of a right angle to abstract algebras. The team asks what a map must look like if it keeps this relation in both ways.
For algebras with a unit, each such linear map is a star map that keeps all the algebra rules, followed by a step that multiplies by a fixed unitary element. For compact algebras, they also pin down the form of such maps when the maps need not be linear. They give a test for compact algebras that uses this relation alone.
Catalog record
| CREATOR | Bojan Kuzma, Srdjan Stefanović, Ryotaro Tanaka |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 114 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-24308
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On preservers of strong Birkhoff-James orthogonality between C ∗ C^{*} -algebras Thanks: The work of the first author is supported in part by the Slovenian Research Agency (research program P1-0285 and research projects J1-50000 N1-0296 and N1-0428). The second author is supported by the Ministry of Science, Technological Development and Innovation of Republic of Serbia: grant number 451-03-47/2023-1/200104 with Faculty of Mathematics. The third author was supported by JSPS KAKENHI Grant Number JP24K06788. Bojan Kuzma, Srdjan Stefanović, Ryotaro Tanaka Abstract . It is shown that every linear strong Birkhoff-James isomorphism between unital C ∗ C^{*} -algebras is a ∗ * -isomorphism followed by a unitary multiplication. Moreover, as a partial extension of this result to the non-unital case, the form of (possibly nonlinear) strong Birkhoff-James isomorphisms between compact C ∗ C^{*} -algebras are determined. A nonlinear characterization of compact C ∗ C^{*} -algebras in terms of strong Birkhoff-James orthogonality is also given. 1991 Mathematics Subject Classification Primary 46L05; Secondary 46L35, 46B20, 46B80, 47B49. 1. Introduction An \mathscr{S} -preserver is a mapping that preserves a certain structure \mathscr{S} of mathematical objects. As basic examples, a linear isomorphism is a linear-structure preserver and a homeomorphism is a topological-structure preserver. In most cases, preservers can be formulated as the class of (iso)morphisms in some concrete category, and the class of objects are chosen between those of Banach spaces and Hilbert C^{*} -modules depending on structures under consideration. It is important to obtain detailed descriptions of various preservers as barometers for amounts of information contained in the corresponding structures. In this direction, the finer the structure \mathscr{S} is, the more restrictive \mathscr{S} -preservers are. A problem that asks for the form of \mathscr{S} -preservers is called a preserver problem of the structure \mathscr{S} . The prototype of the study on preserver problems dates back to 1897, the Frobenius theorem: A linear determinant preserver on a matrix algebra M_{n}(\mathbb{C}) is either of the forms A\mapsto QAP or A\mapsto QA^{T}P , where P and Q are regular matrices such that \det(PQ)=1 . Roughly speaking, up to transpose, the matrix equivalence is controlled by the structure of determinant to some extent. If, further, a linear mapping preserves eigenvalues of matrices with multiplicity, then Q can be chosen P^{-1} ; see [ 49 , Theorem 3] . Here, it is worth noting that the structure of eigenvalues with multiplicity contains more information than that of determinant, since \det(A) is the product of the eigenvalues of A with multiplicity. Up to now, there are many interesting formulations and results on preserver problems concerning algebraic structures in matrix algebras, or more generally, in C^{*} -algebras. We refer to Hou and Petek-Šemrl [ 31 , 57 ] for the structure of preservers of rank-one operators, which is one of the important reduction techniques in preserver problems. Of similarly importance is reduction to singularity preservers [ 22 ] or invertibility preservers; following the idea by Brešar-Šemrl [ 19 ] of using projections as an attempt towards their classification, Aupetit [ 11 ] succeeded to show that unital linear bijections which preserve the invertibility on a (general) von-Neumann algebra, are Jordan maps. Another powerful technique in invertible linear preserver problems on matrices is using overgroups. One can often easily find the natural candidates which preserve \mathscr{S} and it remains to check in the list which of their overgroups still preserves \mathscr{S} ; for more on this see Đoković, Li, and Platonov [ 23 , 24 , 58 ] . We refer also to Li-Pierce [ 47 ] for a brief overview of general techniques. A different approach is using Hermitian operators (in a sense of Vidav-Palmer); this was used by Sourour [ 56 ] to determine that X\mapsto UXV^{\ast} and X\mapsto UX^{t}V^{\ast} ( U,V unitary) are all the isometries of minimal normed ideals of \mathcal{B}(\ell_{2}) , the algebra of bounded operators on a separable Hilbert space. One can also successfully use ring-theoretical techniques in solving (linear) preserver problems; we mention here the theory of functional identities. This was pioneered by Brešar and its early success culminated in a solution [ 17 ] of Herstein conjectures that classify commutativity preserving maps on prime rings essentially as Jordan maps. Later it was fruitfully used by Alaminos, Brešar, and Villena in investigating Lie/Jordan derivations on von-Neumann algebras [ 1 ] . General (unital) C^{*} -algebras lack projections or minimal elements, but one can exploit other properties in solving preserver problems. We mention here Hatori and Molnár’s [ 30 ] classification of isometries of Thompson metric \|\ln(a^{-1/2}ba^{-1/2})\| on a positive cone, from which one can further determine the structure of positive homogeneous preservers isomorphisms of Löwner order (we refer to [ 53 , Theorem 13] for a precise statement). By reducing to order-monotone maps, preservers of various means on a positive cone of a general C^{*} -algebra were classified in [ 25 ] ; see also [ 52 ] . The books by Brešar et al. [ 18 ] and by Molnár [ 51 ] as well as the surveys by Bourhim and Mashreghi [ 16 ] and by Li and Tsing [ 48 ] exhibit plenty of interesting preserver problems on algebraic structures and well-developed techniques thereon. Another important direction of preserver problems is formulated in the normed space settings by using norm-metrics. The simplest (but often highly nontrivial) one is a problem of determining the form of isometric isomorphisms between certain normed spaces. For example, the Banach-Stone theorem states that if T\colon C_{0}(K)\to C_{0}(L) is an isometric isomorphism, then there exists a unimodular continuous function u on L and a homeomorphism \varphi\colon K\to L such that Tf=u(f\circ\varphi^{-1}) for each f\in C_{0}(K) . In particular, the mapping f\mapsto f\circ\varphi^{-1} preserves multiplication (and complex conjugation in the complex case). This means that the Banach space structure of C_{0}(K) controls its whole structure as a Banach (or a C^{*} -) algebra. Moreover, the Banach-Stone theorem has the following generalization: Let \mathfrak{A} and \mathfrak{B} be C^{*} -algebras, and let T\colon\mathfrak{A}\to\mathfrak{B} be an isometric isomorphism. Then there exists a unitary element u\in\mathcal{M}(\mathfrak{B}) and a Jordan * -isomorphism J\colon\mathfrak{A}\to\mathfrak{B} such that Tx=uJx for each x\in\mathfrak{A} , where \mathcal{M}(\mathfrak{B}) is the multiplier algebra of \mathfrak{B} ; see Kadison [ 35 ] for the unital case, and Paterson and Sinclair [ 55 ] for the non-unital case. Therefore, the Banach space structure of a C^{*} -algebra can still control its Jordan * -algebraic structure even if the algebra is non-commutative. Meanwhile, it is well-known as the Mazur-Ulam theorem that every (possibly nonlinear) surjective isometry between normed spaces is affine over the real numbers. Namely, as an affine mapping is a translation of a real-linear mapping, normed spaces are isometrically isomorphic as metric spaces if and only if they are isometrically isomorphic as real normed spaces. Hence, the (nonlinear) metric structure of a normed space solely determines its structure as a real normed space. The combination of the Kadison-Paterson-Sinclair theorem and the Mazur-Ulam theorem suggests a classification of C^{*} -algebras based on substructures of the metric structure. Potential candidates are geometric structures determined by the shape of the unit ball of a normed space. The present paper focuses on the orthogonal structure with respect to (a strengthened version of) the Birkhoff-James (shortly, BJ) orthogonality