A Projection Characterization and Symmetry Bootstrap for Elements of a von Neumann Algebra that are Nearby Commuting Elements

David Herrera

REVIEW 2 major objections 3 minor 31 references

A Projection Characterization and Symmetry Bootstrap for Elements of a von Neumann Algebra that are Nearby Commuting Elements

T0 review · 2 major / 3 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Symmetry maps can be required of nearby commuting approximants, proving Lin's theorem with reflection, rotational, and dihedral symmetries.

desk verdict Solid new bootstrap for order-2 symmetries, but the rotational/dihedral Lin theorem is only proved for invertible elements and the abstract oversells the non-invertible extension. read the letter →

arxiv 2412.20795 v1 pith:VVJRQQCF submitted 2024-12-30 math.OA math.FA

David Herrera This is my paper · ORCID

classification math.OAmath.FA MSC 46L1046L5515A27

keywords Lin'stheoremsymmetrybootstrapalmostcommutingmatricesvonNeumannalgebrasprojectioncharacterizationphasesymmetriesAltland-Zirnbauerclasses

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes that symmetries of an almost normal operator—transpose, conjugation, reflection, rotation, and dihedral phase symmetries—can be forced onto the nearby normal operator whose existence is asserted by Lin's theorem. The core tool is a projection characterization: from a pair of operators that are already close to commuting, the paper extracts spectral projections of the nearby commuting pair, perturbs them into projections subordinate to the original operator, and uses those projections to rebuild a commuting pair that carries the original symmetries. For finite matrices this yields symmetry-preserving versions of Lin's theorem, including reflection and finite rotational/dihedral phase symmetries for invertible elements, and resolves a conjecture about two almost commuting self-adjoint matrices in the Altland–Zirnbauer symmetry classes. The paper also proves bootstrap symmetry results for two and three almost commuting self-adjoint operators.

What carries the argument

The projection characterization: if $U,B$ are close to commuting $U',B'$, then for arcs $\Omega_0 \subset \Omega$ in the spectrum circle there is a projection $F$ satisfying $E_{\Omega_0}(U)\le F\le E_\Omega(U)$, decomposable as $F_- + E_{\Omega_0}(U) + F_+$ with $F_\pm$ subordinate to the complementary subarcs, and with $[F,B]$ small. These projections are assembled into a spectral resolution defining $U''$ and a localization/pinching operation $B\mapsto B''$ that commutes with the symmetries. Symmetry maps—$\mathbb{R}$-linear hermitian maps that are (anti-)multiplicative and linear or conjugate-linear—are the formal language in which 'transpose', 'conjugation', 'reflection', and 'phase rotation' symmetries are expressed.

What would settle it

Take a finite-dimensional matrix $A_0$ with a phase symmetry of order $q>2$ for which $q$ does not divide the matrix dimension $n$; any phase-symmetric matrix in that class has zero as a forced eigenvalue, so $A_0$ cannot be approximated by invertible phase-symmetric matrices with controlled self-commutator. Exhibiting such an $A_0$ with arbitrarily small $\|[A_0^*,A_0]\|$ but no nearby invertible symmetric approximant would disprove the non-invertible version of Theorem 8.2.

Watch

Extended reading notes

Core claim

The central claim is that the existence of nearby commuting approximants $U'$, $B'$ (with $U'$ unitary or self-adjoint) is characterized by the existence of a family of projections $F$ that sit between spectral projections of $U$ and almost commute with $B$; this characterization allows the approximants to be rebuilt as $U''$, $B''$ that inherit any symmetries of $U$ and $B$, provided the symmetries are generated by weakly continuous linear, conjugate-linear, multiplicative or anti-multiplicative maps satisfying mild commuting conditions. As a corollary, the paper proves Lin's theorem with symmetries: an $S$-symmetric almost normal contraction is within distance $O(\sqrt{\epsilon_\ell(\|[A^*,A]\|)})$ of a normal $S$-symmetric element, and with a linear antisymmetry or conjugate-linear symmetry of order two. For a unitary operator with reflection, rotational, or dihedral phase symmetries, the same bootstrap yields nearby commuting unitary and self-adjoint operators preserving those phase symmetries. For invertible elements, the rotational and dihedral versions extend to Lin's theorem directly through a symmetry-respecting polar decomposition lemma.

Load-bearing premise

The advertised rotational and dihedral Lin theorem is proved only for invertible elements; extending it to arbitrary non-invertible elements requires that those elements be approximable by invertible elements with the same phase symmetries and controlled self-commutator, a condition that can fail in finite dimensions (for example, when the order of the phase does not divide the matrix dimension).

Editorial extensions

If this is right

For matrices, any almost normal matrix with a transpose, conjugation, or reflection-type symmetry has a nearby normal matrix with the same symmetry (Lin's theorem with symmetries).
For two almost commuting self-adjoint matrices, the bootstrap preserves arbitrary collections of commuting unitary/anti-unitary symmetries, including all ten Altland–Zirnbauer classes in 1D, resolving Conjecture 2.2 of [21].
For three almost commuting self-adjoint matrices that are already near commuting ones, certain symmetry patterns (one or two antisymmetries) are automatic for the nearby commuting triple, without needing an extra index to vanish.
For unitaries with rotational phase symmetry of finite order $n$ or dihedral symmetry, the error bound must depend on $n$; the paper exhibits a scaling-invariant example showing this dependence is necessary.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The projection/rebuilding method is likely to transfer to other approximation problems where the spectrum lies on a one-dimensional set, yielding symmetry-preserving approximants for almost commuting pairs of unitaries or for normal operators with more general spectral curves.
For non-invertible elements the rotational/dihedral Lin theorem is not proven; if the required invertible approximation fails (e.g., phase order not dividing matrix dimension), the non-invertible statement would be false as stated.
The bootstrap may give a quantitative route to index-theoretic obstructions, since the constructed projections encode the same almost-invariant subspace data used in counterexamples like Voiculescu's unitaries.
The automatic symmetry results for three almost commuting matrices suggest analogous automatic-symmetry phenomena for larger families when the only obstacle is the lack of commutants.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: the self-adjoint symmetry bootstrap is a real new tool, and the AZ two-matrix conjecture is settled. But the advertised rotational/dihedral Lin theorem is only proved for invertible elements; the extension to non-invertibles rests on an approximation condition that fails, for example, for order-3 phase on C^2. The abstract overstates this.

What's actually new: the projection characterization (Lemmas 6.1/6.4, 6.8/6.9) turns near commutation with a unitary/self-adjoint into existence of almost reducing, symmetry-respecting projections; then the bootstrap uses a symmetry-blind commuting pair from external Lin-type theorems to construct a symmetric pair. The self-adjoint version has explicit constants; Theorem 7.3 gives a 90 sqrt(eps) bound, and Theorem 1.9 resolves Conjecture 2.2 of [21] for all ten AZ classes for two self-adjoint matrices. That's a genuine advance, and the proofs are worked out in enough detail that the argument appears sound. The citation pattern looks fair; the reductions in Section 3 are useful.

The soft spots are where the reader said they are. Theorem 7.6 leaves 'Const.' and the smallness threshold on ||[U,B]|| completely unspecified, so the unitary bootstrap is a qualitative statement until someone fills in the arithmetic. More serious, Theorem 8.2 is proved for invertible A, and the final sentence extends to non-invertible A0 only under an approximation hypothesis. The stress-test example is correct: for W = diag(1, e^{2πi/3}) on C^2, the equation W* A W = e^{2πi/3} A forces A = [[0,a],[0,0]], so all such matrices are nilpotent; there are no invertible approximants with that phase. So the non-invertible rotational/dihedral case is not proven, and the abstract's claim needs scaling back. The paper itself states the condition, so I don't call it dishonest — but the advertised theorem is narrower than the abstract suggests.

The order-2 symmetry results and the AZ corollary don't touch this gap; they stand. If I were refereeing, I would ask for the non-invertible gap to be either closed or explicitly stated as open, and for the constants in Theorem 7.6 to be made explicit or the theorem flagged as a qualitative statement. But the core bootstrap and the AZ resolution are solid enough to deserve full referee time.

Recommendation: send to review, likely after minor-to-moderate revision. The paper is for people working on Lin-type theorems, almost commuting matrices with symmetries, and topological insulator classification; they'll want this.

Referee Report

2 major / 3 minor

Summary. The paper introduces a general framework of symmetry maps on unital C*-algebras and uses a projection characterization of pairs (U,B) that are near commuting pairs to "bootstrap" symmetries from an initial nearby commuting pair to a final commuting pair. The main applications are a self-adjoint symmetry bootstrap with explicit constants (Theorem 7.1), Lin's theorem with linear symmetries and with an order-2 conjugate-linear symmetry or linear antisymmetry (Theorems 4.5, 7.3, 7.4), corollaries for two and three almost commuting self-adjoint matrices, a resolution of Conjecture 2.2 of [21] for the Altland-Zirnbauer classes, and a rotational/dihedral version of Lin's theorem in Section 8.

Significance. If the central results stand, the paper gives a genuinely new and useful method: the projection characterization converts an initial nearby commuting pair into symmetry-preserving commuting approximants, and the self-adjoint case is worked out with explicit constants. The proof is not circular: it uses external Lin-type theorems to obtain the initial commuting pair and then constructs the symmetric pair from projections. The resolution of the two-matrix conjecture from [21] is a substantial achievement. However, the Section 8 rotational/dihedral Lin theorem is proved only for invertible elements, and the advertised non-invertible extension rests on an unproved and sometimes false approximation assumption, so that part of the abstract is not supported as stated.

major comments (2)

[Theorem 8.2 and final paragraph; cf. Section 3] The advertised rotational/dihedral Lin theorem is proved only for invertible A. The last sentence of Theorem 8.2 extends it to a non-invertible A0 only under the extra assumption that A0 can be approximated by invertible elements A satisfying the same hypotheses. This approximation is not established, and the paper itself notes the obstruction: Section 3, in the paragraph after Proposition 2.14, explains that for a phase symmetry of order n an invertible approximant may fail to exist in finite dimensions when the order of the phase does not divide the matrix dimension. Concretely, for W = diag(1, e^{2πi/3}) on C^2, the condition W*AW = e^{2πi/3}A forces A to have the form [[0,a],[0,0]], so every such A is nilpotent and no nonzero A0 has an invertible approximant with this phase symmetry. Consequently the abstract's claim of a rotational/dihedral Lin theorem for almost normal matrices is not supported for non-invertible contractions; the paper proves an invertible-element statement plus a conditional assertion. This is load-bearing because Theorem 8.2 is one of the paper's advertised main results.
[Theorem 7.6 and its use in Theorem 8.2] Theorem 7.6 is stated as "for ||[U,B]|| small enough" and with an unspecified universal constant "Const.", while cases (ii) and (iii) say that the constant and the smallness condition depend on n without quantifying either. The proof chooses L = sqrt(Const * eps) and requires L to lie below an L0, but L0 is never computed or bounded, so the theorem does not provide a threshold or a rate. Since Theorem 8.2 invokes Theorem 7.6 to obtain its "Const." bound, the quantitative claim in the rotational/dihedral theorem inherits this unquantified status. The statement also calls Const. universal while later making it depend on n, which should be reconciled. If the intended statement is merely an existence result with some rate, it should be reformulated as such; as written, it gives the appearance of a quantitative theorem without the required data.

minor comments (3)

[Abstract, Introduction, Section 8] "Atland-Zirnbauer" should be "Altland-Zirnbauer" throughout.
[Lemma 6.4, condition 2] The displayed condition "F = F- + E_{Omega_0}(U), + F+" contains a stray comma; it should read "F = F- + E_{Omega_0}(U) + F+".
[Theorem 7.6] The phrase describing Const. as universal should be reconciled with the later statement that the constant depends on n in cases (ii) and (iii), since both appear in the same theorem statement.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the symmetry bootstrap converts external Lin-type approximations into symmetric ones via projection constructions, not by assuming the conclusion.

full rationale

The derivation chain is self-contained in the relevant sense. Section 4 reduces Lin's theorem with linear symmetries to Friis-Rordam (Theorem 4.1) and Loring-Sorensen (Theorem 4.4); the author's role is a reduction argument via Proposition 3.7, not a restatement of the conclusion. The projection characterization (Lemmas 6.1/6.8 and 6.4/6.9) has two directions: given nearby commuting U',B' it constructs spectral-substitute projections with explicit commutator bounds, and conversely given such projections it constructs new commuting U'',B'' by pinching a localized B. The bootstrap theorem (7.1/7.6) applies the first direction to the external Lin approximation and the second to obtain symmetric approximants; the distance estimate O(sqrt(epsilon)) arises from optimizing L, not from assuming the symmetric approximant exists. The main new results (Theorems 7.3, 7.4, 8.2) invoke external theorems (Friis-Rordam, Loring-Sorensen, Kachkovskiy-Safarov, Hastings-Loring) for the unsymmetrized input, and the symmetry preservation is achieved by explicit equivariant projection/localization constructions. Self-citations [16] and [17] appear only in examples and context, e.g., Example 2.5 and Remark 1.12, and are not load-bearing for the central theorems. The conditional final sentence of Theorem 8.2 regarding non-invertible A0 is an unproved approximation condition and a potential correctness gap, but it is not a circular reduction: the theorem explicitly states the extra hypothesis rather than assuming the conclusion. Therefore no circular step is exhibited.

Assumptions & free parameters 0 free parameters · 9 assumptions · 0 invented entities

No free parameters are fitted: the constants CKS, c, and Cρ come from cited theorems or from an explicit numerical bound in Lemma 5.5. The axioms are standard operator-algebraic theorems and explicitly stated domain assumptions. The paper introduces the formal notion of symmetry maps, but these are definitions, not independent postulated entities.

assumptions (9)

standard math Davis-Kahan and Bhatia-Davis-McIntosh spectral projection perturbation estimates (Theorem 5.1).
Used in Lemmas 6.1 and 6.8 to convert nearness of U',B' to U,B into norm bounds on products of spectral projections.
standard math Friis-Rørdam Lin theorem for unital C*-algebras with stable rank 1 (Theorem 4.1).
Base case for Theorem 4.5 with only linear multiplicative symmetries.
standard math Kachkovskiy-Safarov distance to normal elements with explicit CKS bound (Theorem 4.3).
Used in Corollary 1.11 and Theorem 4.7 for stable rank 1 von Neumann algebras.
standard math Loring-Sørensen Lin theorem with a linear anti-multiplicative order-2 symmetry under TR rank 1 (Theorem 4.4).
Base case for Theorem 4.5 with linear anti-multiplicative symmetries.
standard math Hastings-Loring constructions from Theorems 2.3 and 2.4 of [15].
Provides the initial commuting U',P' approximants with S-symmetries in the proof of Theorem 8.2.
standard math Square-root and commutator estimates of Ando [2] and Pedersen [26].
Used in Lemma 8.1 to control the polar decomposition of an almost normal invertible element.
domain assumption TR rank 1 (or stable rank 1) hypothesis on (A,S).
Theorem 4.5 and the bootstrap need invertible S-symmetric elements to be dense; for finite-dimensional algebras this is proved in Theorem 3.4, for general von Neumann algebras it is assumed.
domain assumption Weak continuity and admissible commutation of symmetry maps.
Spectral projections and localization operators respect symmetries only under weak continuity and the commuting conditions in Definition 3.6; used throughout Sections 6 and 7.
domain assumption Spectrum of the reference normal operator lies on a one-dimensional set (unit circle or real line).
The arc/interval decomposition and the unentangled cut in Remark 6.3 require an ordered 1D spectrum; stated in Section 1.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Projection Characterization and Symmetry Bootstrap for Elements of a von Neumann Algebra that are Nearby Commuting Elements." pith.science (2026). https://pith.science/paper/VVJRQQCF

@misc{pith2026241220795,
  author       = {Pith},
  title        = {Pith review of: A Projection Characterization and Symmetry Bootstrap for Elements of a von Neumann Algebra that are Nearby Commuting Elements},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VVJRQQCF}},
  note         = {Machine review of arXiv:2412.20795}
}

abstract

We define a symmetry map $\varphi$ on a unital $C^\ast$-algebra $\mathcal A$ to be an $\mathbb{R}$-linear map on $\mathcal A$ that generalizes transformations on matrices like: transpose, adjoint, complex-conjugation, conjugation by a unitary matrix, and their compositions. We include an overview of such symmetry maps on unital $C^\ast$-algebras. We say that $A\in\mathcal A$ is $\varphi$-symmetric if $\varphi(A)=A$, $A$ is $\varphi$-antisymmetric if $\varphi(A)=-A$, and $A$ has a $\zeta=e^{i\theta}$ $\varphi$-phase symmetry if $\varphi(A)=\zeta A$. Our main result is a new projection characterization of two operators $U$ (unitary), $B$ that have nearby commuting operators $U'$ (unitary), $B'$. This can be used to ``bootstrap'' symmetry from operators $U, B$ that are nearby some commuting operators $U', B'$ to prove the existence of nearby commuting operators $U'', B''$ which satisfy the same symmetries/antisymmetries/phase symmetries as $U, B$, provided that the symmetry maps and symmetries/antisymmetries/phase symmetries satisfy some mild conditions. We also prove a version of this for $X=U$ self-adjoint instead of unitary. As a consequence of the prior literature and the results of this paper, we prove Lin's theorem with symmetries: If a $\varphi$-symmetric matrix $A$ is almost normal ($\|[A^\ast, A]\|$ is small), then it is nearby a $\varphi$-symmetric normal matrix $A'$. We also extend this further to include rotational and dihedral symmetries. We also obtain bootstrap symmetry results for two and three almost commuting self-adjoint operators. As a corollary, we resolve a conjecture of arXiv:1502.03498 for two almost commuting self-adjoint matrices in the Atland-Zirnbauer symmetry classes related to topological insulators.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages

[21]

, K-theory and pseudospectra for topological insulators , Annals of Physics 356 (2015), 383–416

work page 2015
[1]

A. B. Aleksandrov and V. V. Peller, Operator lipschitz functions , Russian Mathematical Surveys 71 (2016), no. 4, 605

work page 2016
[2]

Tsuyoshi Ando, Comparison of norms |||f (A)− f (B)||| and|||f (|A− B|)|||, Mathematische Zeitschrift 197 (1988), 403–409

work page 1988
[3]

, Michigan Mathematical Journal 21 (1975), no

ID Berg, On approximation of normal operators by weighted shifts. , Michigan Mathematical Journal 21 (1975), no. 4, 377–383

work page 1975
[4]

Rajendra Bhatia, Chandler Davis, and Alan McIntosh, Perturbation of spec- tral subspaces and solution of linear operator equations , Linear algebra and its applications 52 (1983), 45–67

work page 1983
[5]

Rajendra Bhatia and Peter Rosenthal, How and why to solve the operator equation AX− XB = Y , Bulletin of the London Mathematical Society 29 (1997), no. 1, 1–21

work page 1997
[6]

symmetry groups

Ola Bratteli and Derek William Robinson, Operator algebras and quantum statistical mechanics: Volume 1: C ∗-and W ∗-algebras. symmetry groups. de- composition of states , Springer Science & Business Media, 1979

work page 1979
[7]

2, 222–240

Kenneth R Davidson, Almost commuting Hermitian matrices , Mathematica Scandinavica 56 (1985), no. 2, 222–240

work page 1985

Show all 31 references

[8]

iii , SIAM Journal on Numerical Analysis 7 (1970), no

Chandler Davis and William Morton Kahan, The rotation of eigenvectors by a perturbation. iii , SIAM Journal on Numerical Analysis 7 (1970), no. 1, 1–46

1970
[9]

4, 913–915

Ruy Exel and Terry A Loring, Almost commuting unitary matrices , Pro- ceedings of the American Mathematical Society 106 (1989), no. 4, 913–915

1989
[10]

10, 2902–2932

N Filonov and Yu Safarov, On the relation between an operator and its self- commutator, Journal of Functional Analysis 260 (2011), no. 10, 2902–2932

2011
[11]

Peter Friis and Mikael Rørdam, Almost commuting self-adjoint matrices-a short proof of Huaxin Lin ’s theorem. (1996)

1996
[12]

Brian C Hall, Lie groups, Lie algebras, and representations , Springer, 2015

2015
[13]

PR Halmos, Quasitriangular operators, Acta Sci. Math. (Szeged) 29 (1968), 283––293

1968
[14]

2, 321–345

Matthew B Hastings, Making almost commuting matrices commute , Com- munications in Mathematical Physics 291 (2009), no. 2, 321–345

2009
[15]

Matthew B Hastings and Terry A Loring, Almost commuting matrices, lo- calized Wannier functions, and the quantum Hall eﬀect , Journal of mathe- matical physics 51 (2010), no. 1. 63

2010
[16]

David Herrera, On Hastings’ approach to Lin ’s theorem for almost commut- ing matrices , arXiv preprint arXiv:2011.11800 (2020)

2020 arXiv
[17]

, Constructing nearby commuting matrices for reducible repr esen- tations of su(2) with an application to Ogata’s theorem , arXiv preprint arXiv:2212.06012 (2022)

2022 arXiv
[18]

1, 61–80

Ilya Kachkovskiy and Yuri Safarov, Distance to normal elements in C ∗- algebras of real rank zero , Journal of the American Mathematical Society 29 (2016), no. 1, 61–80

2016
[19]

Huaxin Lin, Almost commuting selfadjoint matrices and applications , Oper- ator algebras and their applications (1996), 193–233

1996
[20]

Symmetry, Integrability and Geometry: Methods and Applic ations 10 (2014), 077

Terry A Loring, Quantitative K-theory related to spin Chern numbers , SIGMA. Symmetry, Integrability and Geometry: Methods and Applic ations 10 (2014), 077

2014
[22]

Terry A Loring and Adam PW Sørensen, Almost commuting unitary matri- ces related to time reversal , Communications in Mathematical Physics 323 (2013), 859–887

2013
[23]

2, 1051–1068

, Almost commuting orthogonal matrices , Journal of Mathematical Analysis and Applications 420 (2014), no. 2, 1051–1068

2014
[24]

07, 1650017

, Almost commuting self-adjoint matrices: the real and self- dual cases, Reviews in Mathematical Physics 28 (2016), no. 07, 1650017

2016
[25]

9, 2005–2033

Yoshiko Ogata, Approximating macroscopic observables in quantum spin sys - tems with commuting matrices , Journal of Functional Analysis 264 (2013), no. 9, 2005–2033

2013
[26]

Gert K Pedersen, A commutator inequality, Operator algebras, mathematical physics, and low dimensional topology, 1993, pp. 233–235

1993
[27]

Peter Bech Møller Sørensen, On almost commuting operators and matrices , Master’s Thesis, 2012

2012
[28]

Erling Størmer, Positive linear maps of operator algebras , Springer, 2013

2013
[29]

Karen R Strung, An introduction to C ∗-algebras and the classiﬁcation pro- gram, Springer, 2021

2021
[30]

Dan Voiculescu, Remarks on the singular extension in the C ∗-algebra of the Heisenberg group, Journal of Operator Theory (1981), 147–170

1981
[31]

Math.(Szeged) 45 (1983), no

, Asymptotically commuting ﬁnite rank unitary operators wit hout commuting approximants, Acta Sci. Math.(Szeged) 45 (1983), no. 1-4, 429– 431. 64

1983

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [21]

, K-theory and pseudospectra for topological insulators , Annals of Physics 356 (2015), 383–416

work page 2015

[2] [1]

A. B. Aleksandrov and V. V. Peller, Operator lipschitz functions , Russian Mathematical Surveys 71 (2016), no. 4, 605

work page 2016

[3] [2]

Tsuyoshi Ando, Comparison of norms |||f (A)− f (B)||| and|||f (|A− B|)|||, Mathematische Zeitschrift 197 (1988), 403–409

work page 1988

[4] [3]

, Michigan Mathematical Journal 21 (1975), no

ID Berg, On approximation of normal operators by weighted shifts. , Michigan Mathematical Journal 21 (1975), no. 4, 377–383

work page 1975

[5] [4]

Rajendra Bhatia, Chandler Davis, and Alan McIntosh, Perturbation of spec- tral subspaces and solution of linear operator equations , Linear algebra and its applications 52 (1983), 45–67

work page 1983

[6] [5]

Rajendra Bhatia and Peter Rosenthal, How and why to solve the operator equation AX− XB = Y , Bulletin of the London Mathematical Society 29 (1997), no. 1, 1–21

work page 1997

[7] [6]

symmetry groups

Ola Bratteli and Derek William Robinson, Operator algebras and quantum statistical mechanics: Volume 1: C ∗-and W ∗-algebras. symmetry groups. de- composition of states , Springer Science & Business Media, 1979

work page 1979

[8] [7]

2, 222–240

Kenneth R Davidson, Almost commuting Hermitian matrices , Mathematica Scandinavica 56 (1985), no. 2, 222–240

work page 1985