Comparison of Levi-Civita connections in noncommutative geometry

Adam Rennie; Alexander Flamant; Bram Mesland

REVIEW 3 major objections 4 minor 23 references

Comparison of Levi-Civita connections in noncommutative geometry

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Centred noncommutative calculi admit one Levi-Civita connection

desk verdict A valuable translation between the three Levi-Civita frameworks, with a correctable but load-bearing typo in Lemma A.9 that must be fixed before publication. read the letter →

arxiv 2505.18692 v1 pith:ACPZFJXS submitted 2025-05-24 math.QA math-phmath.MPmath.OA

Alexander Flamant , Bram Mesland , Adam Rennie This is my paper · ORCID

classification math.QAmath-phmath.MPmath.OA MSC 58B3453C0546L87

keywords noncommutativegeometryLevi-CivitaconnectioncentredbimoduleHermitiantorsion-freecanonicalbraidingpseudo-Riemanniancalculusdifferentialformsquantummetric

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 three different constructions of Levi-Civita connections in noncommutative geometry—one on vector fields, one on differential one-forms via spectral data, and one on one-forms via centred bimodules—describe the same object whenever their hypotheses overlap. For any centred Hermitian differential calculus, meaning a module of one-forms generated by its centre, with a strongly non-degenerate inner product invariant under the canonical braiding, there is a unique Hermitian torsion-free $\sigma$-$\dagger$-bimodule connection, and the paper writes it down explicitly. This explicitly constructed connection dualises to a pseudo-Riemannian calculus on noncommutative vector fields, and its curvature tensor agrees with the vector-field curvature. The result therefore gives a common language and a purely algebraic existence theorem that unifies earlier existence results.

What carries the argument

The load-bearing object is the centred Hermitian differential calculus: a first-order differential structure $(\Omega_d^1(A), \dagger)$ whose one-form bimodule is generated by its centre, carrying a strongly non-degenerate Hermitian inner product whose bilinear quantum metric $g(\omega\otimes\eta)=-\langle\omega^\dagger\mid\eta\rangle$ is invariant under the canonical braiding $\sigma_{\mathrm{can}}$. The canonical braiding, which flips central one-forms, supplies the projection $\Psi=\frac{1}{2}(1+\sigma_{\mathrm{can}})$ and therefore the exterior derivative, torsion, and curvature in the form framework. The existence proof replaces analytic frames by purely algebraic generating pairs $\{(\omega_i,\eta_i)\}$ of central elements satisfying $\omega=\sum_i \omega_i\langle\eta_i\mid\omega\rangle=\sum_i \eta_i\langle\omega_i\mid\omega\rangle$, which exist under strong non-degeneracy; the correction term $\alpha((1+4PQ)W)$ is the musical-isomorphism manipulation that removes the torsion left by the Hermitian ansatz.

What would settle it

Compute formula (4.1) on a fixed centred Hermitian calculus using two different central dagger-invariant generating pairs, for instance on the noncommutative two-torus with its standard metric; if the two resulting connections differ, the explicit formula is not well-defined. Alternatively, exhibit a finitely generated projective one-form module with strongly non-degenerate inner product that is not centred and for which no Hermitian torsion-free sigma-dagger-bimodule connection exists, which would confirm the paper's own caveat that the algebraic existence proof does not extend beyond centred bimodules.

Watch

Extended reading notes

Core claim

The central claim is Theorem 4.24: on a centred Hermitian calculus $(\Omega_d^1(A), \dagger, \langle\cdot\mid\cdot\rangle, \sigma_{\mathrm{can}})$, where the one-form bimodule is generated by central elements, the inner product is strongly non-degenerate, and the quantum metric is invariant under the canonical braiding, there exists exactly one Hermitian torsion-free $\sigma$-$\dagger$-bimodule connection $\nabla_G$ on one-forms. For any finite central $\dagger$-invariant generating pair $\{(\omega_i,\eta_i)\}$, it is given by $\nabla_G(\omega)=\frac{1}{2}\big(\sum_i \omega_i\otimes d\langle\eta_i\mid\omega\rangle + \eta_i\otimes d\langle\omega_i\mid\omega\rangle\big)-\alpha((1+4PQ)W)$. The same connection induces right and left pseudo-Riemannian calculi on the dual modules of vector fields (Corollary 4.25), and its curvature agrees with the curvature of the induced affine connection (Theorem 4.26).

Load-bearing premise

The one-form bimodule must be centred, meaning every one-form is a combination of elements that commute with the algebra, because the canonical braiding, the coincidence of the two musical isomorphisms, and the algebraic existence of the connection all depend on that property.

Editorial extensions

If this is right

Every centred Hermitian calculus with a strongly non-degenerate metric invariant under the canonical braiding has a unique Hermitian torsion-free connection on one-forms, explicitly computable from any central $\dagger$-invariant generating pair.
The corresponding dual vector-field modules carry right and left real pseudo-Riemannian calculi with strongly non-degenerate inner products, so curvature tensors can be studied in either formalism and agree.
The analytic locality hypothesis used in earlier differential-form constructions can be dropped in the centred case: strong non-degeneracy alone suffices for existence, replacing the need for pre-$C^*$-inner products and frames.
Isospectral deformations coming from free torus actions, including the noncommutative two-torus and three-sphere, admit these connections, recovering earlier examples as a special case.

Reading between the lines

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

Beyond the paper: the explicit generating-pair formula should make the connection computable by linear algebra on any finite-dimensional centred calculus, for example matrix algebras with a differential calculus generated by central one-forms, without first solving the metric-compatibility and torsion equations.
Beyond the paper: the paper's caveat that the algebraic existence proof does not extend to non-centred bimodules suggests that non-centred calculi are where genuinely new behaviour can be sought; a systematic search for non-centred strongly non-degenerate calculi with no or with multiple Hermitian torsion-free connections would map the true boundary of the theorem.
Beyond the paper: since the curvature comparison is proven only when two of the three vector-field slots are bilinear, a natural next step is to test whether the third slot can be freed, which would let the full Riemann tensor and scalar curvature of a pseudo-Riemannian calculus be computed entirely from the one-form formula.

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

Bottom line: this is a real contribution to noncommutative differential geometry, but there is a concrete flaw in Lemma A.9 that the authors need to fix before the paper goes out. The displayed generating set in the proof of Lemma A.9 does not satisfy the stated reconstruction identity. From the proof, averaging gives x = 1/2(Σ x_i⟨x'_i|x⟩ + x_i†⟨x'_i†|x⟩). To package this as Σ u⟨v|x⟩ over †-invariant pairs, the correct pairs are (x_i/√2, x'_i/√2) and (x_i†/√2, x'_i†/√2). The display instead lists (x_i/√2, x_i†/√2) and (x'_i/√2, x'_i†/√2), which does not reconstruct x. Since Proposition 4.23 and Theorem 4.24 import this lemma verbatim, the error is load-bearing, not cosmetic. The good news is the correction is immediate and the surrounding arguments — the σ-braiding computations, the (1+4PQ)W correction term, the torsion calculation — all appear to go through unchanged with the repaired pairs.

What the paper does well: it gives a clean common language for the Arnlind–Wilson derivation-based approach, the BGM20 spectral-triple approach, and the MR24a differential-forms approach. The translation theorems (3.21, 3.24, 4.16, 4.20) are useful and carefully stated. Theorem 4.24 genuinely extends existence to all centred Hermitian calculi with strongly non-degenerate inner product, replacing the analytic frame assumption of MR24a with the algebraic generating-pair technique of Appendix A. The explicit Levi-Civita formula and the curvature comparison (4.26) are valuable. The paper is also honest about the scope: centredness is essential, and the purely algebraic existence argument does not cover non-centred bimodules.

Minor soft spots: a few proofs say "identical to Theorem 4.16" without showing details, and the step in Lemma A.3 / Definition A.1 where functionals are written as ⟨x'|·⟩ rather than ⟨x'†|·⟩ can confuse readers. Neither is serious. The stress-test note's secondary concern about the notational mismatch is reconcilable — the g∘σ = g hypothesis justifies the relevant equality.

Who this is for: specialists in noncommutative geometry who want to compare or import results across the three existing frameworks. It deserves a serious referee, but I would require the Lemma A.9 correction before recommending acceptance.

Referee Report

3 major / 4 minor

Summary. The paper compares three frameworks for Levi-Civita connections in noncommutative geometry: the derivation-based approach of Arnlind–Wilson, the differential-form approach of Bhowmick–Goswami–Mukhopadhyay, and the recent Hermitian framework of Mesland–Rennie. It introduces the notion of a centred Hermitian differential calculus, proves that such calculi induce real metric calculi on dual vector-field modules, and states an existence and uniqueness theorem (Theorem 4.24) with an explicit formula for a Hermitian torsion-free σ-†-bimodule connection, together with a curvature comparison (Theorem 4.26). The algebraic replacement of analytic frames by generating pairs for strongly non-degenerate inner products is developed in Appendix A.

Significance. If Theorem 4.24 is established, the paper makes a substantive contribution: it replaces the analytic locality and frame assumptions in [MR24a] by the algebraic assumption of strong non-degeneracy for centred bimodules, extends the existence result of [BGM20] to arbitrary centred Hermitian calculi, and provides a usable dictionary between connections on differential forms and affine connections on vector fields. The curvature comparison of Theorem 4.26 is also valuable. However, the central existence argument rests on Lemma A.9, whose proof is incorrect as written. Until that lemma is repaired, the main theorem is not supported.

major comments (3)

[Appendix A, Lemma A.9] The displayed †-invariant set is not the set that encodes the averaged identity. From x = 1/2Σ_i(x_i⟨x′_i|x⟩ + x_i†⟨x′†_i|x⟩), packaging this as Σ u⟨v|x⟩ requires the pairs (x_i/√2, x′_i/√2) and (x†_i/√2, x′†_i/√2). The display instead lists (x_i/√2, x†_i/√2) and (x′_i/√2, x′†_i/√2), whose left and right entries do not match the averaged summands. As written, the displayed set does not satisfy the reconstruction identity x = Σ u⟨v|x⟩. Since Proposition 4.23 and Theorem 4.24 import exactly this lemma, the standing hypothesis of the main existence theorem is not proved.
[Appendix A, Lemma A.9] Even after correcting the pairs as described above, the proof only establishes the first reconstruction identity for the new set. The required second identity x = Σ v⟨u|x⟩ would need a proof of x = Σ_i x′†_i⟨x†_i|x⟩, and no such argument is given. In addition, the proof asserts without justification that a centred module yields central sequences {(x_i, x′_i)} satisfying Lemma A.3; Lemma A.3 only produces x′_i as representatives of right-module functionals and does not imply their centrality. Both gaps are load-bearing for Proposition 4.23 and Theorem 4.24.
[Appendix A, Lemma A.3 and Definition A.1] There is a notational mismatch in the strong non-degeneracy condition. Definition A.1 condition 4 represents right-module functionals as ⟨x†|y⟩, while Lemma A.3 writes s_i(x) = ⟨x′_i|x⟩. The mismatch is not purely cosmetic because the dagger computation in Lemma A.9 is applied to the x′_i. The authors should fix a convention (for instance, x′_i chosen so that s_i(x) = ⟨x′†_i|x⟩) and then recompute the identities in Lemma A.9.

minor comments (4)

[Section 4.6, before Theorem 4.26] The text says 'For bilinear vector fields Z1, Z2 ∈ D', but D consists of derivations; this should read Z1, Z2 ∈ φ(D).
[Theorem 4.27] The claim that the constructed connection restricts to the classical Levi-Civita connection on Ω^1(M) is stated without proof; a reference or a short argument should be supplied.
[Lemma A.3] The map π: A^n → X is described as a 'surjective bimodule map', but the construction only guarantees right A-linearity when X is not assumed centred; replace 'bimodule' by 'right module' or add a centrality hypothesis.
[Introduction] There is a typo: 'machinary' should read 'machinery'.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular reduction of the central theorem: Theorem 4.24 is an explicit algebraic construction, and its uniqueness proof rests on [AW17], not on a self-citation chain. The score reflects only the paper's minor, non-load-bearing self-citations to [MR24a]. A separate correctness gap in Lemma A.9's displayed generating set is flagged but is not circularity.

full rationale

I find no step in the paper where a claimed prediction is equivalent to its input by construction. Theorem 4.24 builds ∇_G explicitly from a †-invariant central generating pair, then adds the correction term α((1+4PQ)W) and verifies Hermiticity, torsion-freeness and the σ-†-bimodule condition by direct computation using only the centred-Hermitian axioms, the canonical braiding, and the strong-nondegeneracy isomorphisms of Appendix A. Uniqueness follows from Corollary 4.21, whose proof uses the external uniqueness result [AW17, Theorem 3.4]; the alternative appeal to [MR24a, Theorem 5.14] in Remark 4.22 is explicitly 'alternative and independent', so the same-authors citation is not load-bearing. The many [MR24a] references are for definitions, notation and θ-deformation examples, not for the new theorem. I do flag a non-circular but load-bearing proof gap: Lemma A.9 states 'Thus the set {( xi√2 , x†i√2),( x′i√2 , x′†i√2)} satisfies the conclusion of the Lemma', but the preceding identity x = 1/2(Σ x_i⟨x′_i|x⟩ + x_i†⟨x′_i†|x⟩) requires instead the pairs (x_i/√2, x′_i/√2) and (x_i†/√2, x′_i†/√2). As written, the displayed set gives 1/2(Σ x_i⟨x_i†|x⟩ + x′_i⟨x′_i†|x⟩), which is not shown to equal x. Since Proposition 4.23 and Theorem 4.24 import Lemma A.9, this is a gap in the written existence proof; it is an indexing/display error, not a circular definition.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central existence theorem rests on standard module theory, Skeide's braiding result, and the Arnlind-Wilson uniqueness theorem. No free parameters or invented physical entities appear. The key structural hypotheses (centredness, strong non-degeneracy, and g∘σ=g) are stated as definitions and do not introduce unexplained quantities.

assumptions (4)

standard math The algebra A is a unital complex ∗-algebra.
The paper states this in the Introduction and refers to [MR24a, Section 7.1] for removing unitality.
domain assumption Skeide's theorem: a centred †-bimodule admits a unique braiding σcan that flips central elements.
Used as Proposition A.8 throughout; proof is sketched with a citation to [S96].
domain assumption Arnlind-Wilson uniqueness theorem: a right real metric calculus has at most one pseudo-Riemannian connection.
Quoted as Theorem 4.5 [AW17, Theorem 3.4]; used in Corollary 4.21 and Theorem 4.24 for uniqueness.
domain assumption For θ-deformed spectral triples, existence of Levi-Civita connections from [BGM20, Theorem 5.4] and [MR24a, Theorem 6.12].
Used in Section 4.7 for the examples; not load-bearing for the general algebraic existence theorem.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Comparison of Levi-Civita connections in noncommutative geometry." pith.science (2026). https://pith.science/paper/ACPZFJXS

@misc{pith2026250518692,
  author       = {Pith},
  title        = {Pith review of: Comparison of Levi-Civita connections in noncommutative geometry},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ACPZFJXS}},
  note         = {Machine review of arXiv:2505.18692}
}

read the original abstract

We compare the constructions of Levi-Civita connections for noncommutative algebras developed in arXiv:1505.07330, arXiv:1809.06721, arXiv:2403.13735. The assumptions in these various constructions differ, but when they are all defined, we provide direct translations between them. An essential assumption is that the (indefinite) Hermitian inner product on differential forms/vector fields provides an isomorphism with the module dual. By exploiting our translations and clarifying the simplifications that occur for centred bimodules, we extend the existence results for Hermitian torsion-free connections in arXiv:1505.07330, arXiv:1809.06721.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

[1]

On the existence of noncommutative Levi-Civita connections in derivation based calculi

J. Arnlind, V. Hildebrandsson, On the existence of noncommutative Levi-Civita connections in derivation based calculi , arXiv:2505.13984

work page Pith review arXiv
[2]

Arnlind, M

J. Arnlind, M. Wilson, Riemannian curvature of the noncommutative 3-sphere , J. Noncommut. Geom., 11 (2017), 507--536

work page 2017
[3]

Arnlind, M

J. Arnlind, M. Wilson, On the Chern-Gauss-Bonnet theorem for the noncommutative 4-sphere , J. Geom. Phys., 111 (2017), 126--141

work page 2017
[4]

Beggs, S

E. Beggs, S. Majid, * -compatible connections in noncommutative Riemannian geometry , J. Geom. Phys. 61 (2011), 95-124

work page 2011
[5]

Beggs, S

E. Beggs, S. Majid, Quantum Riemannian Geometry , Springer Cham, 2020

work page 2020
[6]

Bhowmick, D

J. Bhowmick, D. Goswami, S. Joardar, A new look at Levi-Civita connection in noncommutative geometry , Int. J. Geom. Methods Mod. Phys., 18 (07), (2021) 2150105

work page 2021
[7]

Bhowmick, D

J. Bhowmick, D. Goswami, G. Landi, Levi-Civita connections and vector fields for noncommutative differential calculi, Int. J. Math., 31 (08) (2020), 2050065

work page 2020
[8]

Bhowmick, D

J. Bhowmick, D. Goswami and S. Mukhopadhyay, Levi-Civita connections for a class of spectral triples , Lett. Math. Phys., 110 (2020), 835--884

work page 2020

Show all 23 references

[9]

Blackadar, K-theory for operator algebras, 2nd edition,

B. Blackadar, K-theory for operator algebras, 2nd edition,
[10]

Carey, J

A. Carey, J. Phillips, A. Rennie, Spectral triples: examples and index theory , in `Noncommutative Geometry and Physics: Renormalisation, Motives, Index Theory', Ed. A. Carey, EMS, 2011, 175--265

2011
[11]

Connes, Noncommutative Geometry , Academic Press, 1994

A. Connes, Noncommutative Geometry , Academic Press, 1994

1994
[12]

Connes, G

A. Connes, G. Landi, Noncommutative manifolds, the instanton algebra, and isospectral deformations, Commun. Math. Phys., 221, 141--159 (2001)

2001
[13]

Cuntz, D

J. Cuntz, D. Quillen, Algebra extensions and nonsingularity, J. Amer. Math. Soc., 8 (1995), 251--289

1995
[14]

Dabrowski, P

L. Dabrowski, P. Hajac, G. Landi, P. Siniscalco,
[15]

Frank, D

M. Frank, D. Larson, Frames in Hilbert C^* -modules and C^* -algebras , J. Operator Th., 48 (2) (2002), 273--314

2002
[16]

Kajiwara, C

T. Kajiwara, C. Pinzari, Y. Watatani,
[17]

Lance, Hilbert C^* -modules, A toolkit

E.C. Lance, Hilbert C^* -modules, A toolkit
[18]

Landi, An Introduction to noncommutative spaces and their topology , Lecture Notes in Physics: Monographs, m51, Springer-Verlag, Berlin Heidelberg, 1997

G. Landi, An Introduction to noncommutative spaces and their topology , Lecture Notes in Physics: Monographs, m51, Springer-Verlag, Berlin Heidelberg, 1997

1997
[19]

Mesland and A

B. Mesland and A. Rennie, Existence and uniqueness of the Levi-Civita connection on noncommutative differential forms , Adv. Math., 468 (2025), 110207

2025
[20]

Mesland and A

B. Mesland and A. Rennie, Curvature and Weitzenbock formula for spectral triples , Math. Nachr. 297 (12) (2024), 4582--4604

2024
[21]

Mesland, A

B. Mesland, A. Rennie, Levi-Civita connection, curvature and Weitzenbock formula for the Podle\'s sphere , to appear in Commun. Math. Phys., arXiv:2406.18483

arXiv
[22]

Rosenberg, Levi-Civita's theorem for noncommutative tori , SIGMA, 9 (2013) 071, 9 pages

J. Rosenberg, Levi-Civita's theorem for noncommutative tori , SIGMA, 9 (2013) 071, 9 pages

2013
[23]

Skeide, Hilbert modules in quantum electro dynamics and quantum probability , Commun

M. Skeide, Hilbert modules in quantum electro dynamics and quantum probability , Commun. Math. Phys. 192 (1998) 569--604

1998

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

On the existence of noncommutative Levi-Civita connections in derivation based calculi

J. Arnlind, V. Hildebrandsson, On the existence of noncommutative Levi-Civita connections in derivation based calculi , arXiv:2505.13984

work page Pith review arXiv

[2] [2]

Arnlind, M

J. Arnlind, M. Wilson, Riemannian curvature of the noncommutative 3-sphere , J. Noncommut. Geom., 11 (2017), 507--536

work page 2017

[3] [3]

Arnlind, M

J. Arnlind, M. Wilson, On the Chern-Gauss-Bonnet theorem for the noncommutative 4-sphere , J. Geom. Phys., 111 (2017), 126--141

work page 2017

[4] [4]

Beggs, S

E. Beggs, S. Majid, * -compatible connections in noncommutative Riemannian geometry , J. Geom. Phys. 61 (2011), 95-124

work page 2011

[5] [5]

Beggs, S

E. Beggs, S. Majid, Quantum Riemannian Geometry , Springer Cham, 2020

work page 2020

[6] [6]

Bhowmick, D

J. Bhowmick, D. Goswami, S. Joardar, A new look at Levi-Civita connection in noncommutative geometry , Int. J. Geom. Methods Mod. Phys., 18 (07), (2021) 2150105

work page 2021

[7] [7]

Bhowmick, D

J. Bhowmick, D. Goswami, G. Landi, Levi-Civita connections and vector fields for noncommutative differential calculi, Int. J. Math., 31 (08) (2020), 2050065

work page 2020

[8] [8]

Bhowmick, D

J. Bhowmick, D. Goswami and S. Mukhopadhyay, Levi-Civita connections for a class of spectral triples , Lett. Math. Phys., 110 (2020), 835--884

work page 2020