Pith. sign in

REVIEW 3 major objections 4 minor 31 references

On some conjectures by Lu and Wenzel

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The complex Lu-Wenzel conjecture, which would unify the BW and DDVV inequalities, is proved here for normal matrices, rank-one matrices, and dimensions 2 and 3.

desk verdict Solid special-case results undermined by a false isotropy step in the advertised equivalence, so the main claim overreaches. read the letter →

arxiv 1908.06624 v1 pith:2C5SMMMN submitted 2019-08-19 math.DG math.OA

Jianquan Ge , Fagui Li , Zhiqin Lu , Yi Zhou This is my paper · ORCID
classification math.DGmath.OA MSC 15A4515B5753C42
keywords Lu-WenzelconjectureBöttcher-WenzelinequalityDDVVcommutatorFrobeniusnormweakmajorizationeigenvaluebound
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 attacks the Lu-Wenzel conjectures, a family of matrix inequalities proposed as a common generalization of two classical bounds: the Böttcher-Wenzel (BW) inequality on the Frobenius norm of a commutator of two matrices, and the DDVV inequality on sums of commutators of many matrices. The central object is the operator $T_X(Y)=[X^*, [X,Y]]$ for a unit-norm matrix $X$; the conjectures assert sharp upper bounds on the sums of the largest eigenvalues of $T_X$. The paper proves these bounds for three classes of $X$: normal matrices, rank-one matrices, and matrices of size $2$ or $3$. For general $X$ it establishes non-sharp bounds, including $\lambda_1(T_X)+\lambda_3(T_X)\le 2+\sqrt{10}/2$ and a bound for all $2k$-term eigenvalue sums, and it gives new proofs of the complex BW inequality and its equality condition. The importance is that a complete proof would unify the BW and DDVV inequalities into one sharp statement.

What carries the argument

Central machinery: the operator $T_X(Y)=[X^*, [X,Y]]$ acting on $M(n,\mathbb{C})$. It is Hermitian positive semidefinite with $\langle T_X Y,Y\rangle=\|[X,Y]\|^2$, so its largest eigenvalues measure commutator sizes. Proposition 2.6 shows every positive eigenvalue has even multiplicity, $\lambda_{2i-1}(T_X)=\lambda_{2i}(T_X)$; paired with a skew-symmetric operator $S_X$ satisfying $T_X=S_X^*S_X=-S_X^2$, this reduces the conjectures to trace comparisons over subspaces whose images under $S_X$ are orthogonal to the subspace itself (Lemma 2.3). The proofs further use the Kronecker-product representation $T_X=K_X^*K_X$ with $K_X=I\otimes X-X^t\otimes I$ and weak-majorization eigenvalue comparison (Lemma 2.11).

What would settle it

In Proposition 4.1, test the asserted isotropy condition directly: for $X=\mathrm{diag}(1,0,0)$ and $B_2=E_{12}+E_{21}$, $B_3=i(E_{12}-E_{21})$, the trace conditions (i) and (ii) of Conjecture 2 hold, but $\langle S_X B_3, B_2\rangle=-2i\neq 0$, so $S_X(W)$ is not contained in $W^\perp$. This counterexample to the proof's key step shows the equivalence between Conjecture 2 and Conjecture 4 is not established as written.

Watch

Extended reading notes

Core claim

The paper's main theorem (Theorem 1.2) states that the complex Lu-Wenzel Conjectures 7 and 8 hold whenever $X\in M(n,\mathbb{C})$ with $\|X\|=1$ is normal, has rank one, or $n\in\{2,3\}$. These conjectures assert, respectively, that $\sum_{i=1}^{2k}\lambda_i(T_X)\le 2k+2$ for $k=1,\dots,\lfloor n^2/2\rfloor$, and that the eigenvalue list of $T_X$ is weakly majorized by the multiset $\{2^2,1^{2n-4},0^{(n-1)^2+1}\}$. The authors also prove an equivalence theorem (Theorem 1.1, complex version): Conjectures 2, 4, 5, and 6 are equivalent, and each implies Conjectures 1 and 3. Since Conjecture 2 is the fundamental Lu-Wenzel conjecture, these equivalences identify the whole family with one assertion about the eigenvalue distribution of $T_X$. For the unresolved general case, the paper obtains $\lambda_1(T_X)+\lambda_3(T_X)\le 2+\sqrt{10}/2$ (Theorem 1.3) and $\sum_{i=1}^{2k}\lambda_i(T_X)\le 2k+1+2\sqrt{k}$ (Theorem 1.4), together with new proofs of the complex BW inequality and the condition for equality.

Load-bearing premise

The load-bearing step is the claim that the trace conditions in Conjecture 2 make the span of the matrices $B_2,\ldots,B_m$ orthogonal to its image under the operator $S_X$; this is asserted without proof and is the link that turns eigenvalue bounds into the inequality for arbitrary families.

Editorial extensions

If this is right

  • For normal $X$, rank-one $X$, and $n=2,3$, the complex Lu-Wenzel conjecture holds, so in those cases the BW and DDVV inequalities are unified under a single sharp statement.
  • The weak-majorization form fixes the sharp constants in the eigenvalue sums: for $k\ge n$, $\sum_{i=1}^{2k}\lambda_i(T_X)\le 2n$, so the eigenvalue mass is sharply bounded in every initial segment.
  • The partial bound $\lambda_1(T_X)+\lambda_3(T_X)\le 2+\sqrt{10}/2$ gives a concrete non-sharp answer to Conjecture 3 for every $X$, and the general bound $\sum_{i=1}^{2k}\lambda_i(T_X)\le 2k+1+2\sqrt{k}$ answers Question 1 in weaker form.
  • The characterization of equality in the complex BW inequality (top eigenvalue 2 iff $X$ is unitarily similar to $\mathrm{diag}(X_0,O_{n-2})$ with $\mathrm{Tr}(X_0)=0$) provides a clean criterion for maximal commutator pairs.
  • Since Conjecture 2 implies Conjecture 1, the complex DDVV-type inequality with sharp constant 1 follows in the special cases.

Reading between the lines

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

  • If the weak-majorization form were established for all $X$, the sharp bound for Question 1 would be $\sum_{i=1}^k \lambda_{2i-1}(T_X)\le k+1$; the partial results are consistent with that target, and the remaining difficulty appears to lie in the isotropy step rather than in the final inequality.
  • The special-case proofs in Section 5 bypass the trace-condition equivalence, so they may remain valid even if the equivalence proof's isotropy assumption fails for general complex matrices; one natural next step is to determine whether the full conjecture needs a Hermitian rather than complex-linear trace condition.
  • The eigenvalue operator $T_X$ and the even-multiplicity structure suggest the same conjectures could be studied for unitarily invariant norms other than the Frobenius norm, since weak majorization would transfer the bounds to any such norm.
  • The numerical example in Remark 5.14 shows the constant $2+\sqrt{10}/2$ is not optimal; a sharper analysis of the four largest eigenvalues might reach the conjectured constant 3.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper studies the Lu-Wenzel conjectures, which aim at a unified generalization of the Bottcher-Wenzel (BW) inequality and the DDVV inequality. The authors propose several equivalent formulations of the 'real and complex LW Conjecture' (Conjecture 2), prove the complex LW conjecture in special cases (X normal, rank X = 1, n = 2, 3), and give partial bounds for the general case (Theorems 1.3 and 1.4). The paper also provides new proofs of the complex BW inequality and a characterization of equality. The central structural claim is Theorem 1.1, which asserts that Conjectures 2, 4, 5, and 6 are equivalent, so that eigenvalue bounds for the operator T_X would imply the full LW inequality.

Significance. If the equivalence of Theorem 1.1 and the special-case proofs of Theorem 1.2 were correct, the paper would unify two well-known commutator inequalities and resolve several open conjectures of Lu and Wenzel. The new proofs of the BW inequality and the equality characterization are genuinely interesting and appear self-contained. The special-case eigenvalue bounds for normal matrices, rank-one matrices, and n = 2, 3 also seem to be correct and could be useful in their own right. However, the central equivalence is invalid as written, and the advertised reduction of all conjectures to eigenvalue estimates is not established. The paper therefore does not deliver its main claim, although several of its components remain valuable.

major comments (3)
  1. [§4, Proposition 4.1] The proof of Proposition 4.1 claims that conditions (i) and (ii) of Conjecture 2 imply that the subspace W = Span_C{B_2,...,B_m} is isotropic about S_X, i.e., S_X(W) ⊂ W^⊥. This assertion is false in the complex case. The trace condition (ii) controls the anti-linear map ~S_X(Y) = [X,Y]^*, not the linear map S_X defined in (2.3). A concrete counterexample is X = diag(1,0,0), B_2 = (E_12 + i E_21)/√2. Then Tr(B_2 [X,B_2]) = 0, so condition (ii) holds for m = 2 (condition (i) is vacuous), but S_X(B_2) = -i B_2, hence ⟨S_X B_2, B_2⟩ = -i ≠ 0. Thus W is not isotropic about S_X, and the application of Lemma 2.3 in the proof of Conjecture 4 ⇒ Conjecture 2 is invalid. This breaks Theorem 1.1(1).
  2. [§1, Theorem 1.2] Because Proposition 4.1 is false, the statement that the special-case eigenvalue bounds for Conjecture 7 imply 'hence all conjectures of this paper are true' is unsupported. The proofs in Section 5 establish Conjecture 7 (equivalently Conjecture 8) for normal X, rank-one X, and n = 2, 3, and Conjecture 3 follows from Conjecture 7. However, Conjecture 2, and consequently Conjecture 1, are not implied by those eigenvalue bounds without the missing equivalence. The advertised resolution of the fundamental LW Conjecture in these cases is therefore not obtained.
  3. [§4, Proposition 4.1 (second direction)] The confusion between S_X and ~S_X appears systematically in the proof of Proposition 4.1. Since S_X is the complex-linear extension of the anti-linear map ~S_X on the eigenbasis, the identity S_X(B) = [X,B]^* holds only for real scalar multiples of the chosen eigenvectors, not for general complex linear combinations. The trace conditions in Conjecture 2 are complex-linear trace conditions, whereas the isotropy required by Lemma 2.3 is with respect to the Hermitian inner product evaluated on S_X. These are different objects, and the manuscript does not provide a bridge between them.
minor comments (4)
  1. [Abstract] The abstract states 'we prove Conjecture 2 and hence all conjectures in some special cases.' Given the failure of Proposition 4.1, the paper actually proves Conjecture 7 (and hence Conjecture 8 and Conjecture 3) in those cases, not Conjecture 2. The wording should be adjusted to reflect the weaker claim.
  2. [Theorem 1.4] The displayed inequality is written as '2k∑_{t=1} λ_i(T_X)', but the index in the summand should be λ_t(T_X), not λ_i(T_X).
  3. [Proof of Theorem 1.3] In the final displayed inequality, '≤ (4 + √10)|X‖^2' should read '≤ (4 + √10)‖X‖^2'.
  4. [Section 5, Remark 5.14] The phrase 'non-sharp upper bounds' is slightly informal; 'non-sharp' would be clearer as 'not sharp' or 'nonoptimal.'

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper's special-case proofs come from direct eigenvalue estimates; the disputed isotropy step is a correctness gap, not a circular reduction.

full rationale

After walking the derivation chain, I find no circular reductions in the sense of assuming the target inequality or renaming a fit as a prediction. The central reduction (Theorem 1.1, Propositions 4.1–4.3) is an algebraic equivalence between conjectures: Conjecture 2 is assumed only on the forward side and derived from Conjecture 4 via Lemma 2.3 on the reverse side; neither direction presupposes the inequality being proved. The special-case proofs (Theorem 5.1/Corollary 5.3 for normal X, Theorem 5.7 for rank-one X, Theorem 5.12 for n = 2,3) compute or bound eigenvalues of T_X directly, using standard results such as Bhatia's eigenvalue lemma and Kronecker-product spectral facts. The imported lemma of Lu [23] (Lemma 5.2) is parameter-free, has assumptions (normalization and trace-zero) that do not include the LW conjecture, and is therefore independent support rather than a self-citation chain. The BW inequality is re-derived in Section 3 from Proposition 2.8, not assumed as input; Proposition 4.2's use of the BW inequality for the first eigenvalue is justified by Theorem 3.1, proved earlier in the paper. The notable weakness is Proposition 4.1's assertion that 'the conditions (i,ii) of Conjecture 2 show that the subspace W := Span C{Bα }m α=2 is isotropic about SX, i.e., SX(W )⊥CW.' This is asserted without proof and, as the accompanying analysis shows, is false in general because condition (ii) controls the anti-linear map S~X(Y)=[X,Y]^* while S_X is its linear extension. That is a correctness gap in the claimed Conjecture 2 ⇔ Conjecture 4 equivalence, not a circular argument: the target inequality is never assumed as an input, and the gap can be checked by explicit counterexamples rather than by tracing the argument back to its own conclusion. The paper also openly records non-sharpness of its general bounds (Theorems 1.3 and 1.4, Remark 5.13), which further supports that the special-case results are genuine partial proofs rather than repackaged assumptions. Accordingly, the circularity score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central results are parameter-free matrix inequalities; no free constants or new entities are introduced. The paper relies on standard matrix analysis results and its own structural lemmas about T_X.

assumptions (3)
  • standard math Standard eigenvalue majorization results (Bhatia Lemma 5.6, Zhang Lemmas 2.10-2.11, Ando-Bhatia [1])
    Used to bound eigenvalues of Kronecker products and sums; standard results from cited textbooks.
  • standard math Proposition 2.6: positive eigenvalues of T_X have even multiplicity
    Proven in the paper (and earlier in Lu [23]) via the conjugate-linear map ~S_X; this underpins the pairing λ_{2i-1}=λ_{2i}.
  • standard math Lu's lemma (Lemma 5.2) bounding sums of |η_i-η_j|^2
    Cited from Lu [23] and used for the normal eigenvalue bound.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On some conjectures by Lu and Wenzel." pith.science (2026). https://pith.science/paper/2C5SMMMN

@misc{pith2026190806624,
  author       = {Pith},
  title        = {Pith review of: On some conjectures by Lu and Wenzel},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2C5SMMMN}},
  note         = {Machine review of arXiv:1908.06624}
}
abstract

In order to give a unified generalization of the BW inequality and the DDVV inequality, Lu and Wenzel proposed three Conjectures 1, 2, 3 and an open Question 1 in 2016. In this paper we discuss further these conjectures and put forward several new conjectures which will be shown equivalent to Conjecture 2. In particular, we prove Conjecture 2 and hence all conjectures in some special cases. For Conjecture 3, we obtain a bigger upper bound $2+\sqrt{10}/2$, and we also give a weaker answer for the more general Question 1. In addition, we obtain some new simple proofs of the complex BW inequality and the condition for equality.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages

  1. [1]

    Ando and R

    T. Ando and R. Bhatia, Eigenvalue inequalities associated with the cartesian dec omposition, Lin- ear and Multilinear Algebra 22 (2) (1987), 15

    work page 1987

  2. [2]

    K. M. R. Audenaert, Variance bounds, with an application to norm bounds for comm utators, Linear Algebra Appl. 432 (5) (2009), 1126-1143

    work page 2009

  3. [3]

    Bhatia, Matrix analysis , Springer, New York (1997)

    R. Bhatia, Matrix analysis , Springer, New York (1997)

    work page 1997

  4. [4]

    B¨ ottcher and D

    A. B¨ ottcher and D. Wenzel, How big can the commutator of two matrices be and how big is it typically? Linear Algebra Appl. 403 (2005), 216-228

    work page 2005

  5. [5]

    B¨ ottcher and D

    A. B¨ ottcher and D. Wenzel, The Frobenius norm and the commutator , Linear Algebra Appl. 429 (8-9) (2008), 1864-1885

    work page 2008

  6. [6]

    C. M. Cheng, K. S. Fong and W. F. Lei, On some norm inequalities involving the commutator and XY − Y X T , Linear Algebra Appl. 438 (6) (2013), 2793-2807

    work page 2013

  7. [7]

    C. M. Cheng, S. W. Vong and D. Wenzel, Commutators with maximal Frobenius norm , Linear Algebra Appl. 432 (2010), 292-306

    work page 2010

  8. [8]

    C. M. Cheng and Y. Liang, Some sharp bounds for the commutator of real matrices , Linear Algebra Appl. 521 (2017), 263-282

    work page 2017

Show all 31 references
  1. [9]

    Choi and Z.Lu, On the DDVV Conjecture and the Comass in Calibrated Geometry (I), Math

    T. Choi and Z.Lu, On the DDVV Conjecture and the Comass in Calibrated Geometry (I), Math. Z. 260 (2008), 409-429

    2008

  2. [10]

    P. J. De Smet, F. Dillen, L. Verstraelen and L. Vrancken, A pointwise inequality in submanifold theory, Arch. Math. (Brno), 35 (1999), 115-128

    1999

  3. [11]

    Dillen, J

    F. Dillen, J. Fastenakels and J. Veken, Remarks on an inequality involving the normal scalar curvature, Proceedings of the International Congress on Pure and Appl ied Differential Geometry PADGE, Brussels, edited by F. Dillen and I. Van deWoestyne, S haker, Aachen, (2007), 83-92

    2007

  4. [12]

    Dillen, S

    F. Dillen, S. Haesen, M. Petrovi´ c-Torgaˇ sev and L. Ver straelen, An inequality between intrinsic and extrinsic scalar curvature invariants for codimension 2 embeddings , J. Geom. Physics 52 (2004), 101-112

    2004

  5. [13]

    K. S. Fong, C. M. Cheng and I. K. Lok, Another unitarily invariant norm attaining the minimum norm bound for commutators , Linear Algebra Appl. 433 (11) (2010), 1793-1797

    2010

  6. [14]

    J. Q. Ge, DDVV-type inequality for skew-symmetric matrices and Simo ns-type inequality for Riemannian submersions , Advances in Math. 251 (2014), 62-86

    2014

  7. [15]

    J. Q. Ge and Z. Z. Tang, A proof of the DDVV conjecture and its equality case , Pacific J. Math. 237(1) (2008), 87-95

    2008

  8. [16]

    J. Q. Ge and Z. Z. Tang, A survey on the DDVV conjecture , Harmonic maps and differential geometry, 247–254, Contemp. Math. 542, Amer. Math. Soc., Providence, RI, 2011

    2011

  9. [17]

    J. Q. Ge, S. Xu, H. Y. You and Y. Zhou, DDVV-type inequality for Hermitian matrices , Linear Algebra Appl. 529 (2017), 133-147

    2017

  10. [18]

    J. Q. Ge, F. G. Li and Y. Zhou, Some generalizations of the DDVV-type inequalities , arXiv:1807.07307

    arXiv

  11. [19]

    R. A. Horn and C. R. Johnson, Topics in Matrix Analysis , Cambridge University Press, New York (1991)

    1991

  12. [20]

    R. A. Brualdi, Introductory Combinatorics, 5 Edition , Prentice-Hall, Upper Saddle(2009)

    2009

  13. [21]

    L´ aszl´ o,Proof of B¨ ottcher and Wenzel’s conjecture on commutator no rms for 3-by-3 matrices , Linear Algebra Appl

    L. L´ aszl´ o,Proof of B¨ ottcher and Wenzel’s conjecture on commutator no rms for 3-by-3 matrices , Linear Algebra Appl. 422 (2) (2007), 659-663

    2007

  14. [22]

    Lu, Recent developments of the DDVV conjecture , Bull

    Z. Lu, Recent developments of the DDVV conjecture , Bull. Transilv. Univ. Bra¸ sov Ser. B (N.S.) 14(49) (2007), 133-143. ON SOME CONJECTURES BY LU AND WENZEL 31

    2007

  15. [23]

    Lu, Normal scalar curvature conjecture and its applications , J

    Z. Lu, Normal scalar curvature conjecture and its applications , J. Funct. Anal. 261 (2011), 1284- 1308

    2011

  16. [24]

    Lu, Remarks on the B¨ ottcher-Wenzel inequality , Linear Algebra Appl

    Z. Lu, Remarks on the B¨ ottcher-Wenzel inequality , Linear Algebra Appl. 436 (2012), no. 7, 2531-2535

    2012

  17. [25]

    Lu and D

    Z. Lu and D. Wenzel, The normal Ricci curvature inequality , Recent advances in the geometry of submanifolds—dedicated to the memory of Franki Dillen (1 963–2013), Contemporary Mathe- matics 674 (2016), 99-110

    2016

  18. [26]

    Lu and D

    Z. Lu and D. Wenzel, Commutator estimates comparising the Frobenius norm - Look ing back and forth, Oper. Theory: Adv. and Appl. 259 (2017), 533-559

    2017

  19. [27]

    D. A. Trifonov, Generalizations of Heisenberg uncertainty relation , The European Physical Jour- nal B, 29 (2002), 349-353

    2002

  20. [28]

    S. W. Vong and X. Q. Jin, Proof of B¨ ottcher and Wenzel’s conjecture, Oper. Matrices 2 (2008), no. 3, 435-442

    2008

  21. [29]

    Wenzel, Dominating the Commutator , Topics in Operator Theory, Oper

    D. Wenzel, Dominating the Commutator , Topics in Operator Theory, Oper. Theory: Adv. and Appl. 202 (2010), 579-600

    2010

  22. [30]

    Wenzel and K

    D. Wenzel and K. M. R. Audenaert, Impressions of convexity: an illustration for commutator bounds, Linear Algebra Appl. 433 (11) (2010), 1726-1759

    2010

  23. [31]

    Zhang, Matrix theory : basic results and techniques , Springer, Berlin (2011)

    F. Zhang, Matrix theory : basic results and techniques , Springer, Berlin (2011). School of Mathematical Sciences, Laboratory of Mathematics and Complex Systems, Beijing Normal University, Beijing 100875, P.R. CHINA. E-mail address : jqge@bnu.edu.cn School of Mathematical Scie...

    2011

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.