REVIEW 1 major objections 6 minor 32 references
Lie ideals in properly infinite C*-algebras
T0 review · 1 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read The paper proves that every Lie ideal in a unital, properly infinite C*-algebra is commutator equivalent to a unique two-sided ideal, and that this answers an open problem in this setting.
desk verdict A strong, well-written resolution of Robert's Problem B for properly infinite C*-algebras; the only real soft spot is a compressed citation in Proposition 2.10 that a referee should check. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The workhorse is a unital $*$-homomorphism $M_2(\mathbb{C}) \oplus M_3(\mathbb{C}) \to A$, the direct sum of the $2\times 2$ and $3\times 3$ complex matrix algebras, which exists in every unital properly infinite C*-algebra via the Cuntz algebra $O_\infty$ and also in unital real rank zero algebras without characters. The homomorphism supplies two full complementary projections $g$ and $h$, cutting $A$ into four corners $A_{ij}$, and Lemma 2.2 proves that the corner products $A_{11}LA_{22}$, $A_{22}LA_{11}$, $A_{12}LA_{12}$, $A_{21}LA_{21}$ all lie in $[A,[A,L]]$ for any Lie ideal $L$. From these containments, Lemmas 2.3–2.6 and Proposition 2.5 force $[A,L] \subseteq A g L h A = A[A,L]A = I$ and $[A,I] \subseteq [A,[A,L]]$, giving Theorem 2.7: $L$ is related to the unique ideal $I := A[A,L]A$. Theorem 2.9 then shows commutator equivalence is equivalent to the equality $[A,L] = [A,[A,L]]$. The final step for properly infinite algebras is Lemma 5.1: because such algebras have no tracial states, a classical result gives $A = [A,A]$, hence $A = Z(A) + [A,A]$, which forces $[A,L] = [A,[A,L]]$ and upgrades relatedness to commutator equivalence.
What would settle it
Look directly at the proof of Proposition 2.10: the cited lemma [KR15] produces a separable subalgebra $A_0$, and the text then asserts that a larger separable real rank zero subalgebra $B \supseteq A_0$ has no one-dimensional representations. Since a one-dimensional representation of $B$ restricts to one of $A_0$, that assertion requires $A_0$ itself to have no such representations, which the paper does not state. A counterexample would be a unital real rank zero C*-algebra without characters whose separable subalgebras all admit characters; it would remove the embedding and cut off Theorem 5.3. Alternatively, find a Lie ideal $L$ in $O_\infty$ with $[O_\infty,L] \neq [O_\infty,[O_\infty,L]]$; the theorem predicts equality for every $L$.
Extended reading notes
Core claim
The central discovery is Theorem 5.3. Given a unital properly infinite C*-algebra $A$ and any Lie ideal $L\subseteq A$, set $I := A[A,L]A$, the two-sided ideal generated by all commutators $[a,x]$ with $a\in A$ and $x\in L$. The theorem asserts the double equality $[A,L] = [A,[A,L]] = [A,I]$. This means $L$ and $I$ generate exactly the same commutator subspace, so $L$ is commutator equivalent to $I$; it is also embraced by $I$, in the sense that $[A,I] \subseteq L \subseteq T([A,I])$, where $T(K)$ collects the elements whose commutators with $A$ all lie in $K$; and $I$ is the only two-sided ideal to which $L$ is related. Corollary 5.4 turns this into a complete classification: the Lie ideals of $A$ are exactly the additive subgroups sandwiched between $[A,I]$ and $T([A,I])$ for a unique two-sided ideal $I$. The same classification, with the uniqueness clause modified, holds for von Neumann algebras.
Load-bearing premise
The classification rests on the existence of a unital $*$-homomorphism from $M_2(\mathbb{C}) \oplus M_3(\mathbb{C})$ into $A$; for properly infinite algebras this is obtained through a chain of reductions whose final step — that a certain separable subalgebra has no one-dimensional representations — is not fully justified, and the whole theorem collapses if that step fails.
Editorial extensions
If this is right
- For every unital properly infinite C*-algebra $A$, the map $I \mapsto [A,I]$ is a bijection from two-sided ideals onto the Lie ideals satisfying $L = [A,L]$, with inverse $L \mapsto A[A,L]A$ (Theorem 3.10).
- Every Lie ideal $L$ is commutator equivalent to, embraced by, and related to the unique two-sided ideal $A[A,L]A$; in particular, $L$ is determined up to the sandwich $[A,I] \subseteq L \subseteq T([A,I])$ (Theorem 5.3 and Corollary 5.4).
- Closed Lie ideals in these algebras correspond exactly to closed two-sided ideals, since $A[A,L]A$ is closed whenever $L$ is closed (Corollary 2.8).
- In any von Neumann algebra $M$, every Lie ideal $L$ is commutator equivalent to $M[M,L]M$; if $M$ has zero commutative summand, that ideal is the unique two-sided ideal related to $L$ (Theorem 6.1).
- Two-sided ideals in algebras with the $M_2 \oplus M_3$ unit are generated by their commutators: $I = A[A,I]A$ and $[A,I] = [A,[A,I]] = \operatorname{span}_{\mathbb{C}} FN_2(A,I)$, describing $[A,I]$ as the span of orthogonally factorizable square-zero elements (Theorem 3.8).
Reading between the lines
- If Theorem 5.3 is correct, the same uniqueness argument should apply to any unital C*-algebra that admits a unital $M_2(\mathbb{C}) \oplus M_3(\mathbb{C})$ subalgebra and satisfies $A = Z(A) + [A,A]$; testing pure C*-algebras with the Dixmier property would either confirm or bound the scope of Remark 5.5.
- The equality $[A,L] = [A,[A,L]]$ for all Lie ideals means the commutator operation is idempotent on the set of commutator subspaces, which could be exploited to define a lattice-theoretic rank or dimension for Lie ideals of properly infinite algebras.
- The square-zero technology in Sections 3 and 4 suggests a quantitative version: if every element of a Dixmier ideal $I$ is a sum of products of pairs of square-zero elements from $I^{1/2}$, then commutator norms of elements in $I$ might be controlled by the number of such factors, a property that could be tested on operator ideals in $B(H)$.
- Resolving Question 2.12 positively would require extending the $M_2 \oplus M_3$ embedding to all unital character-free C*-algebras, not just those of real rank zero; a counterexample there would show that the real rank zero assumption is essential, not merely technical.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies Lie ideals (not necessarily closed) in unital C*-algebras that admit a unital *-homomorphism from M2(C) ⊕ M3(C). Through a series of matrix-unit computations, the author proves in Theorem 2.7 that every such Lie ideal L is related to the unique two-sided ideal I := A[A,L]A, with [A,I] = [A,[A,L]] ⊆ [A,L] ⊆ I. Theorem 2.9 characterizes commutator equivalence to I in terms of the equality [A,L] = [A,[A,L]]. These results are then applied in Theorem 5.3 to unital, properly infinite C*-algebras, proving that every Lie ideal is commutator equivalent to a unique two-sided ideal, and in Theorem 6.1 to von Neumann algebras, with uniqueness of the related ideal when the algebra has no commutative summand. The latter is presented as a solution to BKS Problem 5.21, and Theorem 5.3 is presented as a positive answer to Robert's Problem B in the properly infinite setting.
Significance. If the main results are correct, they give a clean structural description of all Lie ideals in unital properly infinite C*-algebras: after Corollary 5.4, every Lie ideal is uniquely sandwiched between [A,I] and T([A,I]) for a two-sided ideal I, so the complicated family of Lie ideals is encoded by the lattice of two-sided ideals. This is a substantial positive answer to a question of Robert. The technical core is presented in unusually full detail: the commutator identities in Lemma 2.2, the factorization arguments with v1,w1,v2,w2, and the uniqueness argument in Theorem 2.7 are all written out and, on spot-checking, are correct. The paper also contains several useful byproducts, including Theorem 3.8, Theorem 4.6, and Proposition 6.2. The main caveat is a gap in the proof of Proposition 2.10, which is the bridge needed for the applications to real rank zero algebras and to von Neumann algebras; this is discussed below.
major comments (1)
- [§2, Proposition 2.10] The proof of Proposition 2.10 is incomplete as written. The text invokes [KR15, Lemma 3.5] only to obtain a separable sub-C*-algebra A0 ⊆ A containing the unit, then extends A0 to a separable, real rank zero subalgebra B, and asserts 'It follows that B has no characters.' This inference requires that A0 itself has no characters, because a character of B restricts to a character of A0. The manuscript does not state that [KR15, Lemma 3.5] supplies a character-free A0. If the cited lemma does include that conclusion, the statement should be amended accordingly; otherwise the forward implication of Proposition 2.10 is not established. Since Proposition 2.10 is the bridge producing the unital *-homomorphism M2(C) ⊕ M3(C) → A for the algebras in Theorem 5.3 and Theorem 6.1, this is a load-bearing point that needs to be fixed or clarified.
minor comments (6)
- [§2, Lemma 2.3] The proof uses that g and h are full in A. This is true because g and h are full in M2(C) ⊕ M3(C) and the image of a full projection under a unital *-homomorphism is full, but the manuscript does not say this; a one-sentence justification would remove a hidden step.
- [§5, Proposition 5.2(1)] The sentence 'Combining both results, we get A = Z(A) + [A,A]' is unclear: the Dixmier-property argument already gives the equality, and the closedness of [A,A] from [NR16] is not evidently needed for it. Please clarify the role of the second sentence, or remove it.
- [§4, Lemma 4.2] In the statement and proof of Lemma 4.2, the notation 'span_C(N2(I^{1/2}), N2(I^{1/2}))' appears; this should presumably be 'span_C[ N2(I^{1/2}), N2(I^{1/2}) ]' (the subspace generated by commutators), in line with the proof's conclusion. The same notational issue appears in Theorem 4.3.
- [§2.1] Typo: 'decomposition induces by g and h' should be 'decomposition induced by g and h'.
- [§5, Remark 5.5] Typo: 'tracial stetes' should be 'tracial states'.
- [§1] In the introductory discussion of simple unital algebras, the sentence 'a Lie ideal it is commutator equivalent to A' contains a stray 'it'.
Circularity Check
The central derivation is internal: I := A[A,L]A is defined from [A,L], the nontrivial equality [A,I] = [A,[A,L]] is proved from the M2⊕M3 hypothesis, and no fitted parameter or self-validating definition appears.
full rationale
The core claim (Theorem 5.3) rests on Theorem 2.7, the equality [A,L] = [A,[A,L]] = [A,I] with I := A[A,L]A. I is defined as the two-sided ideal generated by [A,L], and the key equality [A,I] = [A,[A,L]] is proved (Lemma 2.6 with Proposition 2.5), not assumed. Lemma 2.2 establishes the containments A11LA22, A22LA11, A12LA12, A21LA21 ⊆ [A,[A,L]] from the Lie ideal property and the mixed projection decomposition; Lemma 2.4 and Proposition 2.5 give I = AgLhA = A[A,L]A; Lemma 2.6 gives [A,I] ⊆ [A,[A,L]]. Together with the immediate converse inclusion, the equality chain is a genuine proof. Theorem 2.9's equivalence '(1) iff (3)' follows from these proved equalities, and Lemma 5.1 supplies [A,L] = [A,[A,L]] from A = Z(A) + [A,A], which holds for properly infinite A by Pop's theorem (no tracial states). The M2⊕M3 hypothesis for properly infinite A comes from a standard fact (O∞ embeds unitally by [Bla06, Proposition III.1.3.3]; O∞ admits M2⊕M3 by Proposition 2.10). The framework notions (commutator equivalent, related, embraced) and the target problems (Robert's Problem B, BKS Problem 5.21) are external, not redefined to make the conclusion tautological. Theorem 6.1 uses the same internal route with the BKS fact M = Z(M) + [M,M]. Self-citations ([GT24], [GKT23], [GT23], [GT25]) are used as tools in auxiliary remarks, and the load-bearing equalities of Section 2 are proved here. The flagged issue in Proposition 2.10 (whether [KR15, Lemma 3.5] explicitly guarantees a character-free separable A0 so that 'B has no characters' follows) is a possible gap or missing-hypothesis problem in a cited lemma, not a circular reduction: no conclusion is assumed through the citation. Even on the skeptic's reading, the paper would be incomplete, not circular, because Theorem 2.7's internal proof is independent of that bridge. Consequently no step reduces to its own input by construction, and the appropriate score is 0.
Assumptions & free parameters
assumptions (7)
- domain assumption A admits a unital *-homomorphism M2(C) ⊕ M3(C) → A (hypothesis of Theorem 2.7; guaranteed for unital properly infinite A via O∞ and Proposition 2.10).
- standard math Pop's theorem [Pop02, Theorem 1]: a unital C*-algebra with no tracial states satisfies A = [A,A].
- standard math Perera-Rordam [PR04, Proposition 5.7]: unital, separable, real rank zero C*-algebras without characters admit a unital *-homomorphism from M2(C) ⊕ M3(C).
- standard math The separable subalgebra A0 from [KR15, Lemma 3.5] can be chosen without characters, so the larger separable algebra B in the proof of Proposition 2.10 also has no characters.
- standard math Every von Neumann algebra M satisfies M = Z(M) + [M,M], as shown in the proof of [BKS08, Theorem 5.19].
- standard math Semiprime ideal structure theory of [GKT23, Theorem A] (semiprime iff idempotent iff Dixmier with I = I^{1/2}) and zero-product balance of [GT23] and [Bre21, Theorem 2.3].
- standard math Every von Neumann algebra has real rank zero, and O∞ has real rank zero and no characters (folklore facts used in Theorem 6.1 and Theorem 5.3).
Cite this review
Pith. "Pith review of Lie ideals in properly infinite C*-algebras." pith.science (2026). https://pith.science/paper/64ZLCS2R
@misc{pith2026241216087,
author = {Pith},
title = {Pith review of: Lie ideals in properly infinite C*-algebras},
year = {2026},
howpublished = {\url{https://pith.science/paper/64ZLCS2R}},
note = {Machine review of arXiv:2412.16087}
}
read the original abstract
We show that every Lie ideal in a unital, properly infinite C*-algebra is commutator equivalent to a unique two-sided ideal. It follows that the Lie ideal structure of such a C*-algebra is concisely encoded by its lattice of two-sided ideals. This answers a question of Robert in this setting. We obtain similar structure results for Lie ideals in unital, real rank zero C*-algebras without characters. As an application, we show that every Lie ideal in a von Neumann algebra is related to a unique two-sided ideal, which solves a problem of Bre\v{s}ar, Kissin, and Shulman.
Reference graph
Works this paper leans on
-
[1]
T. Amrutam , D. Gao , S. Kunnawalkam Elayavalli , and G. Patchell , Strict comparison in reduced group s , preprint (arXiv:2412.06031 [math.OA]), 2024
arXiv 2024
-
[2]
Antoine , F
R. Antoine , F. Perera , and H. Thiel , Tensor products and regularity properties of C untz semigroups, Mem. Amer. Math. Soc. 251 (2018), viii+191
2018
-
[3]
R. Antoine , F. Perera , H. Thiel , and E. Vilalta , Pure s , preprint (arXiv:2406.11052 [math.OA]), 2024
arXiv 2024
-
[4]
R. Archbold , L. Robert , and A. Tikuisis , The D ixmier property and tracial states for s , J. Funct. Anal. 273 (2017), 2655--2718
work page 2017
-
[5]
R. J. Archbold , I. Gogi\' c , and L. Robert , Local variants of the D ixmier property and weak centrality for s , Int. Math. Res. Not. IMRN (2023), 1483--1513
work page 2023
-
[6]
B. Blackadar , Operator algebras, Encyclopaedia of Mathematical Sciences 122, Springer-Verlag, Berlin, 2006, Theory of s and von Neumann algebras, Operator Algebras and Non-commutative Geometry, III
2006
-
[7]
E. Breuillard , M. Kalantar , M. Kennedy , and N. Ozawa , C^* -simplicity and the unique trace property for discrete groups, Publ. Math. Inst. Hautes \'Etudes Sci. 126 (2017), 35--71
work page 2017
-
[8]
M. Bre s ar , Zero product determined algebras, Frontiers in Mathematics, Birkh\" a user/Springer, Cham, 2021
work page 2021
Show all 32 references
-
[9]
Bre s ar , E
M. Bre s ar , E. Kissin , and V. S. Shulman , Lie ideals: from pure algebra to s , J. Reine Angew. Math. 623 (2008), 73--121
2008
-
[10]
Cuntz and G
J. Cuntz and G. K. Pedersen , Equivalence and traces on s , J. Funct. Anal. 33 (1979), 135--164
1979
-
[11]
Dykema , T
K. Dykema , T. Figiel , G. Weiss , and M. Wodzicki , Commutator structure of operator ideals, Adv. Math. 185 (2004), 1--79
2004
-
[12]
Dykema and A
K. Dykema and A. Krishnaswamy-Usha , Nilpotent elements of operator ideals as single commutators, Proc. Amer. Math. Soc. 146 (2018), 3031--3037
2018
-
[13]
C. K. Fong , C. R. Miers , and A. R. Sourour , Lie and J ordan ideals of operators on H ilbert space, Proc. Amer. Math. Soc. 84 (1982), 516--520
1982
-
[14]
Gardella , K
E. Gardella , K. Kitamura , and H. Thiel , Semiprime ideals in s , J. Eur. Math. Soc. (to appear), preprint (arXiv:2311.17480), 2023
2023 arXiv
-
[15]
Gardella , T.-K
E. Gardella , T.-K. Lee , and H. Thiel , Fully noncentral L ie ideals and invariant additive subgroups in rings, J. London Math. Soc. (to appear), preprint (arXiv:2409.03362), 2024
2024 arXiv
-
[16]
Gardella and H
E. Gardella and H. Thiel , Zero-product balanced algebras, Linear Algebra Appl. 670 (2023), 121--153
2023
-
[17]
Gardella and H
E. Gardella and H. Thiel , Prime ideals in s and applications to L ie theory, Proc. Amer. Math. Soc. 152 (2024), 3647--3656
2024
-
[18]
Gardella and H
E. Gardella and H. Thiel , Rings and s generated by commutators, J. Algebra 662 (2025), 214--241
2025
-
[19]
Haagerup , Quasitraces on exact s are traces, C
U. Haagerup , Quasitraces on exact s are traces, C. R. Math. Acad. Sci. Soc. R. Can. 36 (2014), 67--92
2014
-
[20]
I. N. Herstein , On the L ie and J ordan rings of a simple associative ring, Amer. J. Math. 77 (1955), 279--285
1955
-
[21]
Kirchberg , On the existence of traces on exact stably projectionless simple s , in Operator algebras and their applications ( W aterloo, ON , 1994/1995) , Fields Inst
E. Kirchberg , On the existence of traces on exact stably projectionless simple s , in Operator algebras and their applications ( W aterloo, ON , 1994/1995) , Fields Inst. Commun. 13, Amer. Math. Soc., Providence, RI, 1997, pp. 171--172
1994
-
[22]
Kirchberg and M
E. Kirchberg and M. R rdam , When central sequence s have characters, Internat. J. Math. 26 (2015), 1550049, 32
2015
-
[23]
T. Y. Lam , A first course in noncommutative rings, second ed., Graduate Texts in Mathematics 131, Springer-Verlag, New York, 2001
2001
-
[24]
L. W. Marcoux , Projections, commutators and L ie ideals in s , Math. Proc. R. Ir. Acad. 110A (2010), 31--55
2010
-
[25]
L. W. Marcoux and G. J. Murphy , Unitarily-invariant linear spaces in s , Proc. Amer. Math. Soc. 126 (1998), 3597--3605
1998
-
[26]
Marcoux , On the closed L ie ideals of certain s , Integral Equations Operator Theory 22 (1995), 463--475
L. Marcoux , On the closed L ie ideals of certain s , Integral Equations Operator Theory 22 (1995), 463--475
1995
-
[27]
P. W. Ng , Commutators in the J iang- S u algebra, Internat. J. Math. 23 (2012), 1250113, 29
2012
-
[28]
P. W. Ng and L. Robert , Sums of commutators in pure s , M\" u nster J. Math. 9 (2016), 121--154
2016
-
[29]
Perera and M
F. Perera and M. R rdam , A F -embeddings into s of real rank zero, J. Funct. Anal. 217 (2004), 142--170
2004
-
[30]
Pop , Finite sums of commutators, Proc
C. Pop , Finite sums of commutators, Proc. Amer. Math. Soc. 130 (2002), 3039--3041
2002
-
[31]
Robert , On the L ie ideals of s , J
L. Robert , On the L ie ideals of s , J. Operator Theory 75 (2016), 387--408
2016
-
[32]
Winter , Nuclear dimension and Z -stability of pure s , Invent
W. Winter , Nuclear dimension and Z -stability of pure s , Invent. Math. 187 (2012), 259--342
2012
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.