Pith. sign in

REVIEW 2 major objections 6 minor 39 references

Diagonal dimension and intermediate sub-C*-algebras

T0 review · 2 major / 6 minor · reviewed 2026-07-31 · grok-4.5

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

Pith's one-line read Finiteness of diagonal dimension passes to every intermediate sub-C*-algebra sitting over a C*-diagonal.

desk verdict Solid permanence result for diagonal dimension that cleanly upgrades Archbold–Kumjian; the product factor with dim ˆD is the only real soft spot, and the author flags it. read the letter →

arxiv 2607.24699 v1 pith:BSI2EQYS submitted 2026-07-27 math.OA math.FA

Grigoris Kopsacheilis This is my paper · ORCID
classification math.OAmath.FA MSC 46L0546L35
keywords diagonaldimensionC*-diagonalsintermediatesubalgebrasnuclearAF-diagonalsorder-zeromapscovering
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 proves that if a C*-pair has finite diagonal dimension, then every intermediate algebra that still contains the diagonal also has finite diagonal dimension. The precise bound multiplies the ambient diagonal dimension by one plus the covering dimension of the diagonal’s spectrum. When that spectrum is totally disconnected the dimension cannot increase at all, recovering and strengthening the classical fact that intermediate algebras over AF-diagonals remain AF. The result matters because finite diagonal dimension forces finite nuclear dimension, so many intermediate algebras become classifiable once the ambient pair is. The argument works by building approximate conditional expectations that preserve intermediate algebras and then computing the exact diagonal dimension of the “one-map” building blocks that appear in any finite-dimensional approximation.

What carries the argument

Approximate conditional expectations Ψ : A → C*(D, φ(F)) that fix D, preserve every intermediate algebra, and approximate the identity on the image of a single normalizer-preserving order-zero map φ; combined with the exact computation that any intermediate algebra inside such a C*(D, φ(F)) has diagonal dimension equal to dim ˆD.

What would settle it

Exhibit a concrete C*-diagonal whose spectrum has positive covering dimension together with an intermediate algebra whose diagonal dimension strictly exceeds that of the ambient pair, or prove that no such example exists and the factor dim⁺¹(ˆD) can be dropped.

Watch

Extended reading notes

Core claim

For a C*-diagonal (D ⊂ A) with D separable and any intermediate D ⊂ B ⊂ A, one has dim⁺¹_diag(D ⊂ B) ≤ dim⁺¹(ˆD) · dim⁺¹_diag(D ⊂ A). In particular, when ˆD is totally disconnected the inequality simplifies to dim_diag(D ⊂ B) ≤ dim_diag(D ⊂ A), and the zero-dimensional case recovers that intermediate algebras of AF-diagonals are AF.

Load-bearing premise

The product bound relies on the exact one-map calculation that intermediate algebras inside C*(D, φ(F)) always have diagonal dimension equal to the covering dimension of the diagonal’s spectrum; the paper itself notes that it is unknown whether that factor can be removed.

Editorial extensions

If this is right

  • Any intermediate algebra over a pair of finite diagonal dimension itself has finite nuclear dimension.
  • When that intermediate algebra is simple it falls inside the classification theorem for unital separable nuclear UCT C*-algebras.
  • When the diagonal has totally disconnected spectrum, diagonal dimension is monotone under passage to intermediate algebras.
  • The zero-dimensional case strengthens Archbold–Kumjian: intermediate algebras of AF-diagonals remain AF-diagonals, not merely AF algebras.

Reading between the lines

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

  • The same permanence could be used to distinguish non-conjugate diagonals with homeomorphic spectra by comparing the diagonal dimensions of their intermediate algebras.
  • A matching lower-bound example or a proof that dynamic asymptotic dimension is monotone for wide open subgroupoids would decide whether the factor dim⁺¹(ˆD) is essential.
  • The coordinate-system description of intermediate algebras inside one-map images may extend to give explicit nuclear-dimension bounds for more general Cartan inclusions.
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.

Referee Report

2 major / 6 minor

Summary. The paper studies the diagonal dimension of Li–Liao–Winter for intermediate sub-C*-algebras of a C*-diagonal (D⊂A). The main result (Theorem A/4.1) shows that for D separable and D⊂B⊂A intermediate, dim+1_diag(D⊂B) ≤ dim+1(Ď)·dim+1_diag(D⊂A); in particular finite diagonal dimension passes to intermediate subalgebras, and when Ď is totally disconnected the bound simplifies to dim_diag(D⊂B) ≤ dim_diag(D⊂A), recovering and strengthening the Archbold–Kumjian theorem that intermediate subalgebras of AF-diagonals are AF. The proof proceeds by (i) constructing approximate conditional expectations Ψ: A → C*(D, φ(F)) that preserve every intermediate subalgebra and fix D (Proposition 2.3), (ii) computing exactly the diagonal dimension of any intermediate pair inside a C*-algebra generated by D and a single normalizer-preserving order zero map: dim_diag(D⊂B) = dim Ď (Theorem C/3.6), via a coordinate description of B∩J (Lemma 3.1, Corollary 3.2) and a synchronized (d+1)-colored partition of unity on Ď (Proposition 3.4), and (iii) assembling the (d+1) order-zero pieces in Theorem 4.1. A corollary (Corollary B) gives finite nuclear dimension, and classifiability in the simple unital separable case, for all such intermediate algebras.

Significance. If correct — and I find no load-bearing error — this is a solid and useful contribution to the structure theory of C*-diagonals. It is the natural dimension-theoretic generalization of the Archbold–Kumjian theorem, gives a concrete, falsifiable numerical bound (Theorem A), and pins down exactly where the dim+1(Ď) factor comes from via the sharp computation dim_diag(D⊂B) = dim Ď for intermediate algebras inside C*(D, φ(F)) (Theorem C). Corollary B has immediate classification consequences (finite nuclear dimension, and classifiability in the simple case, for all intermediate subalgebras), and the paper openly identifies the sharpness question (Remark 4.2) and its relation to monotonicity of dynamic asymptotic dimension, which frames a clear direction for future work. The techniques — the explicit approximate expectations of Proposition 2.3, the coordinate description of Lemma 3.1/Corollary 3.2, and the synchronized (d+1)-colored partition of unity of Proposition 3.4 — are likely to be reusable elsewhere.

major comments (2)
  1. [Theorem 4.1, footnote 2] Footnote 2 (p. 23): the proof of Theorem 4.1 requires a contraction y^(i)_{d_i} lying in the finite-dimensional masa ⊕_j D_{\tilde F(i,j)}, and the footnote asserts this is possible 'by inspection of the proof' of the relevant direction of [24, Proposition 2.3] without further detail. Since [24] is a preprint and this point is load-bearing for the final estimate (4.1) — the element y^(i)_{d_i} plays the role of the approximate unit image in the colored decomposition — the manuscript should either include the short argument (presumably: the contraction produced by [24] can be replaced by its image under the conditional expectation onto the masa without harming (4.6)–(4.7)) or state a small lemma. This is a local fix, but as written the proof of Theorem 4.1 has a citation-level gap.
  2. [Theorem 3.6] In the proof of Theorem 3.6 (unital case), the verification of the criterion of [24, Proposition 2.3] via (3.35)–(3.36) uses only the maps ϑ^(c) and the elements x^(c)_b, with no explicit compression map ψ: B → ⊕_c E^(c) constructed in the text. Presumably [24, Proposition 2.3] is stated in exactly this 'elements in the finite-dimensional algebra' form (as is also used in Theorem 4.1, eq. (4.2)–(4.3)), but a one-sentence reminder of the precise hypothesis of that criterion at its first use would remove any ambiguity about what is being verified.
minor comments (6)
  1. [Title page] The running head/title in the arXiv rendering reads 'DIAGONAL DIMENSION AND INTERMEDIA TE SUB-C*-ALGEBRAS' — a line-break artifact in the typesetting of the title page that should be fixed.
  2. [Introduction / Theorem A] The notation dim+1(·) for dim(·)+1 is economical but easy to misread in displayed formulas such as (4.1); consider defining it once more near Theorem A (it is defined in the introduction) or writing the bound as dim_diag(D⊂B)+1 ≤ (dim Ď+1)(dim_diag(D⊂A)+1).
  3. [Theorem 3.6, proof] p. 19: the appeal to [25, Lemma 1.4] for the statement that a sum of normalizer-preserving order zero maps with pairwise orthogonal ranges is normalizer-preserving should be double-checked for an exact match of hypotheses; a brief parenthetical indication of which lemma in [25] is being used (and in what form) would help the reader.
  4. [Proposition 3.4] Proposition 3.4, proof: the continuity argument for g_(W,j) after (3.17) ('if x ∈ D_i then D_i ∩ ∂D_j is nonempty') implicitly uses that X is (locally) connected or that the D_j are chosen with disjoint closures; as stated, D_i ∩ D_j ≠ ∅ only follows if D_i meets ∂D_j, which need not imply D_i ∩ D_j ≠ ∅ without an extra step. The conclusion is surely fine since f_W(q(x)) = 0 on the boundary of the saturated preimage by continuity of f_W ∘ q and the fact that q^{-1}(U_W) is open with f_W ∘ q vanishing off it; a cleaner phrasing (e.g., 'f_W ∘ q vanishes on X \ q^{-1}(U_W), hence on ∂D_j ⊂ X \ q^{-1}(U_W)') would avoid the detour.
  5. [Remark 4.2] Remark 4.2: the sentence 'dynamic asymptotic dimension seems to be monotone with respect to wide open subgroupoids [5]' cites [5] for a statement the author describes only as apparent; please either give a precise reference (theorem number) or soften to 'is expected to be'.
  6. [References] Reference [24] (Li–Liao–Winter) is the foundation for Definition 1.2 and several cited results; if it has appeared or been updated since arXiv:2303.16762, the bibliographic data should be refreshed at proof stage.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the bound is derived from the definition of diagonal dimension via independent constructions

full rationale

The paper proves an upper bound on diagonal dimension for intermediate pairs by assembling approximate conditional expectations (Prop. 2.3), a coordinate description of intermediate algebras inside C*(D,φ(F)) (Cor. 3.2 / Lem. 3.1), a synchronized colored partition of unity (Prop. 3.4), and an exact computation that those single-map intermediate pairs have dim_diag equal to dim ÊD (Thm. 3.6 / Thm. C). These steps are constructive and self-contained against the definition of diagonal dimension (Def. 1.2) and standard tools (order-zero calculus, unique conditional expectations on C*-diagonals, covering-dimension partitions). Prior results (Li–Liao–Winter, Archbold–Kumjian, Donsig–Pitts, Brown–Exel–Fuller–Pitts–Reznikoff) are used as external inputs, not as restatements of the target inequality. Remark 4.2 explicitly flags that the factor dim+1(ÊD) may not be sharp and that no example of dimension increase is known; that is a sharpness limitation, not a circular reduction. Nothing in the derivation chain equals its inputs by construction.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

Pure operator-algebra theorem. No fitted parameters. Load-bearing inputs are the definition and basic permanence properties of diagonal dimension (Li–Liao–Winter), the classical theory of C*-diagonals/Cartan masas, order-zero functional calculus, covering dimension of compact metric spaces, and the Donsig–Pitts normalizer lemma used to keep intermediate algebras invariant under the approximate expectations.

assumptions (7)
  • domain assumption Definition and basic properties of diagonal dimension dim_diag(D⊂A), including the approximation criterion of Li–Liao–Winter Prop. 2.3 and the fact that dim_diag=0 characterises AF-diagonals.
    Invoked throughout; the ambient approximation by d+1 normalizer-preserving order-zero maps is the starting point of the proof of Theorem 4.1.
  • domain assumption A C*-diagonal admits a unique faithful conditional expectation onto D, and normalizers generate A; pure states on D extend uniquely.
    Definition 1.1; used for Φ in Prop. 2.3 and for normalizer calculus.
  • standard math Order-zero maps admit a supporting *-homomorphism and functional calculus that preserves the normalizer-preserving property (Winter–Zacharias; [19, Rem. 4.2]).
    Used to build Ψ in (2.1) and to keep φ^β normalizer-preserving.
  • domain assumption Donsig–Pitts Prop. 3.10: Φ(bv)v* lies in B for b∈B and v∈N_A(D), for intermediate D⊂B⊂A.
    Critical for Ψ(B)⊂B in Prop. 2.3(i); without it the approximate expectations would not preserve intermediates.
  • standard math Ostrand’s theorem / standard covering-dimension partitions of unity for compact metric spaces of dimension ≤d.
    Used in Prop. 3.4 to produce the (d+1)-colorable partition of unity on the quotient space Y.
  • domain assumption Intermediate C*-algebras of C*-diagonals are again C*-diagonals (Brown–Exel–Fuller–Pitts–Reznikoff).
    Cited in the introduction and Remark 3.3; ensures the intermediate pairs are legitimate diagonal pairs, though the dimension bound is proved directly.
  • domain assumption D separable (and A separable in Cor. B); spectra treated as compact/locally compact metric spaces.
    Stated in Theorem A and used to apply metric covering-dimension technology.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Diagonal dimension and intermediate sub-C*-algebras." pith.science (2026). https://pith.science/paper/BSI2EQYS

@misc{pith2026260724699,
  author       = {Pith},
  title        = {Pith review of: Diagonal dimension and intermediate sub-C*-algebras},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BSI2EQYS}},
  note         = {Machine review of arXiv:2607.24699}
}
read the original abstract

We show that finiteness of diagonal dimension, in the sense of Li--Liao--Winter, passes to intermediate sub-C*-algebras. More specifically, we obtain an upper bound for the diagonal dimension of the pair of an intermediate sub-C*-algebra in a C*-diagonal together with the diagonal, which we estimate in terms of the diagonal dimension of the ambient pair and the covering dimension of the diagonal's spectrum. This generalizes a theorem of Archbold and Kumjian stating that intermediate sub-C*-algebras of AF-diagonals are AF C*-algebras.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 2 linked inside Pith

  1. [24]

    K. Li, H. Liao and W. Winter. The diagonal dimension of sub-C ∗-algebras. Preprint, arXiv:2303.16762 (2023)

    arXiv 2023

  2. [1]

    R. J. Archbold. Extensions of pure states and projections of norm one.J. Funct. Anal.165(1) (1999), 24–43

    1999

  3. [2]

    R. J. Archbold and A. Kumjian. Nuclear subalgebras of UHF algebras.Proc. Edinb. Math. Soc.29(1) (1986), 97–100

    1986

  4. [3]

    Barlak and X

    S. Barlak and X. Li. Cartan subalgebras and the UCT problem.Adv. Math.316 (2017), 758–769

    2017

  5. [4]

    Blackadar.Operator Algebras: Theory ofC ∗-algebras and von Neumann algebras

    B. Blackadar.Operator Algebras: Theory ofC ∗-algebras and von Neumann algebras. Encyclopaedia of Mathematical Sciences122. Springer, Berlin (2006)

    2006

  6. [5]

    B¨ onicke

    C. B¨ onicke. On the dynamic asymptotic dimension of ´ etale groupoids.Math. Z. 307(1) (2024), 16

    2024

  7. [6]

    B¨ onicke and K

    C. B¨ onicke and K. Li. Nuclear dimension of subhomogeneous twisted groupoid C ∗- algebras and dynamic asymptotic dimension.Int. Math. Res. Not. IMRN2024(16) (2024), 11597–11610

    2024

  8. [7]

    Brown, R

    J. Brown, R. Exel, A. Fuller, D. R. Pitts and S. Reznikoff. Intermediate C ∗-algebras of Cartan embeddings.Proc. Amer. Math. Soc. Ser. B8(3) (2021), 27–41

    2021

Show all 39 references
  1. [8]

    Inter- mediate C ∗-algebras of Cartan embeddings

    J. Brown, R. Exel, A. Fuller, D. R. Pitts and S. Reznikoff. Corrigendum to “Inter- mediate C ∗-algebras of Cartan embeddings”.Proc. Amer. Math. Soc. Ser. B11(7) (2024), 75–76

    2024

  2. [9]

    Crytser and G

    D. Crytser and G. Nagy. Traces arising from regular inclusions.J. Aust. Math. Soc. 103(2) (2017), 190–230

    2017

  3. [10]

    R. J. Deeley, I. F. Putnam and K. R. Strung. Classifiable C ∗-algebras from minimal Z-actions and their orbit-breaking subalgebras.Math. Ann.388(1) (2024), 703–729

    2024

  4. [11]

    A. P. Donsig and D. R. Pitts. Coordinate systems and bounded isomorphisms.J. Operator Theory59(2) (2008), 359–416

    2008

  5. [12]

    Engelking.Dimension Theory

    R. Engelking.Dimension Theory. North-Holland Mathematical Library19, North- Holland Publishing Co., Amsterdam (1978)

    1978

  6. [13]

    Engelking.General Topology

    R. Engelking.General Topology. Sigma Series in Pure Mathematics,6Heldermann, Berlin, Germany (1989)

    1989

  7. [14]

    Guentner, R

    E. Guentner, R. Willett and G. Yu. Dynamic asymptotic dimension: relation to dynamics, topology, coarse geometry, and C ∗-algebras.Math. Ann.367(1) (2017), 785–829

    2017

  8. [15]

    Hirshberg, W

    I. Hirshberg, W. Winter and J. Zacharias. Rokhlin dimension and C ∗-dynamics. Comm. Math. Phys.335(2) (2015), 637–670

    2015

  9. [16]

    D. Kerr. Dimension, comparison and almost finiteness.J. Eur. Math. Soc. (JEMS) 22(11) (2020), 3697–3745

    2020

  10. [17]

    Kirchberg and W

    E. Kirchberg and W. Winter. Covering dimension and quasidiagonality.Int. J. Math. 15(1) (2004), 63–85

    2004

  11. [18]

    F. Komura. A correspondence between inverse subsemigroups, open wide sub- groupoids and Cartan intermediate C ∗-subalgebras.Proc. Edinb. Math. Soc.65(3) (2022), 861–879. 26 GRIGORIS KOPSACHEILIS

    2022

  12. [19]

    Kopsacheilis and W

    G. Kopsacheilis and W. Winter. Diagonal comparison of ample C ∗-diagonals.Int. Math. Res. Not. IMRN2025(10) (2025), rnaf113

    2025

  13. [20]

    Kopsacheilis and W

    G. Kopsacheilis and W. Winter. Paper-folding models for the CAR algebra.Ergod. Theory Dyn. Syst., First View (2026), 1–25, doi:10.1017/etds.2026.10312

    2026

  14. [21]

    A. Kumjian. On C ∗-diagonals.Canad. J. Math.38(4) (1986), 969–1008

    1986

  15. [22]

    E. C. Lance.HilbertC ∗-modules: a toolkit for operator algebraists. London Mathe- matical Society Lecture Note Series,210. Cambridge University Press, Cambridge (1995)

    1995

  16. [23]

    X. Li. Every classifiable simple C ∗-algebra has a Cartan subalgebra.Invent. Math. 219(2) (2020), 653–699

    2020

  17. [25]

    Liao and A

    H. Liao and A. Tikuisis. Almost finiteness, comparison, and tracialZ-stability.J. Funct. Anal.282(3) (2022), 109309

    2022

  18. [26]

    J. R. Munkres.Topology(2nd edition). Prentice Hall, Upper Saddle River, NJ (2000)

    2000

  19. [27]

    D. R. Pitts. Normalizers and approximate units for inclusions of C ∗-algebras.Indiana Univ. Math. J.72(5) (2023), 1849–1866

    2023

  20. [28]

    I. F. Putnam. The C ∗-algebras associated with minimal homeomorphisms of the Cantor set.Pac. J. Math.136(2) (1989), 329–353

    1989

  21. [29]

    J. Renault. Cartan subalgebras in C ∗-algebras.Irish Math. Soc. Bull.61(2008), 29–63

    2008

  22. [30]

    Schafhauser, A

    C. Schafhauser, A. Tikuisis and S. White. Nuclear C ∗-algebras: 99 problems. To ap- pear in Special Issue of M¨ unster J. Math. in memory of Eberhard Kirchberg. Preprint, arXiv:2506.10902 (2025)

    2025 arXiv

  23. [31]

    A. Sims, G. Szab´ o and D. P. Williams.Operator Algebras and Dynamics: Groupoids, Crossed Products, and Rokhlin Dimension. Springer International Publishing, 2020

    2020

  24. [32]

    Str˘ atil˘ a and D

    S. Str˘ atil˘ a and D. Voiculescu.Representations of AF algebras and of the GroupU(∞). Lecture Notes in Math.486, Springer-Verlag, Berlin-Heidelberg-New York, 1975

    1975

  25. [33]

    G. Szab´ o. The Rokhlin dimension of topologicalZ m-actions.Proc. Lond. Math. Soc. (3)110(3) (2015), 673–694

    2015

  26. [34]

    S. White. Abstract classification theorems for amenable C ∗-algebras.Proc. Int. Congr. Math.,4(2022), 3314–3338. EMS Press (2023)

    2022

  27. [35]

    W. Winter. Covering dimension for nuclear C ∗-algebras.J. Funct. Anal.199(2003), 535–556

    2003

  28. [36]

    W. Winter. Decomposition rank of subhomogeneous C∗-algebras.Proc. London Math. Soc.89(2) (2004), 427–456

    2004

  29. [37]

    W. Winter. Structure of nuclear C ∗-algebras: from quasidiagonality to classification and back again.Proc. Int. Congr. Math.2(2018), 1797–1820. World Scientific, Sin- gapore (2019)

    2018

  30. [38]

    Winter and J

    W. Winter and J. Zacharias. Completely positive maps of order zero.M¨ unster J. Math.2(2009), 311–324

    2009

  31. [39]

    Winter and J

    W. Winter and J. Zacharias. The nuclear dimension of C∗-algebras.Adv. Math.224(2) (2010), 461–498. Grigoris Kopsacheilis, KU Leuven Department of Mathematics, Celestij- nenlaan 200B box 2400, BE-3001 Leuven, Belgium. Email address:gkopsach@kuleuven.be URL:https://sites.google....

    2010

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.