REVIEW 1 major objections 3 minor 3 cited by
Paper-folding models for the CAR algebra
T0 review · 1 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read The CAR algebra contains countably many non-standard Cantor-spectrum diagonals, one for each diagonal dimension from 0 to infinity.
desk verdict Genuinely new construction: paper-folding subshift yields non-AF Cantor diagonals in the CAR algebra and countably many diagonal-dimension values; two small technical glitches, both repairable. 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 paper-folding subshift $X$: the closure of bi-infinite binary sequences whose finite words all appear in the regular paper-folding sequence $\mathbf{t}$. Its two features doing all the work are (1) invariance of the language under anti-reversal, so $\sigma(x)_j=1-x_{-j}$ maps $X$ to itself and anti-commutes with the shift, and (2) the absence of anti-palindromes longer than 6, which makes the $\mathbb{Z}\rtimes\mathbb{Z}_2$ action free. K-theory is computed by a four-letter substitution $\varrho$ (the block code of 2-bit blocks), via dimension-group methods, and then by a theorem from [47] for crossed products of Cantor minimal $\mathbb{Z}\rtimes\mathbb{Z}_2$ systems. Classification theo
What would settle it
Scan the initial segments of the paper-folding sequence for an 8-letter anti-palindrome, i.e. a word $w$ equal to its bit-swapped reversal, such as $01101001$. The proof asserts that none occurs and that the absence propagates under the recursive rule (1.5); finding one as a subword of any $t_n$ would give a fixed point of $\varphi\sigma$, making the action non-free and invalidating the construction of the diagonal.
Extended reading notes
Core claim
The central discovery is that the language of the paper-folding sequence $\mathbf{t}$ is closed under anti-reversal (bit-swap plus reversal) and contains no anti-palindromic word of length exceeding 6. Consequently the shift $\varphi$ together with the anti-reversal involution $\sigma$ gives a free minimal action $\mathbb{Z}\rtimes\mathbb{Z}_2\curvearrowright X$ on the paper-folding subshift $X\subset\{0,1\}^{\mathbb{Z}}$, which is a Cantor space. The authors compute the ordered $K_0$ of the crossed product $C(X)\rtimes(\mathbb{Z}\rtimes\mathbb{Z}_2)$ as $\mathbb{Z}_2\oplus\mathbb{Z}[\tfrac12]$ with the positive cone determined by the dyadic rationals, and $K_1=0$, using a four-letter substi
Load-bearing premise
The load-bearing premise is that the paper-folding sequence contains no anti-palindromic subword of length greater than 6, so the anti-reversal involution has no fixed point in the subshift and the $\mathbb{Z}\rtimes\mathbb{Z}_2$ action is free; if arbitrarily long anti-palindromes appeared, the pair would still be Cartan but not a $C^*$-diagonal.
Editorial extensions
If this is right
- The CAR algebra contains at least countably many pairwise non-conjugate $C^*$-diagonals with Cantor spectrum, indexed by diagonal dimension $n\in\{0,1,\dots,\infty\}$.
- A UHF algebra, which cannot be an integer crossed product because $K_1$ would be nonzero, can nonetheless harbour diagonals built from dynamical systems—here the infinite dihedral group acting on the paper-folding subshift.
- The example answers negatively the question from [43, Problems XLVII–XLVIII] whether every Cantor-spectrum $C^*$-diagonal in $M_{2^\infty}$ is conjugate to the AF diagonal.
- It resolves the problem raised in [31, Remark 6.10]: an AF inclusion can have nonzero diagonal dimension even when the diagonal itself is AF.
- Tensor products of the non-AF diagonal with the standard diagonal give diagonals inside $M_{2^\infty}$ with arbitrary finite diagonal dimension.
Reading between the lines
- The same substitution/anti-reversal construction may work for other self-similar binary sequences with bounded anti-palindrome length, producing further non-AF Cantor-spectrum diagonals in $M_{2^\infty}$ or in other UHF algebras; this is a natural next step suggested by the method.
- Diagonal dimension is a computable numerical invariant that separates the constructed diagonals, but it is unlikely to be a complete invariant; finer invariants such as groupoid cohomology (along the lines of Matui's HK conjecture) might classify the full family.
- The construction indicates that strongly self-absorbing algebras generally may carry many distinct Cantor-spectrum diagonals; if the same phenomenon holds for the Jiang–Su algebra or Cuntz algebras with a prescribed spectrum, it would reinforce the view that existence of Cartan subalgebras is a subtle, non-unique feature even in classifiable algebras.
- A testable extension: replace the regular paper-folding sequence by other folding sequences (left/right patterns) and check whether the anti-palindrome bound persists; each such sequence would give a candidate free $\mathbb{Z}\rtimes\mathbb{Z}_2$ action and hence a new diagonal in the CAR algebra, possibly with different K-theoretic data.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper constructs, for each n in {0,1,2,...,∞}, a C*-diagonal (D ⊂ M_{2^∞}) with Cantor spectrum and diagonal dimension n. For n=1 this gives a Cantor-spectrum diagonal in the CAR algebra that is not conjugate to the standard AF diagonal, answering a question of Blackadar and, for general n, a problem raised in [31, Remark 6.10]. The construction uses the paper-folding subshift X, the free minimal Z⋊Z2 action generated by the shift and anti-reversal, and a substitution subshift X_ϱ used to compute the K-theory of C(X)⋊Z. The crossed product A=C(X)⋊(Z⋊Z2) is shown to be classifiable with K0(A)≅Z2⊕Z[1/2]; tensoring with M_{2^∞} and applying classification gives B≅M_{2^∞}, while the intermediate non-AF subalgebra obstruction shows the diagonal is not AF. Diagonal dimension theory then distinguishes the tensor powers D^{⊗n}. The main technical proof is the K-theory computation in Section 2, and the main gap is in Lemma 2.3, where the clopen partition X=X^(0)⊔X^(1) is not justified as written.
Significance. If the technical issue in Lemma 2.3 is repaired, the results are significant. They provide the first Cantor-spectrum C*-diagonals in the CAR algebra that are not AF diagonals, and they produce countably many pairwise non-conjugate such diagonals distinguished by diagonal dimension. The construction is explicit, the dependence on classification theory is carefully documented, and the use of the paper-folding sequence is natural and self-contained. The paper also gives a satisfying resolution of [31, Remark 6.10]. The proofs are largely built from established theorems (classification, Künneth, diagonal dimension, Thomsen's K-theory of Z⋊Z2 crossed products), and the paper is careful about pointing out where quoted results are used. The main issue is local but load-bearing: Lemma 2.3 must be rewritten before the K-theory computation is rigorous.
major comments (1)
- [§2, Lemma 2.3 (Eqs. (2.17)-(2.18), claims (ii)-(iii))] The sets X^(0) and X^(1) are defined as raw φ²-orbits, hence are countable. Claim (ii) asserts X^(0)⊔X^(1)=X with both clopen; a countable subset of a Cantor space is not clopen unless finite, and a non-periodic orbit is not closed. Claim (iii) asserts β(X^(0))=X_ϱ, impossible since β(X^(0)) is countable while X_ϱ is a Cantor space. The proof should pass to orbit closures. This matters because Proposition 2.4 uses p=χ_{X^(0)} as a clopen projection and identifies A≅M_2(B); all subsequent K-theory and Theorems 3.1 and 3.4 rely on this step.
minor comments (3)
- [Theorem 3.1] K_1((C(X)⋊_φ Z)⊗M_{2^∞}) is Z[1/2], not Z, by the Künneth formula and K_1(C(X)⋊_φ Z)≅Z; since the argument only needs nonvanishing K_1, this is a local error.
- [Lemma 2.3] After the orbit-closure correction, the reduction to k∈N in (2.19) should be justified; citing [39, Lemma 5.2] is insufficient in the new formulation.
- [Proposition 1.2(iii)] The finite checks for t4 and t̂3 1 t3 are asserted as 'directly seen'; an explicit verification table would help.
Circularity Check
No circularity: the construction and computations are self-contained; the only flagged concern is a non-circular technical gap in Lemma 2.3.
full rationale
The derivation chain is self-contained relative to external results. Proposition 1.2 proves the anti-palindrome bound in the text, and Corollary 1.5 follows from Lemma 1.1 and Proposition 1.3. The K-theory of the auxiliary substitution subshift is computed from the explicit matrix M via Durand–Host–Skau [18, Cor. 33] and the inductive limit is identified directly in Proposition 2.2. The paper-folding subshift is connected to the substitution subshift through the explicit coding β in Lemma 2.3/Proposition 2.4. The Z⋊Z2 crossed-product K-theory uses Thomsen [47], Putnam [40], and Bratteli–Evans–Kishimoto [12]/Natsume [36], all external; Lemma 2.6 establishes nontriviality of σ by a direct orbit-sum argument. The final tensor-product step uses the Künneth formula and classification theory [53,51], plus Power [38] for AF-diagonal uniqueness and Archbold–Kumjian [1] for intermediate subalgebras; none of these presupposes the theorem. The self-citation to Li–Liao–Winter [31] supplies the definition and general structural properties of diagonal dimension; it is a framework, not a fitted input tailored to force the conclusion, and the constructed pairs are new. I therefore find no circular reduction. Separately, I flag a non-circular correctness concern: Lemma 2.3 defines X^(0) and X^(1) as literal φ²- and φ^{2k+1}-orbits of a point and claims they are clopen and union to X; as written, orbits in a Cantor space are countable and not closed, so the proof must intend orbit closures. This affects Proposition 2.4's clopen projection p and hence the K-theory chain, but it is a missing justification rather than a circularity, so it does not change the circularity score.
Assumptions & free parameters
assumptions (6)
- domain assumption Elliott classification theorem: simple, separable, unital, nuclear, Z-stable C*-algebras satisfying UCT are classified by ordered K0 with order unit and tracial states (Tikuisis-White-Winter [51], Winter [53]).
- standard math Künneth formula for K-theory of minimal tensor products (Blackadar [5], Schochet [44]).
- domain assumption Thomsen's theorem: K0 of a crossed product by a free minimal Z ⋊ Z2 action has the form Z2 ⊕ (1+σ_*)(C(Ω,Z)/im(1−φ_*)) (Thomsen [47, Theorem 4.42]).
- domain assumption Durand-Host-Skau computation of K0 of primitive substitution subshifts (Corollary 33 in [18]).
- domain assumption Archbold-Kumjian: intermediate sub-C*-algebras of an AF diagonal are AF ([1]).
- domain assumption Li-Liao-Winter diagonal dimension theory, including equality with dynamic asymptotic dimension for Cantor spaces ([31, Theorem 5.4]) and monotonicity under subgroups.
Cite this review
Pith. "Pith review of Paper-folding models for the CAR algebra." pith.science (2026). https://pith.science/paper/HFBC7GWE
@misc{pith2026250804837,
author = {Pith},
title = {Pith review of: Paper-folding models for the CAR algebra},
year = {2026},
howpublished = {\url{https://pith.science/paper/HFBC7GWE}},
note = {Machine review of arXiv:2508.04837}
}
read the original abstract
We show that the CAR algebra admits a Cantor spectrum C*-diagonal that is not conjugate to the standard AF diagonal. We obtain this by classification theory of C*-algebras, and the diagonal arises by realising the CAR algebra as the crossed product of a free minimal action on the Cantor space, where the acting group is the product of a locally finite group with the infinite dihedral group. The main ingredient in the construction is a binary subshift associated to the well-known regular paper-folding sequence. Moreover, we show that the CAR algebra in fact admits countably many, pairwise non-conjugate, Cantor spectrum diagonals which are distinguished by the different values of their diagonal dimension, as defined by Li, Liao and the second named author.
Forward citations
Cited by 3 Pith papers
-
A C*-diagonal in the Jiang-Su algebra via entangled matrix cones
An explicit inductive system of entangled dimension-drop algebras realises Z and yields a C*-diagonal with one-dimensional non-locally-connected spectrum, via a new normaliser characterisation by state excision.
-
Functoriality and Weyl Groupoids of Ample C*-Diagonal Pairs
The paper initiates a functorial study by constructing induced partial morphisms on Weyl groupoids from morphisms of ample C*-diagonal pairs and proves applications including tensor product identification and subaddit...
-
On a C*-Diagonal Generated by the Toric Code
The toric-code stabilizer algebra is a C*-diagonal of M_{2^∞} equivalent to the canonical diagonal.
Reference graph
Works this paper leans on
-
[1]
R. J. Archbold and A. Kumjian. Nuclear subalgebras of UHF algebras. Proc. Edinb. Math. Soc. 29(1) (1986), 97–100
work page 1986
-
[2]
R. J. Archbold and J. S. Spielberg. Topologically free actions and ideals in discrete C∗-dynamical systems. Proc. Edinb. Math. Soc. 37(1) (1994), 119–124
work page 1994
-
[3]
S. Barlak and X. Li. Cartan subalgebras and the UCT problem. Adv. Math. 316 (2017), 758–769
work page 2017
-
[4]
Witten multiple zeta values attached to sl(4)
J. Berstel, L. Boasson, O. Carton and I. Fagnot. Infinite words without palindrome. ArXiv preprint, arXiv:0903.2383 (2009)
work page Pith review arXiv 2009
- [5]
- [6]
- [7]
- [8]
Show all 53 references
-
[9]
B¨ onicke, C
C. B¨ onicke, C. Dell’Aiera, J. Gabe and R. Willett. Dynamic asymptotic dimension and Matui’s HK conjecture. Proc. Lond. Math. Soc. 126(4) (2023), 1182–1253
2023
-
[10]
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. IMRN 2024(16) (2024), 11597–11610
2024
-
[11]
Boyle and D
M. Boyle and D. Handelman. Entropy versus orbit equivalence for minimal homeo- morphisms. Pac. J. Math. 164(1) (1994), 1–13
1994
-
[12]
Bratteli, D
O. Bratteli, D. E. Evans and A. Kishimoto. Crossed products of totally disconnected spaces by Z2 ∗ Z2. Ergod. Theory Dyn. Syst. 13(3) (1993), 445–484
1993
-
[13]
H. Bruin. Topological and ergodic theory of symbolic dynamics. American Mathemat- ical Society 228, Providence, RI (2022)
2022
-
[14]
Castillejos, S
J. Castillejos, S. Evington, A. Tikuisis, S. White and W. Winter. Nuclear dimension of simple C ∗-algebras. Invent. Math. 224(1) (2021), 245–290
2021
-
[15]
R. J. Deeley, I. F. Putnam and K. R. Strung. Constructing minimal homeomorphisms on point-like spaces and a dynamical presentation of the Jiang–Su algebra. J. Reine Angew. Math. 742 (2018), 241–261
2018
-
[16]
F. M. Dekking, M. M. France and A. J. van der Poorten. Folds! I–III. Math. Intell. (1982) 4(3) 130–138, 4(4) 173–181, 4(4) 190–195
1982
-
[17]
Downarowicz and G
T. Downarowicz and G. Zhang. Symbolic extensions of amenable group actions and the comparison property. Mem. Amer. Math. Soc. 281 (2023), 1–108. 24 GRIGORIS KOPSACHEILIS AND WILHELM WINTER
2023
-
[18]
Durand, B
F. Durand, B. Host and C. Skau. Substitutional dynamical systems, Bratteli diagrams and dimension groups. Ergod. Theory Dyn. Syst. 19(4) (1999), 953–993
1999
-
[19]
E. G. Effros and C. Shen. Approximately finite C ∗-algebras and continued fractions. Indiana Univ. Math. J. 29(2) (1980), 191–204
1980
-
[20]
D. G. Evans and A. Sims. When is the Cuntz–Krieger algebra of a higher-rank graph approximately finite-dimensional? J. Funct. Anal. 263(1) (2012), 183–215
2012
-
[21]
Evington and P
S. Evington and P. Sibbel. C ∗-diagonals with Cantor spectrum in Cuntz algebras. ArXiv preprint, arXiv:2506.22163 (2025)
2025
-
[22]
Feldman and C
J. Feldman and C. Moore. Ergodic equivalence relations, cohomologies, von Neumann algebras. I, II. Trans. Amer. Math. Soc. (1977) 234 289–324, 234(2) 325–359
1977
-
[23]
N. P. Fogg. Substitution in dynamics, arithmetics and combinatorics. Lecture Notes in Mathematics 1794. Springer, Berlin (2002)
2002
-
[24]
L. Fuchs. Abelian groups. Pergamon Press, Oxford (1960)
1960
-
[25]
Giordano, I
T. Giordano, I. F. Putnam and C. F. Skau. Topological orbit equivalence and C ∗- crossed products. J. Reine Angew. Math. 469 (1995), 51–112
1995
-
[26]
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
-
[27]
D. Kerr. Dimension, comparison and almost finiteness. J. Eur. Math. Soc. (JEMS) 22(11) (2020), 3697–3745
2020
-
[28]
A. Kumjian. On C ∗-diagonals. Canad. J. Math. 38(4) (1986), 969–1008
1986
-
[29]
A. Kumjian. On the K-theory of the symmetrized non-commutative torus. C. R. Math. Rep. Acad. Sci. Canada 12(3) (1990), 87–90
1990
-
[30]
Kumjian and D
A. Kumjian and D. Pask. Higher rank graph C∗-algebras. New York J. Math. 6 (2000), 1–20
2000
-
[31]
K. Li, H. Liao and W. Winter. The diagonal dimension of sub-C ∗-algebras. ArXiv preprint, arXiv:2303.16762 (2023)
2023 arXiv
-
[32]
Li and J
X. Li and J. Renault. Cartan subalgebras in C ∗-algebras. Existence and uniqueness. Trans. Amer. Math. Soc. 372(3) (2019), 1985–2010
2019
-
[33]
H. Lin. The tracial topological rank of C ∗-algebras. Proc. Lond. Math. Soc. 83(1) (2001), 199–234
2001
-
[34]
H. Matui. Homology and topological full groups of ´ etale groupoids on totally discon- nected spaces. Proc. Lond. Math. Soc. 104(1) (2012), 27–56
2012
-
[35]
Mitscher and J
I. Mitscher and J. Spielberg. AF C ∗-algebras from non-AF groupoids. Trans. Amer. Math. Soc. 375(10) (2022), 7372–7371
2022
-
[36]
K-theory for certain group C ∗- algebras
T. Natsume. On K ∗(C∗(SL2(Z))) (Appendix to “K-theory for certain group C ∗- algebras” by E. C. Lance). J. Oper. Th. 13 (1985), 103–118
1985
-
[37]
Ortega and E
E. Ortega and E. Scarparo. Almost finiteness and homology of certain non-free ac- tions. Groups Geom. Dyn. 17(1) (2023), 77–90
2023
-
[38]
S. C. Power. Limit algebras: an introduction to subalgebras of C∗-subalgebras. Pitman Research Notices in Mathematics 278. Longman Scientific & Technical, Harlow
-
[39]
I. F. Putnam. Cantor minimal systems . American Mathematical Society 77, Provi- dence, RI (2018)
2018
-
[40]
I. F. Putnam. The C ∗-algebras associated with minimal homeomorphisms of the Cantor set. Pac. J. Math. 136(2) (1989), 329–353
1989
-
[41]
J. Renault. Cartan subalgebras in C ∗-algebras. Irish Math. Soc. Bull. 61 (2008), 29–63
2008
-
[42]
Rørdam and E
M. Rørdam and E. Størmer. Classification of Nuclear C∗-algebras. Entropy in Oper- ator Algebras. Encyclopaedia of Mathematical Sciences 126. Springer, Berlin (2002)
2002
-
[43]
Schafhauser, A
C. Schafhauser, A. Tikuisis and S. White. Nuclear C ∗-algebras: 99 problems. ArXiv preprint, arXiv:2506.10902 (2025)
2025 arXiv
-
[44]
Schochet
C. Schochet. Topological methods for C ∗-algebras. II. Geometry resolutions and the K¨ unneth formula.Pac. J. Math. 98(2) (1982), 443–458
1982
-
[45]
Sibbel and W
P. Sibbel and W. Winter. A Cantor spectrum diagonal in O2. Proc. Am. Math. Soc., to appear. ArXiv preprint, arXiv:2409.03511 (2024)
2024 arXiv
-
[46]
Tabachnikov
S. Tabachnikov. Dragon curves revisited. Math. Intell. 36 (2014), 13–17. PAPER-FOLDING MODELS FOR THE CAR ALGEBRA 25
2014
-
[47]
K. Thomsen. The homoclinic and heteroclinic C ∗-algebras of a generalized one- dimensional solenoid. Ergod. Theory Dyn. Syst. 30(1) (2010), 263–308
2010
-
[48]
Tikuisis, S
A. Tikuisis, S. White and W. Winter. Quasidiagonality of nuclear C ∗-algebras. Ann. of Math. (2) 185(1) (2017), 229–284
2017
-
[49]
A. S. Toms and W. Winter. Minimal dynamics and K-theoretic rigidity: Elliott’s conjecture. Geom. Funct. Anal. 23(1) (2013), 467–481
2013
-
[50]
A. Vershik. Nonmeasurable decompositions, orbit theory, algebras of operators. Dokl. Akad. Nauk. 199 (1971), 1004–1007
1971
-
[51]
S. White. Abstract classification theorems for amenable C ∗-algebras. Proc. Int. Congr. Math. , 4 (2022), 3314–3338. EMS Press, 2023
2022
-
[52]
W. Winter. Nuclear dimension and Z-stability of pure C ∗-algebras. Invent. Math. 187(2) (2012), 259–342
2012
-
[53]
W. Winter. Structure of nuclear C ∗-algebras: from quasidiagonality to classification and back again. Proc. Int. Congr. Math. 2 (2018), 1797–1820. World Scientific, 2019. Grigoris Kopsacheilis, Mathematisches Institut, F achbereich Mathematik und Informatik der Universit¨at M¨...
2018
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.