REVIEW 2 major objections 4 minor 27 references
The Ellis semigroup of bijective substitutions
T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read For primitive aperiodic bijective substitutions, the Ellis semigroup is a disjoint union of Z and a completely simple kernel described by a finite permutation group and an odometer extension, with one group-extension caveat.
desk verdict First real computation of Ellis semigroups for a broad class of non-tame substitution shifts, with an honest but load-bearing scope gap when generalised height exceeds classical height. 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 central object is the structural semigroup $M^{\mathrm{fib}}_0(X_\theta)$, the restriction of the fibre-preserving Ellis semigroup to the singular fibre over $0\in \mathbb Z_\ell$. The paper proves that this finite semigroup is isomorphic to $M[G_\theta; I_\theta, \{\pm\}, A]$, a Rees matrix semigroup in which $G_\theta$ is the group generated by all permutations appearing in the expansion of powers of $\theta$, $I_\theta$ consists of the ratios $\theta_i\theta_{i-1}^{-1}$, and $A$ is the sandwich matrix; the little structure group is the subgroup generated by the entries of $A$, and the generalised height is the order of the quotient of $G_\theta$ by its normal completion. Since the substitution system has exactly one orbit of singular fibres for the odometer factor, this finite computation is transported along the orbits to recover the full fibre-preserving semigroup, and then the kernel of $E(X_\theta)$ is recovered as an extension of the odometer by that fibre-preserving structure group. The machinery reduces a non-tame, uncountable Ellis semigroup to a finite permutation-group calculation plus one group extension.
What would settle it
Work out the unresolved example of Section 6.2(2) (length 7, $G_\theta=S_3$, $\overline{\Gamma}_\theta=A_3$, generalised height 2, classical height 1) and decide whether the extension $A_3^{\mathbb Z_7/\mathbb Z}\rtimes \mathbb Z/2\mathbb Z \to G_\theta \to \mathbb Z_7$ splits; if it is non-split, then $G_\theta$ is not the semidirect product required by Theorem 4.22 and the claimed Rees-matrix form of $M(X_\theta)$ does not hold for that substitution. More broadly, any primitive aperiodic bijective substitution with generalised height exceeding classical height whose odometer extension does not split would refute the unconditional version of the paper's central formula.
Extended reading notes
Core claim
The paper's central discovery is that the entire Ellis semigroup of a primitive aperiodic bijective substitution is governed by finite data. Concretely, $E(X_\theta)=\mathbb Z \sqcup M(X_\theta)$, with $M(X_\theta)$ algebraically isomorphic to the Rees matrix semigroup $M[G_\theta; I_\theta, \{\pm\}, A]$, where $G_\theta$ is generated by the permutations occurring in the powers of $\theta$, $I_\theta$ is the R-set of successive ratios $\theta_i\theta_{i-1}^{-1}$, and $A$ is the normalised sandwich matrix. The structure group $G_\theta$ is an extension of the odometer $\mathbb Z_\ell$ by the fibre-preserving structure group $G_\theta^{\mathrm{fib}}$, which itself is a semidirect product of a large power of the normal completion of the little structure group with $\mathbb Z/h\mathbb Z$. This description is proved when the generalised height $h$ equals the classical height; in the trivial-height case the isomorphism is topological, while in general it is algebraic. As an application the paper shows that the shift is semi-regular, with $V(X_\theta)\cong \mathrm{Aut}(X_\theta)\cong C_{S_A}(G_\theta)\times \mathbb Z$.
Load-bearing premise
The load-bearing premise is that the group extension which builds the structure group of the kernel can be split; when the substitution's generalised height is strictly larger than its classical height, the paper does not establish that this extension splits, so the claimed description of the Ellis semigroup is not proved in that case.
Editorial extensions
If this is right
- Every primitive aperiodic bijective substitution shift has a completely regular Ellis semigroup: $E(X_\theta)$ is the disjoint union of $\mathbb Z$ and its completely simple kernel $M(X_\theta)$.
- The full semigroup is effectively computable from the substitution's expansion: the finite structural semigroup $M[G_\theta;I_\theta,\{\pm\},A]$ determines the fibre-preserving part, and the kernel is determined up to the odometer extension.
- The substitution shift is semi-regular, so the virtual automorphism group, a group defined through the Ellis semigroup, coincides with the ordinary automorphism group.
- These systems are not tame, and the paper isolates the source: the structure group contains a product over the uncountably many odometer orbits, making $E(X_\theta)$ larger than the continuum even though every fibre restriction is finite.
- When the generalised height equals the classical height, the structure group is a semidirect product $G_\theta^{\mathrm{fib}}\rtimes \mathbb Z_\ell$, so the kernel has the same Rees-matrix form with this larger structure group.
Reading between the lines
- Beyond the paper, the same finite-fibre strategy could be tested on other unique-singular-orbit systems with an odometer-like equicontinuous factor; any such system whose singular-fibre restriction semigroup is generated by idempotents should admit an analogous Rees description.
- Beyond the paper, the generalised height looks like a group-theoretic invariant that can separate bijective substitutions with the same classical height and spectrum; the unresolved example with generalised height 2 and classical height 1 is a concrete place to look for a new dynamical invariant.
- Beyond the paper, the semi-regularity result suggests that virtual automorphism groups of point-distal non-distal systems may often be readable from the Ellis structure group alone, even where the full Ellis semigroup is not yet computed.
- A practical extension would be to implement the computation of $G_\theta$, $I_\theta$, and $A$ from the expansion of $\theta$; the only non-mechanical step in the general case is deciding whether the height extension splits.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper develops a general framework for computing the Ellis semigroup of minimal systems with an equicontinuous factor and a unique orbit of singular fibres, and applies it to primitive aperiodic bijective substitutions. For such a substitution θ, the authors show that E(Xθ) is the disjoint union of the acting group Z and its kernel M(Xθ), which is a completely simple semigroup. The main structural result, Theorem 4.22, describes M(Xθ) as a Rees matrix semigroup M[Gθ; Iθ, {±}, A] where Gθ is an extension of the odometer Zℓ by the fibre-preserving structure group Gθ^fib; when the generalised height equals the classical height, this extension splits and Gθ ≅ Gθ^fib ⋊ Zℓ. As an application, the authors prove that the virtual automorphism group V(Xθ) is isomorphic to C_A(Gθ) × Z, hence to the classical automorphism group Aut(Xθ), so the shift is semi-regular.
Significance. If the main results hold, this is a substantial contribution: it provides one of the few explicit algebraic descriptions of non-tame Ellis semigroups and answers a question of Auslander and Glasner in a broad class of substitution shifts. The paper is careful and honest: proofs of the new structural statements are detailed, several worked examples are given in Section 6, and the limitations of Theorem 4.22 are explicitly labelled. The semi-regularity application, in particular, does not depend on the unresolved extension problem and appears to be on solid ground. The paper also makes good use of imported classical theorems (Rees-Suskevitch, Dekking, Lemanczyk-Mentzen), which is appropriate for this type of work. The main caveat is that the abstract and introductory claims are broader than what the theorems prove.
major comments (2)
- [Abstract; §4.5.2, Theorem 4.22; Example 6.2(2)] The abstract states that the paper determines the Ellis semigroup for primitive aperiodic bijective substitutions, and the introduction repeats this without qualification. However, Theorem 4.22 gives the full algebraic description of E(Xθ) only under the additional assumption that the generalised height h equals the classical height h_cl; in that case a split section s: Zℓ → Gθ exists and Gθ ≅ Gθ^fib ⋊ Zℓ. The paragraph immediately after Theorem 4.22 explicitly says that when h > h_cl the extension problem (1.2) determining the structure group Gθ remains unsolved. Example 6.2(2) realizes h = 2 with h_cl = 1, so for that substitution the claimed full description of E(Xθ) is not established. This is an honest limitation, but it is load-bearing for the central claim of the paper. I therefore ask the authors to either prove the missing extension in the h > h_cl case, or to revise the abstract and Section 1 so that the conditional nature of the determination is stated precisely. I want to emphasize that Corollary 5.11 on semi-regularity is not affected by this scope gap, since its proof uses the finite centraliser C_A(Gθ) and the unique-singular-orbit machinery rather than the unresolved extension.
- [§4.5.2, Proposition 4.21] The proof of Proposition 4.21, which identifies Γθ with Tθ and computes G^{fib}_{θ,0} as Cov(Γθ), is too compressed for a step on which Theorem 4.22 directly depends. In particular, the sentence 'This is possible only if Tθ is a subgroup of Γθ' is not derived from the preceding degree argument; the role of the opposite degrees of s(z) and s(z)^{-1} and the constancy of the class (Φ_z^0)^{-1}(\tilde f(z))Γθ in z need to be spelled out. Since Theorem 4.22 uses this proposition to determine the structure group of M^{fib}(Xθ), I ask the authors to expand this proof so the reader can verify that no additional hypothesis is hidden here.
minor comments (4)
- [§4.5.2, proof of Theorem 4.17] In the proof of Theorem 4.17, the text says that to any f ∈ E(Xθ, Z+) one can assign a map f_z ∈ Iθ; this should be f_z ∈ Gθ, since f_z is generally a product of elements of Iθ and need not itself belong to Iθ.
- [§5.2.3, Corollary 5.11] The phrase 'Since these are finite groups' is ambiguous: the fibre-preserving parts are finite, but the full groups contain an infinite Z factor. Please clarify that the isomorphism between the finite fibre-preserving parts, together with the fact that the Z-factors coincide under the inclusion, forces the inclusion to be an isomorphism.
- [Definition 4.19 and surrounding text] The notation for the little structure group Γθ and its normal completion Γθ is visually very similar and easy to confuse, especially in the statement of the generalised height. I suggest using a distinct notation, for example Γθ^c or Γθ^+, for the normal completion.
- [§4.3, proof of Theorem 4.6] After the definition of the isomorphism (4.4), a one-sentence explanation of why surjectivity is equivalent to Lemma 4.7 would improve readability; as written, the claim is correct but the verification is left entirely to the reader.
Circularity Check
No significant circularity: the Ellis-semigroup computation is derived from the substitution expansion via internal theorems, with an explicit, non-circular scope gap for h > h_cl.
full rationale
The paper's derivation chain is self-contained. The central structural results—Theorem 4.6, Theorem 4.12, and Theorem 4.22—compute the structural semigroup Mf ib_0(Xθ), the fibre-preserving part Ef ib(Xθ), and the kernel M(Xθ) from the substitution's expansion. Definition 4.4 defines Gθ and Iθ purely combinatorially from the bijections (θ^n)_i; Proposition 4.8 constructs the elements f[L·R;±] as pointwise limits of shifts of fixed points; Proposition 4.9 shows every element of Mf ib_0 has that form; and Theorem 4.6 verifies the Rees matrix multiplication directly. Corollary 3.14, which assembles M(Xθ) from Mf ib(Xθ) and the equicontinuous factor Zℓ, is proved in the paper via Proposition 2.6 and Proposition 3.13, with the split section constructed from covariance; it is an internal theorem, not an imported ansatz. The generalized height h is defined as |Gθ/Γθ|, where Γθ is the normal completion of the little structure group; this is an independent algebraic quantity, and the paper explicitly flags that when h > h_cl the extension (1.2) remains unsolved (Section 4.5.2, after Theorem 4.22; Example 6.2(2)). That is an honest scope limitation, not a circular reduction. The applications to automorphism groups use Lemańczyk–Mentzen's external Theorem 5.1, while Corollary 5.10 derives V(Xθ) from the Ellis-semigroup machinery; the semi-regularity conclusion follows by comparing the two without fitting parameters or renaming known results. The only self-citations, to [7] for an adapted argument in Corollary 2.9 and to [2] for background on coincidence rank, are not load-bearing: the needed directions are proved in the text. No step was found in which an output is defined to be an input, or in which a fitted parameter is renamed as a prediction.
Assumptions & free parameters
assumptions (8)
- standard math Rees-Suskevitch theorem: a semigroup is completely simple if and only if it is isomorphic to a matrix semigroup M[G; I, Λ; A] (Theorem 2.1).
- standard math For compact right-topological semigroups the kernel exists and is completely simple; under joint continuity the kernel is topologically isomorphic to a normalized matrix semigroup (Theorems 2.4 and 2.5).
- domain assumption Forward proximality agrees with forward asymptoticity for these systems (Corollary 4.3), so E(Xθ, Z+) has a unique minimal left ideal (Lemma 2.8).
- domain assumption Theorem 5.1 of Lemanczyk and Mentzen [22]: for primitive aperiodic bijective substitutions, Aut^{fib}(Xθ) is isomorphic to C_{SA}(Gθ) and Aut(Xθ) is isomorphic to C_{SA}(Gθ) × Z.
- domain assumption Dekking's theorem on classical height and the maximal equicontinuous factor (Theorem 4.14, cited to [11]).
- standard math Axiom of choice for lifts s: Y → eEe with covariance and inverse conditions (Section 3).
- domain assumption Standing assumption that the substitution θ is simplified, i.e. after passing to a power all periodic points are fixed and each word θ(a) contains all letters (Section 4.1).
- domain assumption The acting group T is abelian (Section 2.7 onward).
invented entities (2)
-
Generalised height h (Definition 4.19): the order of Gθ/Γθ, where Γθ is the normal completion of the little structure group Γθ.
-
Structural semigroup M^{fib}_0(Xθ) (Theorem 4.6): the restriction of the fibre-preserving kernel to the singular fibre over 0 ∈ Zℓ.
Cite this review
Pith. "Pith review of The Ellis semigroup of bijective substitutions." pith.science (2026). https://pith.science/paper/2UM3CBWR
@misc{pith2026190805690,
author = {Pith},
title = {Pith review of: The Ellis semigroup of bijective substitutions},
year = {2026},
howpublished = {\url{https://pith.science/paper/2UM3CBWR}},
note = {Machine review of arXiv:1908.05690}
}
abstract
For topological dynamical systems $(X,T,\sigma)$ with abelian group $T$ which admit an equicontinuous factor $\pi:(X,T,\sigma)\to (Y,T,\delta)$ the Ellis semigroup $E(X)$ is an extension of $Y$ by its subsemigroup $E^{fib}(X)$ of elements which preserve the fibres of $\pi$. We establish methods to compute $E^{fib}(X)$ and use them to determine the Ellis semigroup of dynamical systems arising from primitive aperiodic bijective substitutions. As an application we show that for these substitution shifts, the virtual automorphism group is isomorphic to the classical automorphism group.
Reference graph
Works this paper leans on
-
[22]
M. Lema´ nczyk and M. K. Mentzen. On metric properties of substitutions. Compositio Math., 65(3):241–263, 1988
work page 1988
-
[1]
J.-B. Aujogue. Ellis enveloping semigroup for almost ca nonical model sets of an euclidean space. Algebraic & Geometric Topology , 15(4):2195–2237, 2015
work page 2015
-
[2]
J.-B. Aujogue, M. Barge, J. Kellendonk, and D. Lenz. Equi continuous factors, proximality and Ellis semigroup for Delone sets. In Mathematics of aperiodic order , volume 309 of Progr. Math., pages 137–194. Birkh¨ auser/Springer, Basel, 2015
work page 2015
- [3]
-
[4]
J. Auslander and E. Glasner. On the virtual automorphism group of a minimal flow to appear in Ergodic Theory and Dynamical Systems DOI=10.1017/etds.2020.8
-
[5]
M. Baake and U. Grimm. Aperiodic order. Vol. 1 , volume 149 of Encyclopedia of Mathe- matics and its Applications . Cambridge University Press, Cambridge, 2013. A mathemati cal invitation, With a foreword by Roger Penrose
work page 2013
-
[6]
M. Barge. private communication
-
[7]
M. Barge. and J. Kellendonk. Complete regularity of Ellis semigroups of Z-actions. Preprint arXiv:2001.02751 (2020)
work page Pith review arXiv 2020
Show all 27 references
-
[8]
Blanchard, E
F. Blanchard, E. Glasner, S. Kolyada, A. Maass On Li-Yorke pairs , Journal fur die reine und angewandte Mathematik 547, 51–68, 2002
2002
-
[9]
J. H. Carruth, J. Hildebrant, and R. J. Koch. The theory of topological semigroups. Vol. 75. Marcel Dekker Incorporated, 1983
1983
-
[10]
E. M. Coven, A. Quas, and R. Yassawi. Computing automorp hism groups of shifts using atypical equivalence classes. Discrete Anal. , pages Paper No. 3, 28, 2016
2016
-
[11]
F. M. Dekking. The spectrum of dynamical systems arisin g from substitutions of constant length. Z. Wahrscheinlichkeitstheorie und Verw. Gebiete , 41(3):221–239, 1977/78
1977
-
[12]
R. Ellis. A semigroup associated with a transformation group. Trans. Amer. Math. Soc., 94:272–281, 1960
1960
-
[13]
N. P. Fogg. Substitutions in dynamics, arithmetics and combinatorics , volume 1794 of Lecture Notes in Mathematics . Springer-Verlag, Berlin, 2002. Edited by V. Berth´ e, S. Fe renczi, C. Mauduit and A. Siegel
2002
-
[14]
Fuhrmann, E
G. Fuhrmann, E. Glasner, T. J¨ ager, and C. Oertel. Irreg ular model sets and tame dynamics. arXiv preprint arXiv:1811.06283 , 2018
2018 arXiv
-
[15]
Furstenberg and Y
H. Furstenberg and Y. Katznelson. Idempotents in compa ct semigroups and ramsey theory. Israel Journal of Mathematics , 68(3):257–270, 1989
1989
-
[16]
E. Glasner. On tame dynamical systems. Colloquium Mathematicum , 105(2):284–295, 2006. 40 JOHANNES KELLENDONK AND REEM YASSA WI
2006
-
[17]
E. Glasner. The structure of tame minimal dynamical sys tems for general groups. Inventiones mathematicae, 211(1):213–244, 2018
2018
-
[18]
Hindman and D
N. Hindman and D. Strauss. Algebra in the Stone-Cech com pactification: theory and appli- cations. W alter de Gruyter, 2011
2011
-
[19]
J. M. Howie. Fundamentals of semigroup theory , volume 12. Clarendon Oxford, 1995
1995
-
[20]
W. Huang. Tame systems and scrambled pairs under an abel ian group action. Ergodic Theory Dynam. Systems, 26(5):1549–1567, 2006
2006
-
[21]
T. Kamae. A topological invariant of substitution mini mal sets. J. Math. Soc. Japan , 24:285– 306, 1972
1972
-
[23]
Lind and B
D. Lind and B. Marcus. An introduction to symbolic dynam ics and coding. Cambridge Uni- versity Press, 1995
1995
-
[24]
J. C. Martin. Substitution minimal flows. Amer. J. Math. , 93:503–526, 1971
1971
-
[25]
Muellner and R
C. Muellner and R. Yassawi. Automorphisms of automatic shifts. To appear in Ergodic Theory and Dynamical Systems DOI=10.1017/etds.2020.13
2020 doi
-
[26]
Petrich and N
M. Petrich and N. R. Reilly. Completely regular semigroups , volume 27. John Wiley & Sons, 1999
1999
-
[27]
Staynova
P. Staynova. The Ellis Semigroup of a Generalised Morse System. To appear in Ergodic Theory and Dynamical Systems . DOI=10.1017/etds.2019.75. Institut Camille Jordan, Universit ´e L yon-1, France Email address : kellendonk@math.univ-lyon1.fr School of Mathematics and Statistic...
2019 doi
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.