A Low-Dimensional Counterexample to the HK-Conjecture
Rachel Chaiser · 2025
Read A Low-Dimensional Counterexample to the HK-Conjecture by Rachel Chaiser (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-05425.txt.
Introduction
Rachel Chaiser gives a low dimensional counterexample to the HK-conjecture in two thousand twenty-five. The conjecture links homology and K-theory for groupoid algebras. Deeley's past counterexample used a flat manifold of dimension nine. His odometer build rested on an expansive self cover there.
Chaiser shows each dimension from four up can host such a case. Flat manifolds of that size supply the odometers in her build. Dimension three and below cannot work, which makes four minimal. For those small cases the conjecture holds for the odometer. Stable and unstable groupoids of Smale spaces feel the fallout.
Catalog record
| CREATOR | Rachel Chaiser |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-05425
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Low-Dimensional Counterexample to the HK-Conjecture Thanks: This work was partially supported by NSF Grants 2000057 and 2247424. Rachel Chaiser Abstract . We provide a counterexample to the HK-conjecture using the flat manifold odometers constructed by Deeley. Deeley’s counterexample uses an odometer built from a flat manifold of dimension 9 and an expansive self-cover. We strengthen this result by showing that for each dimension d ≥ 4 d\geq 4 there is a counterexample to the HK-conjecture built from a flat manifold of dimension d d . Moreover, we show that this dimension is minimal, as if d ≤ 3 d\leq 3 the HK-conjecture holds for the associated odometer. We also discuss implications for the stable and unstable groupoid of a Smale space. 2020 Mathematics Subject Classification 46L80, 22A22 1. Introduction Matui’s HK-conjecture [ 20 ] , stated below, predicts a strong relationship between the homology of a certain class of groupoids and the K -theory of the reduced groupoid C^{*} -algebra. While there are a number of positive results [ 4 , 14 , 20 , 21 , 22 , 32 ] , the conjecture is known to fail in general. Conjecture (HK-Conjecture) . Let \mathcal{G} be a groupoid that is second countable, locally compact Hausdorff, étale, essentially principal, minimal, and ample. Then K_{*}(C^{*}_{r}(\mathcal{G}))\cong\bigoplus_{i}H_{2i+*}(\mathcal{G})\qquad\text{for $*=0,1$.} In 2020, Scarparo [ 29 ] provided the first counterexample to the HK-conjecture using a class of groupoids called odometers. While Scarparo’s counterexample is essentially principal, in 2023 Deeley [ 11 ] gave the first counterexample in the principal case. Deeley’s counterexample is an odometer built from dynamics on a flat manifold of dimension 9. We construct a counterexample to the HK-conjecture using the same class of flat manifold odometers but using a manifold of dimension 4. Moreover, we show that this dimension is minimal. We begin with a flat manifold Y and an expansive self-covering map g:Y\to Y . The existence of such a map is guaranteed by Epstein–Shub [ 13 ] and Shub [ 30 ] . We then construct an action groupoid from the action of \pi_{1}(Y) on the Cantor set \Omega=\varprojlim\left(\frac{\pi_{1}(Y)}{g_{*}^{k}\pi_{1}(Y)},\text{ coset inclusion}\right). The groupoid \mathcal{G}=\Omega\rtimes\pi_{1}(Y) is called an odometer. It is principal and satisfies the hypotheses of the HK-conjecture. Using results of [ 6 , 11 , 29 ] , the homology of this odometer and the K -theory of the reduce groupoid C^{*} -algebra can be computed in terms of the corresponding invariants for Y and the transfer map: H_{*}(\mathcal{G})\cong\varinjlim(H_{*}(Y),\operatorname{tr}_{H}),\qquad\text{and}\qquad K_{*}(C^{*}_{r}(\mathcal{G}))\cong\varinjlim(K_{*}(Y),\operatorname{tr}_{K}). One of the main results of the paper is Theorem 5.2 , stated below. Throughout the paper, T(G) denotes the torsion subgroup of a group G . Theorem . Suppose that Y is a flat manifold with T(K^{*}(Y))\not\cong T\left(\bigoplus_{i}H^{2i+*}(Y)\right) . Then there exists an expansive self-cover g:Y\to Y such that the corresponding odometer \mathcal{G} is a counterexample to the HK-conjecture. Moreover, the relevant groupoid is principal. To find such a flat manifold Y , we leverage the Atiyah–Hirzebruch spectral sequence which relates K -theory and cohomology. If \dim(Y)\leq 4 , the spectral sequence collapses to two short exact sequences \displaystyle 0\to H^{4}(Y)\to\widetilde{K}^{0}(Y)\to H^{2}(Y)\to 0, \displaystyle 0\to H^{3}(Y)\to K^{1}(Y)\to H^{1}(Y)\to 0. The second will always split, while the first can be a nontrivial extension. Our key to finding a nontrivial extension is proved in Lemma 5.1 . Our counterexample comes from the class of flat manifolds called real Bott manifolds, for which the cohomology ring H^{*}(-;\mathbb{Z}_{2}) is readily computed using a binary matrix that defines the manifold. Notably, Deeley’s counterexample also uses a real Bott manifold. In Example 5.3 , we present a real Bott manifold of dimension 4 that gives a counterexample to the HK-conjecture. In Theorem 5.4 , we extend this to counterexamples using a manifold of dimension d for each d\geq 4 . This is a strengthening of the results in [ 11 ] , where counterexamples are given in each dimension d\geq 9 . In Theorem 5.6 , we prove that if \dim(Y)\leq 3 then the HK-conjecture holds for the associated odometer. In Lemma 3.4 , we prove that the dynamic asymptotic dimension of our flat manifold odometers, denoted \operatorname{dad}(\mathcal{G}) , satisfies \operatorname{dad}(\mathcal{G})=\dim(Y) . Thus, for each d\geq 4 there is counterexample to the HK-conjecture coming from a groupoid with \operatorname{dad}(\mathcal{G})=d . This result is a companion to Theorem 4.19 of [ 4 ] , which says that the HK-conjecture holds for principal groupoids with \operatorname{dad}(\mathcal{G})\leq 2 and H_{2}(\mathcal{G}) free. We conclude the paper by discussing implications of the main results to the groupoids associated to Smale spaces in Section 6 . This paper is based on the author’s PhD thesis [ 5 ] . We refer the reader to the thesis for examples and for more detail on the content of the paper. 2. Flat Manifolds A flat manifold Y is a manifold of the form \mathbb{R}^{n}/\pi where \pi is a cocompact, torsionfree, discrete subgroup of O(n)\ltimes\mathbb{R}^{n} that acts freely and properly on \mathbb{R}^{n} . This definition agrees with the definition that Y be Riemannian flat (see [ 7 ] Section II.5). The fundamental group of a flat manifold is a Bieberbach group [ 1 ] . Since the action of \pi on \mathbb{R}^{n} is free and proper, \pi=\pi_{1}(Y) and that Y is a model for B\pi . Every flat manifold Y corresponds to a short exact sequence 0\to\mathbb{Z}^{n}\to\pi\to F\to 0 where F is a finite group called the holonomy . It is a result of Epstein–Shub in [ 13 ] that every flat manifold admits a locally expansive self-covering map. Theorem 2.1 . Let Y be a flat manifold associated to the short exact sequence 0\to\mathbb{Z}^{n}\to\pi\to F\to 0. Then for any k\in\mathbb{Z} such that |k|F|+1|>1 (i.e., k>0 or k<-2/|F| ) the map (k|F|+1)I_{n}:\mathbb{R}^{n}\to\mathbb{R}^{n} projects to an affine expanding endomorphism of Y . While this result was originally stated for k>0 , their proof holds more generally for negative values of k such that |k|F|+1|>1 . So we choose to state the theorem in its full generality. By Lemma 1 of [ 30 ] , such an expanding endomorphism is necessarily a covering map. Throughout, we call such a map an expansive self-cover and denote it by g:Y\to Y . See Examples 2.7 and 2.8 of [ 6 ] for explicit examples of expansive self-covers on flat manifolds. 2.1. Real Bott Manifolds To construct our counterexample, we consider a class of flat manifolds called real Bott manifolds because the cohomology ring with \mathbb{Z}_{2} coefficients is easily computed. Our overview follows [ 8 , 18 ] . A real Bott tower is a sequence of \mathbb{R}P^{1} (i.e., S^{1} ) bundles {\lx@inpgf@ignorespaces M_{n}} {\lx@inpgf@ignorespaces M_{n-1}} {\lx@inpgf@ignorespaces\cdots} {\lx@inpgf@ignorespaces M_{2}} {\lx@inpgf@ignorespaces S^{1}} {\lx@inpgf@ignorespaces pt} \scriptstyle{\lx@inpgf@ignorespaces\mathbb{R}P^{1}} \scriptstyle{\lx@inpgf@ignorespaces\mathbb{R}P^{1}} \scriptstyle{\lx@inpgf@ignorespaces\mathbb{R}P^{1}} \scriptstyle{\lx@inpgf@ignorespaces\mathbb{R}P^{1}} \scriptstyle{\lx@inpgf@ignorespaces\mathbb{R}P^{1}} such that at each stage M_{j}\to M_{j-1} is the projectivization of a real vector bundle of the form L_{j-1}\oplus\underline{\mathbb{R}} over M_{j-1} , where \underline{\mathbb{R}} is the trivial line bundle and L_{j-1} is any real line bundle over M_{j-1} . The total space M_{n} is a real Bott manifold . Since real line bundles are classified by their first Stiefel–Whitney class (this is discussed in Section 4.2 ), each L_{j-1} corresponds to an element in H^{1}(M_{j-1};\mathbb{Z}_{2})\cong\mathbb{Z}_{2}^{j-1} . Viewing these elements as column vectors, padding the bottoms with 0’s to create