betaFeedback

WorksMathematics

Criterion for Glasner property and IC-Rigidity for actions on homogeneous space

Abstract

Let GG be a semi-group acting continuously on a compact metric space XX. The action G↷XG\curvearrowright X is called Glasner if for every infinite set A⊆XA\subseteq X and every ε>0\varepsilon>0, there exists g∈Gg\in G such that gAgA is ε\varepsilon-dense in XX. Let K(X)\mathcal{K}(X) denote the space of nonempty compact subsets of XX. A GG-invariant Borel probability measure on K(X)\mathcal{K}(X) is called an invariant random compact (IRC), and the action is IC-rigid if δXδ_X is the only IRC supported on infinite compact subsets of XX. In this paper we give a pair-correlation criterion implying both IC-rigidity and the Glasner property. As an application, we prove both these properties for finitely generated Ad⁡\operatorname{Ad}-Zariski-dense semi-groups acting on compact homogeneous spaces G/ΛG/Λ, where GG is connected and noncompact with simple Lie algebra and Λ<GΛ<G is a cocompact lattice; in particular, this applies to finitely generated Zariski-dense semi-groups of SLd(R)\mathrm{SL}_d(\mathbb{R}) and PSLd(R)\mathrm{PSL}_d(\mathbb{R}). We also derive a Fourier-analytic criterion for semi-group actions on Td\mathbb{T}^d and use it to prove IC-rigidity for actions arising from suitable matrix-valued polynomials.

Introduction

Let GG be a semi-group acting continuously on a compact space XX. The action G↷XG \curvearrowright X is called Glasner if for every infinite set Y⊂XY \subset X and every ε>0\varepsilon> 0 there exists g∈Gg \in G such that gYgY is ε\varepsilon-dense in XX. The terminology originates from a theorem of Glasner [7], which states that the multiplicative action of N\mathbb{N} on T=R/Z\mathbb{T} = \mathbb{R}/\mathbb{Z} has this property. This result has since been extended in several directions. Berend and Peres [3] proved that for every infinite A⊆TA \subseteq\mathbb{T}, the dilates nAnA converge to T\mathbb{T} along a density-one set of integers. Dong [5, 6] established analogous results for higher-dimensional linear actions and actions on compact homogeneous spaces, and more recently Bulinski and Fish [1] obtained a quantitative version of Glasner results for irreducible linear actions with Zariski-connected Zariski closure.

A closely related notion of IC-rigidity was introduced by Kra and Schmieding in [12]. Both properties concern the behavior of infinite subsets under the action: the Glasner property requires that every infinite set has arbitrarily dense images, while IC-rigidity requires that the only invariant measure supported on infinite compact subsets is δX\delta_X. We now recall the necessary background and notation to formulate the notion of IC-rigidity.

Let K(X)K(X) denote the space of nonempty compact subsets of XX with Hausdorff metric. We assume that the action is surjective and finite-to-one. The action of GG on XX induces an action of GG on K(X)K(X). In [12] Kra and Schmieding introduced the notion of invariant random compact. A Borel probability measure μ\mu on K(X)K(X) which is invariant under the action of GG on K(X)K(X) is called an invariant random compact or an IRC (see [12 Definition 3.1]). An IRC μ\mu for which μ\mu-almost every compact subset of XX is infinite is called nonfinitary.

Assuming XX is infinite, every infinite invariant compact set Y⊂XY \subset X yields a nonfinitary IRC δY\delta_Y, in particular there always exists a nonfinitary IRC, δX\delta_X. Kra and Schmieding call an action of GG on XX to be IC-rigid if the only nonfinitary IRC is δX\delta_X (see [12 Definition 3.7]). In this paper we provide a pair-correlation condition for the action G↷XG \curvearrowright X that implies IC-rigidity and the Glasner property. Let (μN)N≥1(\mu_N)_{N \ge1} be a sequence of probability measures on GG. For f∈C(X;C)f \in C(X;\mathbb{C}) define

2020 Mathematics Subject Classification. 37B05, 22F30, 37A50.

Key words and phrases. IC-rigidity, invariant random compact, Glasner property, homogeneous actions, Zariski-dense semi-groups. pair-correlation kernel as follows.

KN,f(x,y):=∫Gf(gx)f(gy)‾ dμN(g),x,y∈X.(1)K_{N,f}(x,y) := \int_G f(gx)\overline{f(gy)}\,d\mu_N(g), \qquad x,y\in X. \tag*{(1)}

Our main result is the following.

Theorem 1.1. Let mm be a fully supported Borel probability measure on XX. Assume that, for every f∈C(X;C)f \in C(X;\mathbb{C}) satisfying ∫Xf dm=0\int_X f\,dm=0, every η>0\eta>0, and every x∈Xx\in X, the set

{y∈X∖{x}:lim sup⁡N→∞∣KN,f(x,y)∣>η}\left\{y\in X\setminus\{x\}:\limsup_{N\to\infty}\left|K_{N,f}(x,y)\right|>\eta\right\}

is finite. Then:

(a) The action G↷XG\curvearrowright X has the Glasner property: for every infinite set A⊆XA\subseteq X and every ε>0\varepsilon>0, there exists g∈Gg\in G such that gAgA is ε\varepsilon-dense in XX.

(b) The action G↷XG\curvearrowright X is IC-rigid. Equivalently, δX\delta_X is its only nonfinitary invariant random compact.

In the context of the Glasner property for actions on compact homogeneous spaces, the best known result is due to Dong [6]. Dong established the Glasner property for actions of epimorphic subgroups, in particular minimal parabolic subgroups, on compact homogeneous spaces. As an application of Theorem 1.1, we prove both the Glasner property and IC-rigidity for finitely generated Ad-Zariski-dense semi-groups acting on compact homogeneous spaces.

Theorem 1.2. Let GG be a connected noncompact Lie group with simple Lie algebra, and let Λ<G\Lambda<G be a cocompact lattice. Let μ\mu be a finitely supported probability measure on GG whose support generates an Ad-Zariski-dense semi-group Γ⊆G\Gamma\subseteq G. Then the action Γ↷G/Λ\Gamma\curvearrowright G/\Lambda has the Glasner property and is IC-rigid.

As a direct corollary of Theorem 1.2, we show that finitely generated Zariski-dense semi-groups in SLd(R)\mathrm{SL}_d(\mathbb{R}) or PSLd(R)\mathrm{PSL}_d(\mathbb{R}) act on their compact homogeneous spaces with the Glasner property and IC-rigidity.

Corollary 1.3. Let d≥2d\geq2, let

G=SLd(R)orG=PSLd(R),G=\mathrm{SL}_d(\mathbb{R}) \qquad\text{or}\qquad G=\mathrm{PSL}_d(\mathbb{R}),

and let Λ<G\Lambda<G be a cocompact lattice. Let μ\mu be a finitely supported probability measure whose support generates a Zariski-dense semi-group Γ<G\Gamma<G. Then the action Γ↷G/Λ\Gamma\curvearrowright G/\Lambda has the Glasner property and is IC-rigid.

In a recent preprint [11], Kra and Schmieding proved IC-rigidity for the action of the multiplicative semi-group {nr:n∈N}\{n^r:n\in\mathbb{N}\}, for r≥1r\geq1, on the torus T\mathbb{T}. They also proved IC-rigidity for the action on Td\mathbb{T}^d of a subgroup Γ<SLd(Z)\Gamma<\mathrm{SL}_d(\mathbb{Z}) whose Zariski closure is Zariski-connected and whose representation on Rd\mathbb{R}^d is irreducible (see [11 Theorem 1.2, Theorem 1.4]). Building upon Theorem 1.1, we provide a Fourier-analytic criterion (see Theorem 1.4 below) for semi-group actions on the dd-dimensional torus, d∈Nd\in\mathbb{N}, which implies both IC-rigidity and the Glasner property. Our Fourier-analytic criterion also recovers the IC-rigidity results of Kra and Schmieding [11 Theorem 1.2, Theorem 1.4], through an independently developed but closely related approach.

Recall that by a semi-group action on a torus we mean the action of a semi-group GG on Td=Rd/Zd\mathbb{T}^d=\mathbb{R}^d/\mathbb{Z}^d by continuous group endomorphism. As before let (μN)N∈N(\mu_N)_{N\in\mathbb{N}} be a sequence of probability measures on GG. The averaged orbit probability measure of z∈Tdz\in\mathbb{T}^d (corresponding to μN\mu_N), is given by

σN,z:=∫Gδgz dμN(g).\sigma_{N,z}:=\int_G\delta_{gz}\,d\mu_N(g).

For t∈Rt \in\mathbb{R}, we set e(t):=e2πite(t) := e^{2\pi i t}. For m∈Zdm \in\mathbb{Z}^{d}, the mm-th Fourier coefficient of the probability measure σN,z\sigma_{N,z} is given by

σ^N,z(m)=∫Tde(⟨m,x⟩) dσN,z(x)=∫Ge(⟨m,gz⟩) dμN(g).\widehat{\sigma}_{N,z}(m)=\int_{\mathbb{T}^{d}} e(\langle m,x\rangle)\,d\sigma_{N,z}(x)=\int_{G}e(\langle m,gz\rangle)\,d\mu_{N}(g).

We show that Theorem 1.1, together with a density argument, gives the following.

Theorem 1.4. Assume that there exists a sequence of probability measures (μN)N≥1(\mu_N)_{N\ge1} on GG such that, for every m∈Zd∖{0}m \in\mathbb{Z}^{d}\setminus\{0\} and every δ>0\delta>0, the set

{z∈Td:lim sup⁡N→∞∣∫Ge(⟨m,gz⟩) dμN(g)∣≥δ}(2)\left\{z\in\mathbb{T}^{d}:\limsup_{N\to\infty}\left|\int_{G}e(\langle m,gz\rangle)\,d\mu_N(g)\right|\ge\delta\right\} \tag*{(2)}

is finite. Then:

(a) The action G↷TdG\curvearrowright\mathbb{T}^{d} has the Glasner property: for every infinite set A⊆TdA\subseteq\mathbb{T}^{d} and every ε>0\varepsilon>0, there exists g∈Gg\in G such that gAgA is ε\varepsilon-dense in Td\mathbb{T}^{d}.

(b) The action G↷TdG\curvearrowright\mathbb{T}^{d} is IC-rigid. Equivalently, δTd\delta_{\mathbb{T}^{d}} is its only nonfinitary invariant random compact.

In [2] Bulinski and Fish proved Glasner property for actions arising from matrix-valued polynomials. They showed that for A(x)∈Md(Z[x])A(x)\in M_d(\mathbb{Z}[x]) the condition

v⋅(A(x)−A(0))w≠0v\cdot(A(x)-A(0))w\ne0

for every v,w∈Zd∖{0}v,w\in\mathbb{Z}^{d}\setminus\{0\} is equivalent to the set {A(n):n∈Z}\{A(n):n\in\mathbb{Z}\} having the uniform Glasner property on Td\mathbb{T}^{d}. As an application of Theorem 1.4 (b), we prove IC-rigidity for the multiplicative semi-group generated by such a matrix-valued polynomial. In private communication, Bryna Kra and Scott Schmieding informed us that this result can also be obtained using their methods in [11].

Theorem 1.5. Let A(x)∈Md(Z[x])A(x)\in M_d(\mathbb{Z}[x]) satisfy

v⋅(A(x)−A(0))w≠0v\cdot(A(x)-A(0))w\ne0

for every v,w∈Zd∖{0}v,w\in\mathbb{Z}^{d}\setminus\{0\}, and suppose that det⁡A(x)≢0\det A(x)\not\equiv0. Let GAG_A be the multiplicative semi-group generated by

{A(n):n∈Z, det⁡A(n)≠0}.\{A(n):n\in\mathbb{Z},\ \det A(n)\ne0\}.

The action GA↷TdG_A\curvearrowright\mathbb{T}^{d} by endomorphisms, given by

g⋅z=gz(modZd),g\cdot z=gz\pmod{\mathbb{Z}^{d}},

is IC-rigid. Equivalently, δTd\delta_{\mathbb{T}^{d}} is the only nonfinitary invariant random compact for this action.

Open questions

We conclude with a few open questions. For relatively prime integers p,q≥2p,q\ge2, let

Sp,q:={paqb:a,b≥0}S_{p,q}:=\{p^{a}q^{b}:a,b\ge0\}

be the multiplicative semi-group generated by pp and qq, acting on T\mathbb{T} by dilations. It is natural to ask whether this action has the Glasner property. A closely related question concerning IC-rigidity was posed by Kra and Schmieding in [12]. In a recent preprint [Sch26], Schmieding proved that if μ\mu is an IRC for this action and μ\mu-almost every compact set YY has positive Hausdorff dimension, then μ\mu-almost every YY is equal to T\mathbb{T}. In other words, μ=δT\mu=\delta_{\mathbb{T}}.

Question 1.6. Let p,q≥2p,q\ge2 be relatively prime. Does the action Sp,q↷TS_{p,q}\curvearrowright\mathbb{T} have the Glasner property? Is this action IC-rigid?

Another question posed by Kra and Schmieding in [12], asks whether there exists an IC-rigid action of an abelian group, and in particular an IC-rigid Z\mathbb{Z}-action. Recently, Kanigowski, Kasprazak and Lorenzo-Laguano [KKLL26] answered this question positively by proving that mixing rank-one Z\mathbb{Z}-actions are IC-rigid. The IC-rigid Z\mathbb{Z}-actions constructed there are realized as subshifts, and hence the underlying spaces are totally disconnected. On the other hand, Dong [5 Theorem 3.1] proved that no Z\mathbb{Z}-action on a locally connected compact metric space has the Glasner property. These results motivate the following questions.

Question 1.7. (a) Does there exist, for some d≥2d \ge2, a Zd\mathbb{Z}^{d}-action on an infinite connected compact metric space with the Glasner property?

(b) Does there exist, for some d≥1d \ge1, an IC-rigid Zd\mathbb{Z}^{d}-action on an infinite connected compact metric space?

Structure of the paper

In § 2, we provide the basic background and definitions. In Section 3, using a fundamental result concerning the density of measures, we prove the pair-correlation criterion, Theorem 1.1. In § 4, we recall an equidistribution result for homogeneous dynamics due to Benoist–Quint [4] (see Theorem 4.1). Using this equidistribution result together with Theorem 1.1, we prove Theorem 1.2. In the same section, as a corollary we establish the Glasner property and IC-rigidity for Zariski-dense semi-groups of SLd(R)\mathrm{SL}_{d}(\mathbb{R}) and PSLd(R)\mathrm{PSL}_{d}(\mathbb{R}) acting on compact homogeneous spaces, as stated in Theorem 1.3. In § 5, using a density argument together with Theorem 1.1, we prove the Fourier-analytic criterion for torus actions, Theorem 1.4. Finally in the same section, we use Theorem 1.4 to prove Theorem 1.5.

Acknowledgement. The authors would like to thank Bryna Kra and Scott Schmieding for communicating their results and for their encouragement. Both authors were supported by the Australian Research Council through grant DP240100472.

AI statement. AI was used in the paper to obtain the relevant version of the equidistribution theorem for homogeneous spaces due to Benoist–Quint [4] (see Theorem 4.1) in a form suitable for our purposes. Apart from this, AI was used to audit the paper, identify relevant references, and correct editorial typos.

Background and notation

Let (X,d)(X,d) be a compact metric space and GG be a countable semi-group or group. By (X,G)(X,G) we denote a topological dynamical system where GG acts on XX continuously. In the case where GG is a semi-group, we assume that the action is surjective and finite-to-one.

Definition 2.1. A continuous action of a semi-group GG on a compact space XX is called Glasner if for every infinite set Y⊂XY \subset X and ε>0\varepsilon> 0 there exists g∈Gg \in G such that gYgY is ε\varepsilon-dense in XX.

Following the work of Glasner [7], this property has been extensively studied in many different contexts. We refer our interested readers to [3, 10, 5, 6].

As mentioned before, we denote by K(X)\mathcal{K}(X) the space of nonempty compact subsets of XX which is endowed with Hausdorff metric. For K,L∈K(X)K,L \in\mathcal{K}(X) set

dH(K,L)=max⁡{sup⁡x∈Kd(x,L),sup⁡y∈Ld(y,K)},(3)d_{H}(K,L)=\max\left\{\sup_{x\in K}d(x,L),\sup_{y\in L}d(y,K)\right\}, \tag*{(3)}

where d(x,L):=inf⁡y∈Ld(x,y)d(x,L):=\inf_{y\in L}d(x,y). The action of GG on XX induces an action of GG on K(X)\mathcal{K}(X) given by g⋅K=g(K)g\cdot K=g(K). Using the notation of Kra and Schmieding for k∈Nk \in\mathbb{N}, we denote by K≤k(X)\mathcal{K}_{\le k}(X) the collection of all subsets of XX of size at most kk and set

Kfin(X)=⋃k=1∞K≤k(X),Kinf(X)=K(X)∖Kfin(X).\mathcal{K}_{\mathrm{fin}}(X)=\bigcup_{k=1}^{\infty}\mathcal{K}_{\le k}(X),\qquad\mathcal{K}_{\mathrm{inf}}(X)=\mathcal{K}(X)\setminus\mathcal{K}_{\mathrm{fin}}(X).

Since K≤k(X)\mathcal{K}_{\le k}(X) is a closed set for each k∈Nk \in\mathbb{N}, it follows that Kfin(X)\mathcal{K}_{\mathrm{fin}}(X) is a FσF_{\sigma} set, hence it is Borel. This implies that Kinf(X)\mathcal{K}_{\mathrm{inf}}(X) is also Borel. Observe that Kfin(X)\mathcal{K}_{\mathrm{fin}}(X) is GG-invariant and the finite-to-one property of the action implies that Kinf(X)\mathcal{K}_{\mathrm{inf}}(X) is GG-invariant as well. Now we formally define:

Definition 2.2. (see [12 Definition 3.1]) An invariant random compact (IRC) for a topological dynamical system (X,G)(X,G), is a Borel probability measure μ\mu on K(X)\mathcal{K}(X) which is invariant under the induced action of GG on K(X)\mathcal{K}(X).

Observe that an invariant random compact μ\mu can be decomposed into its finitary and nonfinitary parts i.e. μ=μF+μI\mu=\mu_F+\mu_I, where the measures μF,μI\mu_F,\mu_I on K(X)\mathcal{K}(X) are defined as,

μF(A)=μ(A∩Kfin(X))andμI(A)=μ(A∩Kinf(X))\mu_F(\mathcal{A})=\mu\left(\mathcal{A}\cap\mathcal{K}_{\mathrm{fin}}(X)\right) \qquad\text{and}\qquad \mu_I(\mathcal{A})=\mu\left(\mathcal{A}\cap\mathcal{K}_{\mathrm{inf}}(X)\right)

for a Borel set A⊆K(X)\mathcal{A}\subseteq\mathcal{K}(X).

Definition 2.3. (see [12 Definition 3.7]) The action of GG on XX is infinite compact rigid (or IC-rigid) if the only nonfinitary IRC is δX\delta_X. In other words for every IRC μ\mu, the decomposition of μ\mu into its finitary and nonfinitary parts has the form

μ=μF+c⋅δX\mu=\mu_F+c\cdot\delta_X

for some c≥0c\geq0.

Proof of thm:main

In this section we prove Theorem 1.1. To do so we first need an elementary result showing that small averages of finitely many mean-zero continuous functions force dense support.

Proposition 3.1. Let mm be a fully supported Borel probability measure on XX. For every ε>0\varepsilon>0, there exist q∈Nq\in\mathbb{N}, zero-mean functions f1,…,fq∈C(X;C)f_1,\ldots,f_q\in C(X;\mathbb{C}), and constants c1,…,cq>0c_1,\ldots,c_q>0 such that the following holds. If ν\nu is a Borel probability measure on XX satisfying

∣∫Xfj dν∣<cjfor every 1≤j≤q,\left|\int_X f_j\,\mathrm{d}\nu\right|<c_j \qquad\text{for every }1\leq j\leq q,

then supp⁡ν\operatorname{supp}\nu is ε\varepsilon-dense in XX.

Proof. Choose y1,…,yq∈Xy_1,\ldots,y_q\in X such that the open balls

Uj:=B(yj,ε/2),j=1,…,q,U_j:=B(y_j,\varepsilon/2),\qquad j=1,\ldots,q,

cover XX. For every j∈{1,…,q}j\in\{1,\ldots,q\}, choose a nonzero non-negative bump function ϕj\phi_j supported in UjU_j. Since mm is fully supported,

aj:=∫Xϕj dm>0.a_j:=\int_X\phi_j\,\mathrm{d}m>0.

For every j∈{1,…,q}j\in\{1,\ldots,q\}, set fj:=ϕj−ajf_j:=\phi_j-a_j and cj:=aj2c_j:=\frac{a_j}{2}. Using the fact that mm is a probability measure, for every j∈{1,…,q}j\in\{1,\ldots,q\}, we get ∫Xfj dm=0\int_X f_j\,\mathrm{d}m=0. Let ν\nu be a Borel probability measure on XX such that

∣∫Xfj dν∣<cjfor every 1≤j≤q.\left|\int_X f_j\,\mathrm{d}\nu\right|<c_j \qquad\text{for every }1\leq j\leq q.

Then

∫Xϕj dν=aj+∫Xfj dν>aj2>0.\int_X\phi_j\,\mathrm{d}\nu=a_j+\int_Xf_j\,\mathrm{d}\nu>\frac{a_j}{2}>0.

Since ϕj\phi_j vanishes outside UjU_j and ∫Xϕj dν>0\int_X\phi_j\,\mathrm{d}\nu>0, it follows that supp⁡ν∩Uj≠∅\operatorname{supp}\nu\cap U_j\ne\varnothing for every j∈{1,…,q}j\in\{1,\ldots,q\}. Since UjU_j are balls of radius ε/2\varepsilon/2 that cover XX, it follows that supp⁡ν\operatorname{supp}\nu is ε\varepsilon-dense in XX. □\square

Now we are ready to prove Theorem 1.1.

Proof of Theorem 1.1. We first establish a pair-correlation estimate that will be used in proving both part (a) and (b) of the statement. Fix ε>0\varepsilon> 0, and let q∈Nq \in\mathbb{N}, zero-mean functions f1,…,fq∈C(X;C)f_1,\ldots,f_q \in C(X;\mathbb{C}), and constants c1,…,cq>0c_1,\ldots,c_q > 0 be as given by Theorem 3.1. Let A⊆XA \subseteq X be infinite and let η>0\eta> 0. Set C:=∑j=1q∥fj∥∞2C := \sum_{j=1}^{q} \lVert f_j\rVert_{\infty}^{2} and δ:=η2q\delta:= \frac{\eta}{2q}. Choose M∈NM \in\mathbb{N} such that M>2CηM > \frac{2C}{\eta}. For 1≤j≤q1 \leq j \leq q and x∈Xx \in X, let

Rj,δ(x):={y∈X∖{x}:lim sup⁡N→∞∣KN,fj(x,y)∣>δ}.R_{j,\delta}(x) := \left\{y \in X \setminus\{x\} : \limsup_{N \to\infty} \lvert K_{N,f_j}(x,y)\rvert> \delta\right\}.

By the assumption of Theorem 1.1, Rj,δ(x)R_{j,\delta}(x) is finite for every jj and every x∈Xx \in X. Note that, for any x≠y∈Xx \ne y \in X, we have KN,fj(y,x)=KN,fj(x,y)K_{N,f_j}(y,x) = K_{N,f_j}(x,y) and thus ∣KN,fj(y,x)∣=∣KN,fj(x,y)∣\lvert K_{N,f_j}(y,x)\rvert= \lvert K_{N,f_j}(x,y)\rvert. Thus the expression

lim sup⁡N→∞∣KN,fj(x,y)∣>δ\limsup_{N \to\infty} \lvert K_{N,f_j}(x,y)\rvert> \delta

is symmetric in xx and yy. Since AA is infinite and each set Rj,δ(x)R_{j,\delta}(x) is finite, we can choose distinct points x1,…,xM∈Ax_1,\ldots,x_M \in A, inductively so that, for every distinct i,ℓ∈{1,…,M}i,\ell\in\{1,\ldots,M\} and every 1≤j≤q1 \leq j \leq q,

lim sup⁡N→∞∣KN,fj(xi,xℓ)∣≤δ.\limsup_{N \to\infty} \lvert K_{N,f_j}(x_i,x_\ell)\rvert\leq\delta.

We define

τM:=1M∑i=1Mδxi.\tau_M := \frac{1}{M}\sum_{i=1}^{M}\delta_{x_i}.

Note that τM\tau_M is the probability measure on XX obtained by averaging the Dirac measures at x1,…,xMx_1,\ldots,x_M. For g∈Gg \in G, the pushforward of τM\tau_M is given by

g∗τM=1M∑i=1Mδgxi.g_*\tau_M = \frac{1}{M}\sum_{i=1}^{M}\delta_{gx_i}.

For 1≤j≤q1 \leq j \leq q, set

Fj(g):=∫Xfj d(g∗τM)=1M∑i=1Mfj(gxi).F_j(g) := \int_X f_j\,d(g_*\tau_M) = \frac{1}{M}\sum_{i=1}^{M}f_j(gx_i).

Expanding the square and integrating over GG with respect to μN\mu_N gives

∫G∣Fj(g)∣2 dμN(g)=1M2∑i,ℓ=1M∫Gfj(gxi)fj(gxℓ)‾ dμN(g)=1M2∑i,ℓ=1MKN,fj(xi,xℓ).\begin{aligned} \int_G \lvert F_j(g)\rvert^2\,d\mu_N(g) = \frac{1}{M^2}\sum_{i,\ell=1}^{M}\int_G f_j(gx_i)\overline{f_j(gx_\ell)}\,d\mu_N(g) \\ &= \frac{1}{M^2}\sum_{i,\ell=1}^{M}K_{N,f_j}(x_i,x_\ell). \end{aligned}

For the diagonal terms i=ℓi=\ell,

∣KN,fj(xi,xi)∣=∫G∣fj(gxi)∣2 dμN(g)≤∥fj∥∞2.\lvert K_{N,f_j}(x_i,x_i)\rvert= \int_G \lvert f_j(gx_i)\rvert^2\,d\mu_N(g) \leq\lVert f_j\rVert_{\infty}^{2}.

Since there are MM diagonal terms, using triangle inequality we get

∫G∣Fj(g)∣2 dμN(g)≤∥fj∥∞2M+1M2∑i≠ℓ∣KN,fj(xi,xℓ)∣.\int_G \lvert F_j(g)\rvert^2\,d\mu_N(g) \leq\frac{\lVert f_j\rVert_{\infty}^{2}}{M}+\frac{1}{M^2}\sum_{i\ne\ell}\lvert K_{N,f_j}(x_i,x_\ell)\rvert.

By the choice of x1,…,xMx_1,\ldots,x_M, it follows that

lim sup⁡N→∞∫G∣Fj(g)∣2 dμN(g)≤∥fj∥∞2M+M(M−1)M2δ≤∥fj∥∞2M+δ.\limsup_{N\to\infty}\int_G \lvert F_j(g)\rvert^2\,d\mu_N(g) \leq\frac{\lVert f_j\rVert_{\infty}^{2}}{M}+\frac{M(M-1)}{M^2}\delta \leq\frac{\lVert f_j\rVert_{\infty}^{2}}{M}+\delta.

Summing over 1≤j≤q1 \le j \le q yields

lim sup⁡N→∞∫G∑j=1q∣Fj(g)∣2 dμN(g)≤∑j=1qlim sup⁡N→∞∫G∣Fj(g)∣2 dμN(g)≤CM+qδ<η.(4)\limsup_{N\to\infty}\int_G \sum_{j=1}^{q}|F_j(g)|^2\,d\mu_N(g)\le\sum_{j=1}^{q}\limsup_{N\to\infty}\int_G |F_j(g)|^2\,d\mu_N(g)\le\frac{C}{M}+q\delta<\eta. \tag*{(4)}

Thus, for every infinite A⊆XA\subseteq X, every finite collection of zero-mean functions f1,…,fqf_1,\ldots,f_q, and every η>0\eta>0, there is a finite set of distinct points x1,…,xM∈Ax_1,\ldots,x_M\in A for which (4) holds.

We now show that the action G↷XG\curvearrowright X has the Glasner property. Fix an infinite set A⊆XA\subseteq X and ε>0\varepsilon>0. Let q∈Nq\in\mathbb{N}, zero-mean functions f1,…,fq∈C(X;C)f_1,\ldots,f_q\in C(X;\mathbb{C}), and constants c1,…,cq>0c_1,\ldots,c_q>0 be as given by Theorem 3.1, and set

c:=min⁡1≤j≤qcj>0.c:=\min_{1\le j\le q}c_j>0.

For η=c2/2\eta=c^2/2, choose x1,…,xM∈Ax_1,\ldots,x_M\in A such that, for the corresponding τM\tau_M and FjF_j, (4) holds. Thus, for a sufficiently large NN,

∫G∑j=1q∣Fj(g)∣2 dμN(g)<c2.\int_G\sum_{j=1}^{q}|F_j(g)|^2\,d\mu_N(g)<c^2.

Since μN\mu_N is a probability measure, there exists g∈Gg\in G such that ∑j=1q∣Fj(g)∣2<c2\sum_{j=1}^{q}|F_j(g)|^2<c^2 otherwise the integral would be at least c2c^2. Hence

∣∫Xfj d(g∗τM)∣=∣Fj(g)∣<c≤cjfor every 1≤j≤q.\left|\int_X f_j\,d(g_*\tau_M)\right|=|F_j(g)|<c\le c_j\qquad\text{for every }1\le j\le q.

Thus Theorem 3.1 implies that

supp⁡(g∗τM)={gx1,…,gxM}⊆gA\operatorname{supp}(g_*\tau_M)=\{gx_1,\ldots,gx_M\}\subseteq gA

is ε\varepsilon-dense in XX. Hence gAgA is also ε\varepsilon-dense in $X. This proves the Glasner property.

We now show that the action G↷XG\curvearrowright X is IC-rigid. Let ν\nu be a nonfinitary invariant random compact. For ε>0\varepsilon>0, set

Kε:={K∈K(X):dH(K,X)<ε}.K_\varepsilon:=\{K\in\mathcal{K}(X):d_H(K,X)<\varepsilon\}.

Note that the map K↦dH(K,X)K\mapsto d_H(K,X) is continuous hence KεK_\varepsilon is an open subset of K(X)\mathcal{K}(X). We show that for every infinite K∈K(X)K\in\mathcal{K}(X),

lim⁡N→∞μN{g∈G:gK∈Kε}=1.(5)\lim_{N\to\infty}\mu_N\{g\in G:gK\in K_\varepsilon\}=1. \tag*{(5)}

Fix such a KK and let q∈Nq\in\mathbb{N}, zero-mean functions f1,…,fq∈C(X;C)f_1,\ldots,f_q\in C(X;\mathbb{C}), and constants c1,…,cq>0c_1,\ldots,c_q>0 be as given by Theorem 3.1. We set c:=min⁡1≤j≤qcj>0c:=\min_{1\le j\le q}c_j>0. For any η>0\eta>0, choose x1,…,xM∈Kx_1,\ldots,x_M\in K such that, for the corresponding τM\tau_M and FjF_j, (4) holds. Thus,

lim sup⁡N→∞∫G∑j=1q∣Fj(g)∣2 dμN(g)<η.(6)\limsup_{N\to\infty}\int_G\sum_{j=1}^{q}|F_j(g)|^2\,d\mu_N(g)<\eta. \tag*{(6)}

As we noted earlier, if ∑j=1q∣Fj(g)∣2<c2\sum_{j=1}^{q}|F_j(g)|^2<c^2, then

∣∫Xfj d(g∗τM)∣=∣Fj(g)∣<c≤cjfor every 1≤j≤q.\left|\int_X f_j\,d(g_*\tau_M)\right|=|F_j(g)|<c\le c_j\qquad\text{for every }1\le j\le q.

Hence, by Theorem 3.1, supp⁡(g∗τM)={gx1,…,gxM}⊆gK\operatorname{supp}(g_*\tau_M)=\{gx_1,\ldots,gx_M\}\subseteq gK is ε\varepsilon-dense in XX. In other words, gK∈KεgK\in K_\varepsilon. Thus,

{g∈G:gK∉Kε}⊆{g∈G:∑j=1q∣Fj(g)∣2≥c2}.\{g\in G:gK\notin K_\varepsilon\}\subseteq\left\{g\in G:\sum_{j=1}^{q}|F_j(g)|^2\ge c^2\right\}.

Taking measure μN\mu_N of both sides and using Markov’s inequality we conclude

μN{g∈G:gK∉Kε}≤1c2∫G∑j=1q∣Fj(g)∣2 dμN(g).\mu_N\{g \in G : gK \notin\mathcal{K}_\varepsilon\} \le\frac{1}{c^2}\int_G \sum_{j=1}^{q} \lvert F_j(g)\rvert^2\,d\mu_N(g).

Taking lim sup⁡N→∞\limsup_{N\to\infty} of both sides and using (6), we get

lim sup⁡N→∞μN{g∈G:gK∉Kε}≤ηc2.\limsup_{N\to\infty}\mu_N\{g \in G : gK \notin\mathcal{K}_\varepsilon\} \le\frac{\eta}{c^2}.

Since η>0\eta> 0 is arbitrary, letting η→0\eta\to0 we obtain

lim sup⁡N→∞μN{g∈G:gK∉Kε}=0,\limsup_{N\to\infty}\mu_N\{g \in G : gK \notin\mathcal{K}_\varepsilon\}=0,

which proves (5). For K∈K(X)K \in\mathcal{K}(X), N∈NN \in\mathbb{N} and ε>0\varepsilon> 0, set

HN,ε(K):=μN{g∈G:gK∈Kε}=∫G1Kε(gK) dμN(g).H_{N,\varepsilon}(K):=\mu_N\{g \in G:gK \in\mathcal{K}_\varepsilon\}=\int_G 1_{\mathcal{K}_\varepsilon}(gK)\,d\mu_N(g).

Since μN\mu_N is a probability measure, 0≤HN,ε(K)≤10 \le H_{N,\varepsilon}(K) \le1 for K∈K(X)K \in\mathcal{K}(X). Also, using the fact that ν\nu is nonfinitary, (5) implies that HN,ε(K)→1H_{N,\varepsilon}(K) \to1 for ν\nu-almost every K∈K(X)K \in\mathcal{K}(X). Hence, by the dominated convergence theorem we get,

lim⁡N→∞∫K(X)HN,ε(K) dν(K)=1.(7)\lim_{N\to\infty}\int_{\mathcal{K}(X)}H_{N,\varepsilon}(K)\,d\nu(K)=1. \tag*{(7)}

Using Fubini’s theorem and the GG-invariance of ν\nu, we get

∫K(X)HN,ε(K) dν(K)=∫G∫K(X)1Kε(gK) dν(K) dμN(g)=∫Gν{K∈K(X):gK∈Kε} dμN(g)=∫Gν(Kε) dμN(g)=ν(Kε).\begin{aligned} \int_{\mathcal{K}(X)}H_{N,\varepsilon}(K)\,d\nu(K) &=\int_G\int_{\mathcal{K}(X)}1_{\mathcal{K}_\varepsilon}(gK)\,d\nu(K)\,d\mu_N(g)\\ &=\int_G\nu\{K\in\mathcal{K}(X):gK\in\mathcal{K}_\varepsilon\}\,d\mu_N(g)\\ &=\int_G\nu(\mathcal{K}_\varepsilon)\,d\mu_N(g)=\nu(\mathcal{K}_\varepsilon). \end{aligned}

Letting N→∞N \to\infty on both sides and using (7), we get ν(Kε)=1\nu(\mathcal{K}_\varepsilon)=1, for every ε>0\varepsilon> 0. In particular,

ν(⋂n=1∞K1/n)=1.\nu\left(\bigcap_{n=1}^{\infty}\mathcal{K}_{1/n}\right)=1.

But

⋂n=1∞K1/n={K∈K(X):dH(K,X)=0}={X}.\bigcap_{n=1}^{\infty}\mathcal{K}_{1/n}=\{K\in\mathcal{K}(X):d_H(K,X)=0\}=\{X\}.

Hence ν({X})=1\nu(\{X\})=1, thus ν=δX\nu=\delta_X. Thus the action G↷XG\curvearrowright X is IC-rigid.

Zariski-dense actions on compact homogeneous spaces

In this section, we prove Theorem 1.2. Let X=G/ΛX=G/\Lambda and let mm denote the GG-invariant probability measure on XX. For a semi-group Γ⊆G\Gamma\subseteq G, we denote by

ΓΔ:={(γ,γ):γ∈Γ}⊆G×G,\Gamma^\Delta:=\{(\gamma,\gamma):\gamma\in\Gamma\}\subseteq G\times G,

the diagonal semi-group associated with Γ\Gamma. To prove Theorem 1.2, we will apply an equidistribution result of Benoist–Quint [4] to the action ΓΔ↷X×X\Gamma^\Delta\curvearrowright X\times X where Γ⊆G\Gamma\subseteq G is as in the statement and ΓΔ‾\overline{\Gamma^\Delta} denotes the closure of ΓΔ\Gamma^\Delta in G×GG\times G. Using the equidistribution result, we will show that the action Γ↷X\Gamma\curvearrowright X satisfies the hypothesis of Theorem 1.1, hence it is Glasner and IC-rigid.

Equidistribution result of Benoist–Quint

We now recall an equidistribution result due to Benoist–Quint that will be used in the proof of Theorem 1.2. Let HH be a real Lie group, let L<HL < H be a cocompact lattice, and let ϑ\vartheta be a compactly supported Borel probability measure on HH. Let

Σ:=⟨supp⁡ϑ⟩‾+⊆H\Sigma:= \overline{\langle\operatorname{supp}\vartheta\rangle}_{+} \subseteq H

be the closed sub-semi-group of HH generated by supp⁡ϑ\operatorname{supp}\vartheta. Thus Σ\Sigma is a compactly generated closed sub-semi-group of HH. Assume that the Zariski closure of Ad⁡(Σ)\operatorname{Ad}(\Sigma) is semisimple with no compact factor. For k≥1k \ge1, let ϑ∗k\vartheta^{*k} denote the kk-fold convolution of ϑ\vartheta, that is, the distribution of a product hk⋯h1h_k \cdots h_1 of kk independent elements of HH chosen according to ϑ\vartheta. For completeness we set ϑ∗0:=δe\vartheta^{*0} := \delta_e, where ee is the identity element of HH. For z∈H/Lz \in H/L and k≥0k \ge0, define

ϑ∗k∗δz:=∫Hδhz dϑ∗k(h).\vartheta^{*k} * \delta_z := \int_H \delta_{hz}\,\mathrm{d}\vartheta^{*k}(h).

Thus ϑ∗k∗δz\vartheta^{*k} * \delta_z is a probability measure on H/LH/L. It is the kk-step random-walk distribution starting from zz. Recall that a closed subset YY of a homogeneous space H/LH/L is called a finite-volume homogeneous subspace if its stabilizer HYH_Y acts transitively on YY and there exists an HYH_Y-invariant Borel probability measure νY\nu_Y on YY. We call νY\nu_Y the homogeneous probability measure on YY. We use the following consequence of results of Benoist–Quint [4 Theorems 1.1, 1.2 and 1.7], stated in the form adapted to the probability measure ϑ\vartheta above.

Theorem 4.1. [4] Let HH, LL, ϑ\vartheta and Σ\Sigma be as above and for z∈H/Lz \in H/L, set Yz:=Σz‾Y_z := \overline{\Sigma z}. Then YzY_z is a Σ\Sigma-invariant, Σ\Sigma-ergodic finite-volume homogeneous subspace. Let νz\nu_z denote the homogeneous probability measure on YzY_z. Then

1N∑k=0N−1ϑ∗k∗δz⟶νz\frac{1}{N}\sum_{k=0}^{N-1}\vartheta^{*k} * \delta_z \longrightarrow\nu_z

in the weak-* topology. If, in addition, the centralizer of Σ\Sigma in HH is discrete, then for every sequence (zn)n∈N(z_n)_{n\in\mathbb{N}} in H/LH/L, the corresponding homogeneous subspaces Yzn:=Σzn‾Y_{z_n}:=\overline{\Sigma z_n} have a subsequence (Yznm)m∈N(Y_{z_{n_m}})_{m\in\mathbb{N}} converging to a Σ\Sigma-invariant, Σ\Sigma-ergodic finite-volume homogeneous subspace YY. Moreover, Yznm⊆YY_{z_{n_m}} \subseteq Y for all sufficiently large m∈Nm \in\mathbb{N}.

Proof of Theorem 1.2

In this subsection we prove Theorem 1.2 using Theorem 1.1 and Theorem 4.1.

Proof of Theorem 1.2. Let μ\mu be a finitely supported probability measure on GG whose support generates an Ad-Zariski-dense semi-group Γ⊆G\Gamma\subseteq G. For k≥1k \ge1, let μ∗k\mu^{*k} denote the kk-fold convolution of μ\mu, that is, the distribution of a product gk⋯g1g_k \cdots g_1 of kk independent elements chosen according to μ\mu. For N≥1N \ge1, define

μN:=1N∑k=1Nμ∗k.\mu_N := \frac{1}{N}\sum_{k=1}^{N}\mu^{*k}.

Thus μN\mu_N is a probability measure on Γ\Gamma. We will show that the sequence (μN)N≥1(\mu_N)_{N\ge1} satisfies the hypothesis of Theorem 1.1. We set

H:=G×G,L:=Λ×Λ,H := G \times G,\qquad L := \Lambda\times\Lambda,

and identify H/L=X×X=(G×G)/(Λ×Λ)H/L = X \times X = (G \times G)/(\Lambda\times\Lambda). Since Λ\Lambda is cocompact in GG, the lattice LL is cocompact in HH. Let Δ\Delta be the diagonal map given by

Δ:G⟶G×G,Δ(g):=(g,g).\Delta: G \longrightarrow G \times G,\qquad\Delta(g):=(g,g).

Let ϑ:=Δ∗μ\vartheta:= \Delta_*\mu be the pushforward of the measure μ\mu under the diagonal map. Since μ\mu is finitely supported probability measure on GG, ϑ\vartheta is a compactly supported probability measure on HH. Moreover, its support generates the diagonal semi-group ΓΔ\Gamma^\Delta. Hence, using the notation of Theorem 4.1, we have Σ=ΓΔ\Sigma= \Gamma\Delta. Observe that, for k∈Nk \in\mathbb{N}, ϑ∗k=Δ∗(μ∗k)\vartheta^{*k} = \Delta_{*}(\mu^{*k}). In other words ϑ∗k\vartheta^{*k} is the pushforward of μ∗k\mu^{*k}.

Let g:=Lie⁡(G)\mathfrak{g} := \operatorname{Lie}(G), which is simple by assumption. Hence it follows that Lie⁡(H)=g⊕g\operatorname{Lie}(H) = \mathfrak{g} \oplus\mathfrak{g} and for (γ,γ)∈H(\gamma,\gamma) \in H, Ad⁡H(γ,γ)=(Ad⁡(γ),Ad⁡(γ))\operatorname{Ad}_{H}(\gamma,\gamma) = (\operatorname{Ad}(\gamma),\operatorname{Ad}(\gamma)). Since Σ=ΓΔ\Sigma= \Gamma\Delta, the Zariski closure of Ad⁡H(Σ)\operatorname{Ad}_{H}(\Sigma) is the diagonal copy of the Zariski closure of Ad⁡(Γ)\operatorname{Ad}(\Gamma). Since Γ\Gamma is Ad-Zariski dense in GG, this is the diagonal copy of the Zariski closure of Ad⁡(G)\operatorname{Ad}(G). Since g\mathfrak{g} is simple and GG is noncompact, the Zariski closure of Ad⁡H(Σ)\operatorname{Ad}_{H}(\Sigma) is semisimple with no compact factor. Hence the Zariski-closure hypothesis of Theorem 4.1 is satisfied.

We claim that the centralizer of Σ\Sigma in HH is discrete. Let (U,V)∈Lie⁡(ZH(Σ))(U,V) \in\operatorname{Lie}(Z_{H}(\Sigma)), since (U,V)(U,V) belongs to the Lie algebra of the centralizer it is fixed under the adjoint action of every element of Σ\Sigma, that is, Ad⁡H(σ)(U,V)=(U,V)\operatorname{Ad}_{H}(\sigma)(U,V) = (U,V) for every σ∈Σ\sigma\in\Sigma. In particular, since ΓΔ⊆Σ\Gamma\Delta\subseteq\Sigma, for every γ∈Γ\gamma\in\Gamma we have

Ad⁡H(γ,γ)(U,V)=(Ad⁡(γ)U,Ad⁡(γ)V)=(U,V).\operatorname{Ad}_{H}(\gamma,\gamma)(U,V) = (\operatorname{Ad}(\gamma)U,\operatorname{Ad}(\gamma)V) = (U,V).

Hence Ad⁡(γ)U=U\operatorname{Ad}(\gamma)U = U and Ad⁡(γ)V=V\operatorname{Ad}(\gamma)V = V for every γ∈Γ\gamma\in\Gamma. By the Ad-Zariski density of Γ\Gamma, the same holds for every g∈Gg \in G. Now, for Z∈gZ \in\mathfrak{g}, taking g=exp⁡(tZ)g = \exp(tZ) gives

Ad⁡(exp⁡(tZ))U=U,Ad⁡(exp⁡(tZ))V=V.\operatorname{Ad}(\exp(tZ))U = U,\qquad\operatorname{Ad}(\exp(tZ))V = V.

for every tt. Thus Ad⁡(exp⁡(tZ))U\operatorname{Ad}(\exp(tZ))U is constant in tt. Differentiating at t=0t = 0, and using the fact that

ddt∣t=0Ad⁡(exp⁡(tZ))U=[Z,U]\left.\frac{d}{dt}\right|_{t=0} \operatorname{Ad}(\exp(tZ))U = [Z,U]

we get [Z,U]=0[Z,U] = 0 for every Z∈gZ \in\mathfrak{g}. Similarly we get [Z,V]=0[Z,V] = 0 for every Z∈gZ \in\mathfrak{g}. In other words UU and VV belong to the center of g\mathfrak{g}. Since the center of the simple Lie algebra g\mathfrak{g} is trivial, U=V=0U = V = 0. Therefore Lie⁡(ZH(Σ))=0\operatorname{Lie}(Z_{H}(\Sigma)) = 0. Since ZH(Σ)Z_{H}(\Sigma) is a closed subgroup of the Lie group HH, this implies that ZH(Σ)Z_{H}(\Sigma) is discrete. Hence the hypothesis of Theorem 4.1 holds.

Now fix f∈C(X;C)f \in C(X;\mathbb{C}) satisfying ∫Xf dm=0\int_X f\,dm = 0, and for u,v∈Xu,v \in X, set Φf(u,v):=f(u)f(v)‾\Phi_f(u,v) := f(u)\overline{f(v)} for convenience. For x,y∈Xx,y \in X, set

Yx,y:=Σ(x,y)‾Y_{x,y} := \overline{\Sigma(x,y)}

and let νx,y\nu_{x,y} denote its homogeneous probability measure. By Theorem 4.1,

1N∑k=0N−1ϑ∗k∗δ(x,y)⟶νx,y(8)\frac{1}{N}\sum_{k=0}^{N-1}\vartheta^{*k} * \delta_{(x,y)} \longrightarrow\nu_{x,y} \tag*{(8)}

in the weak-* topology. Hence we get,

KN,f(x,y)=∫Gf(gx)f(gy)‾ dμN(g)=∫X×XΦf d(1N∑k=1Nϑ∗k∗δ(x,y))⟶N→∞∫Yx,yΦf dνx,y.(9)K_{N,f}(x,y) = \int_G f(gx)\overline{f(gy)}\,d\mu_N(g) = \int_{X \times X}\Phi_f\,d\left(\frac{1}{N}\sum_{k=1}^{N}\vartheta^{*k} * \delta_{(x,y)}\right) \underset{N\to\infty}{\longrightarrow} \int_{Y_{x,y}}\Phi_f\,d\nu_{x,y}. \tag*{(9)}

Here we used the fact that averages from k=1k = 1 to NN have the same limit as (8), since the two averages differ only in the k=0k = 0 and k=Nk = N terms, each divided by NN. Fix x∈Xx \in X and η>0\eta> 0. We claim that

Ef,η(x):={y∈X∖{x}:lim sup⁡N→∞∣KN,f(x,y)∣>η}E_{f,\eta}(x) := \left\{y \in X \setminus\{x\} : \limsup_{N\to\infty} \lvert K_{N,f}(x,y)\rvert> \eta\right\}

is finite. To see this assume by contradiction that Ef,η(x)E_{f,\eta}(x) is infinite. For n∈Nn \in\mathbb{N}, choose distinct yn∈Ef,η(x)y_n \in E_{f,\eta}(x) and set

Yn:=Yx,yn=Σ(x,yn)‾,νn:=νx,yn.Y_n := Y_{x,y_n} = \overline{\Sigma(x,y_n)},\qquad\nu_n := \nu_{x,y_n}.

Here for n∈Nn \in\mathbb{N}, νn\nu_n denotes the homogeneous probability measure on YnY_n. Then by definition of Ef,η(x)E_{f,\eta}(x) and (9) we have

∣∫YnΦf dνn∣>ηfor every n.\left|\int_{Y_n}\Phi_f\,d\nu_n\right| > \eta\qquad\text{for every } n.

By Theorem 4.1, after passing to a subsequence, there exists a Σ\Sigma-invariant, Σ\Sigma-ergodic finite-volume homogeneous subspace Y⊆X×XY \subseteq X \times X such that Yn⊆YY_n \subseteq Y for all sufficiently large nn, and νn⟶νY\nu_n \longrightarrow\nu_Y in the weak-* topology. This implies

∣∫YΦf dνY∣≥η.(10)\left|\int_Y \Phi_f\,d\nu_Y\right| \geq\eta. \tag*{(10)}

Since (x,yn)∈Yn⊆Y(x,y_n) \in Y_n \subseteq Y for all sufficiently large nn, the fiber F:=Y∩({x}×X)F := Y \cap(\{x\} \times X) is infinite. Let HYH_Y be the stabilizer of YY in HH. Since YY is a finite-volume homogeneous subspace, HYH_Y acts transitively on YY and preserves the homogeneous probability measure νY\nu_Y. We set

Gx:={g∈G:gx=x}.G_x := \{g \in G : gx = x\}.

Since X=G/ΛX = G/\Lambda, the group GxG_x is conjugate to Λ\Lambda, and hence is discrete. The group

K:=HY∩(Gx×G)K := H_Y \cap(G_x \times G)

acts transitively on FF. Assume that KK were discrete. Since HH is a second-countable Lie group, KK would be countable. As KK acts transitively on FF, the set FF would be countable and homogeneous under the action of KK. Moreover, F=Y∩({x}×X)F = Y \cap(\{x\} \times X) is compact, hence FF would be a countable compact homogeneous space. Recall that, by the Baire category theorem, every countable compact homogeneous space is finite. Hence we reach a contradiction. Thus KK is not discrete. Since KK is a closed subgroup of the Lie group HH, it is itself a Lie group, hence Lie⁡(K)≠0\operatorname{Lie}(K) \ne0. Since GxG_x is discrete, Lie⁡(Gx)=0\operatorname{Lie}(G_x) = 0, and K⊆HY∩(Gx×G)K \subseteq H_Y \cap(G_x \times G), we get

0≠Lie⁡(K)⊆hY∩(0⊕g),0 \ne\operatorname{Lie}(K) \subseteq\mathfrak{h}_Y \cap(0 \oplus\mathfrak{g}),

where hY:=Lie⁡(HY)\mathfrak{h}_Y := \operatorname{Lie}(H_Y). In other words, there exists Z∈g∖{0}Z \in\mathfrak{g} \setminus\{0\}, such that (0,Z)∈hY∩(0⊕g)(0,Z) \in\mathfrak{h}_Y \cap(0 \oplus\mathfrak{g}). Thus the set

i:={Z∈g:(0,Z)∈hY}\mathfrak{i} := \{Z \in\mathfrak{g} : (0,Z) \in\mathfrak{h}_Y\}

is nonzero. Since YY is a finite-volume homogeneous subspace and is Σ\Sigma-invariant, Σ\Sigma-ergodic, by definition Σ\Sigma is a sub-semi-group of the stabilizer HYH_Y and acts ergodically on (Y,νY)(Y,\nu_Y). In particular, Σ⊆HY\Sigma\subseteq H_Y. Hence it follows that, (γ,γ)∈HY(\gamma,\gamma) \in H_Y for every γ∈Γ\gamma\in\Gamma. Thus the adjoint action of (γ,γ)(\gamma,\gamma) preserves the Lie algebra hY\mathfrak{h}_Y. In other words, if Z∈iZ \in\mathfrak{i}, then (0,Z)∈hY(0,Z) \in\mathfrak{h}_Y and

Ad⁡H(γ,γ)(0,Z)=(0,Ad⁡(γ)Z)∈hY.\operatorname{Ad}_H(\gamma,\gamma)(0,Z) = (0,\operatorname{Ad}(\gamma)Z) \in\mathfrak{h}_Y.

Hence Ad⁡(γ)Z∈i\operatorname{Ad}(\gamma)Z \in\mathfrak{i} for every γ∈Γ\gamma\in\Gamma, and therefore i\mathfrak{i} is invariant under Ad⁡(Γ)\operatorname{Ad}(\Gamma). By the Ad-Zariski density of Γ\Gamma, i\mathfrak{i} is invariant under Ad⁡(G)\operatorname{Ad}(G). Differentiating the adjoint action gives

[g,i]⊆i.[\mathfrak{g},\mathfrak{i}] \subseteq\mathfrak{i}.

Thus i\mathfrak{i} is an ideal of g\mathfrak{g}. Since g\mathfrak{g} is simple and i≠0\mathfrak{i} \ne0, we get i=g\mathfrak{i} = \mathfrak{g}. Hence by definition of i\mathfrak{i}, we have

0⊕g⊆hY.(11)0 \oplus\mathfrak{g} \subseteq\mathfrak{h}_Y. \tag*{(11)}

Since GG is a connected Lie group, (11) implies

{e}×G⊆HY.\{e\} \times G \subseteq H_Y.

Since the homogeneous measure νY\nu_Y is invariant under the action of HYH_Y, it follows that νY\nu_Y is invariant under the action of {e}×G\{e\} \times G. We define a finite complex measure ρ\rho on XX as follows. Let

∫Xψ(v) dρ(v):=∫X×Xf(u)ψ(v) dνY(u,v),\int_X \psi(v)\,d\rho(v) := \int_{X \times X} f(u)\psi(v)\,d\nu_Y(u,v),

for every ψ∈C(X)\psi\in C(X). This defines ρ\rho by the Riesz representation theorem. The invariance of νY\nu_Y under {e}×G\{e\} \times G implies that ρ\rho is GG-invariant. Since every finite GG-invariant complex measure on G/ΛG/\Lambda is a scalar multiple of mm, it follows that ρ=cm\rho= cm for some c∈Cc \in\mathbb{C}. Taking ψ=f‾\psi= \overline{f}, we obtain

∫YΦf dνY=∫X×Xf(u)f(v)‾ dνY(u,v)=∫Xf(v)‾ dρ(v)=c∫Xf‾ dm=0.\int_Y \Phi_f\,d\nu_Y = \int_{X \times X} f(u)\overline{f(v)}\,d\nu_Y(u,v) = \int_X \overline{f(v)}\,d\rho(v) = c\int_X \overline{f}\,dm = 0.

This contradicts (10), hence Ef,η(x)E_{f,\eta}(x) is finite. Thus the sequence (μN)N≥1(\mu_N)_{N\geq1} satisfies the hypothesis of Theorem 1.1. Hence Theorem 1.1 implies that Γ↷X\Gamma\curvearrowright X has the Glasner property and is IC-rigid. □

Note that the Theorem 1.3 follows immediately from Theorem 1.2.

Proof of Theorem 1.3. For G=SLd(R)G=\mathrm{SL}_d(\mathbb{R}) or G=PSLd(R)G=\mathrm{PSL}_d(\mathbb{R}), the group GG is a connected noncompact Lie group and

Lie⁡(G)=sld(R),\operatorname{Lie}(G)=\mathfrak{sl}_d(\mathbb{R}),

which is simple for d≥2d\geq2. Since Γ\Gamma is Zariski dense in GG, and the adjoint representation is an algebraic homomorphism it follows that Ad⁡(Γ)\operatorname{Ad}(\Gamma) is Zariski dense in Ad⁡(G)\operatorname{Ad}(G). Thus Γ\Gamma is Ad-Zariski dense in GG. Therefore all the hypotheses of Theorem 1.2 are satisfied. Hence the action Γ↷G/Λ\Gamma\curvearrowright G/\Lambda has the Glasner property and is IC-rigid. □

Glasner Property and IC-rigidity of Torus actions

In this section we prove Theorem 1.4 which is a Fourier-analytic condition that implies IC-rigidity and the Glasner property for semi-group actions on dd-dimensional torus. Using this criterion we prove Theorem 1.5, i.e. the matrix-valued polynomial actions on the dd-dimensional torus are IC-rigid. The Glasner property for these actions was previously established by Bulinski and Fish [2], who in fact proved the stronger uniform Glasner property.

To do so we need an elementary result showing that small Fourier coefficients for a measure on the torus force dense support (see Theorem 5.1). Note that this result can be derived from Theorem 3.1 but we provide a simple standalone proof here. Recall that for a Borel probability measure μ\mu on Td\mathbb{T}^d, and m∈Zdm\in\mathbb{Z}^d, we denote the mm-th Fourier coefficient of μ\mu by

μ^(m):=∫Tde(⟨m,x⟩) dμ(x).\widehat{\mu}(m):=\int_{\mathbb{T}^d}e(\langle m,x\rangle)\,d\mu(x).

Proposition 5.1. For every ε>0\varepsilon>0, there exist a constant cε>0c_\varepsilon>0 and a finite nonempty set Λε⊆Zd∖{0}\Lambda_\varepsilon\subseteq\mathbb{Z}^d\setminus\{0\} such that the following holds. If μ\mu is a Borel probability measure on Td\mathbb{T}^d satisfying

∣μ^(m)∣<cεfor every m∈Λε,|\widehat{\mu}(m)|<c_\varepsilon\qquad\text{for every }m\in\Lambda_\varepsilon,

then supp⁡μ\operatorname{supp}\mu is ε\varepsilon-dense in Td\mathbb{T}^d.

Proof. Let λ\lambda denote normalized Haar measure on Td\mathbb{T}^d. Choose y1,…,ys∈Tdy_1,\ldots,y_s\in\mathbb{T}^d such that the collection of open balls {Uj:=B(yj,ε/2):1≤j≤s}\{U_j:=B(y_j,\varepsilon/2):1\leq j\leq s\}, cover Td\mathbb{T}^d. For each 1≤j≤s1\leq j\leq s, choose a non-negative continuous function fjf_j such that supp⁡fj⊆Uj\operatorname{supp}f_j\subseteq U_j and ∫Tdfj dλ=1\int_{\mathbb{T}^d}f_j\,d\lambda=1. Such a function can be obtained by taking a non-negative bump function supported in a smaller ball around yjy_j and dividing by its integral. Using the fact that trigonometric polynomials are uniformly dense in C(Td)C(\mathbb{T}^d), for every 1≤j≤s1\leq j\leq s, we can choose a trigonometric polynomial PjP_j such that

∥Pj−fj∥∞<18.\|P_j-f_j\|_\infty<\frac{1}{8}.

Note that there exists a common finite set Λε⊆Zd∖{0}\Lambda_\varepsilon\subseteq\mathbb{Z}^d\setminus\{0\} such that for every 1≤j≤s1\leq j\leq s we can write

Pj(x)=aj,0+∑m∈Λεaj,me(⟨m,x⟩)P_j(x)=a_{j,0}+\sum_{m\in\Lambda_\varepsilon}a_{j,m}e(\langle m,x\rangle)

for x∈Tdx\in\mathbb{T}^d. We denote by

C:=max⁡1≤j≤s∑m∈Λε∣aj,m∣,andcε:=14(1+C).C:=\max_{1\leq j\leq s}\sum_{m\in\Lambda_\varepsilon}|a_{j,m}|,\qquad\text{and}\qquad c_\varepsilon:=\frac{1}{4(1+C)}.

Since λ\lambda is a Haar measure, for m∈Zd∖{0}m\in\mathbb{Z}^d\setminus\{0\}, ∫Tde(⟨m,x⟩) dλ(x)=0\int_{\mathbb{T}^d}e(\langle m,x\rangle)\,d\lambda(x)=0. Thus for every 1≤j≤s1\leq j\leq s,

∫TdPj dμ−∫TdPj dλ=∑m∈Λεaj,mμ^(m).\int_{\mathbb{T}^d}P_j\,d\mu-\int_{\mathbb{T}^d}P_j\,d\lambda=\sum_{m\in\Lambda_\varepsilon}a_{j,m}\widehat{\mu}(m).

Using the fact that ∣μ^(m)∣<cε|\widehat{\mu}(m)| < c_{\varepsilon} it follows that for every 1≤j≤s1 \le j \le s,

∣∫Tdfj dμ−1∣=∣∫Tdfj dμ−∫Tdfj dλ∣≤2∥fj−Pj∥∞+∑m∈Λε∣aj,m∣∣μ^(m)∣<14+cεC<12.\left|\int_{\mathbb{T}^{d}} f_j\,d\mu-1\right|=\left|\int_{\mathbb{T}^{d}} f_j\,d\mu-\int_{\mathbb{T}^{d}} f_j\,d\lambda\right|\le2\|f_j-P_j\|_{\infty}+\sum_{m\in\Lambda_{\varepsilon}}|a_{j,m}||\widehat{\mu}(m)|<\frac{1}{4}+c_{\varepsilon}C<\frac{1}{2}.

Thus ∫Tdfj dμ>1/2\int_{\mathbb{T}^{d}} f_j\,d\mu>1/2 for every j∈{1,…,s}j\in\{1,\ldots,s\}. Since fjf_j vanishes outside UjU_j and ∫Tdfj dμ>0\int_{\mathbb{T}^{d}} f_j\,d\mu>0, it follows that supp⁡μ∩Uj≠∅\operatorname{supp}\mu\cap U_j\ne\varnothing for every j∈{1,…,s}j\in\{1,\ldots,s\}. Since UjU_j are balls of radius ε/2\varepsilon/2 that cover Td\mathbb{T}^{d}, it follows that supp⁡μ\operatorname{supp}\mu is ε\varepsilon-dense in Td\mathbb{T}^{d}.

Now we are ready to prove Theorem 1.4.

Proof of Theorem 1.4. Let λ\lambda denote normalized Haar measure on Td\mathbb{T}^{d}. Fix ε>0\varepsilon>0 and let cε>0c_{\varepsilon}>0 and Λε⊆Zd∖{0}\Lambda_{\varepsilon}\subseteq\mathbb{Z}^{d}\setminus\{0\} be as in Theorem 5.1. We set q:=∣Λε∣q:=|\Lambda_{\varepsilon}| and enumerate Λε={m1,…,mq}\Lambda_{\varepsilon}=\{m_1,\ldots,m_q\}. For every 1≤j≤q1\le j\le q we define fj(x):=e(⟨mj,x⟩)f_j(x):=e(\langle m_j,x\rangle). Since mj≠0m_j\ne0, it follows that ∫Tdfj dλ=0\int_{\mathbb{T}^{d}}f_j\,d\lambda=0, for every 1≤j≤q1\le j\le q. In other words f1,…,fq∈C(Td;C)f_1,\ldots,f_q\in C(\mathbb{T}^{d};\mathbb{C}), is a collection of zero-mean functions. Note that for every Borel probability measure ν\nu on Td\mathbb{T}^{d},

∫Tdfj dν=∫Tde(⟨mj,x⟩) dν(x)=ν^(mj).\int_{\mathbb{T}^{d}} f_j\,d\nu=\int_{\mathbb{T}^{d}}e(\langle m_j,x\rangle)\,d\nu(x)=\widehat{\nu}(m_j).

Hence, by Theorem 5.1, if

∣∫Tdfj dν∣<cεfor every 1≤j≤q,\left|\int_{\mathbb{T}^{d}} f_j\,d\nu\right|<c_{\varepsilon}\qquad\text{for every }1\le j\le q,

then supp⁡ν\operatorname{supp}\nu is ε\varepsilon-dense in Td\mathbb{T}^{d}. Thus the functions f1,…,fqf_1,\ldots,f_q, with the common constant cεc_{\varepsilon}, have precisely the ε\varepsilon-density property used in the proof of Theorem 1.1. Recall that in the proof of Theorem 1.1, for each ε>0\varepsilon>0, we used Theorem 3.1 to obtain a finite collection of zero-mean functions and positive constants which detect ε\varepsilon-density. Here Theorem 5.1 provides such a collection, namely f1,…,fqf_1,\ldots,f_q, with common constant cεc_{\varepsilon}. The remainder of the proof of Theorem 1.1 uses the pair-correlation condition only for this finite collection. We now verify the pair-correlation condition for this finite collection of functions using the hypothesis of Theorem 1.4. For 1≤j≤q1\le j\le q and x,y∈Tdx,y\in\mathbb{T}^{d} we have

KN,fj(x,y)=∫Ge(⟨mj,gx⟩)e(⟨mj,gy⟩)‾ dμN(g)=∫Ge(⟨mj,g(x−y)⟩) dμN(g).K_{N,f_j}(x,y)=\int_G e(\langle m_j,gx\rangle)\overline{e(\langle m_j,gy\rangle)}\,d\mu_N(g)=\int_G e(\langle m_j,g(x-y)\rangle)\,d\mu_N(g).

Fix x∈Tdx\in\mathbb{T}^{d} and η>0\eta>0, and set

Emj,η:={z∈Td:lim sup⁡N→∞∣∫Ge(⟨mj,gz⟩) dμN(g)∣≥η}.E_{m_j,\eta}:=\left\{z\in\mathbb{T}^{d}:\limsup_{N\to\infty}\left|\int_G e(\langle m_j,gz\rangle)\,d\mu_N(g)\right|\ge\eta\right\}.

By the hypothesis of Theorem 1.4, Emj,ηE_{m_j,\eta} is finite. If

lim sup⁡N→∞∣KN,fj(x,y)∣>η,\limsup_{N\to\infty}|K_{N,f_j}(x,y)|>\eta,

then x−y∈Emj,ηx-y\in E_{m_j,\eta}. Hence it follows that

Rj,η(x):={y∈Td∖{x}:lim sup⁡N→∞∣KN,fj(x,y)∣>η}⊆x−Emj,η.R_{j,\eta}(x):=\left\{y\in\mathbb{T}^{d}\setminus\{x\}:\limsup_{N\to\infty}|K_{N,f_j}(x,y)|>\eta\right\}\subseteq x-E_{m_j,\eta}.

Thus Rj,η(x)R_{j,\eta}(x) is finite for every 1≤j≤q1\le j\le q. Hence the proof of Theorem 1.1, applied with the finite collection of functions above, gives that the action G↷TdG\curvearrowright\mathbb{T}^{d} has the Glasner property and is IC-rigid. □\square

Matrix-valued polynomial actions are IC-rigid

Proof of Theorem 1.5. Let A(x)∈Md(Z[x])A(x) \in M_d(\mathbb{Z}[x]) be as in the statement of the theorem. We consider the action of the semi-group

GA={A(n1)A(n2)⋯A(nℓ):ℓ≥1, n1,…,nℓ∈Z, det⁡A(ni)≠0 for all 1≤i≤ℓ}.G_A=\{A(n_1)A(n_2)\cdots A(n_\ell):\ell\geq1,\ n_1,\ldots,n_\ell\in\mathbb{Z},\ \det A(n_i)\ne0\text{ for all }1\leq i\leq\ell\}.

on Td\mathbb{T}^d by endomorphisms. Since det⁡A(x)\det A(x) is a nonzero polynomial, there are only finitely many integers nn for which det⁡A(n)=0\det A(n)=0. Thus, for all sufficiently large NN, let

IN:={n∈Z:∣n∣≤N, det⁡A(n)≠0}I_N:=\{n\in\mathbb{Z}:|n|\leq N,\ \det A(n)\ne0\}

and define a probability measure on GAG_A by

μN:=1∣IN∣∑n∈INδA(n).\mu_N:=\frac{1}{|I_N|}\sum_{n\in I_N}\delta_{A(n)}.

For the finitely many remaining NN, define μN\mu_N arbitrarily. Since only finitely many integers are excluded, it follows that ∣IN∣2N+1→1\frac{|I_N|}{2N+1}\to1. For z∈Tdz\in\mathbb{T}^d, the average orbit probability measure σN,z\sigma_{N,z} (corresponding to μN\mu_N) is given by

σN,z:=∫GAδgz dμN(g)=1∣IN∣∑n∈INδA(n)z.\sigma_{N,z}:=\int_{G_A}\delta_{gz}\,d\mu_N(g)=\frac{1}{|I_N|}\sum_{n\in I_N}\delta_{A(n)z}.

For m∈Zdm\in\mathbb{Z}^d, the mm-th Fourier coefficient of σN,z\sigma_{N,z} is given by

σ^N,z(m)=∫Tde(⟨m,x⟩) dσN,z(x)=∫GAe(⟨m,gz⟩) dμN(g)=1∣IN∣∑n∈INe(⟨m,A(n)z⟩).\widehat{\sigma}_{N,z}(m)=\int_{\mathbb{T}^d}e(\langle m,x\rangle)\,d\sigma_{N,z}(x)=\int_{G_A}e(\langle m,gz\rangle)\,d\mu_N(g)=\frac{1}{|I_N|}\sum_{n\in I_N}e(\langle m,A(n)z\rangle).

Fix m∈Zd∖{0}m\in\mathbb{Z}^d\setminus\{0\} and δ>0\delta>0, and set

Em,δ:={z∈Td:lim sup⁡N→∞∣1∣IN∣∑n∈INe(⟨m,A(n)z⟩)∣≥δ}.E_{m,\delta}:=\left\{z\in\mathbb{T}^d:\limsup_{N\to\infty}\left|\frac{1}{|I_N|}\sum_{n\in I_N}e(\langle m,A(n)z\rangle)\right|\geq\delta\right\}.

To prove that the action is IC-rigid, by Theorem 1.4 (b), it is enough to show that Em,δE_{m,\delta} is finite.

We first consider a non-torsion point z∈Tdz\in\mathbb{T}^d. For m∈Zdm\in\mathbb{Z}^d let mtm^t denote the corresponding row vector. We write

mt(A(x)−A(0))=(P1(x),…,Pd(x)),m^t(A(x)-A(0))=(P_1(x),\ldots,P_d(x)),

where P1,…,Pd∈Z[x]P_1,\ldots,P_d\in\mathbb{Z}[x]. The polynomials P1,…,PdP_1,\ldots,P_d are linearly independent over Z\mathbb{Z}. Indeed, otherwise there would exist w∈Zd∖{0}w\in\mathbb{Z}^d\setminus\{0\} such that

m⋅(A(x)−A(0))w=0,m\cdot(A(x)-A(0))w=0,

contrary to the hypothesis of the theorem. Since z=(z1,…,zd)z=(z_1,\ldots,z_d) is non-torsion, at least one ziz_i is irrational. We claim that the polynomial ∑i=1dziPi(x)\sum_{i=1}^d z_iP_i(x) has an irrational coefficient. To see this note that, since P1,…,PdP_1,\ldots,P_d are linearly independent over Z\mathbb{Z} they are also linearly independent over Q\mathbb{Q}. Let DD be the maximum of the degrees of the entries of A(x)A(x). Since Pi(0)=0P_i(0)=0 and deg⁡Pi≤D\deg P_i\leq D for every ii, we may write Pi(x)=∑j=1DbjixjP_i(x)=\sum_{j=1}^D b_{ji}x^j, where bji∈Zb_{ji}\in\mathbb{Z}.

Let B∈MD×d(Z)B\in M_{D\times d}(\mathbb{Z}) be the matrix whose ii-th column is the coefficient vector of PiP_i. Since P1,…,PdP_1,\ldots,P_d are linearly independent over Q\mathbb{Q}, the matrix BB has full column rank over Q\mathbb{Q}. Assume by contradiction that polynomial ∑i=1dziPi(x)\sum_{i=1}^d z_iP_i(x) has only rational coefficients then Bz∈QDBz\in\mathbb{Q}^D. Since BB has full column rank over Q\mathbb{Q}, this would imply that z∈Qdz\in\mathbb{Q}^d, contradicting the assumption that zz is non-torsion. Hence ∑i=1dziPi(x)\sum_{i=1}^{d} z_iP_i(x) has an irrational coefficient. Since Pi(0)=0P_i(0)=0 for every ii, this coefficient is necessarily non-constant. Note that

⟨m,A(x)z⟩=mtA(0)z+mt(A(x)−A(0))z=⟨m,A(0)z⟩+∑i=1dziPi(x),\langle m,A(x)z\rangle= m^tA(0)z+m^t(A(x)-A(0))z=\langle m,A(0)z\rangle+\sum_{i=1}^{d}z_iP_i(x),

and the first term is constant in xx. Hence ⟨m,A(x)z⟩\langle m,A(x)z\rangle has at least one irrational non-constant coefficient. By Weyl’s equidistribution theorem, we get

lim⁡N→∞12N+1∑∣n∣≤Ne(⟨m,A(n)z⟩)=0.\lim_{N\to\infty}\frac{1}{2N+1}\sum_{|n|\leq N}e(\langle m,A(n)z\rangle)=0.

Removing the finitely many nn for which det⁡A(n)=0\det A(n)=0 does not affect the limit. Therefore

lim⁡N→∞1∣IN∣∑n∈INe(⟨m,A(n)z⟩)=0.\lim_{N\to\infty}\frac{1}{|I_N|}\sum_{n\in I_N}e(\langle m,A(n)z\rangle)=0.

Thus every point of Em,δE_{m,\delta} is a torsion point. For a torsion point z∈Tdz\in\mathbb{T}^{d}, let

z=aq+Zd,a=(a1,…,ad)∈Zd,gcd⁡(a1,…,ad,q)=1z=\frac{a}{q}+\mathbb{Z}^{d},\qquad a=(a_1,\ldots,a_d)\in\mathbb{Z}^{d},\qquad\gcd(a_1,\ldots,a_d,q)=1

be its reduced form. Since A(x)A(x) has integer polynomial entries and m,a∈Zdm,a\in\mathbb{Z}^{d}, the quantity m⋅A(x)am\cdot A(x)a is a polynomial in Z[x]\mathbb{Z}[x] of degree at most DD. Write

m⋅A(x)a=b0+b1x+⋯+bDxD,m\cdot A(x)a=b_0+b_1x+\cdots+b_Dx^D,

where b0,…,bD∈Zb_0,\ldots,b_D\in\mathbb{Z}. Using the fact that m⋅A(0)a=b0m\cdot A(0)a=b_0, we get

(mtA(x)−mtA(0))a=b1x+⋯+bDxD.(12)(m^tA(x)-m^tA(0))a=b_1x+\cdots+b_Dx^D. \tag*{(12)}

On the other hand, since mt(A(x)−A(0))=(P1(x),…,Pd(x))m^t(A(x)-A(0))=(P_1(x),\ldots,P_d(x)) the left-hand side in (12) is equal to ∑i=1daiPi(x)\sum_{i=1}^{d}a_iP_i(x). By the definition of BB, the coefficient vector of this polynomial is BaBa. Therefore

Ba=(b1,…,bD)t.Ba=(b_1,\ldots,b_D)^t.

Since BB has full column rank over Q\mathbb{Q}, there exists a matrix L∈Md×D(Q)L\in M_{d\times D}(\mathbb{Q}) such that LB=IdLB=I_d. Clearing denominators, there exist C∈Md×D(Z)C\in M_{d\times D}(\mathbb{Z}) and an integer Qm≥1Q_m\geq1, depending only on A(x)A(x) and mm, such that CB=QmIdCB=Q_mI_d. Let r:=gcd⁡(b1,…,bD,q)r:=\gcd(b_1,\ldots,b_D,q). Since rr divides every entry of BaBa and also divides qq, it follows from CBa=QmaCBa=Q_ma that r∣gcd⁡(Qma1,…,Qmad,q)r\mid\gcd(Q_ma_1,\ldots,Q_ma_d,q). Since gcd⁡(a1,…,ad,q)=1\gcd(a_1,\ldots,a_d,q)=1, we have gcd⁡(Qma1,…,Qmad,q)∣Qm\gcd(Q_ma_1,\ldots,Q_ma_d,q)\mid Q_m. Hence r≤Qmr\leq Q_m. We set

q0:=qr,bj′:=bjr,1≤j≤D.q_0:=\frac{q}{r},\qquad b'_j:=\frac{b_j}{r},\qquad1\leq j\leq D.

Then

gcd⁡(b1′,…,bD′,q0)=1.(13)\gcd(b'_1,\ldots,b'_D,q_0)=1. \tag*{(13)}

Since m⋅A(x)am\cdot A(x)a has integer coefficients, m⋅A(n+q)a−m⋅A(n)am\cdot A(n+q)a-m\cdot A(n)a is divisible by qq. Hence the sequence

e(⟨m,A(n)z⟩)=e(m⋅A(n)aq)e(\langle m,A(n)z\rangle)=e\left(\frac{m\cdot A(n)a}{q}\right)

is periodic in nn with period qq. Since only finitely many terms are removed from the symmetric averages, we obtain

lim⁡N→∞1∣IN∣∑n∈INe(⟨m,A(n)z⟩)=1q∑n=1qe(m⋅A(n)aq).\lim_{N\to\infty}\frac{1}{|I_N|}\sum_{n\in I_N}e(\langle m,A(n)z\rangle)=\frac{1}{q}\sum_{n=1}^{q}e\left(\frac{m\cdot A(n)a}{q}\right).

Since bj=rbj′b_j=rb'_j for 1≤j≤D1\le j\le D and q=rq0q=rq_0, we get

m⋅A(n)aq=b0q+b1′n+⋯+bD′nDq0.\frac{m\cdot A(n)a}{q}=\frac{b_0}{q}+\frac{b'_1n+\cdots+b'_Dn^D}{q_0}.

Hence

e(m⋅A(n)aq)=e(b0q)e(b1′n+⋯+bD′nDq0).e\left(\frac{m\cdot A(n)a}{q}\right)=e\left(\frac{b_0}{q}\right)e\left(\frac{b'_1n+\cdots+b'_Dn^D}{q_0}\right).

The first factor is independent of nn and has absolute value one. Note that the numerator of the second factor is a polynomial with integer coefficients, hence the second factor is periodic in nn with period q0q_0. Since q=rq0q=rq_0, the sum over 1≤n≤q1\le n\le q consists of rr identical blocks of length q0q_0. Therefore

∣1q∑n=1qe(m⋅A(n)aq)∣=∣1q0∑n=1q0e(b1′n+⋯+bD′nDq0)∣.(14)\left|\frac{1}{q}\sum_{n=1}^{q}e\left(\frac{m\cdot A(n)a}{q}\right)\right|=\left|\frac{1}{q_0}\sum_{n=1}^{q_0}e\left(\frac{b'_1n+\cdots+b'_Dn^D}{q_0}\right)\right|. \tag*{(14)}

Recall that by Hua’s estimate for exponential sums [8 Chapter I, Theorem 1], if

f(x)=Akxk+⋯+A1x+A0∈Z[x]f(x)=A_kx^k+\cdots+A_1x+A_0\in\mathbb{Z}[x]

and gcd⁡(Ak,…,A1,q)=1\gcd(A_k,\ldots,A_1,q)=1 then for every η>0\eta>0, there exists Ck,η>0C_{k,\eta}>0 such that

∣∑n=1qe(f(n)q)∣≤Ck,ηq1−1k+η.\left|\sum_{n=1}^{q}e\left(\frac{f(n)}{q}\right)\right|\le C_{k,\eta}q^{1-\frac{1}{k}+\eta}.

We apply this with f(x)=(b1′x+⋯+bD′xD)f(x)=(b'_1x+\cdots+b'_Dx^D) and q=q0q=q_0. Let k≤Dk\le D be the degree of ff. Using (13) and dividing both sides by q0q_0, we get

∣1q0∑n=1q0e(b1′n+⋯+bD′nDq0)∣≤Ck,ηq0−1k+η.\left|\frac{1}{q_0}\sum_{n=1}^{q_0}e\left(\frac{b'_1n+\cdots+b'_Dn^D}{q_0}\right)\right|\le C_{k,\eta}q_0^{-\frac{1}{k}+\eta}.

Let η=1/(2D)\eta=1/(2D). Since k≤Dk\le D, we have −1k+12D≤−12D-\frac{1}{k}+\frac{1}{2D}\le-\frac{1}{2D}. Absorbing the finitely many possible values of kk into a constant CDC_D, and applying (14) we get

∣1q∑n=1qe(m⋅A(n)aq)∣≤CDq0−12D.\left|\frac{1}{q}\sum_{n=1}^{q}e\left(\frac{m\cdot A(n)a}{q}\right)\right|\le C_Dq_0^{-\frac{1}{2D}}.

Since q0=q/rq_0=q/r and r≤Qmr\le Q_m,

∣1q∑n=1qe(m⋅A(n)aq)∣≤CD(Qmq)12D.\left|\frac{1}{q}\sum_{n=1}^{q}e\left(\frac{m\cdot A(n)a}{q}\right)\right|\le C_D\left(\frac{Q_m}{q}\right)^{\frac{1}{2D}}.

Thus, if z∈Em,δz\in E_{m,\delta}, then δ≤CD(Qmq)12D\delta\le C_D\left(\frac{Q_m}{q}\right)^{\frac{1}{2D}} and consequently

q≤Qm(CDδ)2D.q\le Q_m\left(\frac{C_D}{\delta}\right)^{2D}.

Hence the order of every point in Em,δE_{m,\delta} is bounded in terms of mm, δ\delta, and A(x)A(x). There are only finitely many torsion points of Td\mathbb{T}^d of bounded order, and therefore Em,δE_{m,\delta} is finite. The hypothesis of Theorem 1.4 is satisfied. Hence, by Theorem 1.4 (b), the action GA↷TdG_A\curvearrowright\mathbb{T}^d is IC-rigid. □\square

References

School of Mathematics and Statistics, University of Sydney, NSW 2006, Australia

Email address: [email protected]

School of Mathematics and Statistics, University of Sydney, NSW 2006, Australia

Email address: [email protected]

References

  1. [1]Kamil Bulinski and Alexander Fish. Glasner property for linear group actions and their products. Math. Z., 303(4):Paper No. 99, 12, 2023.arxiv.org/abs/2210.16973
  2. [2]Kamil Bulinski and Alexander Fish. Glasner property for unipotently generated group actions on tori. Israel J. Math., 255(1):109–122, 2023.arxiv.org/abs/2103.08912
  3. [3]Daniel Berend and Yuval Peres. Asymptotically dense dilations of sets on the circle. J. London Math. Soc. (2), 47(1):1–17, 1993.DOI
  4. [4]Yves Benoist and Jean-François Quint. Stationary measures and invariant subsets of homogeneous spaces (III). Ann. of Math. (2), 178(3):1017–1059, 2013.
  5. [5]Changguang Dong. On density of infinite subsets I. Discrete Contin. Dyn. Syst., 39(5):2343–2359, 2019.arxiv.org/abs/1709.05591
  6. [6]Changguang Dong. On density of infinite subsets II: Dynamics on homogeneous spaces. Proc. Amer. Math. Soc., 147(2):751–761, 2019.arxiv.org/abs/1709.05593
  7. [7]Shmuel Glasner. Almost periodic sets and measures on the torus. Israel J. Math., 32(2-3):161–172, 1979.DOI
  8. [8]L. K. Hua. Additive theory of prime numbers. Translations of Mathematical Monographs, Vol. 13. American Mathematical Society, Providence, RI, 1965.
  9. [10]Michael Kelly and Thái Hoàng Lê. Uniform dilations in higher dimensions. J. Lond. Math. Soc. (2), 88(3):925–940, 2013.DOI
  10. [11]Bryna Kra and Scott Schmieding. IC-rigidity and generalizations of a theorem of Mahler, 2026. arXiv:2611.05607.arxiv.org/abs/2611.05607
  11. [12]Bryna Kra and Scott Schmieding. Invariant random compacts, 2026. arXiv:2605.03993.arxiv.org/abs/2605.03993

Paper details

vlnc:W_twkk54ae
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

AI

View statement
Looking for related papers…
Contents