Papers
Topics
Authors
Recent
Assistant
AI Research Assistant
Well-researched responses based on relevant abstracts and paper content.
Custom Instructions Pro
Preferences or requirements that you'd like Emergent Mind to consider when generating responses.
GPT-5.1
GPT-5.1 114 tok/s
Gemini 3.0 Pro 53 tok/s Pro
Gemini 2.5 Flash 132 tok/s Pro
Kimi K2 176 tok/s Pro
Claude Sonnet 4.5 37 tok/s Pro
2000 character limit reached

Self-Composition Deviation in Self-Affine Sets

Updated 18 November 2025
  • Self-Composition Deviation (SCD) is a phenomenon describing the asymptotic behavior of ergodic integrals in self-affine Delone sets, governed by a hierarchy of cohomological eigenexponents.
  • The method uses expansive linear maps and pattern-equivariant cohomology to derive detailed expansions of ergodic averages with subleading corrections beyond classical boundary terms.
  • Illustrative examples such as the Penrose, Chair, and Ammann–Beenker tilings demonstrate SCD's significance in linking dynamical fluctuations with topological and spectral invariants.

Self-Composition Deviation (SCD) refers to asymptotic phenomena in the growth of ergodic integrals for translation actions on pattern spaces arising from self-affine Delone sets. Under iteration by an expansive linear map, the rate of deviation of ergodic averages is governed by the spectrum of the induced action on the pattern-equivariant cohomology of the tiling/delone set. The deviation is characterized by a hierarchy of exponents determined by the eigenvalues of this cohomological action, leading to a detailed expansion of ergodic integrals with subleading cohomological corrections beyond classical boundary terms. This phenomenon was rigorously developed in the context of self-affine Delone set dynamics by Schmieding–Treviño (Schmieding et al., 2015).

1. Ergodic Integrals and Renormalization for Self-Affine Pattern Spaces

Let ΛRd\Lambda \subset \mathbb{R}^d be a repetitive, finite local complexity (FLC) Delone set, whose pattern space Ω=ΩΛ\Omega = \Omega_\Lambda supports a unique translation-invariant probability measure μ\mu. For any bounded averaging set BRdB \subset \mathbb{R}^d, and any fC0(Ω)f \in C^0(\Omega), the ergodic integral is defined as

I(B,f;Λ0)=Bf(φs(Λ0))ds,I(B, f; \Lambda_0) = \int_B f(\varphi_s(\Lambda_0))\,ds,

where φs\varphi_s denotes translation by ss in Rd\mathbb{R}^d. If Λ\Lambda is self-affine, there exists an expanding AGL+(d,R)A \in GL^+(d,\mathbb{R}) and a measure-preserving homeomorphism ΦA:ΩΩ\Phi_A:\Omega \to \Omega such that ΦAφt=φAtΦA\Phi_A \circ \varphi_t = \varphi_{A t} \circ \Phi_A.

A renormalized family of averaging sets is chosen as BT=gσlogT(B0)B_T = g_{\sigma \log T}(B_0), with gt=exp(ta)g_t = \exp(t a), exp(a)=A\exp(a) = A, and σ=d/logdetA\sigma = d/\log \det A, ensuring VolBT=VolB0Td\operatorname{Vol} B_T = \operatorname{Vol} B_0\,T^d. Under iteration, AnA^n “magnifies” averaging, enabling comparison of ergodic integrals at scales TT and T/τT/\tau under the induced homeomorphism.

2. Deviation Hierarchy and Main Theorem

For a Delone set with “RFT” property (pattern-equivariant cohomology in top degree is finite-dimensional), the automorphism ΦA\Phi_A induces a linear map A=ΦAA^* = \Phi_A^* on Hd(Ω;C)H^d(\Omega;\mathbb{C}) with eigenvalues ν1ν2νr>0|\nu_1| \geq |\nu_2| \geq \cdots \geq |\nu_r| > 0 and ν1=detA\nu_1 = \det A. Writing any ff as a dd-form using f1f \star 1, a Jordan basis decomposition yields

f=i,j,kαi,j,k(f)gi,j,k+dω,\star f = \sum_{i,j,k} \alpha_{i,j,k}(f)\,g_{i,j,k} + \star d\omega,

where gi,j,kg_{i,j,k} are dual basis functions. The coefficients αi,j,k(f)\alpha_{i,j,k}(f) are determined by the cohomology class of f1f \star 1.

The rapidly-expanding subspace E+Hd(Ω)E^+ \subset H^d(\Omega) is defined as the sum of generalized eigenspaces EiE_i for those νi\nu_i with

logνilogν11logλdlogν1,\frac{\log |\nu_i|}{\log \nu_1} \geq 1 - \frac{\log |\lambda_d|}{\log \nu_1},

where λd\lambda_d is the smallest-modulus eigenvalue of AA. The main SCD theorem states there are exactly dimE+\dim E^+ distributions Di,j,kD_{i,j,k} such that:

  • If Di,j,k(f)=0D_{i',j',k'}(f)=0 for all (i,j,k)<(i,j,k)(i',j',k') < (i,j,k) but Di,j,k(f)0D_{i,j,k}(f) \neq 0, then for T>3T>3

I(BT,f;Λ0)CL(i,j,T)Td(logνi)/(logν1)f|I(B_T, f; \Lambda_0)| \leq C\,L(i,j,T)\,T^{d(\log |\nu_i|)/(\log \nu_1)}\, \|f\|_\infty

where L(i,j,T)=(logT)j1L(i,j,T) = (\log T)^{j-1} or (logT)j(\log T)^j depending on boundary-exponent equality, and CC depends only on B0B_0 and AA.

  • If all Di,j,k(f)=0D_{i,j,k}(f)=0, the boundary term dominates:

I(BT,f)=O(Td(1(logλd)/(logν1))).|I(B_T,f)| = O\Big(T^{d(1 - (\log |\lambda_d|)/(\log \nu_1))}\Big).

In the pure-dilation case A=τIdA = \tau \operatorname{Id}, the deviation reduces to classical estimates of the form

I(BR,f)=C(f)Rd+O(Rd1+ϵ),I(B_R, f) = C(f)\,R^d + O(R^{d-1+\epsilon}),

with C(f)=D1,1,1(f)C(f)=D_{1,1,1}(f).

3. Cohomological Structure and Interpretation

The leading constant C(f)=D1,1,1(f)C(f)=D_{1,1,1}(f) is interpreted topologically and dynamically using the asymptotic-cycle current CΛ(Hd(Ω;R))\mathcal{C}_\Lambda \in (H^d(\Omega;\mathbb{R}))', defined by

CΛ([η])=Ωfηdμ,ηΔΛd.\mathcal{C}_\Lambda([\eta]) = \int_\Omega f_\eta\,d\mu, \quad \eta \in \Delta_\Lambda^d.

Via the Hodge star, CΛ\mathcal{C}_\Lambda corresponds to a transverse invariant measure on the canonical transversal. The pure-point part of the diffraction spectrum is the evaluation of CΛ\mathcal{C}_\Lambda on classes [λx][\lambda_x], xRdx \in \mathbb{R}^d, supported on ΛΛ\Lambda - \Lambda, reproducing known spectral measures via Dworkin’s argument.

Diagonalizing AA^* on Hd(Ω)H^d(\Omega), the eigenvector in the unstable line E1,1,1E_{1,1,1} with eigenvalue ν1\nu_1 is [1][\star 1]. The dual distribution D1,1,1D_{1,1,1} recovers the density term:

D1,1,1(f)=CΛ(f1)=Ωfdμ.D_{1,1,1}(f) = \mathcal{C}_\Lambda(f \star 1) = \int_\Omega f\,d\mu.

4. Explicit Examples

Several notable self-affine tilings and Delone sets exhibit the SCD phenomenon:

  • Penrose tiling (inflation τ=(1+5)/2\tau = (1+\sqrt{5})/2): Substitution spectrum τ2,τ1,τ1,0,{\tau^2, -\tau^{-1}, -\tau^{-1}, 0,\dots}, with leading cohomological deviation τ2\tau^2, boundary exponent $1/2$, and subleading terms at boundary order O(R1)O(R^1).
  • Chair tiling (A=2A=2Id): Top cohomology eigenvalues {4,0,0,...}\{4,0,0,...\}; only the 4n4^n term is non-boundary.
  • Ammann–Beenker tiling (octagonal): ASL(4,Z)A \in \mathrm{SL}(4, \mathbb{Z}) with spectrum {1+2,1+2,12,12}\{1+\sqrt{2},1+\sqrt{2},1-\sqrt{2},1-\sqrt{2}\}; two eigenvalues in E+E^+, the leading strictly and another at equality, yielding log corrections.
  • Self-affine codim-1 cut-and-project sets: Top Hd(Ω)H^d(\Omega) splits into a torus-factor and a singularity-part; products iλi>detA/λd\prod_i \lambda_i > \det A / |\lambda_d| lie strictly in E+E^+.

5. Structural Propositions and Asymptotic Expansions

Key results summarizing SCD include:

  • Rapidly-expanding subspace:

E+=i:(logνi)/(logν1)1(logλd)/(logν1)EiE^+ = \bigoplus_{i\,:\,(\log |\nu_i|)/(\log \nu_1) \geq 1 - (\log |\lambda_d|)/(\log \nu_1)} E_i

  • Boundary estimate: If ψ1=dω\psi \star 1 = d \omega is a coboundary,

BTψφsdsCTd(1(logλd)/(logν1)).\left|\int_{B_T} \psi \circ \varphi_s\,ds \right| \leq C\,T^{d(1 - (\log |\lambda_d|)/(\log \nu_1))}.

  • Cohomological decomposition: For ηΔΛd\eta \in \Delta_\Lambda^d,

η=i,j,kαi,j,k(η)gi,j,k+dω.\star\eta = \sum_{i,j,k} \alpha_{i,j,k}(\eta)\,g_{i,j,k} + \star d\omega.

  • Induced-action estimate:

Υn(ηi,j,k)={O(nj1νin),νi>ν1/λd O(njνin),if equality.|\Upsilon^n(\eta_{i,j,k})| = \begin{cases} O(n^{j-1}|\nu_i|^n), & |\nu_i| > \nu_1/|\lambda_d|\ O(n^j|\nu_i|^n), & \text{if equality.} \end{cases}

Combining these elements, the full expansion of ergodic integrals is:

BTfφsds=(i,j,k)Di,j,k(f)Ψi,j,k(T)L(i,j,T)Td(logνi)/(logν1)+O(Td(1(logλd)/(logν1))).\int_{B_T} f \circ \varphi_s\,ds = \sum_{(i,j,k)} D_{i,j,k}(f)\,\Psi_{i,j,k}(T)\,L(i,j,T)\,T^{d(\log |\nu_i|)/(\log \nu_1)} + O\left(T^{d(1 - (\log |\lambda_d|)/(\log \nu_1))}\right).

In the dilation (self-similar) case, this specializes to

I(R,f)=C(f)Rd+O(Rd1+ϵ),I(R, f) = C(f)\,R^d + O(R^{d-1+\epsilon}),

with C(f)C(f) representing pairing with the unstable top class.

6. Broader Implications and Context

Self-Composition Deviation provides a generalization of classical deviation theorems and establishes a bridge between dynamical renormalization, cohomological invariants, and statistical properties of self-affine aperiodic structures. The SCD theorem extends beyond leading-order density terms by predicting intermediate orders of fluctuation (“Zorich–Forni” type exponents) directly from the induced cohomological dynamics and Jordan decomposition. It unifies understanding of deviation phenomena in substitution tilings, cut-and-project sets, and more general Delone sets, linking spectral, cohomological, and dynamical characteristics in a precise quantitative framework (Schmieding et al., 2015).

Definition Search Book Streamline Icon: https://streamlinehq.com
References (1)
Forward Email Streamline Icon: https://streamlinehq.com

Follow Topic

Get notified by email when new papers are published related to Self-Composition Deviation (SCD).