A criterion for the simplicity of the Lyapunov spectrum of square-tiled surfaces (Q887948): Difference between revisions

From MaRDI portal
Importer (talk | contribs)
Created a new Item
 
Added link to MaRDI item.
links / mardi / namelinks / mardi / name
 

Revision as of 15:56, 30 January 2024

scientific article
Language Label Description Also known as
English
A criterion for the simplicity of the Lyapunov spectrum of square-tiled surfaces
scientific article

    Statements

    A criterion for the simplicity of the Lyapunov spectrum of square-tiled surfaces (English)
    0 references
    0 references
    0 references
    0 references
    4 November 2015
    0 references
    A holomorphic 1-form \(\omega\) with zero set \(\Sigma\) on a compact Riemann surface \(M\) defines coordinate charts on \(M - \Sigma\) whose overlap maps are translations in \(\mathbb{C}\). The Teichmüller flow on the moduli space \(\mathcal{H}^{(1)}\) of unit area translation surfaces is related to another flow, \(G_{t}^{KZ}\), the Kontsevich-Zorich cocycle (defined below) that describes how the homology of the surface evolves along orbits of the Teichmüller flow. The connected components of \(\mathcal{H}^{(1)}\) support natural probability measures \(\mu_{MV}\), called Masur-Veech measures, that are invariant under the Teichmüller flow. Based on numerical evidence Kontsevich and Zorich conjectured that the Lyapunov spectrum of \(G_{t}^{KZ}\) is simple with respect to the measures \(\mu_{MV}\). A. Avila and M. Viana proved this conjecture after major prior contributions by G. Forni. Forni also constructed two examples of translation surfaces where the Lyapunov exponents of \(G_{t}^{KZ}\) are all zero with respect to measures (different from \(\mu_{MV}\)) that are invariant under the \(\mathrm{SL}(2,\mathbb{R})\) action on \(\mathcal{H}^{(1)}\). In the present article the authors study the Lyapunov exponents of \(G_{t}^{KZ}\) relative to natural \(\mathrm{SL}(2,\mathbb{R})\) invariant probability measures in a special case that \((M,\omega)\) is a reduced, unit area, square tiled translation surface. In particular they restrict their attention to Riemann surfaces \(M\) of genus 3 with a single zero of order four. They denote this moduli space by \(\mathcal{H}(4)\). Recall that a translation surface \((M,\omega)\) is ``square tiled'' if \(\int_{\gamma} \omega \in \mathbb{Z} \oplus i \mathbb{Z}\) for every relative homology class \(\gamma \in H_{1}(M , \Sigma, \mathbb{Z})\), and \((M,\omega)\) is ``reduced'' if the group of relative periods is equal to \( \mathbb{Z} \oplus i \mathbb{Z}\). In \(\mathcal{H}(4)\) the authors derive a Galois theoretic criterion for the simplicity of the Lyapunov of \(G_{t}^{KZ}\), and they determine some cases where the criterion is satisfied. In particular, if a certain conjecture of Delecroix-Lelievre is true, then their criterion shows that all except possibly finitely many reduced, unit area, square tiled translation surfaces in \(\mathcal{H}(4)\) have simple Lyapunov spectrum for \(G_{t}^{KZ}\). We define some of the objects used in the discussion above, beginning with translation surface. If \(\omega\) is a holomorphic 1-form (i.e., an abelian differential) on a compact Riemann surface M, then \(d \omega = 0\). It follows that if \(\Sigma\) denotes the set of zeros of \(\omega\) on M, then \(M - \Sigma\) can be covered by coordinate charts \(f_{\alpha} : O_{\alpha} \rightarrow \mathbb{C}\) such that \(df_{\alpha} = \omega\) on \(O_{\alpha}\). If \(O_{\alpha} \cap O_{\beta}\) is nonempty, then \(\varphi_{\alpha \beta} = f_{\alpha} \circ f_{\beta}^{-1} : f_{\beta}(O_{\alpha} \cap O_{\beta}) \rightarrow f_{\alpha}(O_{\alpha} \cap O_{\beta})\) is a translation in \(\mathbb{C}\) since \(d \varphi_{\alpha \beta} = \mathrm{Id}\). Conversely, if \(\Sigma\) is a finite set in a Riemann surface \(M\) such that \(M - \Sigma\) can be covered by coordinate charts \(\{f_{\alpha}, O_{\alpha} \}\) whose overlap maps \(\{\varphi_{\alpha \beta} \}\) are translations, then these charts arise from an appropriate abelian differential \(\omega\). There is a natural action of \(\mathrm{SL}(2, \mathbb{R})\) on the moduli space \(\mathcal{H}^{(1)}\) of unit area translation surfaces \((M , \omega)\). The restriction of this action to the subgroup \(g_{t} = \mathrm{diag}(e^{t}, e^{-t})\) is called the Teichmüller flow on \(\mathcal{H}^{(1)}\). To define the KZ-cocycle \(G_{t}^{KZ}\) one first introduces the mapping class group \(\Gamma(M,\Sigma, \kappa)\), where \(\Sigma = \{p_{1}, \dots, p_{N} \}\) is the set of zeros of \(\omega\) and \(\kappa = (\kappa_{1},\dots, \kappa_{N})\) denotes the multiplicities of the corresponding zeros. Define \(\mathrm{Diff}^{+}(M , \Sigma , \kappa)\) to be the group of orientation preserving homeomorphisms f of \(M\) that preserve \(\Sigma\) and \(\kappa\). Define the mapping class group \(\Gamma(M,\Sigma, \kappa) = \mathrm{Diff}^{+}(M , \Sigma , \kappa) / \mathrm{Diff}^{+}_{0}(M , \Sigma , \kappa)\). The Teichmüller space \(\mathcal{T}(M , \Sigma , \kappa)\) is the quotient of the set of structures of translation surfaces on \((M , \Sigma , \kappa)\) by the action of the group \(\mathrm{Diff}^{+}_{0}(M , \Sigma , \kappa)\). The moduli space \(\mathcal{H}(M , \Sigma , \kappa)\) is the quotient of the same set by the action of \(\mathrm{Diff}^{+}(M , \Sigma , \kappa)\). Hence \(\mathcal{H} (M , \Sigma , \kappa) = \mathcal{T}(M , \Sigma , \kappa) / \Gamma(M , \Sigma , \kappa)\). Let \(\mathcal{T}^{(1)}(M , \Sigma , \kappa)\) denote the set of unit area translation surfaces in \(\mathcal{T}(M , \Sigma , \kappa)\). The \(\mathrm{SL}(2,\mathbb{R})\) action, and in particular the Teichmüller flow \(g_{t}\), leaves \(\mathcal{T}^{(1)}(M , \Sigma , \kappa)\) invariant. Finally, define \(\tilde{G_{t}}^{KZ} = g_{t} \times \mathrm{Id} : \mathcal{T}^{(1)}(M, \Sigma , \kappa) \times H_{1}(M,\mathbb{R}) \rightarrow \mathcal{T}^{(1)}(M \Sigma , \kappa) \times H_{1}(M,\mathbb{R})\). The diagonal action of the mapping class group \(\Gamma(M , \Sigma , \kappa)\) on this product space is nontrivial on each factor, and \(\Gamma(M , \Sigma , \kappa)\) commutes with \(\tilde{G_{t}}^{KZ}\) since it commutes with \(g_{t}\). Hence \(\tilde{G_{t}}^{KZ}\) descends to a flow \(G_{t}^{KZ} : H_{1}^{\mathbb{R}} \rightarrow H_{1}^{\mathbb{R}}\), where \(H_{1}^{\mathbb{R}} = (\mathcal{T}^{(1)}(M, \Sigma , \kappa) \times H_{1}(M,\mathbb{R})) / \Gamma(M,\Sigma, \kappa)\). This flow \(G_{t}^{KZ}\) is the KZ-cocycle. The space \(H^{(1)}_{\mathbb{R}}\) is a nontrivial bundle over \(\mathcal{H}^{(1)}(M , \Sigma , \kappa)\) called the ``real Hodge bundle''.
    0 references
    Lyapunov spectrum
    0 references
    Kontsevich-Zorich cocyle
    0 references
    Teichmüller flow
    0 references
    Masur-Veech measure
    0 references
    Delecroix-Lelievre conjecture
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references