Minimal asymptotic translation lengths on curve complexes and homology of mapping tori
Abstract: Let $S_g$ be a closed orientable surface of genus $g > 1$. Consider the minimal asymptotic translation length $L_{\mathcal{T}}(k, g)$ on the Teichm\"uller space of $S_g$, among pseudo-Anosov mapping classes of $S_g$ acting trivially on a $k$-dimensional subspace of $H_1(S_g)$, $0 \le k \le 2g$. The asymptotics of $L_{\mathcal{T}}(k, g)$ for extreme cases $k = 0, 2g$ have been shown by several authors. Jordan Ellenberg asked whether there is a lower bound for $L_{\mathcal{T}}(k, g)$ interpolating the known results on $L_{\mathcal{T}}(0, g)$ and $L_{\mathcal{T}}(2g, g)$, which was affirmatively answered by Agol, Leininger, and Margalit. In this paper, we study an analogue of Ellenberg's question, replacing Teichm\"uller spaces with curve complexes. We provide lower and upper bound on the minimal asymptotic translation length $L_{\mathcal{C}}(k, g)$ on the curve complex, whose lower bound interpolates the known results on $L_{\mathcal{C}}(0, g)$ and $L_{\mathcal{C}}(2g, g)$. Finally, for each $g$, we construct a non-Torelli pseudo-Anosov $f_g \in \operatorname{Mod}(S_g)$ which does not normally generates $\operatorname{Mod}(S_g)$ and so that the asymptotic translation length of $f_g$ on curve complexes decays more quickly than a constant multiple of $1/g$ as $g \to \infty$. From this, we provide a restriction on how small the asymptotic translation lengths on curve complexes should be if the similar phenomenon as in the work of Lanier and Margalit on Teichm\"uller spaces holds for curve complexes.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.