Isometric embeddings of surfaces for scl
Alexis Marchand
TL;DR
The paper develops a framework to compare stable commutator length and the relative Gromov seminorm under embeddings of surfaces, proving that for a $\pi_1$-injective subsurface $T\subset S$ with appropriate hypotheses, admissible surfaces in $S$ can be homotoped into $T$, yielding isometric embeddings of $H_2(T,c)$ into $H_2(S,c)$ and, consequently, isometric embeddings for $\mathrm{scl}$ and the relative Gromov norm. The core methodology introduces a standard (and perfect standard) form for admissible surfaces, combining incompressibility, transversality, connected links, and a non-folding/orientation-perfect framework to control how surfaces sit inside the ambient space. This approach yields precise isometry results (including in the closed-surface setting) and clarifies the interplay between extremal surfaces and quasimorphisms through Bavard duality, notably showing that rotation quasimorphisms respect restriction to subsurfaces under suitable injectivity hypotheses. By separating computation into the relative Gromov seminorm and a subsequent infimum step, the paper suggests a two-step strategy for scl calculations in more complex groups and connects these ideas to broader questions about the structure of scl norms and their dual objects. Overall, the results advance understanding of scl and relative seminorms in surface groups, with potential implications for extremal objects and dynamics on surface groups.
Abstract
Let $\varphi:F_1\to F_2$ be an injective morphism of free groups. If $\varphi$ is geometric (i.e. induced by an inclusion of oriented compact connected surfaces with nonempty boundary), then we show that $\varphi$ is an isometric embedding for stable commutator length. More generally, we show that if $T$ is a subsurface of an oriented compact (possibly closed) connected surface $S$, and $c$ is an integral $1$-chain on $π_1T$, then there is an isometric embedding $H_2(T,c)\to H_2(S,c)$ for the relative Gromov seminorm. Those statements are proved by finding an appropriate standard form for admissible surfaces and showing that, under the right homology vanishing conditions, such an admissible surface in $S$ for a chain in $T$ is in fact an admissible surface in $T$.
