Double shuffle Lie algebra and special derivations
Benjamin Enriquez, Hidekazu Furusho
TL;DR
This work identifies and analyzes a deep link between Racinet’s double shuffle Lie algebra $\mathfrak{dmr}_0$ and a broader automorphism framework built from tangential automorphisms. By constructing a rich network of groups, semidirect products, and category-theoretic diagrams, the authors show that the stabilizer of a certain universal diagram equals $\mathsf{DMR}_0(\mathbf k)$ and derive a key inclusion and ultimately equality between stabilizers of two fundamental data objects $[\![\Delta^{\mathcal{W}}_{r,l}]\!]$ and $[\![\![\rho_{\mathrm{DT}}]\!]$ in the group $\mathcal{G}$. The approach uses an intricate blend of Lie-theoretic, Hopf-algebraic, and category-theoretic machinery, including PSTGA structures, explicit matrix realizations, and cohomological arguments to track invariants under involutions. The results clarify how senary relations and the Krivine–Kontsevich–Ihara-type structures interact with inertia-preserving automorphisms, and they provide a robust framework for comparing stabilizers across different diagrammatic presentations, contributing to a deeper understanding of the algebraic underpinnings of multiple zeta value relations and associated Lie algebras.
Abstract
Racinet's double shuffle Lie algebra $\mathfrak{dmr}_0$ is a Lie subalgebra of the Lie algebra $\mathfrak{tder}$ of tangential derivations of the free Lie algebra with generators $x_0,x_1$, i.e. of derivations such that $x_1\mapsto 0$ and $x_0\mapsto [a,x_0]$ for some element $a$. We prove: (1) $\mathfrak{dmr}_0$ is contained in the Lie subalgebra $\mathfrak{sder}$ of $\mathfrak{tder}$ of special derivations, i.e. satisfying the additional condition that $x_\infty\mapsto [b,x_\infty]$ for some element $b$, where $x_\infty:=x_1-x_0$; (2) $\mathfrak{dmr}_0$ is stable under the involution of $\mathfrak{sder}$ induced by the exchange of $x_0$ and $x_\infty$. The first statement: (a) says that any element of $\mathfrak{dmr}_0$ satisfies the "senary relation" (a fact announced without proof by Ecalle in 2011); (b) implies the inclusion $\mathfrak{dmr}_0\subset \mathfrak{krv}_2$ (which was proved by Schneps in 2012 only conditionally to the truth of (1)).
