An abstract proof of the L2-singular dichotomy for orbital measures on Lie algebras and groups

TL;DR

This paper provides an abstract, geometry-based proof of the L2-singular dichotomy for orbital measures on Lie algebras and groups, extending previous Lie type-dependent results to a more general symplectic geometry framework.

Contribution

It introduces a new proof leveraging the Duistermaat-Heckman theorem, showing the dichotomy is a consequence of symplectic geometry rather than Lie type specifics.

Findings

The L2-singular dichotomy follows from symplectic geometry principles.

Convolution of orbital measures is either singular or in L^{2+ε}.

The result extends to measures supported on conjugacy classes in G.

Abstract

Let $G$ be a compact, connected simple Lie group and $\mathfrak{g}$ its Lie algebra. It is known that if $\mu $ is any $G$-invariant measure supported on an adjoint orbit in $\mathfrak{g}$, then for each integer $k$, the $k$% -fold convolution product of $\mu $ with itself is either singular or in $% L^{2}$. This was originally proven by computations that depended on the Lie type of $\mathfrak{g}$, as well as properties of the measure. In this note, we observe that the validity of this dichotomy is a direct consequence of the Duistermaat-Heckman theorem from symplectic geometry and that, in fact, any convolution product of (even distinct) orbital measures is either singular or in $L^{2+\varepsilon }$ for some $\varepsilon >0$. An abstract transference result is given to show that the $L^{2}$-singular dichotomy holds for certain of the $G$-invariant measures supported on conjugacy classes in $G.$