The spectral gap of a group acting on a space measures how quickly correlations decay. A strong spectral gap on L²(Γ\G) — the space of square-integrable functions on the quotient — is the sharpest form: it controls not just the spherical functions (on the symmetric space) but all matrix coefficients of the regular representation.
Note (2026-11-10): I checked the abstract and searched the full text of arXiv 2603.20552 (Kelmer, Khalil, Sarkar). The theorem is as the essay states it. The essay's definition of geometrically finite is not the paper's: the paper calls a group geometrically finite when the 1-neighborhood of the convex core of the quotient (the smallest closed convex set containing all closed geodesics) has finite volume. The essay's “finite volume in the thick part”, with compact and cusp-like regions, is its own paraphrase. On the threshold, the paper gives the background that, by Patterson and Sullivan, the bottom eigenvalue of the Laplacian lies below the continuous spectrum if and only if the critical exponent is greater than d/2, and Mohammadi and Oh conjectured that this condition is optimal for the strong gap. The essay's reasons (the Patterson-Sullivan measure having enough mass, a threshold for absolute continuity) are not given in the paper.
The paper resolving a conjecture of Mohammadi and Oh (arXiv: 2603.20552) proves the strong spectral gap for geometrically finite, Zariski dense subgroups of SO(d+1,1) with critical exponent strictly greater than d/2.
Geometrically finite means the quotient has finite volume in the thick part — the group has both compact and cusp-like regions. Zariski dense means the group is as algebraically rich as possible. The critical exponent condition (> d/2) is the threshold where the Patterson–Sullivan measure has enough mass to control the spectrum.
The applications are immediate: decay rates for matrix coefficients and exponential mixing of the frame flow. The strong gap quantifies mixing — it's not just that correlations decay, but they decay at a rate determined by the gap.
The through-claim: the critical exponent is the spectral threshold. Below d/2, the group is too thin for spectral control. Above d/2, the strong gap holds and mixing is exponential. The dividing line is geometric (the growth rate of the orbit) and spectral (the threshold for absolute continuity) simultaneously.
2603.20552. Spectral theory / hyperbolic manifolds / spectral gap / mixing / geometrically finite groups.