An almost-almost-Schur lemma on the 3-sphere
Tobias König, Jonas W. Peteranderl
TL;DR
This work extends a quantitative stability result for reverse curvature inequalities to the 3-sphere, recasting the problem in terms of conformal factors and the conformally invariant functionals $F_1$ and $F_2$. It proves a sharp deficit bound for the Andrews–De Lellis–Topping inequality on $\mathbb S^3$, with proximity measured by a two-term distance after optimizing over scalings and Möbius transformations. The authors also establish a reverse $\sigma_2$-curvature stability and derive an interpolation family of reverse inequalities, all proved via a two-step Bianchi–Egnell strategy that combines global-to-local reduction with a detailed local bound grounded in Hessian analysis and frequency decomposition. The results connect to almost-Schur type questions and yield stability statements for related interpolation inequalities, enriching the landscape of stability in conformal geometry and reverse Sobolev-type inequalities.
Abstract
In a recent preprint, Frank and the second author proved that if a metric on the sphere of dimension $d>4$ almost minimizes the total $σ_2$-curvature in the conformal class of the standard metric, then it is almost the standard metric (up to Möbius transformations). This is achieved quantitatively in terms of a two-term distance to the set of minimizing conformal factors. We extend this result to the case $d=3$. While the standard metric still minimizes the total scalar curvature for $d=3$, it maximizes the total $σ_2$-curvature, which turns the related functional inequality into a reverse Sobolev-type inequality. As a corollary of our result, we obtain quantitative versions for a family of interpolation inequalities including the Andrews--De Lellis--Topping inequality on the $3$-sphere. The latter is itself a stability result for the well-known Schur lemma and is therefore called almost-Schur lemma. This makes our stability result an almost-almost-Schur lemma.
