Geodesic stability, the space of rays and uniform convexity in Mabuchi geometry (Q2211891)

A recent breakthrough in Kähler geometry ensures the existence of a constant scalar curvature Kähler (cscK) metric on a compact Kähler manifold \(X\) if and only if the Mabuchi K-energy \(\mathcal{K}\) is proper w.r.t the metric \(d_1\). This result was proved by \textit{X. X. Chen} and \textit{J. Cheng} [``On the constant scalar curvature Kähler metrics, a priori estimates'', Preprint, \url{arXiv:1712.06697}; ``On the constant scalar curvature Kähler metrics, existence results'', Preprint, \url{arXiv:1801.00656}; ``On the constant scalar curvature Kähler metrics, general automorphism group'', Preprint, \url{arXiv:1801.05907}] and it deeply uses all the previous developments (made by the first author) in the understanding of the completion of the space of Kähler metrics \(\mathcal{H}\) endowed by Finsler structures \(d_p, p\geq 1\). Donaldson's geodesic stability conjecture is situated in this context: it states that to ensure the existence of a cscK metric it is enough to check the properness of the \(K\)-energy along geodesic rays of \(\mathcal{H}\). The first issue one encounters when working on this conjecture is the statement itself since geodesics between points in \(\mathcal{H}\) are not smooth. The optimal regularity result for geodesics (\(C^{1,1}\)) was proved only recently in [\textit{J. Chu} et al., Ann. PDE 3, No. 2, Paper No. 15, 12 p. (2017; Zbl 1397.35050)]. This lack of regularity is the reason why the authors propose an analogue of Donaldson's geodesic stability conjecture (Conjecture 1.7). The main result of the paper solves (almost completely) such a conjecture. In Theorem 1.8 the authors prove that the existence of a cscK metric is equivalent to the properness of \(\mathcal{K}\) along \(C^{1,\bar{1}}\) geodesic rays. The missing step is to show that one can use rays that have potentials with fully bounded Hessian (\(C^{1,1}\) rays) and not just bounded complex Hessian. One of the key results in the paper that allows to prove the above statement is the fact that \(C^{1,\bar{1}}\) geodesic stability is equivalent to geodesic stability for rays with potentials in the energy class \(\mathcal{E}^1\) (Theorem 6.3). The first part of the paper is devoted to the study of the metric geometry of the space of these ``weak'' rays. Along the way the authors prove a nice and interesting result on the uniform convexity of \((\mathcal{E}^p, d_p)\), \(p>1\) that can be seen as a generalization of the CAT(0) property of \((\mathcal{E}^2, d_2)\) previously proved in [\textit{X. Chen}, J. Differ. Geom. 56, No. 2, 189--234 (2000; Zbl 1041.58003)] and [the first author, Adv. Math. 285, 182--219 (2015; Zbl 1327.53093)]. As consequence one gets uniqueness of weak geodesics segments connecting points in \(\mathcal{E}^p\), \(p>1\) (this is not true in \(\mathcal{E}^1\)!).