Sufficient bigness criterion for differences of two nef classes (Q2634886)

Let \(X\) be a compact Kähler manifold (or more generally a compact complex manifold admitting a Hermitian metric such that the associated \((1,1)\)-form \(\omega\) satisfies \(\partial \bar \partial \omega^k=0\) for all \(k\)) of dimension \(n\). If \(X\) is a projective manifold, Demailly's holomorphic Morse inequalities provide an important tool to study the positivity of line bundles on \(X\) and it is expected that these inequalities generalise to the setting of cohomology classes on not necessarily algebraic manifolds. The paper under review realises this expectation in a first important case: given an \(n\)-dimensional manifold \(X\) as above and \(\alpha, \beta\) nef Bott-Chern cohomology classes of degree \((1,1)\), suppose that \[ \alpha^n - n \alpha^{n-1} \beta > 0. \] Then the difference \(\alpha-\beta\) is big, i.e., it contains a Kähler current. This improves a recent result of \textit{J. Xiao} [Ann. Inst. Fourier 65, No. 3, 1367--1379 (2015; Zbl 1333.32025)] who proved the same statement under the stronger assumption \(\alpha^n - 4n \alpha^{n-1} \beta > 0\). If \(X\) is projective, \textit{D. W. Nyström} and \textit{S. Boucksom} [``Duality between the pseudoeffective and the movable cone on a projective manifold'', Preprint, \url{arXiv:1602.03778}] have recently obtained an even stronger result including an optimal estimate of the volume of \(\alpha-\beta\).