A central question in complex geometry concerns how the curvature properties of a geometric space behave when you map that space onto another space in a structured way. Specifically, mathematicians study "nefness," a condition saying that a geometric quantity called the canonical bundle does not curve too negatively, and "pseudo-effectiveness," a related but weaker positivity condition. A long-standing conjecture by Demailly, Peternell, and Schneider predicts that if you start with a space whose canonical bundle is nicely behaved (nef in the negative direction), and you map it onto another space in a sufficiently regular way, then the target space should also inherit a similar, if slightly weaker, positivity property.
Previous work had established this result under the assumption that the map between the two spaces is "projective," meaning the map has a certain algebraic structure that makes it easier to work with using tools from algebraic geometry. The projective setting allows mathematicians to apply a well-developed toolkit, but it excludes many naturally occurring geometric situations, particularly those arising in the broader world of Kahler geometry, which includes spaces that cannot be treated purely algebraically.
This paper removes the projectivity assumption entirely, proving the result in the full Kahler setting. The authors show that if you have a surjective holomorphic map from a log canonical pair whose twisted canonical bundle is nef in the negative sense, then the canonical bundle of the target space must be pseudo-effective, even when the map lacks any projective structure. This requires developing new analytic techniques suited to the Kahler context, where algebraic methods break down, and represents a meaningful step toward fully resolving the Demailly-Peternell-Schneider conjecture.