Let $X$ be a projective manifold and $\Delta$ a divisor with simple normal crossings. Consider $X$ as the compactification of a quasi-projective variety $X_0$ with boundary $\Delta$, i.e. $X_0 = X \backslash \Delta$. Suppose that $(X,\Delta)$ is of log general type, i.e. $K_X+D$ is big.

A theorem of Cadorel [Cad16] tells us that a projective log smooth pair $(X,\Delta)$ with $X_0$ admitting a Kähler metric of non-positive bisectional curvature and negative holomorphic sectional curvature, then the sheaf of logarithmic differentials $\Omega_X(\log(\Delta))$ is big. If one also assumes that $\omega$ is bounded near $D$, this can be strengthened to $\Omega_X$ being big.

Guenancia [Gue18], in a similar spirit, has shown that if $(X, \Delta)$ is log smooth with the holomorphic sectional curvature of $\omega$ on $X_0$ bounded above by a negative constant, then $(X,\Delta)$ is of log general type.

Is the converse true? That is, given a log smooth pair $(X, \Delta)$ of log general type, is the sectional curvature bounded above by a negative (nonpositive?) constant on $X_0$; is the bisectional curvature nonpositive on $X_0$?

[Cad16] -- Cadoral, B., Symmetric differentials on complex hyperbolic manifolds with cusps, arXiv:1606.05470, (2016)

[Gue18] -- Guenancia, H., Quasi-projective manifolds with negative holomorphic sectional curvature, arXiv:1808.01854v3, (2018)


1 Answer 1


No, this is not true, even for $\Delta=\emptyset$. If $X$ admits a Kähler metric with negative holomorphic bisectional curvature, then so do all its subvarieties; in particular, all its subvarieties are of general type.

However, many varieties of general type admit subvarieties with non-maximal Kodaira dimension (blow up any general type variety along a smooth center for instance).

Even in the minimal case, there are hypersurfaces of the projective space with ample canonical bundle that admit rational curves (e.g. some Fermat hypersurfaces).

  • $\begingroup$ Thank you so much! $\endgroup$
    – AmorFati
    Oct 5, 2020 at 10:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.