On the locus of higher order jets of entire curves in complex projective varieties (Q2089928)

The GGL (Green-Griffiths-Lang) conjecture predicts that every complex projective manifold \(X\) has a proper algebraic subvariety \(Y\subsetneq X\), called an exceptional locus, containing all non-constant holomorphic curves \(\mathbb{C}\rightarrow X\), called entire curves. The typical strategy now to attack the GGL conjecture is to (1) show that all entire curves \(f:\mathbb{C}\rightarrow X\) must satisfy some algebraic differential equations \(P(f;f',f'',\ldots, f^{(k)})=0\) which are sections of the Green-Griffiths bundle \(E^{GG}_{k,m}T_X^*\) of polynomials of weighted degree \(m\), (2) show that (up to a twist) there exist many such sections using the holomorphic Morse inequalities, and (3) show that the image of the base loci of these sections when projected to \(X\) is a proper subvariety. In this article, the author proves new results on the existence of algebraic differential equations coming from higher jet bundles that strongly restrain the locus of entire curves. This relies on the use of certain new tautological morphisms (Corollary 6.5) related to induced directed structures on subvarieties of the higher jet bundles. For instance, the author proves the following. Theorem 1.2. Assume that \(\Lambda^2T_X^*\) is a strongly big sheaf. Then there exists a subvariety \(Y\subsetneq X\), an integer \(k\in \mathbb{N}\) and a finite collection of induced directed subvarieties \((Z_\alpha, W_\alpha) \subset (X_k, V_K)\) with \(Z_\alpha\) irreducible, rank \(W_\alpha \leq p-1\), such that all entire curves \(f:(\mathbb{C}, T_{\mathbb{C}})\rightarrow (X,V)\) satisfy either \(f(\mathbb{C})\subset Y\) or have a \(k\)-jet lifting \(f_{[k]}:(\mathbb{C}, T_{\mathbb{C}})\rightarrow (Z_\alpha, W_\alpha)\) for some \(\alpha\). Note that in the above theorem, \(\Lambda^2T_X^*\) being a strongly big sheaf means that for some ample line bundle \(A\) on \(X\), the symmetric powers \(S^m(\Lambda^2T_X^*)\otimes \mathcal{O}_X(-A)\) are globally generated on a Zariski open set \(X\setminus Y\), \(Y\subsetneq X\), for all large \(m>0.\) In Remark 7.5, the author asks whether or not this bigness hypothesis may be relaxed and replaced with some suitable semistability conditions instead.