Essentially large divisors and their arithmetic and function-theoretic inequalities (Q1932241)

We have made a selection of the results proved in the paper. \textit{A. Levin} [Ann. Math. (2) 170, No. 2, 609--655 (2009; Zbl 1250.11067)] introduced the notion of a large divisor. In the paper under review, the authors generalize this to so-called essentially large divisors, which are defined as follows. Let \(K\) be any field of characteristic \(0\), \(X\) a non-singular projective variety defined over \(K\), and \(D\) a divisor on \(X\) defined over \(K\). Then \(D\) is said to be essentially very large over \(K\) if it is effective, and there exists an linear subspace \(V\) of \(L(D)=\{ f\;\text{rational function on}\;X|\text{div}(f)\geq -D\}\) such that for every \(P\in D(\overline{K})\) there exists a basis \(B\) of \(V\) such that \(\text{ord}_E\big(\prod_{f\in B} f\big)>0\) for every irreducible component \(E\) of \(D\) with \(P\in E(\overline{K})\). Further, a divisor \(D\) on \(X\) is called essentially large if it is effective, and has the same support as an essentially very large divisor. The definition of large divisor as given by Levin is the same, except that \(V=L(D)\). The authors prove, among other things the following results, which were proved by Levin for large divisors. First, let \(K\) be a number field, \(M_K\) its set of places, \(S\) a finite subset of \(M_K\) containing all infinite places. For a non-singular variety \(X\) and a divisor \(D\) on \(X\), both defined over \(K\), we call a subset \(M\) of \((X\setminus D)(K)\) \((S,D)\)-integral, if for some choice of local Weil functions \(\lambda_{v,D}\) (\(v\in M_K\)), the functions \(\lambda_{v,D}\) \((v\not\in S)\) are all bounded on \(M\). { Theorem 1.} Let \(X\) be a non-singular variety, and \(D\) an essentially large divisor on \(X\), both defined over \(K\). Then no \((S,D)\)-integral subset of \((X\setminus D)(K)\) is Zariski dense in \(X\). The main tool is Schmidt's subspace theorem, combined with an idea of \textit{P. Corvaja} and \textit{U. Zannier} originating from [C. R., Math., Acad. Sci. Paris 334, No. 4, 267--271 (2002; Zbl 1012.11051)]. The authors also prove the following analogue in Nevanlinna theory, using Cartan's second main theorem: {Theorem 2.} Let \(X\) be a non-singular variety, and \(D\) a large divisor on \(X\), both defined over \({\mathbb C}\). Then there is no holomorphic map \(f:{\mathbb C}\to X\setminus D\) having Zariski dense image in \(X\). Most of the paper is concerned with finding good criteria for divisors to be essentially large. The authors prove the following. {Theorem 3.} Let \(X\subset{\mathbb P}^l\) be a non-singular projective variety of dimension \(q\geq 1\), and \(D\) a divisor on \(X\), both defined over a field \(K\) of characteristic zero. Let \(q=\dim X\), \(r\geq q+2\), and \(D=D_1+\cdots +D_r\), where each \(D_i\) is defined as the restriction to \(X\) of a homogeneous polynomial from \(K[X_0,\ldots ,X_l]\) and each subset of \(q+1\) divisors from \(D_1,\ldots ,D_r\) has empty intersection with \(X\) over \(\overline{K}\). Then \(D\) is essentially large. The proof uses among others inequalities of \textit{R. G. Ferretti} and the reviewer for Hilbert weights and Chow weights [Int. Math. Res. Not. 2002, No. 25, 1295--1330 (2002; Zbl 1073.14521); Dev. Math. 16, 175--198 (2008; Zbl 1153.11032)]. The authors show that the condition that each \(D_i\) be defined by the restriction to \(X\) of a homogeneous polynomial from \(K[X_0,\ldots ,X_l]\) can be replaced by the \(D_i\) being effective if one assumes that \(\text{Pic}(X)={\mathbb Z}\).