On the Schmidt subspace theorem for algebraic points (Q488065)

The author proves various versions of Schmidt's subspace theorem for algebraic points. These are thus higher-dimensional analogs of Wirsing theorem. More specifically, let \(X\) be a projective variety of dimension \(n\) defined over a number field \(k\), and \(S\) be a finite set of places of \(k\). Suppose that \(D_1,\ldots, D_q\) are effective Cartier divisors on \(X\) in \(m\)-subgeneral position. Suppose further that there is an ample Cartier divisor \(A\) so that each \(D_i\) is numerically equivalent to \(d_i A\) for some positive integers \(d_i\). The main theorem (Theorem 5.2) proves the following generalization of Schlickewei's conjecture [\textit{H. P. Schlickewei}, in: Lect. Notes Math. 1819, 107--170 (2003; Zbl 1044.11063), Conjecture 5.1]: given \(\varepsilon >0\) and an integer \(\delta \geq 2\), \[ \sum_{v\in S} \sum_{w\in M_{k(P)}, w|v} \sum_{i=1}^q \frac{\lambda_{D_i, w}(P)}{d_i} < (c+ \varepsilon) h_A(P) + O(1) \tag{1} \] holds for all points \(P\in X(\overline k) \setminus \cup_i \mathrm{Supp} D_i\) satisfying \([k(P):k] \leq \delta\), where \(c = \frac{\delta m (\delta m - 1)(\delta n + 1)}{\delta m + \delta n -2}\), \(\lambda\) is a local height, and \(h\) is a Weil height. When \(X = \mathbb{P}^2\), \(D_i\) lines, and \(m=\delta = 2\), the author improves this to the sharp bound \(c=8\) (Theorem 6.3). In fact, \(c=\frac{15}2\) works if one excludes finitely many lines, and the author derives from Vojta's conjecture that this should be \(5\) if one excludes finitely many curves (and \(5\) is sharp by Example 7.3). The author also notes that the arguments carry over to the Nevanlinna setup. For the proof, the author first analyzes the smallest \(c = c(\ell, m,n)\) such that exceptions to (1) are \(< \ell\)-dimensional. In Theorem 4.1, he proves that such \(c(\ell, m,n)\) for points with \([k(P):k]\leq \delta\) is related to \(c(\ell, \delta m, \delta n)\) for rational points, through the symmetric powers of \(X\). To get a bound on the \(c(\ell, m,n)\) for rational points, he uses then a high enough multiple of \(A\) to embed \(D_i\)'s as hyperplane sections, then the asymptotic Riemann-Roch formula and the usual Schmidt's subspace theorem to prove the inequality of the form (1) for the lcm of two divisors, and the \(m\)-subgeneralness to show combinatorially that the sum of the lcm's over all pairs in \(\{1,\ldots, q\}\) contains a certain multiple of \(\sum D_i\). For the sharp bound in Theorem 6.3, the argument is quite more involved and uses Wirsing-type inequality for curves [\textit{X. Song} and \textit{T. J. Tucker}, Compos. Math. 116, No. 2, 219--238 (1999; Zbl 0933.11033)] as well as explicit equations and geometry of \(\mathrm{Sym}^2 \mathbb{P}^2 \hookrightarrow \mathbb{P}^5\).