Rational versus transcendental points on analytic Riemann surfaces (Q2162761)

The paper continues the investigation of the author [Int. Math. Res. Not. 2020, No. 24, 9844--9886 (2020; Zbl 1464.11073)] into the behaviour of algebraic points and full subsets of points which verify some kind of Liouville inequalities on analytic Riemann surfaces. Let \(X_K\) be a projective variety of dimension \(n > 1\) defined over a number field \(K\), let \(M\) be a non compact Riemann surface. One of the main object of the paper is to estimate the number of points of bounded height \(h\) in a relatively compact open set \(U \subset M\) where \(\varphi : M \to X(\mathbb{C})\) is a holomorphic map with Zariski dense image. To do this, the author first re-proves the beautiful theorem by \textit{E. Bombieri} and \textit{J. Pila} [Duke Math. J. 59, No. 2, 337--357 (1989; Zbl 0718.11048)], which gives the estimation of the number of rational points \((\frac{m}{N},\frac{n}{N})\) on an arc of a convex curve lying in the unit square for \(N\) sufficiently large. Let \((X, L)\) be a polarized projective variety defined over a number field \(K\), let \({\sigma_0}; K \hookrightarrow \mathbb{C}\) be an embedding and \(X(\mathbb{C})\) the complex points of the variety \(X_{\sigma_0}\). In his version (Theorem 4.1.) of the theorem by E. Bombieri and J. Pila the author of the paper under review estimate from above, as a function of \(T\) , the cardinality \(A_U (T )\) of the set \(S_U (T ) := \{z \in U / \varphi(z)\in X( K), h_L (\varphi(z))\le T \}\). Theorem 4.1. For every positive number \(\epsilon\), we have \(A_U (T ) \ll \exp(T )\) where the involved constants depend on \(U, \varphi, \epsilon\) and \( L\) but not on T . By using elements of arithmetic geometry with Arakelov theory, Nevanlinna theory with other parts of complex analysis, and elements of measure theory, the author study and improves the upper bound for \(A_U (T )\). Recall some known definitions that the author uses. A model \({\mathcal X} \to\mathrm{Spec}({\mathcal O}_K )\) of \(X_K\) is a flat projective \({\mathcal O}_K\)-scheme whose generic fiber is isomorphic to \(X_K\). Let \({\mathcal X}\) be a model of \(X_K\). Let \(L_K\) and \({\mathcal X}\) be respectively a line bundle over \(X_K\) and a model of it; it is said that a line bundle \({\mathcal L}\) over \({\mathcal X}\) is a model of \(L_K\) if its restriction to the generic fiber is isomorphic to \(L_K\) . For infinite places \(\sigma, \tau\) of \(K\) a hermitian line bundle \(\overline{\mathcal L}\) is a couple \(({\mathcal L} ,\langle \cdot; \cdot \rangle_{\sigma})\), where \({\mathcal L}\) is a line bundle over \({\mathcal X}\) and \(\langle \cdot; \cdot \rangle_{\sigma}\) is a continuous hermitian product on \(L_{\sigma}\), with the condition that, if \({\sigma} = \overline \tau\), then \(\langle \cdot; \cdot\rangle_{\sigma} = \langle \cdot; \cdot \rangle_{\overline \tau}\). The author supposes that the holomorphic line bundle \(L\) over \(X(\mathbb{C})\) is equipped with a smooth hermitian metric \(\|\cdot\|\) and, for \(s \in H^0(X(\mathbb{C}), L^d)\), he denotes by \(s := \sup_{z \in X(\mathbb{C})} \{\|\varphi^*(s)\|(z)\}\). In this framework and on the base of his investigations the author gives two definitions. Let \({\log}^+x = \max(\log x, 0)\). The notion of a point of type \(S\) is introduced (Definition 5.1.): this is the point \(z \in X_{\sigma_0}(\mathbb{C})\) for which it is possible to find positive constants \(a = a(z, {\mathcal L} ), A = A(z, {\mathcal L})\) depending on \(z, {\mathcal L}\) such that, for every positive integer \( d\) and every non zero global section \(s \in H^0(X({\mathcal X}), {\mathcal L}^d)\) the next is hold: \( \log \|s_{\sigma_0}\|_{\sigma_0}(z) \ge - A(\log^+\|s\| + d)^a. \) Author uses notation \(S(X_K)\) for the subset of \(X_{\sigma_0}(\mathbb{C})\) of points of type \(S\). If \(a_0\) is fixed, he denotes by \(S_{a_0} (X_K )\) the set of \(S\)-points of \(X_{\sigma_0}\) for which the involved constant \(a\) is \(a_0\). Therefore \(S(X_K) = \cup_{a \ge 0} S_a (X_K)\). Theorem 5.3. If, for some real number \(a\) we have \(\varphi^{-1}(S_a) \cap U \neq \emptyset\), then \(A_U (T ) \ll T^a \). Next, the author defines a subset of type \(S_a\) of \(M\) (Definition 5.4.): for a compact subset \(B \subset U\) and a real number \(a\) the subset will be a subset of type \(S_a\) of \(M\) with respect to \(\varphi\) if, we can find a positive constant \(A > 1\) such that, for every positive integer \(d\) and \( s \in H^0({\mathcal X}. {\mathcal L}^d )\setminus \{0\}\) we have that \( \log \|s\|_B \ge - A(\log^+\|s\| + d)^a\). Theorem 5.5. Let \(\varphi : M \to X_{\sigma}(\mathbb{C})\) and \(U\) as before. Suppose that we can find a subset \(B \subset U\) of type \(S_a\) with respect to \(\varphi\). Then \(A_U (T ) \ll T^{2a} \) . The author extends the results on the case of rational points on leaves of one dimensional algebraic foliations. Let \(\Delta_r\) be the disk \({|z| < r}\) in \(\mathbb{C}\). Theorem 5.9. Let \({\mathcal F}\) be a foliation on a smooth quasi projective variety \(Z\) defined over a number field \(K\) . Let \(p \in Z(K)\) be a rational point and \(h : \Delta_1 \to Z_{\sigma}(\mathbb{C})\) be a holomorphic morphism tangent to \({\mathcal F}\) at \(p\). Suppose that the dimension of the Zariski closure of \(h(\Delta_1)\) is \(l > 1\). Let \(0 < r < 1\) be a real number, then, for every \(\epsilon > 0\) we have \(A_{\Delta_r} (T ) \ll_{\epsilon} T^{2l+ \epsilon}\). The paper under review includes other interesting results, the list of questions (section 7) and a useful list of references to related works.