The ABC theorem for higher-dimensional function fields

From MaRDI portal





The main theorem of this paper is an \(abc\)-like theorem for Diophantine approximation to hyperplanes in projective space, in the context of higher-dimensional function fields of arbitrary characteristic. Fix an algebraically closed field \(k\) of characteristic \(p\geq0\) and a projective variety \(V\) over \(k\) of dimension \(N\). Assume that \(V\) is nonsingular in codimension one and that its function field \(K=K(V)\) is separably generated over \(k\). Let \(H_1,\dots,H_q\), \(q\geq n+1\), be hyperplanes in \(\mathbb P^n_K\) in (linearly) general position and defined by linear forms with coefficients in \(k\), fix a finite set \(S\) of prime divisors of \(V\), and let \(\mathfrak X=[x_0:\dots:x_n]\in\mathbb P^n(K)\) be such that \(x_0,\dots,x_n\) are linearly independent over \(K^{p^m}\) for some \(m\in\mathbb N\). Then \[ \begin{multlined}(q-n-1)h({\mathfrak X})\leq\sum_{i=1}^q \sum_{\mathfrak p\notin S} N^{(np^{m-1})}_S (H_i,{\mathfrak X}) +\\ (n(n+1)/2)p^{m-1} \biggl(\deg\mathcal K_V+N\deg V+\sum_{\mathfrak p\in S}\deg\mathfrak p\biggr).\end{multlined} \] Here the counting functions are the ``obvious ones, using the fact that the hyperplanes are defined over \(k\), normalized using \(\deg\mathfrak p\), and the superscript \((np^{m-1})\) means that the multiplicities are truncated above \(np^{m-1}\). Also, if \(p=0\) then change \(p^{m-1}\) to \(1\). Moreover, for arbitrary \(p\), if \(N=1\) then the term \(N\deg V\) can be omitted. Finally, \(\mathcal K_V\) is the canonical divisor class of \(V\). The main ingredient in the proof is the following lemma on Wronskian determinants. Fix a separating transcendence basis \(\mathbf z=(z_1,\dots,z_N)\) for \(K\) over \(k\). Then Hasse (divided) partial derivatives \(D_{z_j}^\alpha\) are defined on \(k(z_1,\dots,z_n)\) for \(j=1,\dots,N\) and \(\alpha\geq0\). They extend uniquely to \(K\). For an \(N\)-tuple \(\alpha =(\alpha_1,\dots,\alpha_N)\in\mathbb N^N\) we define \(D_{\mathbf z}^{\alpha} =D_{z_1}^{\alpha_1}\dotsm D_{z_N}^{\alpha_N}\) and \(| \alpha| =\alpha_1+\dots+\alpha_N\). Let \(x_0,\dots,x_n\in K\). Then (in characteristic \(p>0\)): (a) \(x_0,\dots,x_N\) are linearly independent over \(K^{p^m}\) if and only if there exist integer vectors \(\mathbf 0={\pmb{\varepsilon}}^0,\dots,{\pmb{\varepsilon}}^n\), all of whose coordinates are \(\leq p^m-1\), such that the Wronskian determinant \(\det(D_{\mathbf z}^{\pmb{\varepsilon}^i}(x_j))_{0\leq i,j\leq n}\) is nonzero; and (b) if \(x_0,\dots,x_n\) are linearly independent over \(K^{p^m}\) then there are integer vectors \(\pmb{\varepsilon}^0,\dots,{\pmb{\varepsilon}}^n\) such that \(|{\pmb{\varepsilon}}^i| \leq ip^{m-1}\) for all \(i\), and such that the above Wronskian is nonzero. The proof uses methods derived from those of \textit{A. Garcia} and \textit{J. F. Voloch}, [Manuscr. Math. 59, 457--469 (1987; Zbl 0637.12015); \textit{J. T.-Y. Wang}, J. Number Theory 58, 139--157 (1996; Zbl 0866.11066); and Manuscr. Math. 98, 255--264 (1999; Zbl 0931.11009)].



Cites work









This page was built for publication: The ABC theorem for higher-dimensional function fields

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4813789)