Generating arithmetically equivalent number fields with elliptic curves

From MaRDI portal





Pursuing earlier work [\textit{B. de Smit} and \textit{R. Perlis}, Bull. Am. Math. Soc. 31, 213-215 (1994; Zbl 0814.11053) and Technical Report 97-10, Universiteit van Amsterdam (1997)], the author constructs examples of arithmetically equivalent number fields \(K\), \(K'\) (i.e., \(K\), \(K'\) have the same zeta-functions) for which the \(p\)-parts of the class numbers are distinct. Here, \(p\) is a given odd prime. \(K\) and \(K'\) belong to subgroups \(H\), \(H'\) of the Galois group \(G\) of their common Galois hull (with respect to \(\mathbb{Q}\)) such that the corresponding permutation representations agree over the rationals; however, they do not so when viewed as integral \(p\)-adic representations: The first statement reflects the equality of the zeta-functions, the second is forced by the distinctness of the \(p\)-parts of the class numbers. In this paper, \(G\) is the group \(\text{GL}_2(\mathbb{F}_p)\), and it is shown that the subgroups \(H\), \(H'\) of index \(2p+2\) consisting of all matrices in \(G\) with first column, respectively second row, like square, zero, respectively zero, square, satisfy the above conditions. Finally, \(G\) is realized as a Galois group over \(\mathbb{Q}\) by considering \(p\)-torsion points on elliptic curves; the fields \(K\), \(K'\) to the above groups \(H\), \(H'\) are explicitly given. Actual computations are made for \(p=3\), and it turns out that here \(K\) and \(K'\) have class numbers differing by a factor of 3. For the entire collection see [Zbl 0891.00022].











This page was built for publication: Generating arithmetically equivalent number fields with elliptic curves

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3840182)