F-split Galois representations are potentially abelian

From MaRDI portal



Abstract: In this note we relate the property of a semisimple l-adic Galois representation being "F-split" for a number field F to its having abelian image. By F-split we mean that the characteristic polynomials of Frobenii all split in F.


Let \((\rho_\ell)_\ell\) be a strictly compatible \(\mathbb{Q}\)-rational system of continuous, semisimple, \(n\)-dimensional \(\ell\)-adic Galois representations of \(G\), the absolute Galois group of \(\mathbb{Q}\). The author shows that there is a finite extension \(F\) of \(\mathbb{Q}\) such that all the images \(\rho_\ell(G|_F)\) are Abelian (i.e. \(\rho\) is ``potentially Abelian) if and only if there are finite extensions \(L\) and \(F\) of \(\mathbb{Q}\) such that for the Frobenius elements \(\Phi_r\) of \(L\) the characteristic polynomial of \(\rho_\ell(\Phi_r)\) (for \(\ell\) coprime to \(r\)) splits over \(F\) (i.e. \(\rho\) is ``potentially \(F\)-split). The particular example which motivated this result was: the properties of being potentially Abelian and of being potentially \(F\)-split both are properties which characterize that a system of Galois representations coming from an elliptic curve actually comes from a CM elliptic curve. The result is proven in a slightly more general situation than stated here. A main ingredient in the proof is a transcendency result of \textit{G. Henniart} [Théorie des nombres, Sémin. Delange-Pisot-Poitou, Paris 1980-81, Prog. Math. 22, 107-126 (1982; Zbl 0498.12012)].











This page was built for publication: $F$-split Galois representations are potentially abelian

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4417192)