Siegel modular forms, \(L\)-functions, and Satake parameters (Q5931953)

Suppose \(F\) is a Siegel modular form of genus \(n\) which is a simultaneous eigenform for all Hecke operators. The Hecke algebra decomposes as \(\bigotimes_p H_{n,p}\), and the map sending a Hecke operator to the eigenvalue gives an element of \(\Hom(H_{n,p},{\mathbb C})\). This is isomorphic to \(({\mathbb C}^\times)^{n+1}/W\), where \(W\) is the Weyl group, so \(F\) and \(p\) determine an \((n+1)\)-tuple of nonzero complex numbers mod \(W\). This is the Satake \(p\)-parameter of \(F\). The main result of this paper is, essentially, that the \(p\)-parameters for almost all primes determine the \(p\)-parameters for all primes. This is well known for \(n=1\) (when almost all the \(p\)-parameters suffice to determine \(F\) itself) and is proved here for \(n=2\); for \(n>2\) it is proved only up to a sign. For the proof, the author takes two eigenforms and compares certain \(L\)-functions derived from them. She shows that the assumption that almost all the \(p\)-parameters agree forces the \(L\)-functions to agree also, for reasons of elementary complex analysis. For \(n=2\) she uses the spinor zeta function as her chosen \(L\)-function and she can recover all the \(p\)-parameters (but not the eigenform) from it; but for \(n>2\) this is not known to have an analytic continuation, so she is obliged to make do with the standard zeta function, which gives slightly less information.