Modularity of the Rankin-Selberg L-series, and multiplicity one for SL(2)

From MaRDI portal
Publication:1589983



Abstract: A fundamental question, first raised by Langlands, is to know whether the Rankin-Selberg product of two (not necessarily holomorphic) cusp forms f and g is modular, i.e., if there exists an automorphic form f box g on GL(4)/Q whose standard L-function equals L^*(s, f x g) after removing the ramified and archimedean factors. The first main result of this paper is to answer it in the affirmative, in fact with the base field Q replaced by any number field F. Our proof uses a mixture of converse theorems, base change and descent, and it also appeals to the local regularity properties of Eisenstein series and the scalar products of their truncations. One of the applications of this result is that the space of cusp forms on SL(2) has multiplicity one. Concretely this means the following: If f, g are newforms of holomorphic or Maass type with trivial character such that the squares of the p-th coeficients of f and g are the same at almost all primes p, then g must be a twist of f by a quadratic Dirichlet character.


The main result of the paper is the following theorem. Let \(\pi\) and \(\pi'\) be two isobaric automorphic representations of \(\mathrm{GL}(2)\) over a number field \(F\). Then there exists an automorphic representation \(\Pi\) of \(\mathrm{GL}(4)\) over \(F\) whose \(L\)- and \(\varepsilon\)-factors are \(L(s,\pi_v \times \pi'_v)\) and \(\varepsilon(s,\pi_v \times \pi'_v)\) (the \(L\)-function of \(\Pi\) is the ``convolution of the \(L\)-functions of \(\pi\) and \(\pi'\)). The theorem is completed by a criterion for the cuspidality of \(\Pi\). The result is derived from a converse theorem for \(\mathrm{GL}(4)\). In order to define candidates for \(\Pi_v\) at all places, the author uses base change to an infinite family of solvable extensions of \(F\). Then the analytic properties of \(L(s,\Pi\times\eta)\), with \(\eta\) cuspidal on \(\mathrm{GL}(1)\) or \(\mathrm{GL}(2)\), have to be proved, so that the converse theorem can be applied. The most difficult part is the required boundedness in vertical strips of these \(L\)-functions (for \(\eta\) on \(\mathrm{GL}(2)\)). Here the author uses the integral representation of the triple \(L\)-function defined by Piatetski-Shapiro and Rallis. That integral involves an Eisenstein series \(E(f_s)\) on \(\mathrm{GSp}(6)\), which has to be estimated. Arthur's truncation functor is used to do that. Finally one descends to the original base field \(F\). Applications of the theorem are, among others, multiplicity one for cusp forms on \(\mathrm{SL}(2)\) and a proof of the Tate conjecture for 4-fold products of modular curves.




Cited in
(only showing first 100 items - show all)








This page was built for publication: Modularity of the Rankin-Selberg \(L\)-series, and multiplicity one for \(\mathrm{SL}(2)\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1589983)