Poles of p p -adic Asai L L -functions and distinguished representations
David Loeffler, Sarah Livia Zerbes · 2025
Read Poles of p p -adic Asai L L -functions and distinguished representations by David Loeffler, Sarah Livia Zerbes (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-11472.txt.
Introduction
David Loeffler and Sarah Livia Zerbes give a test for special symmetry in number theory objects in this 2025 paper. Their objects are automorphic representations, deep patterns that encode number behavior over a real quadratic field. Some are distinguished, picked out by a symmetry of a smaller group.
A classic theorem by Flicker detects that trait with a complex-valued L-function. The pair give a p-adic version, using a p-adic Asai L-function and its poles. Where the function blows up tells whether the trait holds. The result carries a known test into the p-adic world. It serves researchers on the Langlands program's many bridges.
Catalog record
| CREATOR | David Loeffler, Sarah Livia Zerbes |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 42 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-11472
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Poles of p p -adic Asai L L -functions and distinguished representations Thanks: D.L. gratefully acknowledges the support of the European Research Council through the Horizon 2020 Excellent Science programme (Consolidator Grant “ShimBSD: Shimura varieties and the BSD conjecture”, grant ID 101001051) David Loeffler, Sarah Livia Zerbes Abstract . We give a criterion in terms of p p -adic Asai L L -functions for a cuspidal automorphic representation of GL 2 \GL_{2} over a real quadratic field F F to be a distinguished representation, providing a p p -adic counterpart of a well-known theorem of Flicker for the complex Asai L L -function. 2020 Mathematics Subject Classification 11F41, 11F67 1. Introduction 1.1. Background: complex case Let \Pi be a cuspidal automorphic representation of \GL_{2}/F , for a real quadratic field F . Then the Asai , or twisted tensor , L -series L_{\As}(\Pi,s) is defined by an Euler product which converges for \operatorname{Re}(s)>1 . It always has meromorphic continuation to \mathbf{C} , but the continuation is not always an entire function: it is holomorphic away from s=1 , but it can have a simple pole at s=1 . Similar properties hold more generally for the twisted Asai L -series L_{\As}(\Pi,\chi,s) for \chi a Dirichlet character. The occurrence of these poles is governed by the following theorem (which is due to Flicker [ Fli88 ] for \chi=1 , and can be extended to general \chi using results of Lapid–Rogawski [ LR98 ] ): Theorem 1.1 . The following are equivalent: (1) The function L_{\As}(\Pi,\chi,s) has a pole at s=1 . (2) \Pi is \chi -distinguished , i.e., we have \chi^{-2}=(\omega_{\Pi})|_{\mathbf{Q}} (where \omega_{\Pi} is the central character of \Pi ) and the integral \int_{\mathbf{A}^{\times}\GL_{2}(\mathbf{Q})\backslash\GL_{2}(\mathbf{A})}\phi(h)\chi(\det h)\,\mathrm{d}h is non-zero for some \phi\in\Pi . (Note that the condition on \chi is required for the integrand to be well-defined.) Moreover, there exist characters \chi such that \Pi is \chi -distinguished if, and only if, \Pi is a twist of a base-change from \GL_{2}/\mathbf{Q} . Thus the presence or absence of poles of L_{\As}(\Pi,\chi,s) serves to identify representations with a \GL_{2}(\mathbf{A}) -invariant period, and the representations with such a period are precisely those in the image of a specific functorial lifting – both statements which admit a wealth of interesting generalisations to other automorphic forms and their L -series. If \Pi=\BC(\pi)\times\psi is a twist of a base-change, then we have L_{\As}(\Pi,\chi,s)=L_{\As}(\BC(\pi),\chi\psi|_{\mathbf{Q}},s) , so we can suppose \psi=1 without loss of generality. Then we have the factorisation formula ( Corollary 2.1 below) L_{\As}(\Pi,\chi,s)=L(\Sym^{2}\pi,\chi,s)\cdot L(\chi\omega_{\pi}\eta_{F},s), where \eta_{F} is the quadratic character associated to F/\mathbf{Q} . (Compare e.g. [ Gha96 , Ch. 4, Cor. 1] for imaginary quadratic fields.) It follows that \Pi is always \chi -distinguished for \chi=\eta_{F}\omega_{\pi}^{-1} ; and this is the only such character unless \pi is dihedral (in which case there may be further poles coming from the \Sym^{2} term). 1.2. P-adic case In the recent paper [ GLZ23 ] , Grossi and the present authors defined a p -adic counterpart of the Asai L -function, assuming p=\mathfrak{p}_{1}\mathfrak{p}_{2} is split in F and \Pi is generated by a holomorphic eigenform \mathcal{F} which is ordinary at \mathfrak{p}_{1} . However, this function is always a p -adic measure, i.e., an analytic function on the weight space; it never has poles, irrespective of the choice of \mathcal{F} . In this p -adic setting, we do have an analogue of the factorisation formula ({\dagger}) (although this is a significantly deeper theorem than the complex version, and will be established in forthcoming work of D. Krekov); and the p -adic zeta function has a pole at s=1 , like its complex counterpart. However, the p -adic symmetric square L -function has a “trivial zero” at s=1 , forced by the shape of the interpolation factors at p , and this trivial zero cancels out the pole of the p -adic zeta function. Hence the presence or absence of poles of the p -adic L -function can no longer be used to characterize which automorphic representations are distinguished. In this short note, we show that this problem can be circumvented via an “improved” p -adic Asai L -function, interpolating the values L(\As\mathcal{F}_{m},\chi,1-m) as \mathcal{F}_{m} varies through a Hida family of Hilbert eigenforms of weight (k+2m,k) for varying m and fixed k . This function satisfies an interpolation formula, relating it to the Asai L -function of the weight (k+2m,k) specialisation at s=1-m ; and this interpolation formula involves fewer Euler factors than the p -adic L -function of [ GLZ23 ] – in particular, the Euler factor that gives the trivial zero does not appear. We show that this improved p -adic L -function genuinely can have a pole at m=0 ; and, moreover, this pole occurs if and only if the weight k specialisation generates a \chi -distinguished representation, giving a p -adic counterpart of Theorem 1.1 . 2. Complex L -series and period integrals We fix a real quadratic field F of discriminant D , and a numbering of its real embeddings as (\sigma_{1},\sigma_{2}) . 2.1. Asai L -functions Let \Pi be a (unitary) cuspidal automorphic representation of \GL_{2}/F , generated by a cuspidal Hilbert modular newform \mathcal{F} of level \mathfrak{N} and weight (k_{1},k_{2}) . Then we define the Asai L -function L_{\As}(\Pi,s) as in §1.2 of [ GLZ23 ] , with Euler factors at the bad primes determined via the local Langlands correspondence; this extends the original definition due to Asai [ Asa77 ] in the case \mathfrak{N}=1 . We define twisted Asai L -functions L_{\As}(\Pi,\chi,s) similarly. We also define the imprimitive Asai L -function L_{\As}^{\imp}(\Pi,\chi,s) as Definition 5.1.2 of op.cit. , given by L_{\As}^{\imp}(\Pi,s)=L_{N_{\chi}\Nm(\mathfrak{N})}\Big(\chi^{2}\omega_{\Pi}|_{\mathbf{Q}},2s\Big)\cdot\sum_{\begin{subarray}{c}n\geqslant 1\\ (n,N_{\chi})=1\end{subarray}}a^{\circ}_{n\mathcal{O}_{F}}(\mathcal{F})\chi(n)n^{-s}, where a^{\circ}_{\mathfrak{m}}(\mathcal{F}) is the (unitarily-normalised) Hecke eigenvalue of \mathcal{F} at the ideal \mathfrak{m} , and L_{N_{\chi}\Nm(\mathfrak{N})}(-) denotes the Dirichlet L -series with the Euler factors at primes dividing N_{\chi}\Nm(\mathfrak{N}) omitted. These are related by L^{\imp}_{\As}(\Pi,\chi,s)=L_{\As}(\Pi,\chi,s)\cdot\prod_{\ell\mid N_{\chi}\operatorname{Nm}(\mathfrak{N})}C_{\ell}(\Pi,\chi,\ell^{-s}) for some polynomials C_{\ell}(\Pi,\chi,X)\in\overline{\mathbf{Q}}[X] ; moreover, the zeroes of the factors C_{\ell}(\Pi,\chi,\ell^{-s}) have real parts \leqslant 0 ( [ LLZ18 , Proposition 5.1.3] ). Thus L^{\imp}_{\As}(\Pi,\chi,s) , like L_{\As}(\Pi,\chi,s) , has a simple pole at s=1 when \Pi is \chi -distinguished and is an entire function otherwise. Base-change case Given a cuspidal automorphic representation \pi of \GL_{2}/\mathbf{Q} , there is an automorphic representation \Pi=\BC(\pi) of \GL_{2}/F , the base-change (or Doi–Nagunuma lift) of \pi , which is cuspidal unless \pi is induced from a Grössencharacter of F . Note that the central character of \BC(\pi) is given by \omega_{\Pi}=\omega_{\pi}\circ\Nm_{F/\mathbf{Q}} ; and if \pi is generated by a holomorphic modular form of weight k , then \Pi is generated by a holomorphic Hilbert modular form of parallel weight (k,k) . Corollary 2.1 . If \Pi=\operatorname{BC}(\pi) , then there is a factorisation of complex L -functions L_{\As}(\Pi,\chi,s)=L(\Sym^{2}\pi,\chi,s)\cdot L(\chi\eta_{F}\omega_{\pi},s), for any \chi , where L(\Sym^{2}\pi,\chi,s) is the twisted symmetric square L -function. Proof. It is easily seen that the local Euler factors at unramified primes agree on both sides. To check the compatibility at all places, one can make a purely automorphic argument using compatibility of local and global base-change maps; or when \pi is holomorphic (the only case we need below) we can argue that both sides are the L -functions of compatible families of semisimple Galois representations,