Gamma factors for Asai representations of ${\rm GL}_2$

Shih-Yu Chen, Cheng Yao, Isao IshikawaarXiv 2018 · DOI · arXiv

Abstract

Let $E$ be a quadratic semisimple extension of a local field $F$ of characteristic zero. We determine the explicit relation between gamma factors for Asai representations of $R_{E/F}{\rm GL}_{2/E}$ defined by Weil-Deligne representations and local zeta integrals. While results for the split case ($E = F \times F$) were previously established by Henniart and Jacquet, this work completes the theory for general quadratic extensions through explicit calculation and global-to-local arguments.

Introduction

In the Langlands program, local factors—specifically L-functions ($L$), epsilon factors ($\epsilon$), and gamma factors ($\gamma$)—serve as fundamental invariants for studying the symmetry and analytic behavior of representations. These factors are typically defined through two distinct lenses: the arithmetic/algebraic "Galois side" via the Local Langlands Correspondence, and the analytic side via zeta integrals (such as Rankin-Selberg or Piatetski-Shapiro–Rallis methods).

Table of contents outlining the systematic approach to Asai local factors across different field types.

For a quadratic extension of fields $E/F$, the Asai representation provides a mechanism to transfer a representation from ${\rm GL}_n(E)$ to ${\rm GL}_{n^2}(F)$. This paper addresses the "missing link" in the explicit comparison between these two definitions for $n=2$, providing the necessary constants and relations to prove their equivalence across both archimedean and non-archimedean local fields.

Local Factors and the Langlands Program

The core challenge addressed is the reconciliation of the Galois-side Weil-Deligne representations with analytic zeta integrals. For a quadratic extension $E/F$, the Asai representation involves complex interactions between the fields. While the "split" case (where $E = F \times F$) was well-understood, the case where $E$ is a quadratic extension field remained incomplete until this study.

A critical component of this reconciliation is the Langlands constant $\lambda_{E/F}(\psi)$. The authors demonstrate that the analytic epsilon factor $\epsilon_{RS}$ is not identical to the Galois epsilon factor $\epsilon_{Gal}$; rather, they differ by a precise factor involving the central character and this constant, which acts as a correction for different additive character normalizations.

Methodology

The research employs a combination of explicit local calculations and global-to-local bridging arguments:

  • Non-Archimedean Case: The authors compute Rankin-Selberg integrals by selecting specific "good" sections of the representations and performing direct integration over p-adic groups.
  • Archimedean Analysis: For real and complex fields, the theory of zeta integrals is developed for Harish-Chandra representations, ensuring absolute convergence and meromorphic continuation.
  • Multiplicativity: A significant innovation is the proof of the multiplicativity of gamma factors for Asai representations. This allows the reduction of complex representations to simpler "principal series" cases, making the explicit calculations feasible.
  • Global Integration: By utilizing the known functoriality of global Asai transfers, the authors bridge gaps where local calculations are prohibitively complex, ensuring the local theory remains compatible with global automorphic forms.

Key Results

The paper culminates in several landmark theorems that solidify the theory of Asai local factors:

  • Theorem A ($n=1$): Establishes that for generic representations, the Rankin-Selberg L-function matches the Galois L-function ($L_{RS} = L_{Gal}$), providing the explicit proportionality constant for the epsilon factors.
  • Theorem B ($n=2$): Defines the relationship between gamma factors from the Piatetski-Shapiro–Rallis method ($\gamma_{PSR}$) and the Galois version ($\gamma_{Gal}$), incorporating the quadratic character $\omega_{E/F}$ associated with the extension.
  • Theorem C (Trilinear Forms): Connects local factors to the existence of invariant trilinear forms. It proves that the space ${\rm Hom}_{{\rm GL}_2(F)}(\pi \otimes \tau, \mathbb{C})$ is non-zero if and only if a specific epsilon factor condition is satisfied at the center of the critical strip ($s=1/2$).
  • Theorem D (Global Compatibility): Confirms that for every place $v$ of a number field $F$, the local factors align as expected, validating the local-global compatibility of the framework.

Conclusion and Future Directions

By completing the theory for ${\rm GL}_2$, this work provides a rigorous foundation for further exploration of Asai representations. While the results are currently specific to $n=2$, they suggest a path forward for higher ranks, despite the increasing complexity of zeta integrals for $n > 2$. Future research may also extend these findings to more general "Shalika" models or explore the implications of the twisted Asai factors detailed in the latter sections of the paper.

Summary updated

Abstract

Let $E$ be a quadratic semisimple extension of a local field $F$ of characteristic zero. We determine explicit relation between gamma factors for Asai representations of $R_{E/F}{\rm GL}_{2/E}$ defined by the Weil-Deligne representations and local zeta integrals. When $E = F\times F$, the results were due to Henniart and Jacquet. We completed the theory in this article based on explicit calculation.
BibTeX
@misc{chen2018gamma,
  author = {Shih-Yu Chen and Cheng Yao and Isao Ishikawa},
  title = {{Gamma factors for Asai representations of $\{\textbackslash\{\}rm GL\}\_2$}},
  year = {2018},
  eprint = {1810.12561},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1810.12561},
}
Figure