U.K. Anandavardhanan,C.S. Rajan
Abstract: We show the equality of the local Asai L-functions defined via the Rankin-Selberg method and the Langlands-Shahidi method for a square integrable representation of GL(n,E). As a consequence we characterise reducibility of certain induced representations of U(n,n), and the image of the base change map from U(n) to GL(n,E) in terms of GL(n,F)-distinguishedness.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is about the equivalence between different definitions of local Asai L - functions of simple groups and some representation - theory problems arising therefrom. Specifically:
1. **Prove the equality of different definitions of local Asai L - functions**: The author proves, by global methods, that for a square - integrable representation π of GLn(E), the local Asai L - functions defined by the Rankin - Selberg method and the Langlands - Shahidi method are equal, that is, \( L_1(s, \text{As}(\pi)) = L_2(s, \text{As}(\pi)) \).
2. **Characterize the reducibility of certain induced representations**: Based on the above results, the author further studies the reducibility of certain induced representations of U(n,n) and describes the images of these representations under the base - change mapping.
3. **Characterize stable base - change lifts**: The author uses these results to characterize the stable base - change lifts from U(n) to GLn(E), especially for the case of supercuspidal representations.
### Main Conclusions
- **Theorem 1**: For a supercuspidal representation π of GL3(E), it is distinguished with respect to GL3(F) if and only if it is a stable base - change lift from U(3).
\[
\pi \text{ is distinguished with respect to } GL_3(F) \iff \pi \text{ is a stable base - change lift from } U(3)
\]
- **Theorem 2**: For the quasi - split unitary group U(n,n), given a discrete - series representation π, the structure of its R - group \( R(\pi) \) is related to whether πi is an ωE/F - distinguished representation. Specifically,
\[
R(\pi) \simeq (\mathbb{Z}/2\mathbb{Z})^r
\]
where \( r \) is the number of inequivalent πi that are ωE/F - distinguished with respect to GLni(F).
- **Corollary 1**: For the maximal parabolic subgroup P of U(n,n) whose Levi part is isomorphic to GLn(E), if π is a discrete - series representation of GLn(E), then I(π) is reducible if and only if π is ωE/F - distinguished.
- **Theorem 3**: The distinguishability of the Steinberg representation π depends on the parity of n. Specifically, π is distinguished with respect to a character χ◦det of GLn(F) if and only if n is odd and χ is the trivial character, or n is even and χ = ωE/F.
### Methods
The author uses a variety of tools and techniques to solve these problems, including:
- **Rankin - Selberg method**: Used to define and study Asai L - functions.
- **Langlands - Shahidi method**: Provides another way to define Asai L - functions and is related to the base - change theory.
- **Global method**: Solves the problem by embedding local representations as local components of automorphic representations and using the properties of global L - functions.
These results not only deepen the understanding of Asai L - functions but also provide new perspectives for the base - change theory and other problems in representation theory.