Satake isomorphism

Satake isomorphism
Name	Satake isomorphism
Field	Representation theory, Number theory, Algebraic geometry
Introduced by	Ichirō Satake
Year	1963

Satake isomorphism The Satake isomorphism is a fundamental identification between a spherical Hecke algebra for a reductive group over a nonarchimedean local field and a symmetric algebra associated to the Langlands dual group. It lies at the crossroads of Ichirō Satake's theory of spherical representations, the Langlands program, and the structure theory of reductive groups, and it provides a crucial bridge between harmonic analysis on p-adic fields, algebraic geometry, and automorphic form theory.

Introduction

The Satake isomorphism relates the unramified Hecke algebra of a split reductive group over a nonarchimedean local field such as Q_p or a finite extension to the representation ring or coordinate ring of the complex dual group like GL_n(C), SL_n(C), Sp_2n(C), or SO_n(C). It is central to the formulation of the unramified case of the Langlands correspondence and underpins constructions in the theories of automorphic representations, Shimura variety cohomology, and the study of L-functions.

Historical background

Ichirō Satake introduced his isomorphism in 1963 while studying spherical functions and harmonic analysis on p-adic Lie groups, building on earlier work by Harish-Chandra on real groups, and drawing on the structure theory of Chevalley groups, Cartan and Weyl group methods. Later developments connected Satake's construction with the algebraic geometry of Tannakian categorys, the formulation of the Langlands dual group by Robert Langlands and Pierre Deligne, and with geometric reformulations by George Lusztig and work surrounding the geometric Langlands program by Alexander Beilinson and Vladimir Drinfeld. The isomorphism also influenced progress on the Mackey theory for reductive groups and the modern theory of Hecke algebra modules.

Statement of the Satake isomorphism

Let G be a connected split reductive group over a nonarchimedean local field F such as Q_p or a finite extension, with maximal compact subgroup K (for example GL_n(O_F) inside GL_n(F)). The spherical Hecke algebra H(G,K) of compactly supported, bi-K-invariant complex-valued functions on G with convolution is commutative. The Satake isomorphism asserts that H(G,K) is canonically isomorphic to the algebra of W-invariant regular functions on the complex torus given by the complexified dual maximal torus of the Langlands dual group G^, where W denotes the Weyl group of G. Equivalently, H(G,K) ≅ C[X_*(T)]^W after an explicit normalization involving the modular character and choices related to the Iwasawa decomposition and the Cartan decomposition of G.

Proof outline and key ideas

The proof uses the Cartan decomposition G = K T^+ K for a split torus T and positive Weyl chamber T^+, together with the Iwasawa decomposition and the integration formulas developed by Harish-Chandra and George Mackey. One constructs the Satake transform mapping a bi-K-invariant function to a W-invariant function on T by averaging over K-double cosets and normalizing by the modular character (with parallels to the Gindikin–Karpelevich formula). Key algebraic inputs include the description of double cosets by dominant cocharacters, properties of the Hecke algebra as a convolution algebra, and the representation-theoretic link to unramified principal series representations induced from characters of T, as studied by Jacques Tits and Jean-Pierre Serre.

Unramified representations and Hecke algebras

Unramified representations of G(F) are smooth irreducible representations admitting nonzero K-fixed vectors; archetypal examples occur for GL_n(F) and classical groups studied by Iwahori and Matsumoto. The spherical Hecke algebra H(G,K) acts on the K-fixed vectors, and by the Satake isomorphism its spectrum parametrizes semisimple conjugacy classes in the complex dual group G^. This connection underlies the unramified local Langlands correspondence envisioned by Robert Langlands and made explicit in many classical cases by work of Harris and Taylor, Henniart, and others. The identification also yields explicit formulas for eigenvalues of Hecke operators on automorphic representations and for local factors of L-functions.

Satake transform and spherical functions

The Satake transform is the explicit linear map S: H(G,K) → C[X_*(T)]^W sending a compactly supported, bi-K-invariant function to its integral along K-orbits on the torus T, normalized by the modular character. S intertwines convolution with multiplication and sends spherical Hecke operators to symmetric functions in characters of the dual torus; this explains the compatibility of spherical functions with eigenvalues coming from diagonal matrices in examples like GL_2(Q_p). Classical spherical functions on Riemannian symmetric spaces studied by Harish-Chandra have p-adic analogues given by matrix coefficients of unramified principal series, and the Satake transform identifies these with characters or traces on representations of the dual group, linking to the Peter–Weyl theorem flavor in the p-adic setting.

Applications and connections (Langlands correspondence, representation theory)

The Satake isomorphism is a cornerstone in the unramified Langlands correspondence, allowing one to attach to an unramified representation of G(F) a semisimple conjugacy class in the dual group G^ which determines local factors of automorphic L-functions used by Langlands, Gelbart, Jacquet, and Shalika. It informs the construction of Galois representations in the work of Wiles and Taylor–Wiles, plays a role in the cohomology of Shimura varietys and arithmetic geometry via comparison with Frobenius conjugacy classes, and appears in categorical and geometric avatars such as the geometric Satake equivalence developed by George Lusztig, I. Mirković, and K. Vilonen tying perverse sheaves on the affine Grassmannian to representations of G^. The isomorphism also guides harmonic analysis on p-adic groups in the tradition of Bernstein and Zelevinsky and informs explicit computations for classical groups appearing in work by Moeglin and Vignéras.

Category:Representation theory