Number Theory

The Langlands Program

Цели урока

  • Understand the Langlands correspondence: Galois representations <-> automorphic forms with equal L-functions
  • Know the modularity theorem as the complete GL_2 Langlands correspondence over Q
  • Understand the functoriality conjecture and its proved cases
  • Have an overview of the geometric Langlands program and its 2024 breakthrough

Предварительные знания

  • Iwasawa Theory
  • Automorphic Forms
  • Class Field Theory
  • Iwasawa Theory
  • Class Field Theory

Why does a proof about whole numbers (Fermat's Last Theorem) require deep understanding of analytic functions on a complex domain? The Langlands program is the answer.

  • Cryptography: modularity theorem enables verified pairing-based cryptography on concrete curves
  • Machine learning: transformer architecture attention mechanisms are formally analogous to Hecke operators
  • Quantum computing: geometric Langlands is used in topological quantum field theory constructions
  • Theoretical physics: S-duality in N=4 super Yang-Mills is the geometric Langlands correspondence

A 17-page letter that changed mathematics

January 1967. Robert Langlands writes a letter to Andre Weil opening with the suggestion that the text should be read as pure speculation rather than serious mathematics, a framing for the ideas about to be explained. The letter is 17 pages and outlines a program unifying number theory, representation theory, and geometry. Weil politely replied he didn't understand. Over the next 57 years, mathematicians proved pieces: Lafforgue (function fields, Fields Medal 2002), Ngo Bao Chau (fundamental lemma, Fields Medal 2010), and in 2024 - the geometric correspondence for unitary groups, proved by a 9-author team.

The Langlands Correspondence

In 1967, Robert Langlands wrote a 17-page letter to Andre Weil laying out a grand program to unify number theory, geometry, and representation theory. Andrew Wiles's 1995 proof of Fermat's Last Theorem used a special case of this program, and by 2024 the full program for unitary groups was proved by a team of 9 mathematicians.

Class field theory is exactly the GL_1 case of the Langlands program: the correspondence Gal(K^ab/K) ~ A_K^x/K^x is the local Langlands correspondence for GL_1. The program generalizes this to GL_n and arbitrary reductive groups.

In the Langlands correspondence, homomorphisms rho: Gal(K-bar/K) -> GL_n(C) correspond to automorphic forms Pi for GL_n. What is the main consequence?

The GL_2 Case: Modularity Theorem

The most concrete instance of the Langlands correspondence is for GL_2 over Q. Here automorphic representations are modular forms of level N, and the corresponding Galois representations are 2-dimensional l-adic. This case is fully proved and has revolutionary consequences.

Fermat's Last Theorem as a consequence

Frey curve and GL_2 Langlands

Suppose a^n + b^n = c^n for prime n >= 5. Frey attaches a curve E: y^2 = x(x - a^n)(x + b^n). Ribet (1990) proved: if such a curve existed, it could not be modular. Wiles's theorem says all elliptic curves over Q are modular. Contradiction - no Fermat solution exists. The entire proof is an instance of the GL_2 Langlands correspondence.

What does modularity of an elliptic curve E/Q mean in the Langlands correspondence?

Modularity of E: there exists f in S_2(Gamma_0(N_E)) with a_p(E) = a_p(f) for all but finitely many p. This gives L(E,s) = L(f,s). Wiles proved this for all elliptic curves over Q in 1995 (semistable case) and 2001 (general case).

Functoriality and Its Consequences

Functoriality is the central conjecture of the Langlands program. It predicts: a homomorphism of L-groups L_H -> L_G should transfer automorphic forms from H to G while preserving L-functions. Each proved case of functoriality is a major theorem.

The Ramanujan conjecture for GL_2 (|a_p| at most 2p^{(k-1)/2}) follows from functoriality: tempered representations on the analytic side correspond to having all Satake parameters of absolute value 1. Deligne proved this for holomorphic forms via Weil conjectures. For Maass forms it remains open.

Functoriality caseProvedAuthor(s)
Base change for GL_n (cyclic)1989Arthur-Clozel
Symmetric square for GL_21978Gelbart-Jacquet
Sym^3 and Sym^4 for GL_22002Kim-Shahidi
Full program for unitary groups20249-author team
General GL_n functorialityOpen-

What does Langlands functoriality predict for a homomorphism r: L_H -> L_G?

Functoriality: r: L_H -> L_G gives a transfer Pi_H ~> Pi_G with L(s,Pi_H) = L(s,Pi_G). Proved for base change (Arthur-Clozel 1989) and symmetric powers for small n (Kim-Shahidi 2002). The general case is open.

Geometric Langlands Program

The geometric version of the Langlands program replaces number fields with function fields of algebraic curves, and algebraic representations with D-modules and perverse sheaves. This makes the correspondence more tractable and brings in powerful geometric tools.

The geometric Langlands correspondence for GL_n over function fields was proved by Frenkel, Gaitsgory, and others. In 2024, Ben-Zvi, Chen, Helm, Nadler, and others proved a version of the geometric correspondence over the complex numbers. This is a major breakthrough after 30 years of work.

Geometric Langlands and physics

Connection to gauge theory

Kapustin and Witten (2007) showed that the geometric Langlands correspondence is the mathematical content of S-duality in 4-dimensional N=4 super Yang-Mills theory. The Langlands dual group L_G is the electromagnetic dual of G. This gives a physical interpretation of a purely mathematical correspondence - and suggests proof strategies using ideas from quantum field theory.

What corresponds to automorphic forms in the geometric Langlands program?

In the geometric program: local systems (= pi_1 representations) <-> D-modules on Bun_G. Hecke eigensheaves are the geometric analog of automorphic Hecke eigenforms. Proved for GL_n over function fields; proved for complex curves in 2024.

Connections to other topics

The Langlands program is the central unifying program of modern mathematics.

  • Representation theory — Related topic
  • Algebraic geometry — Related topic
  • Theoretical physics — Related topic

Итоги

  • Langlands correspondence: {rho: Gal -> GL_n} <-> {automorphic forms Pi for GL_n}, L(s,rho) = L(s,Pi)
  • Class field theory is the abelian GL_1 case: Gal(K^ab/K) ~ A_K^x/K^x
  • Modularity theorem (Wiles 1995): every elliptic curve over Q has L(E,s) = L(f,s); implies Fermat
  • Functoriality: r: L_H -> L_G transfers automorphic forms preserving L-functions
  • Geometric Langlands: local systems <-> D-modules on Bun_G; proved for unitary groups in 2024

Вопросы для размышления

  • Why is class field theory exactly the GL_1 Langlands program, and what new structure is needed for GL_n with n > 1?
  • How can functoriality be used to prove analytic continuation of L-functions, and why is this a key application?
  • What is the significance of the 2024 proof of geometric Langlands for the classical number-theoretic version of the program?

Связанные уроки

  • nt-25-iwasawa-theory — Iwasawa theory provides the p-adic approximation to the Langlands program
  • nt-28 — Automorphic forms are the main objects on the right side of the Langlands correspondence
  • nt-24-class-field-theory — Class field theory is the abelian (GL_1) case of the Langlands program
The Langlands Program

0

1

Sign In