Number Theory

Automorphic Forms

Цели урока

  • Understand Hecke operators and multiplicativity of coefficients of eigenforms
  • Know the adelic formulation of automorphic representations
  • Distinguish holomorphic modular forms and Maass forms
  • Understand the Atkin-Lehner theory: newforms, oldforms, primitive forms

Предварительные знания

  • Zeta Functions of Varieties
  • Langlands Program
  • Complex analysis
  • Zeta Functions of Varieties
  • Langlands Program

How do the numbers 1, -24, 252, -1472, 4830, ... (Ramanujan tau coefficients) encode the number of points on algebraic varieties over finite fields?

  • Cryptography: L-functions of modular forms verify safety properties of elliptic curves in ECDH
  • Quantum chaos: Maass form eigenvalue distribution matches energy level statistics in chaotic quantum systems
  • Coding theory: extremal lattice codes defined by modular forms achieve Shannon capacity in special cases
  • Particle physics: Hecke operators are formally analogous to Wilson loop operators in conformal field theories

From Ramanujan to Deligne: 58 years for one inequality

In 1916, Srinivasa Ramanujan published a paper on the function tau(n). He computed the first values and made a striking conjecture: |tau(p)| <= 2p^{11/2} for all primes p. No proof idea existed - just intuition. For 58 years, mathematicians tried to understand this inequality. It turned out to be equivalent to the Weil conjectures about zeta functions - also open. In 1974 Pierre Deligne proved both simultaneously, earning the Fields Medal in 1978. The proof technique - etale cohomology - opened a new era in algebraic geometry.

0

1

Sign In

Modular Forms and Hecke Operators

Srinivasa Ramanujan conjectured in 1916 that the coefficients tau(n) of the Delta function Delta(z) = q prod(1-q^n)^24 satisfy |tau(p)| <= 2p^{11/2}. Pierre Deligne proved this in 1974 as a corollary of his Weil conjectures theorem.

Multiplicativity of coefficients a_{mn} = a_m a_n follows from the commutativity of the Hecke algebra: T_m T_n = T_{mn} when gcd(m,n)=1. A simultaneous eigenvector of all T_n automatically has multiplicative eigenvalues.

Why are Hecke operators central to the theory of modular forms?

Adelic Automorphic Forms

Classical modular forms are functions on the upper half-plane. The adelic reformulation packages all local information into a single function on the adeles and allows working with automorphic representations of GL_n(A_K) for arbitrary n and number fields K.

The adelic language is the right setting for the Langlands program: automorphic representations of GL_n(A_K) correspond (conjecturally) to n-dimensional Galois representations of Gal(K-bar/K). The Hecke operators at finite primes become translation operators in the adelic double coset.

Satake parameters and Euler factors

Connecting local representation theory to L-functions

For the modular form f = sum a_n q^n with a_p = tau(p) (Delta form), the local component Pi_p of the adelic form has Satake parameters alpha_p, beta_p satisfying alpha_p + beta_p = tau(p)/p^{11/2} and alpha_p * beta_p = 1 (after normalization). Ramanujan's conjecture |tau(p)| <= 2p^{11/2} is equivalent to |alpha_p| = |beta_p| = 1, which is the tempered condition on Pi_p.

What is an automorphic representation Pi in the adelic formulation?

An automorphic representation Pi is an irreducible constituent of L^2(GL_n(K)\GL_n(A_K)). It decomposes as Pi = restricted tensor product of Pi_v over all places v of K.

Newforms, Oldforms, and the Atkin-Lehner Theory

The space M_k(Gamma_0(N)) contains both 'old' forms (lifts from smaller level) and 'new' forms (genuinely of level N). The Atkin-Lehner theory separates them and identifies primitive forms - one for each isogeny class of elliptic curves in the modularity theorem.

In the modularity theorem, each isogeny class of elliptic curves over Q of conductor N corresponds to exactly one primitive form f in S_2^new(Gamma_0(N)). The Fourier coefficients a_p(f) equal the traces a_p(E) = p+1 - #E(F_p) for all but finitely many primes.

Level 37: the simplest case with two newforms

Multiple newforms at the same level

The space S_2^new(Gamma_0(37)) has dimension 2, spanned by two newforms f_1 and f_2. These correspond to two non-isogenous elliptic curves: 37a (rank 0) and 37b (rank 1). The form f_1 has sign epsilon = +1 (even analytic rank), while f_2 has epsilon = -1 (odd analytic rank). This is consistent with BSD: L(E_{37b}, 1) = 0 since the curve has a rational point of infinite order.

What is a primitive modular form?

Primitive form: newform with a_1 = 1, simultaneous eigenfunction of all T_p. Corresponds bijectively to a primitive automorphic representation of GL_2 of conductor N.

Maass Forms and the Spectrum

Classical modular forms are holomorphic. But there exist real-analytic automorphic forms - Maass forms, discovered in 1949. They correspond to different automorphic representations of GL_2 with non-integer infinitesimal character, and are connected to the open Ramanujan conjecture.

Quantum chaos connects to Maass forms: the distribution of eigenvalues lambda_j of the Laplacian on a hyperbolic surface follows random matrix statistics (GUE conjecture). This is the quantum version of the fact that geodesics on hyperbolic surfaces are 'chaotic' in the sense of ergodic theory.

PropertyHolomorphic formsMaass forms
AnalyticityHolomorphicReal-analytic
Eigenfunction ofNot LaplacianLaplacian: Delta f = lambda f
Archimedean component Pi_infDiscrete seriesPrincipal series
Ramanujan conjectureProved (Deligne 1974)Open
q-expansionf = sum a_n q^nf = sum a_n W_s(y) e^{2pi i n x}

What is the fundamental difference between Maass forms and classical holomorphic modular forms?

Maass forms: real-analytic, not holomorphic. Instead of the Cauchy-Riemann condition, they satisfy Delta f = lambda f. Both are automorphic forms for GL_2 but with different archimedean components Pi_infty.

Connections to other topics

Automorphic forms are central objects in modern number theory.

  • Spectral theory — Related topic
  • Representation theory — Related topic
  • Algebraic geometry — Related topic

Итоги

  • Modular form of weight k: f((az+b)/(cz+d)) = (cz+d)^k f(z) for matrices in SL_2(Z)
  • Hecke operators T_p: eigenfunctions have multiplicative a_{mn} = a_m a_n for gcd(m,n)=1
  • Adelic formulation: Pi = tensor_v Pi_v - automorphic representation of GL_2(A_Q)
  • Atkin-Lehner: S_k(Gamma_0(N)) = S_k^old + S_k^new; primitive forms correspond to isogeny classes
  • Ramanujan conjecture: |a_p| <= 2p^{(k-1)/2} - proved for holomorphic forms (Deligne 1974), open for Maass forms

Вопросы для размышления

  • Why does the commutativity of the Hecke algebra force simultaneous diagonalizability of all T_p, and why does this imply multiplicativity of eigenvalues?
  • How does the adelic reformulation unify holomorphic modular forms and Maass forms into a single category of automorphic representations?
  • Why is the Ramanujan conjecture equivalent to temperedness of automorphic representations, and what makes holomorphic forms (Deligne) easier to handle than Maass forms?

Связанные уроки

  • nt-27 — Zeta functions of varieties - motivation for Hecke operators and L-factors
  • nt-26-langlands — Automorphic forms are the concrete right-hand side of the Langlands correspondence
  • nt-29 — Arithmetic of elliptic curves uses L-functions of automorphic forms via modularity
Automorphic Forms