MTT: cusp forms, modular integrals, periods, and complex critical values
DefinitionMTT_ArithmeticAnalytic cusp forms with explicit nebentypus and normalized Hecke eigenform conditions; actual modular integrals and inverse-character Mellin transforms; and signed algebraic period data with a finitely generated integral lattice. Existence of this period system is a separate theorem target. The signed projections include one half.
import Mathlib.NumberTheory.ModularForms.QExpansion
import Mathlib.NumberTheory.ModularForms.CongruenceSubgroups
import Mathlib.NumberTheory.DirichletCharacter.Basic
import Mathlib.MeasureTheory.Integral.Bochner.Set
import Mathlib.RingTheory.Finiteness.Defs
import Mathlib.FieldTheory.IsAlgClosed.AlgebraicClosure
set_option autoImplicit false
noncomputable section
open scoped BigOperators
open MeasureTheory
namespace MTT
abbrev Qbar := AlgebraicClosure ℚ
abbrev GammaOne (N : ℕ) :=
(CongruenceSubgroup.Gamma1 N).map (Matrix.SpecialLinearGroup.mapGL ℝ)
/-- The classical prime Hecke operator, including U_l when l divides the level. -/
def heckePrime (k : ℕ) (εl : ℂ) (l : ℕ) (f : UpperHalfPlane → ℂ)
(z : UpperHalfPlane) : ℂ :=
(l : ℂ)⁻¹ * ∑ b : Fin l, f (UpperHalfPlane.ofComplex ((z + (b.val : ℂ)) / l)) +
εl * (l : ℂ) ^ (k - 1) * f (UpperHalfPlane.ofComplex ((l : ℂ) * z))
/-- Normalized algebraic cuspidal Hecke eigenforms, with the nebentypus law explicit. -/
structure Eigenform (N k : ℕ) (ι : Qbar →+* ℂ) where
form : CuspForm (GammaOne N) (k : ℤ)
epsilon : DirichletCharacter Qbar N
coeff : ℕ → Qbar
coeff_eq : ∀ n, (UpperHalfPlane.qExpansion 1 form).coeff n = ι (coeff n)
normalized : coeff 1 = 1
character_law : ∀ γ : CongruenceSubgroup.Gamma0 N, ∀ z : UpperHalfPlane,
form ((Matrix.SpecialLinearGroup.mapGL ℝ γ.val) • z) =
ι (epsilon (γ.val 1 1 : ZMod N)) *
(((γ.val 1 0 : ℤ) : ℂ) * z + ((γ.val 1 1 : ℤ) : ℂ)) ^ k * form z
eigen : ∀ l : ℕ, l.Prime → ∀ z : UpperHalfPlane,
heckePrime k (ι (epsilon l)) l form z = ι (coeff l) * form z
/-- MTT I.(1.2), with the polynomial specified by its coefficients. -/
def modularIntegral (f : UpperHalfPlane → ℂ) (P : Polynomial ℂ) (r : ℚ) : ℂ :=
(2 * Real.pi : ℂ) * ∫ t in Set.Ioi (0 : ℝ),
f (UpperHalfPlane.ofComplex ((r : ℂ) + Complex.I * t)) *
P.eval ((r : ℂ) + Complex.I * t)
/-- MTT I.§3(i), for P(X)=X^j. -/
def modularSymbol (f : UpperHalfPlane → ℂ) (j : ℕ) (a m : ℚ) : ℂ :=
modularIntegral f (((m : ℂ) • Polynomial.X + Polynomial.C (a : ℂ)) ^ j) (-a / m)
/-- Boolean true denotes the plus eigenspace, false the minus eigenspace. -/
def sign (s : Bool) : ℤ := if s then 1 else -1
/-- The involution sends (P,r) to (P(-X),-r); the projection contains 1/2. -/
def signedIntegral (f : UpperHalfPlane → ℂ) (s : Bool) (j : ℕ) (r : ℚ) : ℂ :=
(modularIntegral f (Polynomial.X ^ j) r +
(sign s : ℂ) * (-1 : ℂ) ^ j * modularIntegral f (Polynomial.X ^ j) (-r)) / 2
/-- A period system, including a finite integral lattice of normalized signed values. -/
structure Periods (k : ℕ) (ι : Qbar →+* ℂ) (f : UpperHalfPlane → ℂ) where
omega : Bool → ℂ
omega_ne : ∀ s, omega s ≠ 0
value : Bool → ℕ → ℚ → Qbar
comparison : ∀ s j r, j ≤ k - 2 →
ι (value s j r) = signedIntegral f s j r / omega s
lattice_fg : (Submodule.span ℤ {v : Qbar | ∃ s j r,
j ≤ k - 2 ∧ v = value s j r}).FG
/-- Gauss sum with the positive complex exponential, including modulus one. -/
def gaussSum (ι : Qbar →+* ℂ) (m : ℕ) [NeZero m]
(χ : DirichletCharacter Qbar m) : ℂ :=
∑ a : ZMod m, ι (χ a) * Complex.exp (2 * Real.pi * Complex.I * a.val / m)
/-- The inverse-character twist: Birch's finite-translate expression f_{χ^{-1}}. -/
def inverseTwist (ι : Qbar →+* ℂ) (f : UpperHalfPlane → ℂ)
(m : ℕ) [NeZero m] (χ : DirichletCharacter Qbar m)
(z : UpperHalfPlane) : ℂ :=
(gaussSum ι m χ)⁻¹ * ∑ a : ZMod m,
ι (χ a) * f (UpperHalfPlane.ofComplex (z + (a.val : ℂ) / m))
/-- The actual complex critical L-value L(f_{χ^{-1}},j+1), defined by its Mellin integral. -/
def criticalLValue (ι : Qbar →+* ℂ) (f : UpperHalfPlane → ℂ)
(m : ℕ) [NeZero m] (χ : DirichletCharacter Qbar m) (j : ℕ) : ℂ :=
(2 * Real.pi : ℂ) ^ (j + 1) / (j.factorial : ℂ) *
∫ t in Set.Ioi (0 : ℝ),
inverseTwist ι f m χ (UpperHalfPlane.ofComplex (Complex.I * t)) * (t : ℂ) ^ j
end MTT
Read-back
What the Lean code literally says, in plain math · GPT-6 (Codex)
Qbar. This abbreviation denotes the chosen algebraic closure of the rational field, written ; it has no parameters and does not itself choose an embedding into .
GammaOne. For every natural number , this is the subgroup of obtained by mapping the integer determinant-one matrices satisfying and in into real invertible matrices. There is no positivity assumption on : for these residue conditions are the integer equalities , and for all determinant-one integer matrices satisfy them.
heckePrime. For natural numbers , an arbitrary complex number , an arbitrary function , and , this defines , with all scalar arithmetic in . Here is the upper half-plane and is viewed as an upper-half-plane point when , and otherwise a fixed arbitrarily chosen upper-half-plane point. The definition assumes neither primality nor positivity of , and no regularity of . For the sum is empty and complex inversion of zero is zero; the remaining term is if or , and zero if , since .
Eigenform. For arbitrary natural numbers and a unital ring homomorphism , this structure consists of a weight- holomorphic cusp form for the real image of the determinant-one integer matrices with modulo , a Dirichlet character (extended by zero on nonunits), and coefficients for every natural number . Its required identities are that the th coefficient of the width-one -expansion of is for every , that , that for every determinant-one integer matrix with modulo and every , and that for every natural prime and every . A cusp form here is invariant under the weight- slash action, holomorphic on , and vanishes at every cusp of that group; the width-one expansion uses the Taylor coefficients at zero of its cusp function in the parameter . No lower bounds on or are imposed, including and ; the declaration specifies data and conditions without asserting existence for every choice of parameters.
modularIntegral. For every function , complex polynomial , and rational number , this defines the complex number , where the integral is the Bochner integral with respect to real Lebesgue measure restricted to the positive half-line. The point is represented by the upper-half-plane conversion, which agrees with that point for all in the integration domain. No measurability, regularity, or integrability hypothesis is imposed; if the integrand is not integrable, the totalized Bochner integral is zero.
modularSymbol. For every function , natural number , and rationals , this defines with computed in , and real Lebesgue measure. There are no integrability assumptions, and the integral is zero when the integrand is not integrable. The parameters include , when the polynomial is identically one, and , when rational division makes and the polynomial is the constant (including ).
sign. For a Boolean parameter , this defines the integer when is true and the integer when is false.
signedIntegral. For every function , Boolean , natural number , and rational , this defines , where for true and for false. Both integrals use real Lebesgue measure and each is independently zero if its integrand is not integrable; no analytic hypothesis on is required. The definition includes and ; at it equals the single integral multiplied by .
Periods. For a natural number , a unital ring homomorphism , and an arbitrary function , this structure supplies complex numbers for each Boolean and algebraic numbers for every Boolean , natural number , and rational . For every such triple with it requires , where is for true and for false, and each Lebesgue Bochner integral is zero if its integrand is not integrable. It also requires that the -submodule of spanned by all satisfying that same bound on be finitely generated. No other comparison or finite-generation constraint is imposed on values outside that bound. In particular, for the bounded indices consist of , rather than an empty range; no existence assertion or analytic assumption on is part of this structure declaration.
gaussSum. For a unital ring homomorphism , a natural number with a supplied assumption , and a Dirichlet character modulo taking values in and extended by zero on nonunits, this defines , where is the representative in . No primitivity or nonvanishing of this sum is required. Modulus one is included and its sole term gives the value .
inverseTwist. For a unital ring homomorphism , an arbitrary function , a natural modulus with the assumption , a Dirichlet character modulo valued in and extended by zero on nonunits, and , this defines , where and is the standard representative. The translated arguments have positive imaginary part. There is no primitivity assumption and no assumption ; if the definition is zero by totalized complex inversion, while for it equals .
criticalLValue. For a unital ring homomorphism , an arbitrary function , a natural modulus with the assumption , a Dirichlet character modulo valued in and extended by zero on nonunits, and any natural number , this defines , where and is the representative in . The integral is the complex Bochner integral for real Lebesgue measure on the positive half-line and is zero if its integrand is not integrable. No cusp-form, weight, primitivity, convergence, or upper-bound-on- assumption is imposed. This includes and ; is always nonzero, and if the twist and thus this value are zero by totalized inversion.
Confirmed by the mission captain (proposal self-audit).