Prove2Me
Navigate
MissionsFormalpediaBlogsUsersMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

MTT: p-adic measures, disk moments, and interpolation

Definition
MTT_Measures

by davidloeffler · Sep 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

modular-formsnumber-theoryp-adic-l-functions

Cp-valued bounded measures on the actual p-adic unit group, the ordinary-root predicate, algebraic signed symbols, disk moments from MTT (10.2), and the full interpolation predicate. The complex-to-p-adic comparison uses algebraic numbers under two fixed embeddings. Continuous test functions are quantified with exact pointwise specifications. Primitive conductor and critical-exponent bounds are enforced in the interpolation predicate.

Definition code
import Definitions.Def_MTT_Arithmetic
import Mathlib.NumberTheory.Padics.Complex
import Mathlib.NumberTheory.Padics.RingHoms
import Mathlib.NumberTheory.Padics.Measure.Basic

set_option autoImplicit false
noncomputable section
open scoped BigOperators
namespace MTT

variable {p : ℕ} [Fact p.Prime]

/-- A bounded Cp-valued measure on the actual compact group Zp^*. -/
abbrev UnitMeasure (p : ℕ) [Fact p.Prime] :=
  AbstractMeasure (ℤ_[p])ˣ ℂ_[p] ℂ_[p]

/-- The tautological coordinate, via Qp -> Cp. -/
def coordinate (x : (ℤ_[p])ˣ) : ℂ_[p] :=
  algebraMap ℚ_[p] ℂ_[p] (x.val : ℚ_[p])

/-- The ordinary root of MTT (10.1). -/
def IsOrdinaryRoot {N k : ℕ} {ι : Qbar →+* ℂ}
    (f : Eigenform N k ι) (ιp : Qbar →+* ℂ_[p]) (α : ℂ_[p]) : Prop :=
  ‖α‖ = 1 ∧ α ^ 2 - ιp (f.coeff p) * α +
    ιp (f.epsilon p) * (p : ℂ_[p]) ^ (k - 1) = 0

/-- Algebraic signed modular symbols, obtained by expanding (mX+a)^j. -/
def algebraicSymbol {k : ℕ} {ι : Qbar →+* ℂ} {f : UpperHalfPlane → ℂ}
    (P : Periods k ι f) (s : Bool) (j : ℕ) (a m : ℚ) : Qbar :=
  ∑ t ∈ Finset.range (j + 1),
    (j.choose t : Qbar) * (m : Qbar) ^ t * (a : Qbar) ^ (j - t) *
      P.value s t (-a / m)

/-- MTT (10.2), specialized to p-power disks of positive depth; polynomial X^j. -/
def diskMoment {N k : ℕ} {ι : Qbar →+* ℂ} (f : Eigenform N k ι)
    (ιp : Qbar →+* ℂ_[p]) (P : Periods k ι f.form) (α : ℂ_[p])
    (s : Bool) (j n : ℕ) (a : ℤ) : ℂ_[p] :=
  (α ^ n)⁻¹ * ιp (algebraicSymbol P s j a (p ^ n)) -
    (ιp (f.epsilon p) * (p : ℂ_[p]) ^ (k - 2) / α ^ (n + 1)) *
      ιp (algebraicSymbol P s j a (p ^ (n - 1)))

/-- Pointwise description of X^j on a residue disk and zero elsewhere. -/
def diskFunction (n : ℕ) (a : ℤ) (j : ℕ) (x : (ℤ_[p])ˣ) : ℂ_[p] :=
  if PadicInt.toZModPow n x.val = (a : ZMod (p ^ n)) then coordinate x ^ j else 0

/-- The measure realizes every critical polynomial moment on every unit disk. -/
def RealizesMoments {N k : ℕ} {ι : Qbar →+* ℂ} (f : Eigenform N k ι)
    (ιp : Qbar →+* ℂ_[p]) (P : Periods k ι f.form) (α : ℂ_[p])
    (s : Bool) (μ : UnitMeasure p) : Prop :=
  ∀ (n : ℕ), 0 < n → ∀ (a : ℤ), IsCoprime a (p : ℤ) →
    ∀ (j : ℕ), j ≤ k - 2 → ∃ g : C((ℤ_[p])ˣ, ℂ_[p]),
      (∀ x, g x = diskFunction n a j x) ∧ μ g = diskMoment f ιp P α s j n a

/-- The finite-order character times X^j, as a pointwise function. -/
def specialFunction (ιp : Qbar →+* ℂ_[p]) (n : ℕ)
    (χ : DirichletCharacter Qbar (p ^ n)) (j : ℕ) (x : (ℤ_[p])ˣ) : ℂ_[p] :=
  ιp (χ (PadicInt.toZModPow n x.val)) * coordinate x ^ j

/-- Period parity: chi(-1)(-1)^j, true for plus and false for minus. -/
def criticalSign {m : ℕ} (χ : DirichletCharacter Qbar m) (j : ℕ) : Bool :=
  by classical exact decide (χ (-1) * (-1 : Qbar) ^ j = 1)

/-- MTT I.§14 multiplier; characters are evaluated at their primitive modulus. -/
def eulerMultiplier {N k : ℕ} {ι : Qbar →+* ℂ} (f : Eigenform N k ι)
    (ιp : Qbar →+* ℂ_[p]) (α : ℂ_[p]) (n : ℕ)
    (χ : DirichletCharacter Qbar (p ^ n)) (j : ℕ) : ℂ_[p] :=
  (α ^ n)⁻¹ *
    (1 - ιp (χ⁻¹ (p : ZMod (p ^ n))) * ιp (f.epsilon p) *
      (p : ℂ_[p]) ^ (k - 2 - j) / α) *
    (1 - ιp (χ (p : ZMod (p ^ n))) * (p : ℂ_[p]) ^ j / α)

/-- The complex side after division by the appropriate period, using MTT normalization. -/
def normalizedCriticalValue {N k : ℕ} {ι : Qbar →+* ℂ} (f : Eigenform N k ι)
    (omega : Bool → ℂ) (n : ℕ) (χ : DirichletCharacter Qbar (p ^ n)) (j : ℕ) : ℂ :=
  ((p ^ n : ℕ) : ℂ) ^ (j + 1) * (j.factorial : ℂ) /
    ((-2 * Real.pi * Complex.I) ^ j * gaussSum ι (p ^ n) χ⁻¹ *
      omega (criticalSign χ j)) * criticalLValue ι f.form (p ^ n) χ j

/-- Interpolation with an explicit algebraic bridge between C and Cp.
The existential continuous function is forced pointwise to equal chi(x) x^j. -/
def Interpolates {N k : ℕ} {ι : Qbar →+* ℂ} (f : Eigenform N k ι)
    (ιp : Qbar →+* ℂ_[p]) (omega : Bool → ℂ) (α : ℂ_[p])
    (μ : UnitMeasure p) : Prop :=
  ∀ (n : ℕ) (χ : DirichletCharacter Qbar (p ^ n)), χ.IsPrimitive →
    ∀ (j : ℕ), j ≤ k - 2 → ∃ (g : C((ℤ_[p])ˣ, ℂ_[p])) (v : Qbar),
      (∀ x, g x = specialFunction ιp n χ j x) ∧
      ι v = normalizedCriticalValue f omega n χ j ∧
      μ g = eulerMultiplier f ιp α n χ j * ιp v

end MTT
Source
Mazur–Tate–Teitelbaum, On p-adic analogues of the conjectures of Birch and Swinnerton-Dyer, Invent. Math. 84 (1986), https://doi.org/10.1007/BF01388731; Chapter I, §§1–4, 7–14.
Read-back

What the Lean code literally says, in plain math · GPT-6 (Codex)

UnitMeasure. For every natural prime ppp (with a supplied proof of primality), this is the type of abstract Cp\mathbb C_pCp​-valued, Cp\mathbb C_pCp​-linear measures on the group Zp×\mathbb Z_p^\timesZp×​ of units in the ppp-adic integers: bounded continuous linear functionals on continuous Cp\mathbb C_pCp​-valued functions on that group. No positivity or probability normalization is imposed.

coordinate. For a natural prime ppp and a unit x∈Zp×x\in\mathbb Z_p^\timesx∈Zp×​, this is its underlying ppp-adic integer, included first in Qp\mathbb Q_pQp​ and then by the specified algebra map into Cp\mathbb C_pCp​; denote it by c(x)c(x)c(x).

IsOrdinaryRoot. Given a natural prime ppp, natural numbers N,kN,kN,k, a ring embedding ι:Q‾→C\iota:\overline{\mathbb Q}\to\mathbb Cι:Q​→C, an Eigenform datum fff of level NNN and weight kkk, a ring embedding ιp:Q‾→Cp\iota_p:\overline{\mathbb Q}\to\mathbb C_pιp​:Q​→Cp​, and α∈Cp\alpha\in\mathbb C_pα∈Cp​, this predicate is the conjunction ∥α∥=1\|\alpha\|=1∥α∥=1 and α2−ιp(ap)α+ιp(ε(p))pmax⁡(k−1,0)=0\alpha^2-\iota_p(a_p)\alpha+\iota_p(\varepsilon(p))p^{\max(k-1,0)}=0α2−ιp​(ap​)α+ιp​(ε(p))pmax(k−1,0)=0. Here the Eigenform datum consists of a cusp form FFF for the real matrix image of Γ1(N)\Gamma_1(N)Γ1​(N) of integer weight kkk, a Dirichlet character ε\varepsilonε modulo NNN with values in Q‾\overline{\mathbb Q}Q​, and coefficients an∈Q‾a_n\in\overline{\mathbb Q}an​∈Q​ whose images under ι\iotaι are the coefficients of the period-one qqq-expansion of FFF, with a1=1a_1=1a1​=1. It includes the law F(γz)=ι(ε(d))(cz+d)kF(z)F(\gamma z)=\iota(\varepsilon(d))(cz+d)^kF(z)F(γz)=ι(ε(d))(cz+d)kF(z) for every γ=(abcd)∈Γ0(N)\gamma=\left(\begin{smallmatrix}a&b\\c&d\end{smallmatrix}\right)\in\Gamma_0(N)γ=(ac​bd​)∈Γ0​(N) and every upper-half-plane point zzz, and, for every prime ℓ\ellℓ, the identity ℓ−1∑b=0ℓ−1F((z+b)/ℓ)+ι(ε(ℓ))ℓmax⁡(k−1,0)F(ℓz)=ι(aℓ)F(z)\ell^{-1}\sum_{b=0}^{\ell-1}F((z+b)/\ell)+\iota(\varepsilon(\ell))\ell^{\max(k-1,0)}F(\ell z)=\iota(a_\ell)F(z)ℓ−1∑b=0ℓ−1​F((z+b)/ℓ)+ι(ε(ℓ))ℓmax(k−1,0)F(ℓz)=ι(aℓ​)F(z). Neither NNN nor kkk has a positive lower bound in this declaration; there is no condition that ppp be coprime to NNN.

algebraicSymbol. For a natural number kkk, a ring embedding ι:Q‾→C\iota:\overline{\mathbb Q}\to\mathbb Cι:Q​→C, an arbitrary function FFF on the upper half-plane, a Periods datum PPP for (k,ι,F)(k,\iota,F)(k,ι,F), a Boolean sss, natural jjj, and rational a,ma,ma,m, the result is ∑t=0j(jt)mtaj−tV(s,t,−a/m)∈Q‾\sum_{t=0}^j\binom jt m^t a^{j-t}V(s,t,-a/m)\in\overline{\mathbb Q}∑t=0j​(tj​)mtaj−tV(s,t,−a/m)∈Q​. The datum PPP supplies nonzero complex numbers ωs\omega_sωs​ and values V(s,t,r)∈Q‾V(s,t,r)\in\overline{\mathbb Q}V(s,t,r)∈Q​ for every Boolean sss, natural ttt, and rational rrr. For t≤max⁡(k−2,0)t\le\max(k-2,0)t≤max(k−2,0) it requires ι(V(s,t,r))=(It(r)+σs(−1)tIt(−r))/(2ωs)\iota(V(s,t,r))=(I_t(r)+\sigma_s(-1)^tI_t(-r))/(2\omega_s)ι(V(s,t,r))=(It​(r)+σs​(−1)tIt​(−r))/(2ωs​), where σtrue=1\sigma_{\mathrm{true}}=1σtrue​=1, σfalse=−1\sigma_{\mathrm{false}}=-1σfalse​=−1, and It(r)=2π∫0∞F(r+iu)(r+iu)t duI_t(r)=2\pi\int_0^\infty F(r+iu)(r+iu)^t\,duIt​(r)=2π∫0∞​F(r+iu)(r+iu)tdu. The integer submodule spanned by all these values with t≤max⁡(k−2,0)t\le\max(k-2,0)t≤max(k−2,0) must be finitely generated. No comparison is imposed on larger ttt. No bound on jjj or nonzero condition on mmm is required here: m=0m=0m=0 makes −a/m=0-a/m=0−a/m=0 by totalized field division, and zeroth powers, including 000^000, are 111. The integrals are totalized Bochner integrals, equal to zero when the integrand is not integrable.

diskMoment. Given a natural prime ppp, natural N,kN,kN,k, embeddings ι:Q‾→C\iota:\overline{\mathbb Q}\to\mathbb Cι:Q​→C and ιp:Q‾→Cp\iota_p:\overline{\mathbb Q}\to\mathbb C_pιp​:Q​→Cp​, an Eigenform datum f=(F,ε,(ar))f=(F,\varepsilon,(a_r))f=(F,ε,(ar​)) of level NNN, weight kkk and embedding ι\iotaι, a Periods datum PPP for FFF, any α∈Cp\alpha\in\mathbb C_pα∈Cp​, Boolean sss, natural j,nj,nj,n, and integer aaa, this is α−nιp(S(s,j,a,pn))−ιp(ε(p))pmax⁡(k−2,0)(αn+1)−1ιp(S(s,j,a,pmax⁡(n−1,0)))\alpha^{-n}\iota_p(S(s,j,a,p^n))-\iota_p(\varepsilon(p))p^{\max(k-2,0)}(\alpha^{n+1})^{-1}\iota_p(S(s,j,a,p^{\max(n-1,0)}))α−nιp​(S(s,j,a,pn))−ιp​(ε(p))pmax(k−2,0)(αn+1)−1ιp​(S(s,j,a,pmax(n−1,0))), where α−n\alpha^{-n}α−n here means (αn)−1(\alpha^n)^{-1}(αn)−1 and S(s,j,a,m)=∑t=0j(jt)mtaj−tV(s,t,−a/m)S(s,j,a,m)=\sum_{t=0}^j\binom jt m^ta^{j-t}V(s,t,-a/m)S(s,j,a,m)=∑t=0j​(tj​)mtaj−tV(s,t,−a/m). The Eigenform datum requires the normalized algebraic period-one Fourier expansion of a Γ1(N)\Gamma_1(N)Γ1​(N) cusp form, the Γ0(N)\Gamma_0(N)Γ0​(N) transformation law with character ε(d)\varepsilon(d)ε(d), and the prime eigenrelations ℓ−1∑b<ℓF((z+b)/ℓ)+ι(ε(ℓ))ℓmax⁡(k−1,0)F(ℓz)=ι(aℓ)F(z)\ell^{-1}\sum_{b<\ell}F((z+b)/\ell)+\iota(\varepsilon(\ell))\ell^{\max(k-1,0)}F(\ell z)=\iota(a_\ell)F(z)ℓ−1∑b<ℓ​F((z+b)/ℓ)+ι(ε(ℓ))ℓmax(k−1,0)F(ℓz)=ι(aℓ​)F(z). The Periods datum supplies nonzero ωs\omega_sωs​ and algebraic V(s,t,r)V(s,t,r)V(s,t,r) with ι(V(s,t,r))=(It(r)+σs(−1)tIt(−r))/(2ωs)\iota(V(s,t,r))=(I_t(r)+\sigma_s(-1)^tI_t(-r))/(2\omega_s)ι(V(s,t,r))=(It​(r)+σs​(−1)tIt​(−r))/(2ωs​) for t≤max⁡(k−2,0)t\le\max(k-2,0)t≤max(k−2,0), where It(r)=2π∫0∞F(r+iu)(r+iu)tduI_t(r)=2\pi\int_0^\infty F(r+iu)(r+iu)^tduIt​(r)=2π∫0∞​F(r+iu)(r+iu)tdu and σs\sigma_sσs​ is 111 for true and −1-1−1 for false; those bounded-degree values span a finitely generated integer submodule. These are totalized Bochner integrals. This definition imposes no ordinary-root condition, no coprimality on aaa, no degree bound, and no positive-depth condition. In particular n=0n=0n=0 uses p0=1p^0=1p0=1 in both symbols, and α=0\alpha=0α=0 is allowed with 0−1=00^{-1}=00−1=0 and 00=10^0=100=1.

diskFunction. For a natural prime ppp, natural n,jn,jn,j, integer aaa, and unit x∈Zp×x\in\mathbb Z_p^\timesx∈Zp×​, this is c(x)jc(x)^jc(x)j when the underlying ppp-adic integer of xxx reduces to aaa modulo pnp^npn, and is zero otherwise; c(x)c(x)c(x) is its image under Zp↪Qp↪Cp\mathbb Z_p\hookrightarrow\mathbb Q_p\hookrightarrow\mathbb C_pZp​↪Qp​↪Cp​. At n=0n=0n=0 the modulus is one, so the condition holds for every xxx. At j=0j=0j=0 the function is the indicator of that residue class. There is no coprimality assumption on aaa; at positive depth a residue class divisible by ppp contains no such units.

RealizesMoments. For a natural prime ppp, natural N,kN,kN,k, embeddings ι:Q‾→C\iota:\overline{\mathbb Q}\to\mathbb Cι:Q​→C and ιp:Q‾→Cp\iota_p:\overline{\mathbb Q}\to\mathbb C_pιp​:Q​→Cp​, an Eigenform datum f=(F,ε,(ar))f=(F,\varepsilon,(a_r))f=(F,ε,(ar​)), a Periods datum (ω,V)(\omega,V)(ω,V) for FFF, any α∈Cp\alpha\in\mathbb C_pα∈Cp​, Boolean sss, and bounded Cp\mathbb C_pCp​-linear functional μ\muμ on continuous functions Zp×→Cp\mathbb Z_p^\times\to\mathbb C_pZp×​→Cp​, this predicate says: for every positive natural nnn, every integer aaa coprime to ppp, and every natural j≤max⁡(k−2,0)j\le\max(k-2,0)j≤max(k−2,0), there exists a continuous function ggg equal at every unit xxx to c(x)jc(x)^jc(x)j if x≡a(modpn)x\equiv a\pmod{p^n}x≡a(modpn) and zero otherwise, such that μ(g)=(αn)−1ιp(S(s,j,a,pn))−ιp(ε(p))pmax⁡(k−2,0)(αn+1)−1ιp(S(s,j,a,pn−1))\mu(g)=(\alpha^n)^{-1}\iota_p(S(s,j,a,p^n))-\iota_p(\varepsilon(p))p^{\max(k-2,0)}(\alpha^{n+1})^{-1}\iota_p(S(s,j,a,p^{n-1}))μ(g)=(αn)−1ιp​(S(s,j,a,pn))−ιp​(ε(p))pmax(k−2,0)(αn+1)−1ιp​(S(s,j,a,pn−1)). Here c(x)c(x)c(x) is the image of xxx in Cp\mathbb C_pCp​, and S(s,j,a,m)=∑t=0j(jt)mtaj−tV(s,t,−a/m)S(s,j,a,m)=\sum_{t=0}^j\binom jt m^ta^{j-t}V(s,t,-a/m)S(s,j,a,m)=∑t=0j​(tj​)mtaj−tV(s,t,−a/m). The Eigenform datum requires a Γ1(N)\Gamma_1(N)Γ1​(N) cusp form of weight kkk, Fourier coefficients ι(ar)\iota(a_r)ι(ar​) with a1=1a_1=1a1​=1, character law F(γz)=ι(ε(d))(cz+d)kF(z)F(\gamma z)=\iota(\varepsilon(d))(cz+d)^kF(z)F(γz)=ι(ε(d))(cz+d)kF(z) for γ∈Γ0(N)\gamma\in\Gamma_0(N)γ∈Γ0​(N), and prime eigenrelations ℓ−1∑b<ℓF((z+b)/ℓ)+ι(ε(ℓ))ℓmax⁡(k−1,0)F(ℓz)=ι(aℓ)F(z)\ell^{-1}\sum_{b<\ell}F((z+b)/\ell)+\iota(\varepsilon(\ell))\ell^{\max(k-1,0)}F(\ell z)=\iota(a_\ell)F(z)ℓ−1∑b<ℓ​F((z+b)/ℓ)+ι(ε(ℓ))ℓmax(k−1,0)F(ℓz)=ι(aℓ​)F(z). The Periods datum requires nonzero ωs\omega_sωs​ and ι(V(s,t,r))=(It(r)+σs(−1)tIt(−r))/(2ωs)\iota(V(s,t,r))=(I_t(r)+\sigma_s(-1)^tI_t(-r))/(2\omega_s)ι(V(s,t,r))=(It​(r)+σs​(−1)tIt​(−r))/(2ωs​) for t≤max⁡(k−2,0)t\le\max(k-2,0)t≤max(k−2,0), with It(r)=2π∫0∞F(r+iu)(r+iu)tduI_t(r)=2\pi\int_0^\infty F(r+iu)(r+iu)^tduIt​(r)=2π∫0∞​F(r+iu)(r+iu)tdu, σtrue=1\sigma_{\mathrm{true}}=1σtrue​=1, σfalse=−1\sigma_{\mathrm{false}}=-1σfalse​=−1, and a finitely generated integer span of those values; the integrals are totalized Bochner integrals. No ordinary-root or nonzero condition on α\alphaα occurs. Weights zero and one still require degree zero moments. The predicate concerns the supplied μ\muμ, rather than asserting that a measure exists.

specialFunction. Given a natural prime ppp, a ring embedding ιp:Q‾→Cp\iota_p:\overline{\mathbb Q}\to\mathbb C_pιp​:Q​→Cp​, natural n,jn,jn,j, a Q‾\overline{\mathbb Q}Q​-valued Dirichlet character χ\chiχ modulo pnp^npn, and a unit x∈Zp×x\in\mathbb Z_p^\timesx∈Zp×​, this is ιp(χ(x mod pn))c(x)j\iota_p(\chi(x\bmod p^n))c(x)^jιp​(χ(xmodpn))c(x)j, with c(x)c(x)c(x) the image of xxx in Cp\mathbb C_pCp​. There is no primitive-character assumption or positive-depth assumption; n=0n=0n=0 uses the modulus-one character and j=0j=0j=0 uses the zeroth power 111.

criticalSign. For any natural modulus mmm, any Q‾\overline{\mathbb Q}Q​-valued Dirichlet character χ\chiχ modulo mmm, and any natural jjj, this Boolean is true precisely when χ(−1)(−1)j=1\chi(-1)(-1)^j=1χ(−1)(−1)j=1 in Q‾\overline{\mathbb Q}Q​, and is false otherwise. This declaration does not require positive modulus or primitivity; it includes m=0m=0m=0, m=1m=1m=1, and j=0j=0j=0.

eulerMultiplier. For a natural prime ppp, natural N,kN,kN,k, embeddings ι:Q‾→C\iota:\overline{\mathbb Q}\to\mathbb Cι:Q​→C and ιp:Q‾→Cp\iota_p:\overline{\mathbb Q}\to\mathbb C_pιp​:Q​→Cp​, an Eigenform datum f=(F,ε,(ar))f=(F,\varepsilon,(a_r))f=(F,ε,(ar​)), any α∈Cp\alpha\in\mathbb C_pα∈Cp​, natural n,jn,jn,j, and a Dirichlet character χ\chiχ modulo pnp^npn valued in Q‾\overline{\mathbb Q}Q​, this is (αn)−1(1−ιp(χ−1(p mod pn))ιp(ε(p))pmax⁡(k−2−j,0)/α)(1−ιp(χ(p mod pn))pj/α)(\alpha^n)^{-1}(1-\iota_p(\chi^{-1}(p\bmod p^n))\iota_p(\varepsilon(p))p^{\max(k-2-j,0)}/\alpha)(1-\iota_p(\chi(p\bmod p^n))p^j/\alpha)(αn)−1(1−ιp​(χ−1(pmodpn))ιp​(ε(p))pmax(k−2−j,0)/α)(1−ιp​(χ(pmodpn))pj/α). The Eigenform datum consists of a weight-kkk Γ1(N)\Gamma_1(N)Γ1​(N) cusp form with normalized algebraic Fourier coefficients, the Γ0(N)\Gamma_0(N)Γ0​(N) character law F(γz)=ι(ε(d))(cz+d)kF(z)F(\gamma z)=\iota(\varepsilon(d))(cz+d)^kF(z)F(γz)=ι(ε(d))(cz+d)kF(z), and prime eigenrelations ℓ−1∑b<ℓF((z+b)/ℓ)+ι(ε(ℓ))ℓmax⁡(k−1,0)F(ℓz)=ι(aℓ)F(z)\ell^{-1}\sum_{b<\ell}F((z+b)/\ell)+\iota(\varepsilon(\ell))\ell^{\max(k-1,0)}F(\ell z)=\iota(a_\ell)F(z)ℓ−1∑b<ℓ​F((z+b)/ℓ)+ι(ε(ℓ))ℓmax(k−1,0)F(ℓz)=ι(aℓ​)F(z). The character is evaluated at the displayed modulus, with no requirement of primitivity and no change of modulus. For n>0n>0n>0, its value and the inverse character's value at ppp are zero, so the two parenthesized factors are one. For n=0n=0n=0 the modulus is one. Neither a bound on jjj nor an ordinary-root or nonzero condition on α\alphaα is imposed; inverse zero is zero, so α=0\alpha=0α=0 yields zero when n>0n>0n>0 and one when n=0n=0n=0.

normalizedCriticalValue. For a natural prime ppp, natural N,kN,kN,k, an embedding ι:Q‾→C\iota:\overline{\mathbb Q}\to\mathbb Cι:Q​→C, an Eigenform datum f=(F,ε,(ar))f=(F,\varepsilon,(a_r))f=(F,ε,(ar​)), an arbitrary function ω:{false,true}→C\omega:\{\mathrm{false},\mathrm{true}\}\to\mathbb Cω:{false,true}→C, natural n,jn,jn,j, and an algebraic-valued Dirichlet character χ\chiχ modulo m=pnm=p^nm=pn, this is mj+1j! L/((−2πi)jG(χ−1)ωs)m^{j+1}j!\,L/((-2\pi i)^jG(\chi^{-1})\omega_s)mj+1j!L/((−2πi)jG(χ−1)ωs​), where sss is true exactly when χ(−1)(−1)j=1\chi(-1)(-1)^j=1χ(−1)(−1)j=1, G(ψ)=∑a∈Z/mZι(ψ(a))exp⁡(2πia~/m)G(\psi)=\sum_{a\in\mathbb Z/m\mathbb Z}\iota(\psi(a))\exp(2\pi i\widetilde a/m)G(ψ)=∑a∈Z/mZ​ι(ψ(a))exp(2πia/m) with a~∈{0,…,m−1}\widetilde a\in\{0,\ldots,m-1\}a∈{0,…,m−1}, and L=(2π)j+1/j!∫0∞[G(χ)−1∑a∈Z/mZι(χ(a))F(it+a~/m)]tjdtL=(2\pi)^{j+1}/j!\int_0^\infty[G(\chi)^{-1}\sum_{a\in\mathbb Z/m\mathbb Z}\iota(\chi(a))F(it+\widetilde a/m)]t^jdtL=(2π)j+1/j!∫0∞​[G(χ)−1∑a∈Z/mZ​ι(χ(a))F(it+a/m)]tjdt. The Eigenform datum requires a weight-kkk Γ1(N)\Gamma_1(N)Γ1​(N) cusp form, Fourier coefficients ι(ar)\iota(a_r)ι(ar​) with a1=1a_1=1a1​=1, the Γ0(N)\Gamma_0(N)Γ0​(N) law with multiplier ι(ε(d))(cz+d)k\iota(\varepsilon(d))(cz+d)^kι(ε(d))(cz+d)k, and prime eigenrelations ℓ−1∑b<ℓF((z+b)/ℓ)+ι(ε(ℓ))ℓmax⁡(k−1,0)F(ℓz)=ι(aℓ)F(z)\ell^{-1}\sum_{b<\ell}F((z+b)/\ell)+\iota(\varepsilon(\ell))\ell^{\max(k-1,0)}F(\ell z)=\iota(a_\ell)F(z)ℓ−1∑b<ℓ​F((z+b)/ℓ)+ι(ε(ℓ))ℓmax(k−1,0)F(ℓz)=ι(aℓ​)F(z). There is no requirement that ω\omegaω be a period system or nonzero, that χ\chiχ be primitive, or that jjj be in a critical range. The modulus-one case n=0n=0n=0 is included. All divisions use inverse zero equal to zero, and the integral is the totalized Bochner integral, equal to zero for a nonintegrable integrand.

Interpolates. For a natural prime ppp, natural N,kN,kN,k, embeddings ι:Q‾→C\iota:\overline{\mathbb Q}\to\mathbb Cι:Q​→C and ιp:Q‾→Cp\iota_p:\overline{\mathbb Q}\to\mathbb C_pιp​:Q​→Cp​, an Eigenform datum f=(F,ε,(ar))f=(F,\varepsilon,(a_r))f=(F,ε,(ar​)), any function ω:{false,true}→C\omega:\{\mathrm{false},\mathrm{true}\}\to\mathbb Cω:{false,true}→C, any α∈Cp\alpha\in\mathbb C_pα∈Cp​, and a bounded Cp\mathbb C_pCp​-linear functional μ\muμ on continuous functions on Zp×\mathbb Z_p^\timesZp×​, this predicate says that for every natural nnn, every primitive algebraic-valued Dirichlet character χ\chiχ modulo m=pnm=p^nm=pn, and every natural j≤max⁡(k−2,0)j\le\max(k-2,0)j≤max(k−2,0), there exist a continuous function g:Zp×→Cpg:\mathbb Z_p^\times\to\mathbb C_pg:Zp×​→Cp​ and v∈Q‾v\in\overline{\mathbb Q}v∈Q​ such that g(x)=ιp(χ(x mod m))c(x)jg(x)=\iota_p(\chi(x\bmod m))c(x)^jg(x)=ιp​(χ(xmodm))c(x)j for every unit xxx, ι(v)=mj+1j!L/((−2πi)jG(χ−1)ωs)\iota(v)=m^{j+1}j!L/((-2\pi i)^jG(\chi^{-1})\omega_s)ι(v)=mj+1j!L/((−2πi)jG(χ−1)ωs​), and μ(g)=(αn)−1(1−ιp(χ−1(p mod m))ιp(ε(p))pmax⁡(k−2−j,0)/α)(1−ιp(χ(p mod m))pj/α)ιp(v)\mu(g)=(\alpha^n)^{-1}(1-\iota_p(\chi^{-1}(p\bmod m))\iota_p(\varepsilon(p))p^{\max(k-2-j,0)}/\alpha)(1-\iota_p(\chi(p\bmod m))p^j/\alpha)\iota_p(v)μ(g)=(αn)−1(1−ιp​(χ−1(pmodm))ιp​(ε(p))pmax(k−2−j,0)/α)(1−ιp​(χ(pmodm))pj/α)ιp​(v). Here c(x)c(x)c(x) is the image of xxx in Cp\mathbb C_pCp​, sss is true exactly when χ(−1)(−1)j=1\chi(-1)(-1)^j=1χ(−1)(−1)j=1, G(ψ)=∑a mod mι(ψ(a))exp⁡(2πia~/m)G(\psi)=\sum_{a\bmod m}\iota(\psi(a))\exp(2\pi i\widetilde a/m)G(ψ)=∑amodm​ι(ψ(a))exp(2πia/m), and L=(2π)j+1/j!∫0∞[G(χ)−1∑a mod mι(χ(a))F(it+a~/m)]tjdtL=(2\pi)^{j+1}/j!\int_0^\infty[G(\chi)^{-1}\sum_{a\bmod m}\iota(\chi(a))F(it+\widetilde a/m)]t^jdtL=(2π)j+1/j!∫0∞​[G(χ)−1∑amodm​ι(χ(a))F(it+a/m)]tjdt, with least nonnegative representatives a~\widetilde aa and totalized Bochner integration. The Eigenform datum requires a weight-kkk Γ1(N)\Gamma_1(N)Γ1​(N) cusp form with algebraic Fourier coefficients ι(ar)\iota(a_r)ι(ar​), a1=1a_1=1a1​=1, the law F(γz)=ι(ε(d))(cz+d)kF(z)F(\gamma z)=\iota(\varepsilon(d))(cz+d)^kF(z)F(γz)=ι(ε(d))(cz+d)kF(z) for γ∈Γ0(N)\gamma\in\Gamma_0(N)γ∈Γ0​(N), and prime eigenrelations ℓ−1∑b<ℓF((z+b)/ℓ)+ι(ε(ℓ))ℓmax⁡(k−1,0)F(ℓz)=ι(aℓ)F(z)\ell^{-1}\sum_{b<\ell}F((z+b)/\ell)+\iota(\varepsilon(\ell))\ell^{\max(k-1,0)}F(\ell z)=\iota(a_\ell)F(z)ℓ−1∑b<ℓ​F((z+b)/ℓ)+ι(ε(ℓ))ℓmax(k−1,0)F(ℓz)=ι(aℓ​)F(z). The predicate imposes neither a period-system condition on ω\omegaω nor an ordinary-root condition on α\alphaα; both may vanish, and divisions by zero are zero. It includes n=0n=0n=0 and the primitive modulus-one character, and includes degree zero even for k=0k=0k=0 or 111. At positive depth the character values at ppp vanish. The existential algebraic vvv requires the displayed normalized complex expression to lie in the image of ι\iotaι. The statement concerns a supplied μ\muμ, without asserting existence of such a measure, and imposes no condition for nonprimitive characters.

Human review
  • Endorsed by Shuze Chen · Sep 5, 2026

  • Endorsed by davidloeffler · Sep 5, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactJoin Slack© 2026 Prove2Me