Documentation

LeanPool.Chudnovsky.SingularModuli.Rationality

Rationality of j(τ₁₆₃): the three-prime argument (Phase C, §4.2) #

This file implements PhaseC-PLAN.md §4.2, the "three-prime argument" proving that j₀ := j τ₁₆₃ is rational (hence, with statement 1, an integer).

The top-level target is

theorem j_τ₁₆₃_rational_of (hsurj : Function.Surjective (j : ℍ → ℂ))
    (hint : IsIntegral ℤ (j τ₁₆₃)) : ∃ r : ℚ, j τ₁₆₃ = (r : ℂ)

gated on surjectivity of j (proved in parallel in Valence.lean) and on integrality of j τ₁₆₃ (statement 1, proved in parallel).

Structure of the proof (steps (C1)(C7) of the plan) #

Let j₀ := j τ₁₆₃ and x an arbitrary complex root of minpoly ℚ j₀.

(C1)/(C2) The diagonal polynomial G_m(X) = Φ_m(X, X) ∈ ℚ[X] and its evaluation #

The diagonal polynomial G_m(X) := Φ_m(X, X) ∈ ℚ[X], obtained by substituting the outer variable X for the coefficient variable Y (the -analogue of diagPhiZ).

Equations
Instances For

    The diagonal substitution commutes with Y ↦ z: G_m(z) = Φ_m(z, z) for z : ℂ, where the left side is aeval z (diagPhiQ Φ_m) and the right side is (specializeY z Φ_m).eval z.

    theorem Chudnovsky.diagQ_eval_eq_prod {m : } [NeZero m] {PhiQ : Polynomial (Polynomial )} (hPhiQ : ∀ (τ : UpperHalfPlane), orbitPoly m τ = (specializeY (j τ)) PhiQ) (τ : UpperHalfPlane) :
    (Polynomial.aeval (j τ)) (diagPhiQ PhiQ) = i : Option (ZMod m), (j τ - f m i τ)

    G_m(j τ) = ∏_i (j τ − f m i τ): evaluating the diagonal polynomial at j τ recovers the orbit product.

    (C2, input) The CM coset relation ∃ i, f m i τ₁₆₃ = j τ₁₆₃ #

    The CM matrix M_n = [[n+1, −41],[1, n]] fixes τ₁₆₃; it decomposes as γ · Acol m n with γ = [[n+1, −1],[1, 0]] ∈ SL(2,ℤ), so j (Acol m n • τ₁₆₃) = j τ₁₆₃.

    noncomputable def Chudnovsky.cmGL (n : ) :
    GL (Fin 2)

    The CM matrix M_n = [[n+1, −41],[1, n]] as an element of GL (Fin 2) ℝ (determinant n² + n + 41 > 0).

    Equations
    Instances For
      @[simp]
      theorem Chudnovsky.val_cmGL (n : ) :
      (cmGL n) = !![n + 1, -41; 1, n]

      M_n fixes τ₁₆₃ as a Möbius transformation.

      The SL(2,ℤ) cofactor γ = [[n+1, −1],[1, 0]] (determinant 1).

      Equations
      Instances For

        Coset decomposition M_n = γ · Acol m n (as GL (Fin 2) ℝ), valid when m = n² + n + 41.

        theorem Chudnovsky.cm_coset_val {m : } (n : ) [NeZero m] (hm : m = n ^ 2 + n + 41) :

        The b-coset value at τ₁₆₃. j (Acol m n • τ₁₆₃) = j τ₁₆₃.

        theorem Chudnovsky.cm_coset_rel {m : } (n : ) [NeZero m] (hm : m = n ^ 2 + n + 41) :
        ∃ (i : Option (ZMod m)), f m i τ₁₆₃ = j τ₁₆₃

        (C2, input): the CM coset relation. For m = n² + n + 41, some coset value coincides with j τ₁₆₃: f m (some (n : ZMod m)) τ₁₆₃ = j τ₁₆₃.

        (C4)/(C5) Form-theory helpers (binary quadratic forms of a CM point) #

        Added to the Chudnovsky.QF namespace of QuadraticPoints.lean / FormReduction.lean. They provide: a fixing integer matrix of a point yields a definite integer form; two forms with a common root are proportional; a primitive representative exists; and the mod-4 constraint disc ≤ −3.

        theorem Chudnovsky.QF.disc_neg_of_isRoot {F : BQF} {τ : UpperHalfPlane} (ha : F.a 0) (h : IsRoot F τ) :
        disc F < 0

        A binary quadratic form with a ≠ 0 whose root lies in (a non-real number) has negative discriminant.

        theorem Chudnovsky.QF.root_prop {F G : BQF} {τ : UpperHalfPlane} (hF : IsRoot F τ) (hG : IsRoot G τ) :
        F.a * G.b = G.a * F.b F.a * G.c = G.a * F.c

        Two forms with a common root τ ∈ ℍ are proportional: F.a·G.b = G.a·F.b and F.a·G.c = G.a·F.c.

        theorem Chudnovsky.QF.exists_int_mult {F G : BQF} {τ : UpperHalfPlane} (hFp : IsPrimitive F) (hFa : F.a 0) (hF : IsRoot F τ) (hG : IsRoot G τ) :
        ∃ (l : ), G.a = l * F.a G.b = l * F.b G.c = l * F.c

        If F is primitive with a root τ, any form G with the same root is an integer multiple of F.

        theorem Chudnovsky.QF.exists_primitive_isRoot {F : BQF} {τ : UpperHalfPlane} (hpd : IsPosDef F) (hroot : IsRoot F τ) :
        ∃ (F₀ : BQF), IsPrimitive F₀ IsPosDef F₀ IsRoot F₀ τ

        A positive-definite form has a primitive positive-definite representative with the same root.

        theorem Chudnovsky.QF.disc_le_neg_three {F : BQF} (hpd : IsPosDef F) :
        disc F -3

        A positive-definite form has discriminant ≤ −3 (its discriminant is ≡ 0 or 1 mod 4 and negative).

        (C4) From a coset coincidence to a fixing integer matrix and a definite form #

        theorem Chudnovsky.fixes_of_coset {e f g : } (γ : Matrix.SpecialLinearGroup (Fin 2) ) {τ' : UpperHalfPlane} (hW : = (e * τ' + f) / g) (hg : g 0) (hfix : γ = τ') :
        ∃ (p : ) (q : ) (r : ) (s : ), QF.Fixes p q r s τ' p * s - q * r = e * g

        Fixing matrix from a coset factorization. If γ ∈ SL(2,ℤ) sends the coset point (with Aτ = (e·τ' + f)/g) to τ', then the integer matrix γ · [[e,f],[0,g]] fixes τ'; its entries give a QF.Fixes relation of determinant e·g.

        theorem Chudnovsky.exists_fixes_of_coincidence {m : } [Fact (Nat.Prime m)] {τ' : UpperHalfPlane} {i : Option (ZMod m)} (h : f m i τ' = j τ') :
        ∃ (p : ) (q : ) (r : ) (s : ), QF.Fixes p q r s τ' p * s - q * r = m

        (C4). A coset coincidence f m i τ′ = j τ′ yields integers p q r s with a QF.Fixes relation of determinant m.

        theorem Chudnovsky.exists_form_of_coincidence {m : } [Fact (Nat.Prime m)] {τ' : UpperHalfPlane} {i : Option (ZMod m)} (h : f m i τ' = j τ') :
        ∃ (F : QF.BQF) (t : ), QF.IsPosDef F QF.IsRoot F τ' QF.disc F = t ^ 2 - 4 * m

        (C4), packaged. A coset coincidence at τ′ produces a positive-definite integer form rooted at τ′ of discriminant t² − 4m.

        (C5)/(C6) The three-prime bridge: τ′ is Γ-equivalent to τ₁₆₃ #

        theorem Chudnovsky.cm_bridge {τ' : UpperHalfPlane} (h41 : ∃ (i : Option (ZMod 41)), f 41 i τ' = j τ') (h43 : ∃ (i : Option (ZMod 43)), f 43 i τ' = j τ') (h61 : ∃ (i : Option (ZMod 61)), f 61 i τ' = j τ') :

        (C5)+(C6). If τ′ carries a coset coincidence for each of m ∈ {41, 43, 61}, then τ′ is SL(2,ℤ)-equivalent to τ₁₆₃.

        (C1)–(C3) Vanishing of Φ_m(x, x) at roots, and the coincidence at τ′ #

        theorem Chudnovsky.phi_coincidence (n : ) (mm : ) [NeZero mm] [Fact (Nat.Prime mm)] (hm : mm = n ^ 2 + n + 41) {x : } (hx : (Polynomial.aeval x) (minpoly (j τ₁₆₃)) = 0) {τ' : UpperHalfPlane} (hτ' : j τ' = x) :
        ∃ (i : Option (ZMod mm)), f mm i τ' = j τ'

        (C2)+(C3). For a root x of minpoly ℚ j₀ and τ′ with j τ′ = x, each prime m = n² + n + 41 yields a coset coincidence f m i τ′ = j τ′.

        (C7) Rationality of j τ₁₆₃ #

        The three-prime argument (§4.2). Assuming surjectivity of j and integrality of j τ₁₆₃, the value j τ₁₆₃ is rational.