Documentation

GQ2.DeepCount.Bounds

Head, tail, and assembly bounds for the deep count #

The tail survivor, the head bound, and their assembly into the structural inequality.

See GQ2.DeepCount for the paper-facing overview, source citations, and deviations.

The tail survivor: #Dc_{2e} ≥ 2 #

The graded squaring at i = e kills [−1] ≠ [1] (‖−1 − 1‖ = ‖2‖ = ‖π‖^e exactly), so it is NOT injective; on equal-card grs it is then not surjective, and any unit class outside its range is a NONZERO element of Dc_{2e}: were it zero, the Kummer kernel would make the unit a square with w ∈ U_e (the dichotomy), putting it back in the range.

theorem GQ2.neg_one_mem_depthUnits (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) {e : } (he : 2 = π ^ e) :
-1 depthUnits k π e

−1 is a depth-e unit: ‖−1 − 1‖ = ‖2‖ = ‖π‖^e.

theorem GQ2.neg_one_not_mem_depthUnits_succ (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) (hπ0 : π 0) (hπ1 : π < 1) {e : } (he : 2 = π ^ e) :
-1depthUnits k π (e + 1)

−1 is NOT a depth-(e+1) unit (‖π‖^e > ‖π‖^{e+1}).

theorem GQ2.exists_kummerDepth_ne_zero (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) [FiniteDimensional ℚ_[2] k] :
π k∀ (hπ0 : π 0) (hπ1 : π < 1), (∀ xk, x < 1x π)∀ {e : } (he : 2 = π ^ e) {f : } (hcard_e : Nat.card ((depthUnits k π e) (depthUnits k π (e + 1)).subgroupOf (depthUnits k π e)) = 2 ^ f) (hcard_2e : Nat.card ((depthUnits k π (2 * e)) (depthUnits k π (2 * e + 1)).subgroupOf (depthUnits k π (2 * e))) = 2 ^ f), ξkummerDepth k π (2 * e), ξ 0

The tail survivor: some depth-2e Kummer class is nonzero. The graded squaring at i = e has [−1] in its kernel but [−1] ≠ [1], so it is not injective, hence (equal-card grs) not surjective; a unit class [a] outside the range gives kummerClassK a ≠ 0 — were it zero, the Kummer kernel would write a = w² with w ∈ U_e (norm-one by taking norms; depth-e by the dichotomy), and [a] = grSq [w].

The head: #(M ⧸ Dc_1) ≤ 2 #

Two inputs: the level-0 collapse Dc_0 ≤ Dc_1 (the residue group U⁰/U¹ has ODD order 2^f − 1, so squaring is bijective on it — grSq at i = 0 is the squaring map of the gr-group itself), and the π-parity decomposition (every a ∈ k^× is u·π^m with u norm-one, by discreteness), which makes M ⧸ Dc_1 generated by the single 2-torsion class mk [π].

theorem GQ2.exists_nat_val (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) (hπk : π k) (hπ0 : π 0) (hπ1 : π < 1) (hπmax : xk, x < 1x π) {x : AlgebraicClosure ℚ_[2]} (hx : x k) (hx0 : x 0) (hx1 : x 1) :
∃ (m : ), x = π ^ m

The ℕ-valuation from discreteness: a nonzero integral k-element has norm an exact power of ‖π‖. Take the least m with ‖π‖^{m+1} < ‖x‖; then x/π^m is integral of norm > ‖π‖, hence norm one by hπ_max.

theorem GQ2.kummerDepth_zero_collapse (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) (hπle : π 1) {e : } (he : 2 = π ^ e) {f : } (hf_pos : 1 f) (hcard_0 : Nat.card ((depthUnits k π 0) (depthUnits k π 1).subgroupOf (depthUnits k π 0)) = 2 ^ f - 1) :
kummerDepth k π 0 kummerDepth k π 1

The level-0 collapse Dc_0 ≤ Dc_1: the residue group U⁰/U¹ has odd order 2^f − 1 (B13 card_gr_zero, as a hypothesis over the depthUnits 0-form), so squaring is bijective on it and every norm-one unit is a square times a principal unit.

theorem GQ2.kummerClassK_pow (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (a : (↥k)ˣ) (m : ) :
kummerClassK k (a ^ m) = m kummerClassK k a

kummerClassK of a power: [a^m] = m • [a].

theorem GQ2.nsmul_mod_two {G : Type u_1} [AddCommGroup G] (ξ : G) ( : ξ + ξ = 0) (m : ) :
m ξ = (m % 2) ξ

2-torsion nsmul reduction: m • ξ = (m % 2) • ξ.

noncomputable def GQ2.piUnit (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) (hπk : π k) (hπ0 : π 0) :
(↥k)ˣ

The uniformizer as a unit of k.

Equations
  • GQ2.piUnit k π hπk hπ0 = Units.mk0 π, hπk
Instances For
    theorem GQ2.card_quot_kummerDepth_one_le_two (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) [FiniteDimensional ℚ_[2] k] [Finite (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2))] (hπk : π k) (hπ0 : π 0) (hπ1 : π < 1) (hπmax : xk, x < 1x π) {e : } (he : 2 = π ^ e) {f : } (hf_pos : 1 f) (hcard_0 : Nat.card ((depthUnits k π 0) (depthUnits k π 1).subgroupOf (depthUnits k π 0)) = 2 ^ f - 1) :
    Nat.card (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2) kummerDepth k π 1) 2

    The head bound: M ⧸ Dc_1 is generated by the single 2-torsion class of the uniformizer (π-parity via the ℕ-valuation, unit part into Dc_1 by the level-0 collapse), so it has at most 2 elements.

    The assembly: #(M ⧸ Dc_{e+1}) ≤ #Dc_e #

    The paired descent: R(s) : #(M⧸Dc_{e+1})·#Dc_{e+1+s} ≤ #Dc_e·#(M⧸Dc_{e−s}) holds at s = 0 with equality (double Lagrange), and each step trades the level e−s−1 on the right for the level e+1+s on the left — same-parity levels summing to 2e, where the class-gr counts compare (= 2^f odd / = 1 ≤ even). At s = e−1 the head (≤ 2) and the tail survivor (≥ 2) close the inequality.

    theorem GQ2.card_kummerDepth_step (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) (hπ1 : π 1) (j : ) :
    Nat.card (kummerDepth k π j) = Nat.card ((kummerDepth k π j) (kummerDepth k π (j + 1)).addSubgroupOf (kummerDepth k π j)) * Nat.card (kummerDepth k π (j + 1))

    Lagrange step-down for the class filtration: #Dc_j = #(Dc_j/Dc_{j+1}) · #Dc_{j+1}.

    theorem GQ2.card_quot_kummerDepth_step (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) [Finite (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2))] (hπ1 : π 1) (j : ) :
    Nat.card (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2) kummerDepth k π (j + 1)) = Nat.card (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2) kummerDepth k π j) * Nat.card ((kummerDepth k π j) (kummerDepth k π (j + 1)).addSubgroupOf (kummerDepth k π j))

    Lagrange step-up for the ambient quotients: #(M⧸Dc_{j+1}) = #(M⧸Dc_j) · #(Dc_j/Dc_{j+1}).

    theorem GQ2.card_classGr_pair_le (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) [Finite (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2))] [FiniteDimensional ℚ_[2] k] (hπk : π k) (hπ0 : π 0) (hπ1 : π < 1) (hπmax : xk, x < 1x π) {e : } (he : 2 = π ^ e) (he_pos : 1 e) {f : } (hf_pos : 1 f) (hcard_gr : ∀ (i : ), 1 iNat.card ((depthUnits k π i) (depthUnits k π (i + 1)).subgroupOf (depthUnits k π i)) = 2 ^ f) {j : } (hj1 : 1 j) (hje : j e - 1) :
    Nat.card ((kummerDepth k π j) (kummerDepth k π (j + 1)).addSubgroupOf (kummerDepth k π j)) Nat.card ((kummerDepth k π (2 * e - j)) (kummerDepth k π (2 * e - j + 1)).addSubgroupOf (kummerDepth k π (2 * e - j)))

    The paired level comparison: for 1 ≤ j ≤ e − 1, the class-gr at j is at most the class-gr at 2e − j (same parity: odd levels are both 2^f, even levels collapse to 1).

    theorem GQ2.card_quot_deep_le_card_mid (k : IntermediateField ℚ_[2] (AlgebraicClosure ℚ_[2])) (π : AlgebraicClosure ℚ_[2]) [Finite (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2))] [FiniteDimensional ℚ_[2] k] (hπk : π k) (hπ0 : π 0) (hπ1 : π < 1) (hπmax : xk, x < 1x π) {e : } (he : 2 = π ^ e) (he_pos : 1 e) {f : } (hf_pos : 1 f) (hcard_zero : Nat.card ((normUnits k) (depthUnits k π 1).subgroupOf (normUnits k)) = 2 ^ f - 1) (hcard_gr : ∀ (i : ), 1 iNat.card ((depthUnits k π i) (depthUnits k π (i + 1)).subgroupOf (depthUnits k π i)) = 2 ^ f) :
    Nat.card (ContCoh.H1 (↥k.fixingSubgroup) (ZMod 2) kummerDepth k π (e + 1)) Nat.card (kummerDepth k π e)

    THE STRUCTURAL COUNT — the single remaining input of (H4)'s sharpness: #(M ⧸ Dc_{e+1}) ≤ #Dc_e, by the paired descent between the double-Lagrange identity at s = 0 and the head/tail comparison at s = e − 1.