Documentation

TauCeti.Topology.Algebra.Group.Profinite.Free.SuccessiveApproximation.Tail

Successive approximation of a relator carrying the p-power tails #

Let F = freeProP p X be the free pro-p group on a finite linearly ordered type X, with lower p-series λ_k = λ_k(F), and let r, w ∈ λ_1(F) be relators with the same class ρ ∈ gr_1(F). When the basis-modification map δ_ρ is not onto gr_{m+1}(F), the successive-approximation argument of TauCeti.Topology.Algebra.Group.Profinite.Free.SuccessiveApproximation.Basic still runs as soon as the span statement holds up to the tail of a set S of generators, the span TauCeti.freeProP.gradedPowIterSpan of the iterated p-powers π^{m+1} ξ_i of the generators x_i with i ∈ S:

gr_{m+1}(F) = Im δ_ρ + ⟨π^{m+1} ξ_i : i ∈ S⟩ for every m ≥ 1.

The set S is a parameter: for the odd-rank dyadic relators it is the set of generators whose coefficient c_i in ρ vanishes, so that the span is the tail T_{m+1}(ρ) of TauCeti.freeProP.basisModificationTail, while for the even-rank dyadic relators it also contains the generator carrying the square. At each level the discrepancy between the moved relator and the target is the class of a basis modification up to a tail class Σ_{i ∈ S} c_i π^{m+1} ξ_i, which is the class of the product of the powers x_i^{p^{m+1} c_i}. The basis modification is applied and the powers are absorbed into per-generator tail elements t_i, which are carried along instead of being killed: the approximation at level m is a congruence φ r ≡ t_{i₁} ⋯ t_{i_a} * w * t_{j₁} ⋯ t_{j_b} modulo λ_{m+2}(F), where the two lists of indices fix where the tail of each generator is placed, and the tail elements lie in the closed procyclic subgroups ⟨x_i⟩ and in λ_2(F). The tail of each generator is placed at a single position, so the lists are required to be disjoint and duplicate-free and to cover the set S.

The limit is taken through the levelwise comparison schema TauCeti.PLowerCentralSeriesComparison with the finite tail data carried in the comparison data, and the tail elements are recovered from the compatible sequence of their classes by the inverse-limit description of F. Since the closed subgroups ⟨x_i⟩ ∩ λ_2(F) are detected on the finite quotients (TauCeti.IsProP.mem_of_forall_mk_mem_map_pLowerCentralSeries), the limits stay in them. The conclusion is an exact equation e r = t_{i₁} ⋯ t_{i_a} * w * t_{j₁} ⋯ t_{j_b} for a continuous automorphism e of F.

This is the first half of Labute's treatment of the relators with q = 2. For the odd-rank dyadic normal-form word x₁² (x₂, x₃) ⋯ (x_{n-1}, x_n) the tails are the 2-powers x_i^{2^{m+1}} of the generators x₂, …, x_n other than the one carrying the square, and the argument yields the relator in the intermediate form x₁² r₀(x) x₂^{α₂} ⋯ x_n^{α_n} with 2-adic exponents α₂, …, α_n divisible by 4 (TauCeti.Topology.Algebra.Group.Profinite.Demushkin.NormalForm.Two.Odd.Approximation), before the tail relator r₀(x) x₂^{α₂} ⋯ x_n^{α_n} is normalised in its own right. For the even-rank word x₁² (x₁, x₂) (x₃, x₄) ⋯ (x_{n-1}, x_n) the tails are the 2-powers of every generator but x₂, and the tail of x₁ is placed in front of the word, which is why the theorem takes two lists of tail positions; the intermediate form is x₁^{2+α} (x₁, x₂) r₀(x) x₃^{α₃} ⋯ x_n^{α_n} (TauCeti.Topology.Algebra.Group.Profinite.Demushkin.NormalForm.Two.Even.Approximation).

Main results #

References #

theorem TauCeti.freeProP.exists_continuousMonoidHom_inv_mul_apply_mem_of_range_sup_gradedPowIterSpan_eq_top {p : ℕ} [Fact (Nat.Prime p)] {X : Type u} [Finite X] [LinearOrder X] (r w : ↥(pLowerCentralSeries p (freeProP p X) 1)) (h : gradedMk p (freeProP p X) 1 r = gradedMk p (freeProP p X) 1 w) (S : Set X) (l₁ l₂ : List X) (hnd : (l₁ ++ l₂).Nodup) (hl : ∀ i ∈ S, i ∈ l₁ ++ l₂) (k : ℕ) (hspan : ∀ (m : ℕ) (hm : 1 ≤ m), m ≤ k → ((basisModificationDelta p X hm) (gradedMk p (freeProP p X) 1 r)).range ⊔ gradedPowIterSpan p X S (m + 1) = ⊤) :
∃ (φ : freeProP p X →ₜ* freeProP p X) (t : X → freeProP p X), (∀ (g : freeProP p X), g⁻¹ * φ g ∈ pLowerCentralSeries p (freeProP p X) 1) ∧ (∀ (i : X), t i ∈ (Subgroup.closure {of i}).topologicalClosure) ∧ (∀ (i : X), t i ∈ pLowerCentralSeries p (freeProP p X) 2) ∧ (φ ↑r)⁻¹ * ((List.map t l₁).prod * ↑w * (List.map t l₂).prod) ∈ pLowerCentralSeries p (freeProP p X) (k + 2)

Finite successive approximation with tails. Let r, w ∈ λ_1(F) have the same class ρ ∈ gr_1(F), let S be a set of generators and l₁, l₂ disjoint duplicate-free lists of generators containing S, and suppose gr_{m+1}(F) = Im δ_ρ + ⟨π^{m+1} ξ_i : i ∈ S⟩ for 1 ≤ m ≤ k. Then there are a continuous endomorphism φ of F, congruent to the identity modulo λ_1(F), and tail elements t_i ∈ ⟨x_i⟩ ∩ λ_2(F) with φ r ≡ (∏_{i ∈ l₁} t_i) * w * (∏_{i ∈ l₂} t_i) mod λ_{k+2}(F).

theorem TauCeti.freeProP.exists_continuousMulEquiv_apply_eq_of_range_sup_gradedPowIterSpan_eq_top {p : ℕ} [Fact (Nat.Prime p)] {X : Type u} [Finite X] [LinearOrder X] (r w : ↥(pLowerCentralSeries p (freeProP p X) 1)) (h : gradedMk p (freeProP p X) 1 r = gradedMk p (freeProP p X) 1 w) (S : Set X) (l₁ l₂ : List X) (hnd : (l₁ ++ l₂).Nodup) (hl : ∀ i ∈ S, i ∈ l₁ ++ l₂) (hspan : ∀ (m : ℕ) (hm : 1 ≤ m), ((basisModificationDelta p X hm) (gradedMk p (freeProP p X) 1 r)).range ⊔ gradedPowIterSpan p X S (m + 1) = ⊤) :
∃ (e : freeProP p X ≃ₜ* freeProP p X) (t : X → freeProP p X), (∀ (i : X), t i ∈ (Subgroup.closure {of i}).topologicalClosure) ∧ (∀ (i : X), t i ∈ pLowerCentralSeries p (freeProP p X) 2) ∧ e ↑r = (List.map t l₁).prod * ↑w * (List.map t l₂).prod

The successive-approximation theorem with tails. Let r, w ∈ λ_1(F) be relators of the free pro-p group F on a finite linearly ordered type with the same class ρ ∈ gr_1(F), let S be a set of generators and l₁, l₂ disjoint duplicate-free lists of generators containing S, and suppose gr_{m+1}(F) = Im δ_ρ + ⟨π^{m+1} ξ_i : i ∈ S⟩ for every m ≥ 1. Then a continuous automorphism e of F carries r to (∏_{i ∈ l₁} t_i) * w * (∏_{i ∈ l₂} t_i) for tail elements t_i of the closed procyclic subgroups ⟨x_i⟩, all lying in λ_2(F).