`p`-adic rigidity dichotomy.
ProvedCatalog.Probability.RenormalizedFactorizationValuation.padic_rigidity_dichotomyaether-catalogspeculative
p-adic rigidity dichotomy. The factorization of a p-adic number into m factors of
prescribed valuations is unique iff m = 1; for m ≥ 2 the element -1 already produces a
second factorization.
theorem Catalog.Probability.RenormalizedFactorizationValuation.padic_rigidity_dichotomy(k : ℤ) (m : ℕ) (hm : 1 ≤ m) (d : ℕ → ℤ) (g : (ℚ_[p])ˣ)
(hg : Padic.valuation (p := p) (g : ℚ_[p]) = k + ∑ i ∈ range m, d i) :
(DiscreteVal.factorizations (padicVal p) k m d g).Subsingleton ↔ m = 1 := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Speculative/AutoResearch/RenormalizedFactorizationValuation.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Speculative/AutoResearch/RenormalizedFactorizationValuation.lean
import Mathlib
import Definitions.Def_Speculative_AutoResearch_RenormalizedFactorizationValuation
/-
# Renormalized factorization over an arbitrary discretely valued group (Cycle 4)
This file closes two of the open conjectures listed in `FUTURE_DIRECTIONS.md` for the
Conjecture-C thread of `Catalog/Probability/RenormalizedNormalizedFactorization.lean`:
* **C4 (transfer to other valued fields).** Nothing in the Laurent-series proof uses the
coefficientwise structure of `LaurentSeries K`. All that is needed is a commutative group
`G` (the group of units of the field), a homomorphism `val : G → ℤ`, and a uniformizer
`π` with `val π = 1`. This is packaged as `DiscreteVal G`. In this generality:
`realizable_iff` — for every `m ≥ 1`, every integer exponent `k` and every pole profile
`d`, the set of `π ^ k * ∏_{i < m} f i` with `val (f i) = d i` is exactly the level set
`{g | val g = k + ∑_{i<m} d i}`.
* **C1 (rigidity index).** The fibre of the renormalized-product map over a realizable `g`
is a torsor under the group of "twists" (`fibreEquivTwist`), and that twist group is in
bijection with `Fin (m-1)` copies of the valuation-zero subgroup (`twistEquivPi`). Hence
`card_factorizations`: the fibre has exactly `#{u | val u = 0} ^ (m-1)` elements — the
rigidity index is `m - 1`, so the fibre is a singleton **iff** `m = 1`
(`rigidity_dichotomy`).
The two instantiations proved at the end are genuinely different worlds:
* `laurentVal K` — the Laurent series field `LaurentSeries K` over any field `K`, recovering
the results of the companion file;
* `padicVal p` — the `p`-adic numbers `ℚ_[p]`, where the same dichotomy is new.
No `sorry`, no `native_decide`, no new axioms.
-/
open Catalog.Probability.RenormalizedFactorizationValuation
open Finset
variable {G : Type*} [CommGroup G]
/-! ## The abstract setting: a `ℤ`-valued valuation with a uniformizer -/
open DiscreteVal
variable (V : DiscreteVal G)
/-! ## Renormalized products and their fibres -/
/-! ## The fibre is a torsor under the twist group -/
/-! ## Rigidity: the fibre is a singleton exactly when `m = 1` -/
/-! ## The rigidity index: the fibre has `#{val = 0} ^ (m-1)` elements -/
/-! ## Instantiation 1: Laurent series -/
open HahnSeries
/-! ## Instantiation 2: the `p`-adic numbers -/
variable (p : ℕ) [hp : Fact p.Prime]Formal statement
theorem Catalog.Probability.RenormalizedFactorizationValuation.padic_rigidity_dichotomy(k : ℤ) (m : ℕ) (hm : 1 ≤ m) (d : ℕ → ℤ) (g : (ℚ_[p])ˣ)
(hg : Padic.valuation (p := p) (g : ℚ_[p]) = k + ∑ i ∈ range m, d i) :
(DiscreteVal.factorizations (padicVal p) k m d g).Subsingleton ↔ m = 1 := by sorrySource