Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Nonexpansive bijections between balls of spaces between c00c_{00}c00​ and c0c_0c0​ are signed coordinate bijections

Proved
Plasticity.CUB37

by scanstone · Oct 14, 2026 · Mathlib 0df444a (Lean v4.33.1)

Let Γ,Δ\Gamma,\DeltaΓ,Δ be sets, XXX a space between c00(Γ)c_{00}(\Gamma)c00​(Γ) and c0(Γ)c_0(\Gamma)c0​(Γ) and YYY a space between c00(Δ)c_{00}(\Delta)c00​(Δ) and c0(Δ)c_0(\Delta)c0​(Δ), both with the supremum norm, and let F ⁣:BX→BYF\colon B_X\to B_YF:BX​→BY​ be a nonexpansive bijection of their closed unit balls. Suppose that X=c00(Γ)X=c_{00}(\Gamma)X=c00​(Γ), or that YYY realizes all sign patterns (for instance Y=c0(Δ)Y=c_0(\Delta)Y=c0​(Δ)). Then there are a bijection σ ⁣:Γ→Δ\sigma\colon\Gamma\to\Deltaσ:Γ→Δ and signs αn∈{1,−1}\alpha_n\in\{1,-1\}αn​∈{1,−1} such that

F(x)σ(n)=αnxnfor all x∈BX, n∈Γ;F(x)_{\sigma(n)}=\alpha_n x_n\qquad\text{for all }x\in B_X,\ n\in\Gamma;F(x)σ(n)​=αn​xn​for all x∈BX​, n∈Γ;

that is, FFF is the restriction of a signed coordinate bijection. No continuity of F−1F^{-1}F−1 is assumed.

Preamble
import Definitions.Def_Plasticity_SeqSpace
import Mathlib.Topology.EMetricSpace.Lipschitz
Formal statement
namespace Plasticity
theorem CUB37 {Γ Δ : Type*} (X : Submodule ℝ (Linf Γ)) (Y : Submodule ℝ (Linf Δ))
    (hX0 : c00 Γ ≤ X) (hX1 : X ≤ c0 Γ) (hY0 : c00 Δ ≤ Y) (hY1 : Y ≤ c0 Δ)
    (hmatch : X = c00 Γ ∨ RealizesSignPatterns Y)
    (F : Linf Γ → Linf Δ)
    (hbij : Set.BijOn F (ballIn X) (ballIn Y))
    (hlip : LipschitzOnWith 1 F (ballIn X)) :
    ∃ σ : Γ ≃ Δ, ∃ α : Γ → ℝ, (∀ n, α n = 1 ∨ α n = -1) ∧
      ∀ x ∈ ballIn X, ∀ n, F x (σ n) = α n * x n := by sorry
end Plasticity
Source
This repository: claim CUB37 (claims/cub.md), item (iii), in the form of Corollary 4 (i)–(ii) of expositions/c0_between_spaces.md; proof in §§2–4 there
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Setting. Γ\GammaΓ and Δ\DeltaΔ are arbitrary index types (possibly empty, finite, countably infinite or uncountable, and not assumed related to each other). ℓ∞(Γ)\ell_\infty(\Gamma)ℓ∞​(Γ) is the real Banach space of bounded functions Γ→R\Gamma \to \mathbb{R}Γ→R with ∥x∥∞=sup⁡γ∣x(γ)∣\|x\|_\infty = \sup_\gamma |x(\gamma)|∥x∥∞​=supγ​∣x(γ)∣ (and ∥x∥∞=0\|x\|_\infty = 0∥x∥∞​=0 when Γ\GammaΓ is empty); likewise ℓ∞(Δ)\ell_\infty(\Delta)ℓ∞​(Δ). Recall

  • c00(Γ)c_{00}(\Gamma)c00​(Γ): the finitely supported functions Γ→R\Gamma \to \mathbb{R}Γ→R;
  • c0(Γ)c_0(\Gamma)c0​(Γ): the functions x:Γ→Rx : \Gamma \to \mathbb{R}x:Γ→R such that {γ:∣x(γ)∣≥ε}\{\gamma : |x(\gamma)| \ge \varepsilon\}{γ:∣x(γ)∣≥ε} is finite for every ε>0\varepsilon > 0ε>0;

both regarded as linear subspaces of ℓ∞(Γ)\ell_\infty(\Gamma)ℓ∞​(Γ), and similarly for Δ\DeltaΔ. The data are:

  • XXX: a real linear subspace of ℓ∞(Γ)\ell_\infty(\Gamma)ℓ∞​(Γ) with c00(Γ)⊆X⊆c0(Γ)c_{00}(\Gamma) \subseteq X \subseteq c_0(\Gamma)c00​(Γ)⊆X⊆c0​(Γ);
  • YYY: a real linear subspace of ℓ∞(Δ)\ell_\infty(\Delta)ℓ∞​(Δ) with c00(Δ)⊆Y⊆c0(Δ)c_{00}(\Delta) \subseteq Y \subseteq c_0(\Delta)c00​(Δ)⊆Y⊆c0​(Δ).

XXX and YYY need not be closed or complete. Their closed unit balls, taken as subsets of the ambient ℓ∞\ell_\inftyℓ∞​ spaces, are

BX={x∈X:∥x∥∞≤1}⊆ℓ∞(Γ),BY={y∈Y:∥y∥∞≤1}⊆ℓ∞(Δ).B_X = \{x \in X : \|x\|_\infty \le 1\} \subseteq \ell_\infty(\Gamma), \qquad B_Y = \{y \in Y : \|y\|_\infty \le 1\} \subseteq \ell_\infty(\Delta).BX​={x∈X:∥x∥∞​≤1}⊆ℓ∞​(Γ),BY​={y∈Y:∥y∥∞​≤1}⊆ℓ∞​(Δ).

F:ℓ∞(Γ)→ℓ∞(Δ)F : \ell_\infty(\Gamma) \to \ell_\infty(\Delta)F:ℓ∞​(Γ)→ℓ∞​(Δ) is an arbitrary function on the whole of ℓ∞(Γ)\ell_\infty(\Gamma)ℓ∞​(Γ); only its values on BXB_XBX​ matter. For x∈ℓ∞(Γ)x \in \ell_\infty(\Gamma)x∈ℓ∞​(Γ) and δ∈Δ\delta \in \Deltaδ∈Δ, F(x)(δ)F(x)(\delta)F(x)(δ) denotes the δ\deltaδ-th coordinate of F(x)F(x)F(x).

Hypotheses.

  1. (Matching condition) Either X=c00(Γ)X = c_{00}(\Gamma)X=c00​(Γ) exactly, or YYY realizes sign patterns, meaning: for every countable (finite, possibly empty, or countably infinite) set C⊆ΔC \subseteq \DeltaC⊆Δ and every ε:Δ→R\varepsilon : \Delta \to \mathbb{R}ε:Δ→R with ε(d)∈{1,−1}\varepsilon(d) \in \{1, -1\}ε(d)∈{1,−1} for all d∈Cd \in Cd∈C, there is v∈Yv \in Yv∈Y (no norm bound) with ε(d) v(d)>0\varepsilon(d)\, v(d) > 0ε(d)v(d)>0 for all d∈Cd \in Cd∈C. The first alternative constrains only XXX, the second only YYY.
  2. (Bijection of balls) FFF maps BXB_XBX​ into BYB_YBY​, FFF is injective on BXB_XBX​, and every y∈BYy \in B_Yy∈BY​ equals F(x)F(x)F(x) for some x∈BXx \in B_Xx∈BX​. Thus FFF restricts to a bijection BX→BYB_X \to B_YBX​→BY​.
  3. (Nonexpansive) ∥F(x)−F(x′)∥∞≤∥x−x′∥∞\|F(x) - F(x')\|_\infty \le \|x - x'\|_\infty∥F(x)−F(x′)∥∞​≤∥x−x′∥∞​ for all x,x′∈BXx, x' \in B_Xx,x′∈BX​.

Nothing is assumed about the inverse of F∣BXF|_{B_X}F∣BX​​ (no continuity or Lipschitz bound).

Conclusion. There exist a bijection σ:Γ→Δ\sigma : \Gamma \to \Deltaσ:Γ→Δ and signs α:Γ→R\alpha : \Gamma \to \mathbb{R}α:Γ→R with α(n)=1\alpha(n) = 1α(n)=1 or α(n)=−1\alpha(n) = -1α(n)=−1 for every n∈Γn \in \Gamman∈Γ, such that

F(x)(σ(n))=α(n) x(n)for every x∈BX and every n∈Γ.F(x)(\sigma(n)) = \alpha(n)\, x(n) \qquad \text{for every } x \in B_X \text{ and every } n \in \Gamma.F(x)(σ(n))=α(n)x(n)for every x∈BX​ and every n∈Γ.

Since σ\sigmaσ is a bijection, this fixes every coordinate of F(x)F(x)F(x): for x∈BXx \in B_Xx∈BX​ and δ∈Δ\delta \in \Deltaδ∈Δ, F(x)(δ)=α(σ−1δ) x(σ−1δ)F(x)(\delta) = \alpha(\sigma^{-1}\delta)\, x(\sigma^{-1}\delta)F(x)(δ)=α(σ−1δ)x(σ−1δ); i.e. on BXB_XBX​, FFF coincides with a signed coordinate permutation. In particular the conclusion asserts that Γ\GammaΓ and Δ\DeltaΔ are in bijection. Nothing is asserted about FFF outside BXB_XBX​, and no statement relating XXX and YYY beyond this formula is made explicitly.

Edge and satisfiability remarks.

  • The hypotheses are jointly satisfiable, e.g. by Γ=Δ\Gamma = \DeltaΓ=Δ, X=Y=c0(Γ)X = Y = c_0(\Gamma)X=Y=c0​(Γ), FFF the identity.
  • If Γ\GammaΓ is empty, then BX={0}B_X = \{0\}BX​={0} and hypothesis 2 forces BYB_YBY​ to be a single point; since c00(Δ)⊆Yc_{00}(\Delta) \subseteq Yc00​(Δ)⊆Y, this happens only if Δ\DeltaΔ is empty, and then the conclusion holds with the empty bijection. Symmetrically, Δ\DeltaΔ empty and Γ\GammaΓ nonempty is incompatible with injectivity in hypothesis 2.
  • If Δ\DeltaΔ is finite, then Y=ℓ∞(Δ)=RΔY = \ell_\infty(\Delta) = \mathbb{R}^\DeltaY=ℓ∞​(Δ)=RΔ and the second alternative of hypothesis 1 holds automatically. If Δ\DeltaΔ is infinite and Y=c00(Δ)Y = c_{00}(\Delta)Y=c00​(Δ), the second alternative fails (a countably infinite CCC would need vvv with infinite support), so hypothesis 1 then requires X=c00(Γ)X = c_{00}(\Gamma)X=c00​(Γ). If Y=c0(Δ)Y = c_0(\Delta)Y=c0​(Δ), the second alternative holds, so XXX may be any subspace between c00(Γ)c_{00}(\Gamma)c00​(Γ) and c0(Γ)c_0(\Gamma)c0​(Γ).
Human review
  • Endorsed by Shuze Chen · Oct 14, 2026

    Confirmed by the moderator at approval.

  • Endorsed by scanstone · Oct 14, 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, with reuse governed by our licensing terms.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me