OctoInformBeta

OpenAI preprint, September 2026 · Appendix A. An explicit map with positive displacement

Proposition A.1: Negative normalization of a correction along an oscillating curve in L2([0,1];ℝ2)L^2([0,1];\mathbb R^2)

A Lipschitz map with positive displacement.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

The analytic question is whether a Lipschitz self-map of a Hilbert ball can move every point by a fixed positive distance. The ambient plane is the Euclidean space ℝ2\mathbb R^2, and μ\mu is Lebesgue measure restricted to [0,1][0,1]. Let HH be the real Hilbert space of square-integrable plane-valued functions, with functions equal μ\mu-almost everywhere identified.✦ For a seminormed additive commutative group EE, the closed unit ball BEB_E is the subset of EE consisting of elements of norm at most one.✦ The space and ball used here are those in (1). H=L2([0,1],μ;ℝ2),BE={x∈E:∥x∥≤1},B=BH.H=L^2([0,1],\mu;\mathbb R^2),\qquad B_E=\{x\in E:\|x\|\leq1\},\qquad B=B_H.(1)

The Oscillating Curve

The unit circle parametrization p:ℝ→ℝ2p:\mathbb R\to\mathbb R^2 and its tangent p′p' define the oscillator representative wzw_z and velocity representative ẇz\dot w_z on [0,1][0,1] by (2). The Hilbert space oscillator W(z)∈HW(z)\in H and oscillator velocity W′(z)∈HW'(z)\in H are their respective equivalence classes. p(z)=(cos⁡z,sin⁡z),p′(z)=(−sin⁡z,cos⁡z),wz(r)=p(zr),ẇz(r)=rp′(zr),W(z)=[wz],W′(z)=[ẇz].\begin{aligned} p(z)&=(\cos z,\sin z),&p'(z)&=(-\sin z,\cos z),\\ w_z(r)&=p(zr),&\dot w_z(r)&=r p'(zr),\\ W(z)&=[w_z],&W'(z)&=[\dot w_z]. \end{aligned}(2)

The radial profile a:ℝ→ℝa:\mathbb R\to\mathbb R and angular profile θ:ℝ→ℝ\theta:\mathbb R\to\mathbb R, together with their radial and angular derivatives, are the functions in (3). a(t)={t/8t≤2,5/16−(3−t)2/162<t<3,5/16t≥3,a′(t)={1/8t≤2,(3−t)/82<t<3,0t≥3,θ(t)={0t≤1,(t−1)2/21<t<2,t−3/2t≥2,θ′(t)={0t≤1,t−11<t<2,1t≥2.\begin{aligned} a(t)&=\begin{cases} t/8&t\leq2,\\ 5/16-(3-t)^2/16&2<t<3,\\ 5/16&t\geq3, \end{cases} &a'(t)&=\begin{cases} 1/8&t\leq2,\\ (3-t)/8&2<t<3,\\ 0&t\geq3, \end{cases}\\ \theta(t)&=\begin{cases} 0&t\leq1,\\ (t-1)^2/2&1<t<2,\\ t-3/2&t\geq2, \end{cases} &\theta'(t)&=\begin{cases} 0&t\leq1,\\ t-1&1<t<2,\\ 1&t\geq2. \end{cases} \end{aligned}(3)

The radial direction v:ℝ→Hv:\mathbb R\to H and radial oscillating curve γ:ℝ→H\gamma:\mathbb R\to H are defined in (4); the same display defines the curve velocity γ′\gamma' and normalized unit tangent u:ℝ→Hu:\mathbb R\to H. v(t)=W(θ(t)),γ(t)=a(t)v(t),γ′(t)=a′(t)v(t)+a(t)θ′(t)W′(θ(t)),u(t)=∥γ′(t)∥−1γ′(t).\begin{aligned} v(t)&=W(\theta(t)),&\gamma(t)&=a(t)v(t),\\ \gamma'(t)&=a'(t)v(t)+a(t)\theta'(t)W'(\theta(t)),& u(t)&=\|\gamma'(t)\|^{-1}\gamma'(t). \end{aligned}(4)

Regular separated curve data C=(c,w,τ,δ)C=(c,w,\tau,\delta) in a real Hilbert space EE consist of a differentiable curve c:ℝ→Ec:\mathbb R\to E, its velocity ww, a unit tangent τ:ℝ→E\tau:\mathbb R\to E, and a real separation constant δ\delta, subject to (5) for all real s,ts,t. c′(t)=w(t),∥τ(t)∥=1,w(t)=∥w(t)∥τ(t),∥w(t)∥≥18,∥w(t)−w(s)∥≤|t−s|,∥τ(t)−τ(s)∥≤16|t−s|,δ>0,|t−s|≥132⇒∥c(t)−c(s)∥≥δ.\begin{gathered} c'(t)=w(t),\qquad \|\tau(t)\|=1,\qquad w(t)=\|w(t)\|\tau(t),\qquad \|w(t)\|\geq\tfrac18,\\ \|w(t)-w(s)\|\leq|t-s|,\qquad \|\tau(t)-\tau(s)\|\leq16|t-s|,\\ \delta>0,\qquad |t-s|\geq\tfrac1{32}\ \Longrightarrow\ \|c(t)-c(s)\|\geq\delta. \end{gathered}(5)✦

Let σ∈ℝ\sigma\in\mathbb R be the separation constant selected for γ\gamma at parameter scale 1/321/32.✦ The oscillating curve regularity data Cγ=(γ,γ′,u,σ)C_\gamma=(\gamma,\gamma',u,\sigma) are regular separated curve data for γ\gamma satisfying (5).✦

The Tube Correction

For regular separated curve data C=(c,w,τ,δ)C=(c,w,\tau,\delta), the admissible tube radius ρC\rho_C and open curve tube UC⊂EU_C\subset E are defined by (6). The oscillating curve tube radius is ρ=ρCγ\rho=\rho_{C_\gamma}.✦ ρC=min⁡{11024,δ8},UC={x∈E:dist⁡(x,c(ℝ))<ρC}.\rho_C=\min\{\tfrac1{1024},\tfrac\delta8\},\qquad U_C=\{x\in E:\operatorname{dist}(x,c(\mathbb R))<\rho_C\}.(6)

A nearest curve parameter t∈ℝt\in\mathbb R for a curve c:ℝ→Ec:\mathbb R\to E and a point x∈Ex\in E is a real number minimizing ∥x−c(t)∥\|x-c(t)\|.✦ For x∈UCx\in U_C, let pC(x)p_C(x) be the chosen nearest parameter and let eC(x)∈Ee_C(x)\in E be its normal residual, as in (7). ∥x−c(pC(x))∥≤∥x−c(t)∥(t∈ℝ),eC(x)=x−c(pC(x)).\|x-c(p_C(x))\|\leq\|x-c(t)\|\quad(t\in\mathbb R),\qquad e_C(x)=x-c(p_C(x)).(7)

The skew rank-two operator Sh,k:E→ES_{h,k}:E\to E and rotation ℛh,k:E→E\mathcal R_{h,k}:E\to E associated with vectors h,kh,k in a real Hilbert space EE are the continuous real-linear operators in (8), where II is the identity operator on EE. Writing ℒ(H)\mathcal L(H) for the space of continuous real-linear operators on HH, the tangent rotation R:ℝ→ℒ(H)R:\mathbb R\to\mathcal L(H) is the family given there. Sh,k(e)=⟨h,e⟩k−⟨k,e⟩h,ℛh,k=I+Sh,k+(1+⟨h,k⟩)−1Sh,k,2,R(t)=ℛu(t),v(t).\begin{aligned} S_{h,k}(e)&=\langle h,e\rangle k-\langle k,e\rangle h,\\ \mathcal R_{h,k}&=I+S_{h,k}+(1+\langle h,k\rangle)^{-1}S_{h,k}^{,2},\\ R(t)&=\mathcal R_{u(t),v(t)}. \end{aligned}(8)

Radial correction data D=(a,v,q,R)D=(a,v,q,R) for CC consist of real functions a,qa,q, a direction v:ℝ→Ev:\mathbb R\to E, and continuous real-linear operators R(t)R(t) on EE satisfying (9) for all real s,ts,t and all e∈Ee\in E. c(t)=a(t)v(t),a(0)=0,a(t)≤516,t>1⇒a(t)≥18,q(t)=min⁡{a(t),−18}−a(t),|q(t)|≤716,|q(t)−q(s)|≤18|t−s|,∥v(t)∥=1,∥v(t)−v(s)∥≤|t−s|,∥c(t)−c(s)∥≤716|t−s|,∥R(t)e∥=∥e∥,e⟂τ(t)⇒R(t)e⟂v(t),∥R(t)e−R(s)e∥≤238|t−s|∥e∥,t≤1⇒R(t)e=e and τ(t)=v(t).\begin{gathered} c(t)=a(t)v(t),\qquad a(0)=0,\qquad a(t)\leq\tfrac5{16},\qquad t>1\ \Longrightarrow\ a(t)\geq\tfrac18,\\ q(t)=\min\{a(t),-\tfrac18\}-a(t),\qquad |q(t)|\leq\tfrac7{16},\qquad |q(t)-q(s)|\leq\tfrac18|t-s|,\\ \|v(t)\|=1,\qquad\|v(t)-v(s)\|\leq|t-s|,\qquad \|c(t)-c(s)\|\leq\tfrac7{16}|t-s|,\\ \|R(t)e\|=\|e\|,\qquad e\perp\tau(t)\ \Longrightarrow\ R(t)e\perp v(t),\\ \|R(t)e-R(s)e\|\leq238|t-s|\|e\|,\\ t\leq1\ \Longrightarrow\ R(t)e=e\ \text{ and }\ \tau(t)=v(t). \end{gathered}(9)✦

For the oscillating curve, the radial correction q:ℝ→ℝq:\mathbb R\to\mathbb R is q(t)=−max⁡{a(t)+1/8,0}q(t)=-\max\{a(t)+1/8,0\}.✦ The oscillating curve correction data Dγ=(a,v,q,R)D_\gamma=(a,v,q,R) are radial correction data for CγC_\gamma satisfying (9).✦

The inner cutoff β\beta and outer cutoff χ\chi are real functions of a radius 𝜚\varrho and a real variable rr, defined by (10). β(𝜚,r)=min⁡{1,max⁡{0,2−4r/𝜚}},χ(𝜚,r)=min⁡{1,max⁡{0,2−2r/𝜚}}.\beta(\varrho,r)=\min\{1,\max\{0,2-4r/\varrho\}\},\qquad \chi(\varrho,r)=\min\{1,\max\{0,2-2r/\varrho\}\}.(10)

For radial correction data DD along CC, the additive tube correction ΔD:UC→E\Delta_D:U_C\to E and corrected tube point JD:UC→EJ_D:U_C\to E are defined by (11), where p=pC(x)p=p_C(x) and e=eC(x)e=e_C(x). ΔD(x)=β(ρC,∥e∥)q(p)v(p)+χ(ρC,∥e∥)(R(p)e−e),JD(x)=x+ΔD(x).\Delta_D(x)=\beta(\rho_C,\|e\|)q(p)v(p) +\chi(\rho_C,\|e\|)(R(p)e-e),\qquad J_D(x)=x+\Delta_D(x).(11)

The glued correction associated with radial correction data DD along CC sends x∈UCx\in U_C to the corrected tube point JD(x)J_D(x) and fixes every point outside UCU_C, including its boundary.✦ Let G̃:H→H\widetilde G:H\to H be this glued correction for DγD_\gamma. The unit ball correction G:B→HG:B\to H is the restriction of G̃\widetilde G to BB.✦ The correction Lipschitz constant KCK_C associated with radial correction data along CC, and the displacement Lipschitz constant LL, are defined in (12). KC=max⁡{1,max{30+74ρC+3808ρC,5+78ρC}},L=2KCγρ/4.K_C=\max\left\{1,\max\left\{ 30+\frac7{4\rho_C}+3808\rho_C,\;5+\frac7{8\rho_C}\right\}\right\},\qquad L=\frac{2K_{C_\gamma}}{\rho/4}.(12)

Negative radial normalization is the map N:E→EN:E\to E given by (13) at nonzero vectors and taking the value zero at zero. For a map Q:BE→EQ:B_E\to E with ∥Q(x)∥≥ε>0\|Q(x)\|\geq\varepsilon>0 on BEB_E, its normalized nonvanishing map is x↦N(Q(x))x\mapsto N(Q(x)) from BEB_E to BEB_E.✦ The displacement map T:B→BT:B\to B is the normalization of GG in that display. N(z)=−∥z∥−1z,T(x)=N(G(x))(x∈B).N(z)=-\|z\|^{-1}z,\qquad T(x)=N(G(x))\quad(x\in B).(13)

The Claim

Theorem 1. For the displacement map T:B→BT:B\to B, the norm identity (14) and displacement bound (15) hold at every x∈Bx\in B: ∥T(x)∥=1.\|T(x)\|=1.(14)✦ ∥T(x)−x∥≥12.\|T(x)-x\|\geq\tfrac12.(15)✦ The constant LL in (12) has the sign in (16): L>0.L>0.(16)✦ For all x,y∈Bx,y\in B, the Lipschitz bound is (17): ∥T(x)−T(y)∥≤L∥x−y∥.\|T(x)-T(y)\|\leq L\|x-y\|.(17)✦

In the paper, the result reads: for H=L2([0,1];ℝ2)H=L^2([0,1];\mathbb R^2) and B={x∈H:∥x∥≤1}B=\{x\in H:\|x\|\leq1\}, there is a globally Lipschitz map f:B→Bf:B\to B with ∥f(x)∥=1\|f(x)\|=1 and ∥f(x)−x∥≥1/2\|f(x)-x\|\geq1/2 for every x∈Bx\in B.

What Is Checked

The statement is over the Lebesgue space HH and its closed unit ball BB, with the oscillating curve regularity data CγC_\gamma and radial correction data DγD_\gamma determining GG, TT and LL as in (12) and (13).

The oscillator and velocity representatives in (2) are square-integrable for every real zz. The unit ball correction satisfies the nonvanishing estimate (18) for every x∈Bx\in B. ∥G(x)∥≥ρ/4>0.\|G(x)\|\geq\rho/4>0.(18)

The radial profile aa is bounded above by 5/165/16 on the real line.✦ The radial correction qq has absolute value at most 7/167/16 at every real parameter.✦ The radial direction v(t)v(t) has norm one for every real tt.✦ The radial direction v:ℝ→Hv:\mathbb R\to H is 11 Lipschitz.✦ The radial oscillating curve γ:ℝ→H\gamma:\mathbb R\to H is 7/167/16 Lipschitz.✦ At every real parameter, the curve velocity is its speed multiplied by the unit tangent.✦ The speed of the radial oscillating curve is at least 1/81/8 at every real parameter.✦ The curve velocity γ′:ℝ→H\gamma':\mathbb R\to H is 11 Lipschitz.✦ The unit tangent u:ℝ→Hu:\mathbb R\to H is 1616 Lipschitz.✦ The curve separation constant σ\sigma is positive.✦ Values of the radial oscillating curve at real parameters at least 1/321/32 apart have distance at least σ\sigma.✦ Every tangent rotation R(t)R(t) preserves the norm of every vector in HH.✦ For every real tt, the tangent rotation R(t)R(t) sends vectors orthogonal to u(t)u(t) to vectors orthogonal to v(t)v(t).✦ For every vector e∈He\in H, the function t↦R(t)et\mapsto R(t)e is Lipschitz with constant 238∥e∥238\|e\|.✦ For every real t≤1t\leq1, the tangent rotation R(t)R(t) is the identity on HH.✦ For every real t≤1t\leq1, the unit tangent u(t)u(t) equals the radial direction v(t)v(t).✦