OctoInformBeta

A uniformly discrete counterexample to bounded approximation in Lipschitz-free spaces

For a pointed metric space (M,d,0)(M,d,0), let Lip⁡0(M)\operatorname{Lip}_0(M) denote the Banach space of real Lipschitz functions vanishing at 00, with the Lipschitz norm. For x∈Mx\in M, the evaluation functional δx\delta_x is defined by δx(f)=f(x)\delta_x(f)=f(x) for f∈Lip⁡0(M)f\in\operatorname{Lip}_0(M). The real Lipschitz-free space ℱ(M)\mathcal F(M) is the closed linear span of these evaluations in Lip⁡0(M)*\operatorname{Lip}_0(M)^*. A Banach space has the approximation property (AP) if its identity can be approximated uniformly on compact sets by bounded finite-rank operators. For λ≥0\lambda\ge0, the λ\lambda-bounded approximation property (λ\lambda-BAP) requires such approximating operators to have norm at most λ\lambda. The bounded approximation property (BAP) means λ\lambda-BAP for some finite λ≥0\lambda\ge0.

A metric space is uniformly discrete if distances between distinct points have a positive lower bound. A metric space is proper if every closed bounded subset is compact. The question is whether AP implies BAP for real Lipschitz-free spaces over countable uniformly discrete metric spaces.

The preprint constructs a countable uniformly discrete metric space whose real Lipschitz-free space has AP but fails BAP.

The Lean library proves the existence of a countable unbounded nonproper metric space with distances between distinct points at least 11 whose real Lipschitz-free space has AP and does not, for any real λ\lambda, admit approximation of its identity uniformly on compact sets by bounded finite-rank operators of norm at most λ\lambda. For every integer p≥1p\ge1, the Lean library proves the existence of a bounded countable metric space MpM_p with distances between distinct points at least 11 and a finite set Ap⊂MpA_p\subset M_p containing the basepoint such that every bounded finite-rank operator TT on ℱ(Mp)\mathcal F(M_p) with ∥T∥≤p\|T\|\le p satisfies max⁡a∈Ap∥Tδa−δa∥≥12\max_{a\in A_p}\|T\delta_a-\delta_a\|\ge\tfrac12. Lean has checked every result of the Lean library, so the Lean library is the reference. The statements on these pages are those of the Lean library, with any differently stated result from the preprint given beside them. The preprint is a guide only to the order and names of the results.

The Lean library: LipschitzFree

LipschitzFree is a Lean library: 17 items, in six groups. Each has a page written by OctoInform, listed below.

Section 1. Introduction 4 items
  1. Section 1 (setting): real Lipschitz-free spaces and approximation properties Lipschitz-free spaces and finite-rank approximation.
  2. Theorem 1.1 A Lipschitz-free space separates approximation properties.
  3. Theorem 1.2 Blocks obstruct finite-rank approximation at fixed norm.
  4. Corollary 1.3 Equivalent norms with prescribed approximation obstructions.
Section 2. From bounded blocks to AP without BAP 2 items
  1. Proposition 2.1 Bounded spaces with unit separation have summable coordinates.
  2. Section 2 (setting): metric wedges, linearized retractions and AP Metric wedges give contractive retractions and approximation.
Section 3. Colored graphs and the bounded block 3 items
  1. Section 3 (setting): incidences and the finite replacement for Lemma 3.1 Colored graph models from free group balls.
  2. Section 3 (setting): quantitative scales, capped labels and anchors Scales and anchors for a metric block.
  3. Lemma 3.2 The bounded block has controlled neighborhoods.
Section 4. Finite-rank operators and signed measures 3 items
  1. Lemma 4.1 Finite-support truncation of finite-rank operators.
  2. Section 4 (setting): affine signed weights and label bumps Affine weights and Lipschitz label tests.
  3. Lemma 4.2 Signed measures from finite-support operators.
Section 5. Rare matched walks and continuation records 3 items
  1. Section 5 (setting): compatibility, matched walks, records and exceptional states Continuation records define exceptional directions.
  2. Lemma 5.1 Counting matched roots across finite colorings.
  3. Lemma 5.3 Matched edges meet exceptional states.
Section 6. Scalar tests and incidence averaging 2 items
  1. Lemma 6.1 Bounded scalar tests outside exceptional states.
  2. Section 6 (proof of Theorem 1.2): signed pairing and incidence averaging Finite anchor sets obstruct supported operators.

No item matches.