Alphabeta Math
CounterexampleConstruction: AI-generatedVerification: AI-adaptedPipeline-generatedjudge pass (gpt-6.1-sol)audited 2026-10-02
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Ambient-smooth density fails on a slit disc

Statement refuted

The smooth-up-to-the-boundary density conclusion of Ambient smooth restrictions are dense on bounded C^k domains cannot be extended from bounded Ck domains to arbitrary bounded open sets. Assume the Axiom of Choice and let Ω=B(0,1)∖{(x,0):0≤x<1}⊂R2,u(x,y)=arg⁡(x+iy)∈(0,2π). For every 1≤p<2 the branch u belongs to W1,p(Ω), but there is no sequence of functions φj∈C∞(R2) with ∥φj∣Ω−u∥W1,p(Ω)→0. Thus the restrictions of globally smooth functions are not dense in W1,p(Ω) on this bounded open set, although they are dense on every bounded Ck domain. The two one-sided boundary values of u on the slit differ by 2π, and a globally smooth function has equal one-sided values, which is the obstruction.

Facts & Assumptions

Given: the Axiom of Choice; the bounded open slit disc Ω=B(0,1)∖{(x,0):0≤x<1}; the branch u(x,y)=arg⁡(x+iy)∈(0,2π); and 1≤p<2.

[F1]

Classical derivatives of C1 functions are weak derivatives (Classical derivatives agree with weak derivatives).

[F2]

Membership in W1,p(Ω;K) means membership of the class in Lp together with weak first derivatives in Lp, with finite-p norm ∥w∥W1,p(Ω)=(∥w∥Lp(Ω)p+∥∂xw∥Lp(Ω)p+∥∂yw∥Lp(Ω)p)1/p; at p=∞ the norm is the maximum of these three essential bounds (Integer-order Sobolev spaces and their norms).

[F3]

Polar coordinates: ∫R2f dλ2=∫0∞∫S1f(rω) r dσ(ω) dr for Borel measurable f≥0 (Polar coordinates decompose Lebesgue measure into r^{n-1} dr d sigma).

[F4]

Tonelli–Fubini: for nonnegative measurable f on a completed product, the double integral equals the iterated integrals; and λ2 is the completion of λ1×λ1 (Tonelli and Fubini for the completed product, with only almost-everywhere section measurability, The Euclidean Lebesgue measure is the completion of the product of the factor Lebesgue measures).

[F5]

Hölder's inequality: for conjugate exponents and integrable functions, ∫∣fg∣≤∥f∥p∥g∥q (Holder's inequality for integrals, including the endpoint cases).

[F6]

The density theorem: on a bounded Ck domain with k≥1 and 1≤q<∞, the restrictions of Cc∞(Rn;K) functions are dense in Wk,q (Ambient smooth restrictions are dense on bounded C^k domains).

[F7]

Integral over a measurable set: ∫Af dμ is the integral of f1A (Integral over a measurable subset).

Choice use. The Axiom of Choice is assumed; the argument invokes it only through the Countable Choice declared by [F1] and through the choice-bearing product-measure and polar-coordinate interfaces [F3] and [F4]. The contradiction argument itself selects no family.

Counterexample

1.1givenalgebra

The set Ω is open, because it is the intersection of the open disc B(0,1) with the open set {y≠0}∪{x<0}, and it is bounded; the branch u is of class C∞ on Ω with 0<u<2π and ∇u(x,y)=(−yx2+y2, xx2+y2),∣∇u(x,y)∣=1r,r=x2+y2.

1.2F5given

Endpoint estimate. Let g∈C1([0,ε]) and let 1≤p<∞. For every s∈(0,ε) the fundamental theorem gives g(0)=g(s)−∫0sg′(t) dt, so ε∣g(0)∣≤∫0ε∣g∣+ε∫0ε∣g′∣, and [F5] applied to the two summands yields ∣g(0)∣p≤C(ε,p)(∫0ε∣g∣p+∫0ε∣g′∣p) with C(ε,p)=2p−1max⁡{ε−1,εp−1}; the same estimate holds on (−ε,0) for the endpoint 0.

2.1F3F7step 1.1

Integrability. Since ∣u∣≤2π and Ω⊆B(0,1) has finite area, ∫Ω∣u∣p<∞; by [F3], [F7] and ∣∇u∣=1/r of step 1.1, ∫Ω∣∇u∣p=∫Ωr−p≤∫012πr1−p dr=2π2−p<∞, because p<2 makes the one-dimensional integral converge at r=0.

3.1F1F2step 1.1step 2.1

Membership. The function u is C∞ on the open set Ω, so by [F1] its classical partial derivatives of step 1.1 are its weak derivatives; by step 2.1 they lie in Lp(Ω) together with u, and the membership criterion of [F2] gives u∈W1,p(Ω) for every 1≤p<2.

4.1F2F4step 1.2step 3.1

Upper strip. Suppose φj∈C∞(R2) satisfy ∥φj∣Ω−u∥W1,p(Ω)→0. Fix 0<ε<1/4 and put S+=(1/4,3/4)×(0,ε)⊂Ω. The function u=arctan⁡(y/x) extends C1 to the closed strip, with u(x,0)=0. For each j, apply the endpoint estimate of step 1.2 to gj(x,y)=φj(x,y)−u(x,y) on each vertical section of S+ and integrate in x using [F4]. Since ε is fixed, its constant is fixed, and ∫1/43/4∣φj(x,0)∣pdx≤C(ε,p)∫S+(∣φj−u∣p+∣∂yφj−∂yu∣p)≤C(ε,p)∥φj−u∥W1,p(Ω)p⟶0.

4.2F2F4step 1.2step 3.1

Lower strip. On S−=(1/4,3/4)×(−ε,0) the function u=2π−arctan⁡(∣y∣/x) extends C1 to the closed strip with u(x,0)=2π. Applying step 1.2 to gj=φj−u on each vertical section and integrating in x gives, for the same fixed ε, ∫1/43/4∣φj(x,0)−2π∣pdx≤C(ε,p)∫S−(∣φj−u∣p+∣∂yφj−∂yu∣p)≤C(ε,p)∥φj−u∥W1,p(Ω)p⟶0.

5.1F7step 4.1step 4.2

Contradiction. For every j the elementary inequality ∣2π∣p≤2p−1(∣φj(x,0)∣p+∣φj(x,0)−2π∣p) holds pointwise on (1/4,3/4), so 12(2π)p=∫1/43/4∣2π∣pdx≤2p−1(∫1/43/4∣φj(x,0)∣pdx+∫1/43/4∣φj(x,0)−2π∣pdx)→j→∞0 by steps 4.1 and 4.2, which is impossible since (2π)p/2>0.

6.1F6step 3.1step 5.1∎

Therefore no sequence of globally smooth functions converges to u in W1,p(Ω) for any 1≤p<2, so ambient smooth restrictions are not dense on this bounded open set. By contrast [F6] gives that density for every bounded Ck domain, so the conclusion cannot be extended to arbitrary open sets. Indeed Ω is not a bounded C1 domain in the graph sense: near a slit point (x0,0) with 0<x0<1, its complement is only a line segment and has empty interior, so Ω is dense on both sides of that segment. A one-sided graph domain has a nonempty open complementary side in every sufficiently small chart neighbourhood, which rules out such a chart here.

Depends on

Used by

Nothing in the library uses this result yet.

Dependency tree · two levels

56 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources