OpenMath

Proof source

A restricted positive Borel result from a uniform Hausdorff estimate

Complete proof text in the unsigned reading edition.

Download sourceBrowse research records

Reading edition of the complete source note. Original labels and explicit equation tags are retained; references use source titles where compiled numbering is unavailable. Current result status is recorded separately in the project register.

Contents of this source
  1. A restricted positive Borel result from a uniform Hausdorff estimate

A restricted positive Borel result from a uniform Hausdorff estimate

Author and status.

Independent derivation and audit by task /root/b1_positive_routes. PROVED-INTERNAL, conditional only on the already audited original local linear trace lemma. This is a restricted positive result, not the general B1 theorem and not a claim of worldwide novelty.

Retain the fixed original KK, absolute value, original nn-torus with n≥2n\ge2, the strict Laurent box VV, normalized moments τ:V→Q=(−1,1)n\tau:V\to Q=(-1,1)^n, and the smooth strongly positive background S=βn−1[X]S=\beta^{n-1}[X] used in the preceding sections. Thus q=1q=1 and τ∗(Da∧S)=γΔga,γ=(n−1)!λn−2,\tau_*(Da\wedge S)=\gamma\Delta g_a, \qquad\gamma=(n-1)!\lambda^{n-2}, for every original normalized tester 0≤a≤10\le a\le1, with canonical convex trace 0≤ga≤10\le g_a\le1. Fix a compact set P⋐QP\Subset Q and set d=dist⁡(P,Rn∖Q)>0d=\operatorname{dist}(P,\mathbf R^n\setminus Q)>0.

Lemma (Uniform local ball and Hausdorff control). There is a constant Mn,P<∞M_{n,P}<\infty, independent of the original tester, such that every Borel F⊂PF\subset P satisfies CS∗(τ−1F;V)≤Mn,Pγ Hn−1(F).(HR1)C_S^*(\tau^{-1}F;V) \le M_{n,P}\gamma\,\mathcal H^{n-1}(F). \tag{HR1} The Hausdorff measure has any one fixed standard normalization; its normalizing factor is absorbed in Mn,PM_{n,P}.

Proof. Every convex g:Q→[0,1]g:Q\to[0,1] is Lipschitz with constant at most L=2/dL=2/d on the d/2d/2 neighborhood of PP. Indeed, at a point of that neighborhood, moving toward either sign of any unit vector by a distance approaching d/2d/2 remains in QQ; the subgradient inequality bounds every directional component of the subgradient by 2/d2/d.

Choose a fixed smooth nonnegative cutoff η\eta on Rn\mathbf R^n, equal to one on the unit ball and supported in the ball of radius two. For z∈Pz\in P and 0<ρ<d/80<\rho<d/8, let ηz,ρ(x)=η((x−z)/ρ)\eta_{z,\rho}(x)=\eta((x-z)/\rho). Positivity of the Radon measure Δg\Delta g and distributional integration by parts imply Δg(B(z,ρ))≤∫ηz,ρ dΔg=−∫∇g⋅∇ηz,ρ dx≤L∥∇η∥L1(Rn)ρn−1.(HR2)\begin{split} \Delta g(B(z,\rho)) &\le\int\eta_{z,\rho}\,d\Delta g =-\int\nabla g\cdot\nabla\eta_{z,\rho}\,dx\\ &\le L\|\nabla\eta\|_{L^1(\mathbf R^n)}\rho^{n-1}. \end{split} \tag{HR2} All supports remain in the protected region where the common Lipschitz bound is valid. The identity is legitimate for locally Lipschitz gg; no trace theorem on sphere boundaries is used. The original linear trace identity consequently gives CS(τ−1B(z,ρ);V)≤γL∥∇η∥1ρn−1.(HR3)C_S(\tau^{-1}B(z,\rho);V) \le\gamma L\|\nabla\eta\|_1\rho^{n-1}. \tag{HR3}

Cover FF by countably many open balls centered in FF, of radii less than d/8d/8. The original open union of their inverseimages contains τ−1F\tau^{-1}F. For every original tester, its mass on that union is at most the sum of its masses on the individual inverseimages, so (HR3) bounds the raw capacity of this union by the sum of the right sides. Taking the infimum over arbitrarily fine ball covers proves (HR1), using the ball-cover characterization of Hausdorff measure and absorbing its fixed dimensional normalization into Mn,PM_{n,P}. The capacity being estimated is always the original open outer capacity; it is not replaced by a Hausdorff capacity or a smaller tester class. ◻

Theorem (Borel inner regularity for finite-Hausdorff saturated sets). Let B⊂PB\subset P be Borel with Hn−1(B)<∞\mathcal H^{n-1}(B)<\infty, and let E=τ−1BE=\tau^{-1}B. Then CS∗(E;V)=sup⁡H⊂E, H compactCS∗(H;V).(HR4)\boxed{\quad C_S^*(E;V) =\sup_{H\subset E,\ H\text{ compact}}C_S^*(H;V).\quad} \tag{HR4} In fact compact inverseimages H=τ−1FH=\tau^{-1}F, with F⊂BF\subset B compact, already suffice for the supremum.

Proof. The finite Borel measure A↦Hn−1(A∩B)A\mapsto\mathcal H^{n-1}(A\cap B) on Rn\mathbf R^n is Radon. Consequently, for every ϵ>0\epsilon>0, there is a compact F⊂BF\subset B such that Hn−1(B∖F)<ϵ\mathcal H^{n-1}(B\setminus F)<\epsilon. This uses ordinary regularity of a single finite Borel measure on Euclidean space, not regularity of the capacity under investigation.

The moment map is proper over compact subsets of QQ, hence τ−1F\tau^{-1}F is an original compact subset of EE. Countable subadditivity of the original outer capacity and (HR1) give CS∗(E;V)≤CS∗(τ−1F;V)+CS∗(τ−1(B∖F);V)≤CS∗(τ−1F;V)+Mn,Pγϵ.C_S^*(E;V) \le C_S^*(\tau^{-1}F;V) +C_S^*(\tau^{-1}(B\setminus F);V) \le C_S^*(\tau^{-1}F;V)+M_{n,P}\gamma\epsilon. For clarity, the subadditivity used here follows directly from the definition: choose original open supersets of the separate sets with raw capacities arbitrarily close to their outer capacities, take their union, and use countable subadditivity of every tester measure before taking the tester supremum. It is ordinary subadditivity, not the strong-subadditivity assertion refuted above.

Since (HR1) makes the left side finite, letting ϵ↓0\epsilon\downarrow0 proves the upper bound by the compact supremum. The reverse bound is monotonicity. No increasing-sequence continuity of the outer capacity is used. ◻

Exact limitations.

The theorem concerns full moment-saturated sets τ−1B\tau^{-1}B, with BB inside one fixed protected compact stage and of finite Hn−1\mathcal H^{n-1} measure, for this specific smooth strongly positive background and remaining degree one. It does not cover arbitrary Borel subsets of the original Berkovich space, arbitrary closed backgrounds, or higher remaining degrees. In particular a Borel subset of τ−1B\tau^{-1}B need not be the inverseimage of any Borel subset of BB; no such extension is asserted. Likewise σ\sigma-finiteness alone gives no uniform small Hausdorff remainder and cannot be substituted for finiteness without an additional argument. No global exhaustion or boundary-tail theorem is obtained.

The estimate also gives outer capacity zero for every saturated Hn−1\mathcal H^{n-1}-null set in the protected stage. This is compatible with the failure of strong subadditivity: regularity and submodularity are distinct properties. General B1 remains OPEN.

Source text and edition identity

This reading copy omits personal attribution and places existing mathematical question marks over their relation symbols. The mathematical content and source status are retained.

Reading-copy SHA-256 8f9ccf79fe0cbc4f89a58790e10130cf97b5561ba901ea29f371ffb387394236

Frozen-original SHA-256 8f9ccf79fe0cbc4f89a58790e10130cf97b5561ba901ea29f371ffb387394236

\section{A restricted positive Borel result from a uniform Hausdorff estimate}
\paragraph{Author and status.}
Independent derivation and audit by task \texttt{/root/b1\_positive\_routes}.
\texttt{PROVED-INTERNAL}, conditional only on the already audited original
local linear trace lemma. This is a restricted positive result, not the
general B1 theorem and not a claim of worldwide novelty.

Retain the fixed original $K$, absolute value, original $n$-torus with
$n\ge2$, the strict Laurent box $V$, normalized moments
$\tau:V\to Q=(-1,1)^n$, and the smooth strongly positive background
$S=\beta^{n-1}[X]$ used in the preceding sections. Thus $q=1$ and
\[
 \tau_*(Da\wedge S)=\gamma\Delta g_a,
 \qquad\gamma=(n-1)!\lambda^{n-2},
\]
for every original normalized tester $0\le a\le1$, with canonical convex
trace $0\le g_a\le1$. Fix a compact set $P\Subset Q$ and set
$d=\operatorname{dist}(P,\mathbf R^n\setminus Q)>0$.

\begin{lemma}[Uniform local ball and Hausdorff control]
There is a constant $M_{n,P}<\infty$, independent of the original tester,
such that every Borel $F\subset P$ satisfies
\[
 C_S^*(\tau^{-1}F;V)
       \le M_{n,P}\gamma\,\mathcal H^{n-1}(F).
 \tag{HR1}
\]
The Hausdorff measure has any one fixed standard normalization; its
normalizing factor is absorbed in $M_{n,P}$.
\end{lemma}
\begin{proof}
Every convex $g:Q\to[0,1]$ is Lipschitz with constant at most
$L=2/d$ on the $d/2$ neighborhood of $P$. Indeed, at a point of that
neighborhood, moving toward either sign of any unit vector by a distance
approaching $d/2$ remains in $Q$; the subgradient inequality bounds every
directional component of the subgradient by $2/d$.

Choose a fixed smooth nonnegative cutoff $\eta$ on $\mathbf R^n$, equal
to one on the unit ball and supported in the ball of radius two. For
$z\in P$ and $0<\rho<d/8$, let
$\eta_{z,\rho}(x)=\eta((x-z)/\rho)$. Positivity of the Radon measure
$\Delta g$ and distributional integration by parts imply
\[
 \begin{split}
 \Delta g(B(z,\rho))
 &\le\int\eta_{z,\rho}\,d\Delta g
   =-\int\nabla g\cdot\nabla\eta_{z,\rho}\,dx\\
 &\le L\|\nabla\eta\|_{L^1(\mathbf R^n)}\rho^{n-1}.
 \end{split}
 \tag{HR2}
\]
All supports remain in the protected region where the common Lipschitz
bound is valid. The identity is legitimate for locally Lipschitz $g$;
no trace theorem on sphere boundaries is used. The original linear trace
identity consequently gives
\[
 C_S(\tau^{-1}B(z,\rho);V)
       \le\gamma L\|\nabla\eta\|_1\rho^{n-1}.
 \tag{HR3}
\]

Cover $F$ by countably many open balls centered in $F$, of radii less
than $d/8$. The original open union of their inverseimages contains
$\tau^{-1}F$. For every original tester, its mass on that union is at
most the sum of its masses on the individual inverseimages, so (HR3)
bounds the raw capacity of this union by the sum of the right sides.
Taking the infimum over arbitrarily fine ball covers proves (HR1),
using the ball-cover characterization of Hausdorff measure and absorbing
its fixed dimensional normalization into $M_{n,P}$. The capacity being
estimated is always the original open outer capacity; it is not replaced
by a Hausdorff capacity or a smaller tester class.
\end{proof}

\begin{theorem}[Borel inner regularity for finite-Hausdorff saturated sets]
Let $B\subset P$ be Borel with $\mathcal H^{n-1}(B)<\infty$, and let
$E=\tau^{-1}B$. Then
\[
 \boxed{\quad C_S^*(E;V)
      =\sup_{H\subset E,\ H\text{ compact}}C_S^*(H;V).\quad}
 \tag{HR4}
\]
In fact compact inverseimages $H=\tau^{-1}F$, with $F\subset B$ compact,
already suffice for the supremum.
\end{theorem}
\begin{proof}
The finite Borel measure
$A\mapsto\mathcal H^{n-1}(A\cap B)$ on $\mathbf R^n$ is Radon.
Consequently, for every $\epsilon>0$, there is a compact
$F\subset B$ such that $\mathcal H^{n-1}(B\setminus F)<\epsilon$.
This uses ordinary regularity of a single finite Borel measure on
Euclidean space, not regularity of the capacity under investigation.

The moment map is proper over compact subsets of $Q$, hence
$\tau^{-1}F$ is an original compact subset of $E$. Countable
subadditivity of the original outer capacity and (HR1) give
\[
 C_S^*(E;V)
 \le C_S^*(\tau^{-1}F;V)
        +C_S^*(\tau^{-1}(B\setminus F);V)
 \le C_S^*(\tau^{-1}F;V)+M_{n,P}\gamma\epsilon.
\]
For clarity, the subadditivity used here follows directly from the
definition: choose original open supersets of the separate sets with
raw capacities arbitrarily close to their outer capacities, take their
union, and use countable subadditivity of every tester measure before
taking the tester supremum. It is ordinary subadditivity, not the
strong-subadditivity assertion refuted above.

Since (HR1) makes the left side finite, letting $\epsilon\downarrow0$
proves the upper bound by the compact supremum. The reverse bound is
monotonicity. No increasing-sequence continuity of the outer capacity
is used.
\end{proof}

\paragraph{Exact limitations.}
The theorem concerns full moment-saturated sets $\tau^{-1}B$, with $B$
inside one fixed protected compact stage and of finite
$\mathcal H^{n-1}$ measure, for this specific smooth strongly positive
background and remaining degree one. It does not cover arbitrary
Borel subsets of the original Berkovich space, arbitrary closed
backgrounds, or higher remaining degrees. In particular a Borel subset
of $\tau^{-1}B$ need not be the inverseimage of any Borel subset of $B$;
no such extension is asserted. Likewise $\sigma$-finiteness alone gives
no uniform small Hausdorff remainder and cannot be substituted for
finiteness without an additional argument. No global exhaustion or
boundary-tail theorem is obtained.

The estimate also gives outer capacity zero for every saturated
$\mathcal H^{n-1}$-null set in the protected stage. This is compatible
with the failure of strong subadditivity: regularity and submodularity
are distinct properties. General B1 remains OPEN.