NAG-003190 · Historical whole-proof gap
All-test interchange of two declared potentials
Records without a chapter binding
Remaining historical proof gap
The sealed ledger records a gap in the proof of the historical whole statement.
Historical statement · exact TeX
\begin{theorem}[All-test interchange of two declared potentials]\label{f9tf:two}
For $u,v\in\PP(U)$ define
\begin{equation}\label{eq:two}
R(u,v)=\DD(vT_u),\qquad
R(u,v)(\alpha)=\int v\,d\nu_{u,\DD\alpha}
\quad(\alpha\in\E^{n-2,n-2}(U)).
\end{equation}
This is a closed native current of bidimension $(n-2,n-2)$ and
\begin{equation}\label{eq:interchange}
R(u,v)=R(v,u)
\quad\text{on every indicated compact native test}.
\end{equation}
It is bilinear, local, and independent of all decompositions and approximations used in the proof. It is nonnegative on $(n-2)$-plane nonnegative tests for two convex pieces, and locally a finite signed sum of such currents in general. Its support lies in $\supp T_u\cap\supp T_v$; for two finite-max presentations it lies in $Z_u\cap Z_v$. For the affinoid component data of \eqref{eq:components}, it also satisfies
\[
R(u,v)=\sum_i m_i(j_i)_*R(u|Y_i,v|Y_i)
\]
on interior tests, with the same original lengths.
Local potentials in $\PP$ of two actual continuously metrized analytic lines therefore give a frame-independent symmetric two-factor current. Replacing power metrics by positive real roots of exponents $e_1,e_2$ scales it by $1/(e_1e_2)$. No equality with an atomic model measure, divisor-current formula, or general height is asserted.
\end{theorem}