NAG-003179 · Historical whole-proof gap
First curvature and its supported signed test measures
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}[First curvature and its supported signed test measures]\label{f9tf:first}
For $u\in\PP(U)$ define on every $\alpha\in\E^{n-1,n-1}(U)$
\begin{equation}\label{eq:first}
T_u(\alpha)=\int_U u\,d\mu_{U,\DD\alpha},
\end{equation}
using the frozen top-form measure of \NAG{002970}. This is a closed native current of bidimension $(n-1,n-1)$. It has the canonical signed measures $\nu_{u,\alpha}:=\nu_{T_u,\alpha}$ of Theorem~\ref{f9tf:measure}, and all continuous multipliers on it exist with the proved fixed-stage bounds. If $u$ is a convex piece of \eqref{eq:decomp}, $T_u$ is nonnegative on every test nonnegative on $(n-1)$-planes; for general $u$ it is locally a fixed finite signed sum of such currents.
On a local presentation consisting only of finite affine maxima, let $Z_u$ be the union of the inverse images of their non-affine loci in the ambient moment spaces. Then
\begin{equation}\label{eq:support}
\supp T_u\subseteq Z_u,\qquad
\supp\nu_{u,\alpha}\subseteq Z_u\cap\supp\alpha.
\end{equation}
No positivity is asserted for an arbitrary signed or model potential. For a pure affinoid $Y=\mathcal M(A)$ with reduced components $j_i:Y_i\hookrightarrow Y$ and $m_i=\operatorname{length}_{A_{\mathfrak q_i}}A_{\mathfrak q_i}$, on interior tests one has
\begin{equation}\label{eq:components}
T_u=\sum_i m_i(j_i)_*T_{u|Y_i},\qquad
\nu_{u,\alpha}=\sum_i m_i(j_i)_*\nu_{u|Y_i,j_i^*\alpha}.
\end{equation}
Thus the full generic nilpotent lengths, not merely the reduced components, remain in the scalar measures.
\end{theorem}