NAG-003015 · Accepted whole
Exact boundary image and conductor descent
Proper direct image on historical ordinary compact tests
Open in the English reader · PDF page 1882
Statement · exact TeX
\begin{proposition}[Exact boundary image and conductor descent]\label{prop:conductor}
For the $Y,V,B,S$ of Theorem~\ref{thm:lattice}, set
\[
\begin{aligned}
E_S&=\{f\in F^\times: \ord_D(f)=0
\text{ for every codimension-one }D\notin S\},\\
P_S&=\{(\ord_D f)_{D\in S}:f\in E_S\}\subset\Z^S.
\end{aligned}
\]
Then $E_S=B^\times$ and $B^\times/K^\times\simeq P_S$. The exact image of $M_A$ is
\[
L_A=\{(\ord_D f)_{D\in S}: f\in B^\times,\ f,f^{-1}\in A\}.
\tag{2}\label{eq:exactimage}
\]
If $A\ne B$, let $I=\{a\in A:aB\subset A\}$ be the conductor. It is a nonzero common ideal of $A$ and $B$, and (\ref{eq:exactimage}) can equivalently be specified by
\[
f\in B^\times,\qquad
\bar f\in(A/I)^\times\subset(B/I)^\times.
\tag{3}\label{eq:conductor}
\]
For $A=B$, the descent condition is vacuous and $L_A=P_S$.
\end{proposition}Full selected proof · exact TeX
\begin{proof}
The height-one points of $V$ are exactly the height-one points of $Y$ lying in $V$, hence those outside $S$. Apply \tagref{031T}(2) to $B$, both to $f$ and to $f^{-1}$. This proves $E_S=B^\times$. The kernel of the order map is $K^\times$ by the same properness argument as above. Formula (\ref{eq:exactimage}) follows because a function and its inverse descend to $A$ precisely when the function is a unit of $A$.
Since $B$ is finite over $A$ and contained in $F$, a common nonzero denominator of finite $A$-module generators of $B$ belongs to $I$. If $a\in I$ and $b\in B$, then $ab\in A$ and $(ab)B\subset A$, so $I$ is also an ideal of $B$. There is an equality of subrings of $B$,
\[
A=B\mathbin{\times}_{B/I}A/I.
\]
Indeed, a residue class of $b\in B$ lies in $A/I$ if and only if $b$ differs from an element of $A$ by an element of $I\subset A$. For $f\in B^\times$, condition (\ref{eq:conductor}) therefore puts both $f$ and $f^{-1}$ in $A$, and the converse is immediate. Multiplication by a constant does not change this condition, so it is well-defined on boundary vectors.
\end{proof}Typeset reader excerpt

Typed proof inputs
- NAG-003013 · selected whole (earlier same-packet whole)
Reader cross-references
Proof map
Verified map inputs · NAG-003015
Registered proof input
- NAG-003013 · Boundary construction of the actual unit lattice · accepted whole
- Use the very same normalization, affine chart, boundary and injective order map. The proof inherits its model and its constant-kernel argument.
NAG-003015 · Exact boundary image and conductor descent · Chapter argument