let $S$ be a commutative ring and $I_1,...,I_n\unlhd S$, such that $I_i+I_j=S\ \forall i\neq j$.

Let $g_1,...,g_n\in S$.

Why are there $h_1,...,h_n,h'\in S$, such that

$h_i\big|_{V(I_i)}=g_i\big|_{V(I_i)}$ and $h_i\big|_{V(I_j)}=0\ \forall i\neq j$ and $h'-g_i\in I_i\ \forall i$ hold?

Note that $\widetilde{S}$, the étalé space belonging to $S$, is a sheaf.

Thanks for the help.

  • 1
    $\begingroup$ Are you trying to use CRT to prove your statement?. Because if it is, then the existence of $h_i$ and $h'$ follows trivially from CRT applied respectively to the $n$-tuples $(0,\ldots,0,g_i,0,\ldots,0)$ and $(g_1,\ldots,g_n)$ $\endgroup$ – Diego Nov 5 '14 at 21:04
  • $\begingroup$ No, I would like to prove the CRT by using the fact that $\widetilde{S}$ is a sheaf. $\endgroup$ – Leopold Schirmer Nov 5 '14 at 22:55

Here there is an elemental result on sheaves from Godement's book "Topologie algebrique et theorie des faisceaux", Theorem 1.3.1:

$\bf{Theorem}:$ Let $\mathscr{F}$ be a sheaf on a topological space $X$ and let $(M_i)_{i\in I}$ be a locally finte covering of $X$ formed by closed subsets. Suppose we have sections $s_i\in\mathscr{F}(M_i)$ such that $s_i=s_j$ in $M_i\cap M_j$ for all $i,j\in I$. Then there exists a section $s\in\mathscr{F}(X)$ which concides with $s_i$ in $M_i$ for all $i\in I$.

In your case let $X=\cup_{i=1}^n V(I_i)=V(\cap_{i=1}^n I_i)$. The hypothesis on the $I_i$'s implies that $V(I_i)\cap V(I_j)=\emptyset$ for $i\neq j$ and therefore we do not need to check conditions "$s_i=s_j$ in $V(I_i)\cap V(I_j)$" if we want to apply the theorem in this setting. Then CRT follows easily: for the existence of $h'$, let $s\in \tilde{S}(X)$ such that $s=g_i$ in $V(I_i)$ for all $i$ (this follows from the Theorem). The section $s\in \tilde{S}(X)=\tilde{S}(V(\cap_{i=1}^n))$ is the restriction of a section $h'\in \tilde{S}(Spec(S))=S$. It follows that $h'=g_i$ in $V(I_i)$ for all $i$, and this is equivalent to $h'-g'\in I_i$ for all $i$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.