6
$\begingroup$

Let $T_1$ and $T_2$ be two Grothendieck topologies on the same small category $C$, and let $T_3 = T_1 ∪ T_2$ (by which I mean the smallest Grothendieck topology on C containing $T_1$ and $T_2$).

Assume that every $T_3$ cover has a refinement of the form $\stackrel{f_1}{\to} \stackrel{f_2}{\to}$ where $f_2$ is a $T_2$ cover and $f_1$ is a $T_1$ cover (this is true for example, if $T_1$ is the Nisnevich topology, and $T_2$ is the topology consisting of proper cdh covers; in this case $T_3$ is the cdh topology).

If $F$ is a $T_1$-sheaf, is its $T_2$-sheafification still a $T_1$-sheaf (and therefore a $T_3$-sheaf)?

I can show this is true if $F$ is a $T_2$ separated $T_1$ sheaf, but I want to remove this condition.

$\endgroup$
1

1 Answer 1

1
$\begingroup$

How about this for a counter-example: We take the category

$$ A \stackrel{f_2}{\to} B \rightrightarrows C \stackrel{f_1}{\to} D $$

where there is a unique morphism from $A$ to $C$ and from $B$ to $D$. The only $T_1$ cover is $f_1$ and the only $T_2$ cover is $f_2$.

This category and these topologies satisfy my conditions.

Consider the $T_1$ sheaf $F$:

$F(A) = $ a singleton set $\underline{0} = ${$0$}.

$F(B) = $ a two element set $\underline{1} = ${$0, 1$}.

$F(C) = $ a singleton set $\underline{0} = ${$0$}.

$F(D) = \varnothing$ the empty set.

$$ \underline{0} \leftarrow \underline{1} \leftleftarrows \underline{0} \leftarrow \varnothing $$

The two morphisms $F(B) \leftleftarrows F(C)$ are the two distinct morphims $\underline{1} \leftleftarrows \underline{0}$. This is a $T_1$ sheaf.

Then the $T_2$ separated presheaf (which happens to already by a $T_2$ sheaf) associated to this $T_1$ sheaf is

$$ \underline{0} \leftarrow \underline{0} \leftleftarrows \underline{0} \leftarrow \varnothing $$

which is no longer a $T_1$ sheaf.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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