The question is about Proposition 3.8.1 in Laumon and Moret-Bailly book on algebraic stacks.

Let $S$ be a scheme and let $F: \mathscr{X} \rightarrow \mathscr{Y}$ be a morphism of $S$-stacks (for the etale topology, say). Let $F': \mathscr{X}' \rightarrow \mathscr{Y}'$ be the base change of $F$ along an epimorphism of $S$-stacks $Q\colon \mathscr{Y}' \rightarrow \mathscr{Y}$. Suppose that $F'$ is a monomorphism. Is $F$ then necessarily a monomorphism, too?

This is claimed in the reference cited above, but the proof given there is unclear to me. In the proof they seem to use that $\mathscr{X}' \times_{\mathscr{Y}'} \mathscr{X}' \rightarrow \mathscr{X} \times_{\mathscr{Y}} \mathscr{X}$ is an epimorphism, but I cannot see why this is so, the subtlety being that these are $2$-fiber products, so objects take into account isomorphisms in $\mathscr{Y}'$ and $\mathscr{Y}$ and we cannot a priori lift an isomorphism in a fiber category of $\mathscr{Y}$ to $\mathscr{Y}'$.


The proof given in Laumon and Moret-Bailly is clear and does not seem to use your claim.

Denote by $G=\Delta_F :\mathcal X \to \mathcal X \times_{\mathcal Y} \mathcal X$ the diagonal morphism. The main points used are

A. $F$ is representable iff $\Delta_F$ is a mono (not really used, but good to know)

B. $F$ is a mono iff $\Delta_F$ is an iso

C. $\Delta_G$ is always a mono

for proofs of these facts you can have a look at http://stacks.math.columbia.edu/tag/0AHJ

D. The formation of $\Delta_F$ commutes with based change

E. A morphism of stacks is an iso iff it is a mono and an epi

F. $F'$ is an epi iff $F$ is

And this is enough :

First case : assume $F$ is representable, i.e. $\Delta_F$ is a mono. Then if $F'$ is a mono, $\Delta_{F'}$ is an iso, hence an epi, hence $\Delta_F$ is a epi as well, hence an iso, so $F$ is a mono.

General case : First note that since $F'$ is a mono by hypothesis, $G'$ is an iso. Since $G$ is representable, we deduce from the first case that $G$ is an iso as well, that is, $F$ is a mono.

  • $\begingroup$ Thank you, but I don't completely understand your argument for the same reason that made me question the original proof. Namely, your F. is not an 'if and only if' in general (to get $F'$ epi $\Rightarrow$ $F$ epi you need that $Q$ is an epi). This then seems to be a problem at the step "$\Delta_{F'}$ is epi, hence $\Delta_F$ is epi" precisely for the reason that we don't know that the new $Q$ (written out in my question) is an epi! Could you clarify? $\endgroup$ – O-Ren Ishii Apr 25 '15 at 16:07
  • $\begingroup$ You are simply confused, because there is no such new $Q$. That is my point D. $\endgroup$ – Niels Apr 25 '15 at 19:49
  • 1
    $\begingroup$ You are right! The key point that I overlooked is that $\mathscr{X}' \times_{\mathscr{Y}'} \mathscr{X}' \rightarrow \mathscr{X} \times_{\mathscr{Y}} \mathscr{X}$ is itself a base change along $\mathscr{Y}' \rightarrow \mathscr{Y}$. Thank you very much for taking time to help me! This has been extremely useful. (I would vote your answer up if I had enough reputation to do that.) $\endgroup$ – O-Ren Ishii Apr 25 '15 at 20:29

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.