Suppose I have an Artin stack $\mathfrak M$ locally of finite-type over $\mathbb C$ with presentation $M\rightarrow \mathfrak M$. Suppose further that $\mathfrak M$ "represents" (in the stack sense) some set of objects which are bounded, i.e. roughly there is a family of such objects $\mathcal F$ over a scheme $S$ of finite type over $\mathbb C$ such that every element of our set of objects occurs as $\mathcal F_s$ for some closed points $s\in S$. Why does it follow that $\mathfrak M$ is actually of finite-type over $\mathbb C$?

Specifically, I've seen it argued that $S$ induces a surjection $S\rightarrow M$ of schemes over $\mathbb C$, which thus makes $M$ of finite-type over $\mathbb C$. But I don't see why we get an induced map to $M$ instead of just to $\mathfrak M$. Indeed the family $\mathcal F$ should induce a surjective morphism $S\rightarrow \mathfrak M$, but I don't see why this should factor through $M$.

  • $\begingroup$ There is generally no such factorization (think about the case when $\mathfrak{M}$ is a scheme and $S = \mathfrak{M}$!). The argument you've seen should have the source replaced by $S \times_{\mathfrak{M}} M$ (which is, strictly speaking, an algebraic space and not necessarily a scheme, but nonetheless is of finite type over $M$ and has the "expected" set of $\mathbf{C}$-points, so you're good to go). $\endgroup$ – user30379 Apr 4 '13 at 20:49
  • $\begingroup$ Why is the fiber product of finite type over $M$? Are you claiming that since $S$ is of finite-type over $\mathbb C$, it must then also be of finite-type over $\mathfrak M$, and then base-change? $\endgroup$ – HNuer Apr 4 '13 at 21:03
  • 1
    $\begingroup$ Any map $f:X \rightarrow Y$ between quasi-separated algebraic spaces locally of finite type over an affine scheme $B$ with $X$ of finite type over $B$ is itself of finite type. Indeed, $Y$ is covered by q-c opens $U_i$, preimages of which are open in $X$, so finitely many preimages cover $X$. Hence, $f$ lands inside a quasi-compact open $U \subset Y$, and $U \rightarrow B$ is finite type (being q-c and lft), so $X \rightarrow U$ is finite type. Thus, we want $U \hookrightarrow Y$ to be finite type. For $V \rightarrow Y$ with $V$ affine, $U \times_Y V$ is q-c since $Y$ is quasi-separated. $\endgroup$ – user30379 Apr 4 '13 at 21:56

Consider the fiber product $S \times_{\frak M} M$; its projection onto $S$ is smooth and surjective, hence open. Since $S$ is quasi-compact, there exists a quasi-compact open subset $U$ of $S \times_{\frak M} M$ surjecting onto $S$. The image of $U$ in $M$ is quasi-compact; let $V$ be a quasi-compact open subscheme of $M$ containing it. Since $S$ surjects onto $\frak M$, so does $U$. So $V \to \frak M$ is smooth and surjective, and $V$ is of finite type, so $\frak M$ is of finite type.

  • $\begingroup$ How do you see the existence of the open subset $U$ of $M$ with the properties you claim? $\endgroup$ – HNuer Apr 4 '13 at 20:58
  • $\begingroup$ The map from the fiber product onto $S$ is smooth, hence open $\endgroup$ – Angelo Apr 4 '13 at 21:04
  • $\begingroup$ So is your $U$ an open of $M$ or $S\times_{\mathfrak M} M$? $\endgroup$ – HNuer Apr 4 '13 at 21:11
  • $\begingroup$ Yes, you are right, my explanation is incorrect. I edited the answer. $\endgroup$ – Angelo Apr 4 '13 at 21:19
  • $\begingroup$ Got it, nice slick argument. Thanks $\endgroup$ – HNuer Apr 4 '13 at 21:27

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.