The Stacks project

Lemma 102.3.7. Let $S$ be a scheme. Let $\mathcal{X}$ be an algebraic stack over $S$. If $\mathcal{X} \to S$ is locally of finite presentation, then $\mathcal{X}$ is limit preserving in the sense of Artin's Axioms, Definition 98.11.1 (equivalently: the morphism $\mathcal{X} \to S$ is limit preserving).

Proof. Choose a surjective smooth morphism $U \to \mathcal{X}$ for some scheme $U$. Then $U \to S$ is locally of finite presentation, see Morphisms of Stacks, Section 101.27. We can write $\mathcal{X} = [U/R]$ for some smooth groupoid in algebraic spaces $(U, R, s, t, c)$, see Algebraic Stacks, Lemma 94.16.2. Since $U$ is locally of finite presentation over $S$ it follows that the algebraic space $R$ is locally of finite presentation over $S$. Recall that $[U/R]$ is the stack in groupoids over $(\mathit{Sch}/S)_{fppf}$ obtained by stackyfying the category fibred in groupoids whose fibre category over $T$ is the groupoid $(U(T), R(T), s, t, c)$. Since $U$ and $R$ are limit preserving as functors (Limits of Spaces, Proposition 70.3.10) this category fibred in groupoids is limit preserving. Thus it suffices to show that fppf stackyfication preserves the property of being limit preserving. This is true (hint: use Topologies, Lemma 34.13.2). However, we give a direct proof below using that in this case we know what the stackyfication amounts to.

Let $T = \mathop{\mathrm{lim}}\nolimits T_\lambda $ be a directed limit of affine schemes over $S$. We have to show that the functor

\[ \mathop{\mathrm{colim}}\nolimits [U/R]_{T_\lambda } \longrightarrow [U/R]_ T \]

is an equivalence of categories. Let us show this functor is essentially surjective. Let $x \in \mathop{\mathrm{Ob}}\nolimits ([U/R]_ T)$. In Groupoids in Spaces, Lemma 78.24.1 the reader finds a description of the category $[U/R]_ T$. In particular $x$ corresponds to an fppf covering $\{ T_ i \to T\} _{i \in I}$ and a $[U/R]$-descent datum $(u_ i, r_{ij})$ relative to this covering. After refining this covering we may assume it is a standard fppf covering of the affine scheme $T$. By Topologies, Lemma 34.13.2 we may choose a $\lambda $ and a standard fppf covering $\{ T_{\lambda , i} \to T_\lambda \} _{i \in I}$ whose base change to $T$ is equal to $\{ T_ i \to T\} _{i \in I}$. For each $i$, after increasing $\lambda $, we can find a $u_{\lambda , i} : T_{\lambda , i} \to U$ whose composition with $T_ i \to T_{\lambda , i}$ is the given morphism $u_ i$ (this is where we use that $U$ is limit preserving). Similarly, for each $i, j$, after increasing $\lambda $, we can find a $r_{\lambda , ij} : T_{\lambda , i} \times _{T_\lambda } T_{\lambda , j} \to R$ whose composition with $T_{ij} \to T_{\lambda , ij}$ is the given morphism $r_{ij}$ (this is where we use that $R$ is limit preserving). After increasing $\lambda $ we can further assume that

\[ s \circ r_{\lambda , ij} = u_{\lambda , i} \circ \text{pr}_0 \quad \text{and}\quad t \circ r_{\lambda , ij} = u_{\lambda , j} \circ \text{pr}_1, \]

and

\[ c \circ (r_{\lambda , jk} \circ \text{pr}_{12}, r_{\lambda , ij} \circ \text{pr}_{01}) = r_{\lambda , ik} \circ \text{pr}_{02}. \]

In other words, we may assume that $(u_{\lambda , i}, r_{\lambda , ij})$ is a $[U/R]$-descent datum relative to the covering $\{ T_{\lambda , i} \to T_\lambda \} _{i \in I}$. Then we obtain a corresponding object of $[U/R]$ over $T_\lambda $ whose pullback to $T$ is isomorphic to $x$ as desired. The proof of fully faithfulness works in exactly the same way using the description of morphisms in the fibre categories of $[U/T]$ given in Groupoids in Spaces, Lemma 78.24.1. $\square$


Comments (2)

Comment #2251 by Niels Borne on

There is a typo at the start of the third sentence of the proof.


Post a comment

Your email address will not be published. Required fields are marked.

In your comment you can use Markdown and LaTeX style mathematics (enclose it like $\pi$). A preview option is available if you wish to see how it works out (just click on the eye in the toolbar).

Unfortunately JavaScript is disabled in your browser, so the comment preview function will not work.

All contributions are licensed under the GNU Free Documentation License.




In order to prevent bots from posting comments, we would like you to prove that you are human. You can do this by filling in the name of the current tag in the following input field. As a reminder, this is tag 0CMX. Beware of the difference between the letter 'O' and the digit '0'.