## 105.12 Moduli spaces

This section discusses morphisms $f : \mathcal{X} \to Y$ from algebraic stacks to algebraic spaces. Under suitable hypotheses $Y$ is called a moduli space for $\mathcal{X}$. If $\mathcal{X} = [U/R]$ is a presentation, then we obtain an $R$-invariant morphism $U \to Y$ and (under suitable hypotheses) $Y$ is a quotient of the groupoid $(U, R, s, t, c)$. A discussion of the different types of quotients can be found starting with Quotients of Groupoids, Section 82.1.

Definition 105.12.1. Let $\mathcal{X}$ be an algebraic stack. Let $f : \mathcal{X} \to Y$ be a morphism to an algebraic space $Y$.

1. We say $f$ is a categorical moduli space if any morphism $\mathcal{X} \to W$ to an algebraic space $W$ factors uniquely through $f$.

2. We say $f$ is a uniform categorical moduli space if for any flat morphism $Y' \to Y$ of algebraic spaces the base change $f' : Y' \times _ Y \mathcal{X} \to Y'$ is a categorical moduli space.

Let $\mathcal{C}$ be a full subcategory of the category of algebraic spaces.

1. We say $f$ is a categorical moduli space in $\mathcal{C}$ if $Y \in \mathop{\mathrm{Ob}}\nolimits (\mathcal{C})$ and any morphism $\mathcal{X} \to W$ with $W \in \mathop{\mathrm{Ob}}\nolimits (\mathcal{C})$ factors uniquely through $f$.

2. We say is a uniform categorical moduli space in $\mathcal{C}$ if $Y \in \mathop{\mathrm{Ob}}\nolimits (\mathcal{C})$ and for every flat morphism $Y' \to Y$ in $\mathcal{C}$ the base change $f' : Y' \times _ Y \mathcal{X} \to Y'$ is a categorical moduli space in $\mathcal{C}$.

By the Yoneda lemma a categorical moduli space, if it exists, is unique. Let us match this with the language introduced for quotients.

Lemma 105.12.2. Let $(U, R, s, t, c)$ be a groupoid in algebraic spaces with $s, t : R \to U$ flat and locally of finite presentation. Consider the algebraic stack $\mathcal{X} = [U/R]$. Given an algebraic space $Y$ there is a $1$-to-$1$ correspondence between morphisms $f : \mathcal{X} \to Y$ and $R$-invariant morphisms $\phi : U \to Y$.

Proof. Criteria for Representability, Theorem 96.17.2 tells us $\mathcal{X}$ is an algebraic stack. Given a morphism $f : \mathcal{X} \to Y$ we let $\phi : U \to Y$ be the composition $U \to \mathcal{X} \to Y$. Since $R = U \times _\mathcal {X} U$ (Groupoids in Spaces, Lemma 77.22.2) it is immediate that $\phi$ is $R$-invariant. Conversely, if $\phi : U \to Y$ is an $R$-invariant morphism towards an algebraic space, we obtain a morphism $f : \mathcal{X} \to Y$ by Groupoids in Spaces, Lemma 77.23.2. You can also construct $f$ from $\phi$ using the explicit description of the quotient stack in Groupoids in Spaces, Lemma 77.24.1. $\square$

Lemma 105.12.3. With assumption and notation as in Lemma 105.12.2. Then $f$ is a (uniform) categorical moduli space if and only if $\phi$ is a (uniform) categorical quotient. Similarly for moduli spaces in a full subcategory.

Proof. It is immediate from the $1$-to-$1$ correspondence established in Lemma 105.12.2 that $f$ is a categorical moduli space if and only if $\phi$ is a categorical quotient (Quotients of Groupoids, Definition 82.4.1). If $Y' \to Y$ is a morphism, then $U' = Y' \times _ Y U \to Y' \times _ Y \mathcal{X} = \mathcal{X}'$ is a surjective, flat, locally finitely presented morphism as a base change of $U \to \mathcal{X}$ (Criteria for Representability, Lemma 96.17.1). And $R' = Y' \times _ Y R$ is equal to $U' \times _{\mathcal{X}'} U'$ by transitivity of fibre products. Hence $\mathcal{X}' = [U'/R']$, see Algebraic Stacks, Remark 93.16.3. Thus the base change of our situation to $Y'$ is another situation as in the statement of the lemma. From this it immediately follows that $f$ is a uniform categorical moduli space if and only if $\phi$ is a uniform categorical quotient. $\square$

Lemma 105.12.4. Let $f : \mathcal{X} \to Y$ be a morphism from an algebraic stack to an algebraic space. If for every affine scheme $Y'$ and flat morphism $Y' \to Y$ the base change $f' : Y' \times _ Y \mathcal{X} \to Y'$ is a categorical moduli space, then $f$ is a uniform categorical moduli space.

Proof. Choose an étale covering $\{ Y_ i \to Y\}$ where $Y_ i$ is an affine scheme. For each $i$ and $j$ choose a affine open covering $Y_ i \times _ Y Y_ j = \bigcup Y_{ijk}$. Set $\mathcal{X}_ i = Y_ i \times _ Y \mathcal{X}$ and $\mathcal{X}_{ijk} = Y_{ijk} \times _ Y \mathcal{X}$. Let $g : \mathcal{X} \to W$ be a morphism towards an algebraic space. Then we consider the diagram

$\xymatrix{ \mathcal{X}_ i \ar[r] \ar[d] & \mathcal{X} \ar[d] \ar[r]_ g & W \\ Y_ i \ar[r] \ar@{..>}[rru] & Y }$

The assumption that $\mathcal{X}_ i \to Y_ i$ is a categorical moduli space, produces a unique dotted arrow $h_ i : Y_ i \to W$. The assumption that $\mathcal{X}_{ijk} \to Y_{ijk}$ is a categorical moduli space, implies the restriction of $h_ i$ and $h_ j$ to $Y_{ijk}$ are equal. Hence $h_ i$ and $h_ j$ agree on $Y_ i \times _ Y Y_ j$. Since $Y = \coprod Y_ i / \coprod Y_ i \times _ Y Y_ j$ (by Spaces, Section 64.9) we conclude that there is a unique morphism $Y \to W$ through which $g$ factors. Thus $f$ is a categorical moduli space. The same argument applies after a flat base change, hence $f$ is a uniform categorical moduli space. $\square$

## 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 0DUF. Beware of the difference between the letter 'O' and the digit '0'.