5
$\begingroup$

Let $\mathcal{M}$ be a smooth 1-dimensional Deligne-Mumford stack with finite diagonal, and let $M$ be its coarse moduli scheme (all over some field $k$). Suppose $M$ is also smooth over $k$, and that all local deformation problems for $\mathcal{M}$ are pro-representable (I think this is already implied by the smooth + DM conditions?)

Let $x_0 : \text{Spec }\bar{k}\rightarrow\mathcal{M}$ a geometric point, and let $\bar{x_0}\in M$ be its image in $M$.

Let $\mathcal{R}$ be the completion of the etale local ring at $x_0$ (colimit of global sections of scheme etale neighborhoods of $x_0$), and let $R$ be the completion of the etale local ring of $M$ at $\bar{x_0}$.

This yields a natural map $R\rightarrow\mathcal{R}$ (both are isomorphic to a 1-variable power series ring over $\bar{k}$), whence a map $$p : \text{Spec }\mathcal{R}\rightarrow\text{Spec }R$$

I believe the infinitesimal lifting property of etale morphisms then implies that $\mathcal{R}$ is also the universal deformation ring of the object corresponding to $x_0$ (am I missing any conditions here?).

Let $X_0/\bar{k}$ be the object corresponding to $x_0$, and let $X^\text{univ}/\mathcal{R}$ be the universal deformation of $X_0$ over $\mathcal{R}$.

I would like to say something like: $p$ is finite flat totally ramified with ramification index equal the order of the group of automorphisms of $X_0$ modulo the subgroup of those automorphisms which extend to an automorphism of $X^\text{univ}/\mathcal{R}$.

Is this true? What is a good reference? If it's false, what is the relation between the universal deformation ring of $x_0$ and the complete etale local ring of $x_0$?

Actually, some calculations of mine seem to indicate that may be false, though it's possible I've made a mistake somewhere else...

$\endgroup$
2
$\begingroup$

I believe that this is true - that is, "$p$ is finite flat totally ramified with ramification index equal to the order of the group of automorphisms of $X_0$ modulo the subgroup of those which extend to an automorphism of $X^\text{univ}/\mathcal{R}$."

Specifically, let $G$ be the automorphism group of the object $x_0$. Let $\mathcal{R}$ be the etale local ring of $\mathcal{M}$ at $x_0$. Then there is a natural action of $G$ on $\mathcal{R}$. Let $R$ be the etale local ring of $\overline{x_0}\in M$. Let $\mathcal{M}_{(x_0)} := \mathcal{M}\times_M\text{Spec }R$. From the proof of Theorem 11.3.1 of Olsson's book "Algebraic Spaces and Stacks", we find that $$\mathcal{M}_{(x_0)} \cong [\text{Spec }\mathcal{R}/G]$$ and moreover the composition $$\text{Spec }\mathcal{R}\rightarrow[\text{Spec }\mathcal{R}/G]\rightarrow \text{Spec }R$$ is finite.

Since the map $\text{Spec }R\rightarrow M$ is flat, the projection $[\mathcal{R}/G]\cong\mathcal{X}_{(x_0)}\rightarrow\text{Spec }R$ identifies the target with the coarse moduli scheme of $[\mathcal{R}/G]$ (Theorem 11.1.2 of the same book). Thus, we have $R = \mathcal{R}/G = \mathcal{R}^G$. Since $M,\mathcal{M}$ are smooth and 1-dimensional, $R,\mathcal{R}$ are discrete valuation rings, and the map $R\rightarrow\mathcal{R}$ being a finite map between regular schemes must be flat (see Katz-Mazur's book, end of p507).

At this point standard ramification theory tells us that if $K$ is the kernel of the action of $G$ on $\mathcal{R}$, then since we're working with strictly henselian DVRs, the extension $\mathcal{R}/R$ is totally ramified with ramification index $|G/K|$.

Let $X_0$ be the object over $\overline{k}$ corresponding to $x_0$, and let $X/\mathcal{R}$ be the object corresponding to the canonical map $\text{Spec }\mathcal{R}\rightarrow\mathcal{M}$.

It remains to show that $K$ is precisely the subgroup of automorphisms of $X_0$ which extend to automorphisms of $X^\text{univ}/\mathcal{R}$. For this, we use the interpretation of $\mathcal{R}$ as the universal deformation ring which pro-represents the deformation functor $F_{X_0}$ which to every artinian local $\overline{k}$-algebra $A$ associates the set of isomorphism classes $$\{(X/A,\;\;\varphi : X_{\overline{k}}\cong X_0)\}/\cong$$ Then group $G$ acts naturally on these sets by acting on the isomorphism $\varphi$, ie $g\in G$ sends $(X,\varphi)$ to $(X,g\circ\varphi)$, and hence $G$ acts naturally on the functor $F_{X_0}$, and hence on the prorepresenting object $\mathcal{R}$. From this it is clear that an element $g\in G$ fixes $\mathcal{R}$ if and only if for every deformation $X/A$ over Artinian local $\overline{k}$-algebras $A$, there exists an extension of $g$ to an automorphism of $X/A$. Since $\mathcal{M}$ is assumed Deligne-Mumford, the diagonal is unramified (hence the automorphism group schemes of objects are unramified - in particular, formally unramified), so such extensions of $g$, if they exist, are unique. This implies that whenever $g$ extends to every $X/A$, the various extensions are compatible with each other, and hence gives an automorphism of $X^\text{univ}/\mathcal{R}$. Conversely, it's clear that automorphism of $X^\text{univ}/\mathcal{R}$ restricts to an automorphism $g\in G$ of $X_0$ which acts trivially on the deformation functor, and hence on $\mathcal{R}$.

$\endgroup$

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.