The Stacks project

90.18 Schlessinger's theorem on prorepresentable functors

We deduce Schlessinger's theorem characterizing prorepresentable functors on $\mathcal{C}_\Lambda $.

Lemma 90.18.1. Let $F, G: \mathcal{C}_\Lambda \to \textit{Sets}$ be deformation functors. Let $\varphi : F \to G$ be a smooth morphism which induces an isomorphism $d\varphi : TF \to TG$ of tangent spaces. Then $\varphi $ is an isomorphism.

Proof. We prove $F(A) \to G(A)$ is a bijection for all $A \in \mathop{\mathrm{Ob}}\nolimits (\mathcal{C}_\Lambda )$ by induction on $\text{length}_ A(A)$. For $A = k$ the statement follows from the assumption that $F$ and $G$ are deformation functors. Suppose that the statement holds for rings of length less than $n$ and let $A'$ be a ring of length $n$. Choose a small extension $f : A' \to A$. We have a commutative diagram

\[ \xymatrix{ F(A') \ar[r] \ar[d]_{F(f)} & G(A') \ar[d]^{G(f)} \\ F(A) \ar[r]^{\sim } & G(A) } \]

where the map $F(A) \to G(A)$ is a bijection. By smoothness of $F \to G$, $F(A') \to G(A')$ is surjective (Lemma 90.8.8). Thus we can check bijectivity by checking it on fibers $F(f)^{-1}(x) \to G(f)^{-1}(\varphi (x))$ for $x \in F(A)$ such that $F(f)^{-1}(x)$ is nonempty. These fibers are precisely $\text{Lift}(x, A')$ and $\text{Lift}(\varphi (x), A')$ and by assumption we have an isomorphism $d\varphi \otimes \text{id} : TF \otimes _ k \mathop{\mathrm{Ker}}(f) \to TG \otimes _ k \mathop{\mathrm{Ker}}(f)$. Thus, by Lemma 90.17.5 and Remark 90.17.6, for $x \in F(A)$ such that $F(f)^{-1}(x)$ is nonempty the map $F(f)^{-1}(x) \to G(f)^{-1}(\varphi (x))$ is a map of sets commuting with free transitive actions by $TF \otimes _ k \mathop{\mathrm{Ker}}(f)$. Hence it is bijective. $\square$

Note that in case $k' \subset k$ is separable condition (c) in the theorem below is empty.

Theorem 90.18.2. Let $F: \mathcal{C}_\Lambda \to \textit{Sets}$ be a functor. Then $F$ is prorepresentable if and only if (a) $F$ is a deformation functor, (b) $\dim _ k TF$ is finite, and (c) $\gamma : \text{Der}_\Lambda (k, k) \to TF$ is injective.

Proof. Assume $F$ is prorepresentable by $R \in \widehat{\mathcal{C}}_\Lambda $. We see $F$ is a deformation functor by Example 90.16.10. We see $\dim _ k TF$ is finite by Example 90.11.11. Finally, $\text{Der}_\Lambda (k, k) \to TF$ is identified with $\text{Der}_\Lambda (k, k) \to \text{Der}_\Lambda (R, k)$ by Example 90.11.14 which is injective because $R \to k$ is surjective.

Conversely, assume (a), (b), and (c) hold. By Lemma 90.16.6 we see that (S1) and (S2) hold. Hence by Theorem 90.15.5 there exists a minimal versal formal object $\xi $ of $F$ such that ( holds. Say $\xi $ lies over $R$. The map

\[ d\underline{\xi } : \text{Der}_\Lambda (R, k) \to T\mathcal{F} \]

is bijective on $\text{Der}_\Lambda (k, k)$-orbits. Since the action of $\text{Der}_\Lambda (k, k)$ on the left hand side is free by (c) and Lemma 90.12.6 we see that the map is bijective. Thus we see that $\underline{\xi }$ is an isomorphism by Lemma 90.18.1. $\square$

Comments (0)

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 06JK. Beware of the difference between the letter 'O' and the digit '0'.