4
$\begingroup$

In my quest of "understanding" stacks, I recently tried to figure out the structure of a smooth algebraic stack of finite type $\mathcal X$ over $\mathbb C$ with affine diagonal and precisely one $\mathbb C$-object up to isomorphism. Unless I'm not mistaken, it is not hard to see that $\mathcal X$ is isomorphic to $BG$ for some affine finite type (smooth) group scheme $G$ over $\mathbb C$.

On the other hand, it seems natural to wonder about the following "exercise":

What is the structure of a smooth algebraic stack of finite type $\mathcal X$ over $\mathbb C$ with affine diagonal and precisely two $\mathbb C$-objects up to isomorphism?

It could be the disjoint union of $BG$ with another $BG'$, but it could also be a connected stack such as $[\mathbb A^1/\mathbb G_m]$. Are these the only possibilities up to isomorphism? Or is there more?

$\endgroup$
3
  • 4
    $\begingroup$ What do you mean by "such as"? There are many smooth varieties with an action of a linear algebraic group having precisely two orbits. $\endgroup$ Feb 16, 2016 at 19:49
  • 1
    $\begingroup$ You can say something if you add a couple of additional hypotheses to your $\mathcal X$. For instance, if $\mathcal X$ has a dense open non-stacky $\mathbb C$-point and has reductive stabilizers, then $\mathcal X$ is isomorphic to $[X/T]$ with $X$ an affine toric variety and $T$ its torus. (The hypotheses on $\mathcal X$ are equivalent to saying that $\mathcal X$ is generically an algebraic space and that the unique stacky point has a reductive stabilizer.) This is a consequence of Theorem 4.5 of Geraschenko-Satriano's Toric Stacks II; see arxiv.org/abs/1107.1907 $\endgroup$ Feb 17, 2016 at 9:11
  • 1
    $\begingroup$ I think that an $\mathcal X$ as in the question is isomorphic to a gerbe over $[\mathbb A^1/\mathbb G_m]$. Indeed, if the generic stabilizer is trivial and the geometric stabilizers are reductive, then one can use (up to checking the details) Geraschenko-Satriano's main theorem to see that $\mathcal X$ is isomorphic to $[\mathbb A^1/\mathbb G_m]$. $\endgroup$ Feb 17, 2016 at 18:26

1 Answer 1

2
$\begingroup$

Here's what you can say (following my comments above).

Lemma. Let $X$ be a smooth finite type connected algebraic stack over $\mathbb C$ with precisely two $\mathbb C$-objects (up to isomorphism). Suppose that

1) the diagonal of $X$ is affine;

2) the stabilizers of all geometric points of $X$ are reductive; and

3) the stack $X$ has a dense open non-stacky point.

Then $X\cong [\mathbb A^1/\mathbb G_m]$.

Proof. This follows from the proof of the main theorem of Geraschenko-Satriano given in Toric Stacks II; see arxiv.org/abs/1107.1907. Indeed, under our assumptions, $X$ has precisely one divisor. The natural open immersion associated to this Cartier divisor $X\to [\mathbb A^1/\mathbb G_m]$ is shown to be an isomorphism in their proof by using the local structure theorem of their paper (Theorem 4.5 in loc. cit.). QED

Remark. $[\mathbb P^1/\mathbb G_a]$ has precisely two objects, and it's not isomorphic to $[\mathbb A^1/\mathbb G_m]$ because condition 2) fails.

Remark. If you relax condition 3, then you can use "rigidification" to see that $X$ is a gerbe over $[\mathbb A^1/\mathbb G_m]$ for some reductive group $G$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.