14
$\begingroup$

I have asked this question on Math.StackExchange, but haven't got any reply.

Suppose $X$ is a toric variety with fan $\Sigma$, and the lattice of one-parameter subgroups of its torus is $N$. Suppose $\phi$ is a representation of finite group $G$ in $N \otimes_{\mathbb{Z}} \mathbb{R}$, \begin{equation} \phi:G \rightarrow \text{GL}( N \otimes_{\mathbb{Z}} \mathbb{R}) \end{equation} which sends the lattice $N$ to $N$, and the cones of $\Sigma$ to cones of $\Sigma$, therefore each $\phi(g)$ defines a toric automorphism of $X$.

Question 1: Is the quotient variety $X/G$ (which exists from GIT) still a toric variety? Is the quotient morphism $X \rightarrow X/G$ toric?

Question 2: If so, how to construct the fan of $X/G$ and the toric morphism $X \rightarrow X/G$?

Any references on the two questions?

Improve this question
$\endgroup$
3
  • 5
    $\begingroup$ The quotient morphism is certainly never toric unless $\phi$ is trivial. Toric morphisms between toric varieties are those that commute with the torus action, while $\phi$ obviously doesn't. If $X/G$ is toric, it may be toric in more than one way, so there might be no construction of the toric fan. $\endgroup$
    – Will Sawin
    Apr 10, 2018 at 12:28
  • $\begingroup$ @WillSawin Sorry for bothering you with a question again after 10 days. Since $X$ is toric, it's defined over $\mathbb{Q}$. Even the quotient $X/G$ may not be toric, I guess it is still a variety defined over $\mathbb{Q}$? But I don't know how to prove this! $\endgroup$
    – Wenzhe
    Apr 20, 2018 at 11:56
  • $\begingroup$ Take a $G$-invariant ample line bundle. This is also defined over $\mathbb Q$. Then the quotient is the Proj of the ring of $G$-invariant sections. Because everything in sight is defined over $\mathbb Q$, this ring is defined over $\mathbb Q$. $\endgroup$
    – Will Sawin
    Apr 20, 2018 at 13:49

1 Answer 1

Reset to default
36
$\begingroup$

No. If $X$ is the 47-dimensional torus and $G=\mathbb{Z}/47\mathbb{Z}$ acting by the permutation representation, then $X/G$ is not a rational variety (R. Swan, Inv. math. 7, 148-158 (1969)), therefore not toric.

Improve this answer
$\endgroup$
1
  • 2
    $\begingroup$ The title of the cited paper is "Invariant rational functions and a problem of Steenrod" in case anybody wants to look it up $\endgroup$
    – Yosemite Stan
    Nov 26, 2019 at 7:54

Your Answer

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

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