I had a quick question about quotients of varieties. I am still not very good at them. Let $T$ be a torus, $\alpha$ a nontrivial character of $T$, and $S = (\textrm{Ker } \alpha)^0$. Since $T$ is commutative, $S$ is automatically normal in $T$, so I know that the quotient variety $T/S$ is an affine algebraic group. It is also connected, as a continuous image of $T$.

Let $G_m$ be the variety $k \setminus \{0\}$. Since $\alpha$ is a morphism of algebraic groups $T \rightarrow G_m$, we have that $G_m$ is a $T$-space, by the action $(a,x) \mapsto \alpha(a)x$. Since our field is algebraically closed, and $\alpha$ is nontrivial, we can argue further that $G_m$ is an equivariant $T$-space. Now $1 \in G_m$ has the property that $S$ is contained in its isotropy group (that is, $\textrm{Ker } \alpha$), so the universal property of quotient varieties implies that there is a unique equivariant morphism $\phi: T/S \rightarrow G_m$ such that $\phi(S) = 1$.

$\phi$ is given by the formula $\phi(aS) = \alpha(a)$, and also $\phi$ is surjective. Now I know that the dimension of $T$ is equal to the dimension of the kernel of $\alpha$ plus the dimension of the image of $\alpha$. And, the dimension of the kernel of $\alpha$ is the same as the dimension of $S$. Moreover, you can calculate the dimension of a quotient of varieties by subtracting the dimensions. So it follows that $\textrm{Dim } T/S = 1$.

But what I really want to conclude is that $T/S \cong G_m$ as algebraic groups. Should this be the case? This is what Springer says in 7.1.4.

  • $\begingroup$ Why are you taking the connected component? What is true is that $T/\ker \alpha\cong \mathbf{G}_m$, not modding out by the connected component. Did you consider the case when $T=\mathbf{G}_m$ (let's say in characteristic $0$)? Then for all non-trivial characters $\alpha$ one has that $(\ker \alpha)^\circ$ is trivial! $\endgroup$ – Alex Youcis May 7 '16 at 0:42
  • $\begingroup$ I am not sure why, that is just what Springer does in 7.1.4. He says "Fix $\alpha \in P'$. The group $G_{\alpha}$ contains the torus $S = (\textrm{Ker } \alpha)^0$...since $T/S$ is isomorphic to $G_m$ one has that $W_{\alpha}$ has order $\leq 2$." $\endgroup$ – D_S May 7 '16 at 2:20

Yes, since the only connected 1-dim algebraic groups are $G_a$ and $G_m$, and $G_a$ is unipotent, $G_m$ "semisimple".

  • $\begingroup$ Thanks. But why does $T/S$ consist of semisimple elements (e.g. why is $(T/S)_s = T/S$)? $\endgroup$ – D_S May 17 '16 at 0:04
  • 1
    $\begingroup$ canonical morphism $T \to T/S$ maps semisimple elements to semisimple elements. $\endgroup$ – yisishoujo May 17 '16 at 2:42
  • $\begingroup$ That makes sense, thanks very much for answering. $\endgroup$ – D_S May 17 '16 at 3:18

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.