Let $G$ be a semisimple algebraic group over the complex numbers and we fix a maximal torus $T$. Let $w\in W$ be an element in the Weyl group, and let $T^{w}$ be the elements in $T$ that are fixed by $w$. Is $T^{w}$ still a torus? Is the Lie algebra of $T^{w}$ going to be $t^{w}$, the fixed points of $w$ acting on the Cartan subalgebra?


The first statement amounts to determining whether $T^w$ is connected. In the simplest case $G={\rm SL}_2$, $T$ may be identified with the multiplicative group ${\rm G}_m=\Bbb{C}^{\times}$, the Weyl group is cyclic of order 2 and the simple reflection $s$ acts by $x\mapsto x^{-1}$. Therefore, its fixed point set is $\{x: x^2=1\}=\{1,-1\}$, which is disconnected and hence not a torus.

The second statement is true and holds in a more general setting of an automorphism of a Lie group: If $H$ is a connected Lie group with Lie algebra ${\frak{h}}={\rm Lie}(H)$ and $\alpha: H\to H$ is an automorphism, then ${\rm Lie}(H^{\alpha})={\rm Lie}(H)^{\alpha}={\frak{h}}^{\alpha}$. This follows from the basic Lie correspondence between Lie groups and Lie algebras.

By the way, $T$ is a torus; "tori" is the plural form.

| cite | improve this answer | |
  • $\begingroup$ It may be just barely worth noting that the fact about Lie-groups and automorphisms is very much a characteristic-0 thing, that doesn't extend 'for free' to algebraic groups over imperfect fields. $\endgroup$ – LSpice Sep 5 '18 at 18:43

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.