Suppose $G$ is a connected complex algebraic group with lie algebra $\mathfrak{g}$ and $V$ a (potentially infinite-dimensional) vector space with actions of $G$ and $\mathfrak{g}$ which are 'compatible', i.e. for all $g\in G$, $X\in\mathfrak{g}$, $v\in V$, we have $$g\cdot X\cdot v=Ad(g)(X)\cdot g\cdot v$$ Note that I am not assuming that $V$ is a 'nice' representation of $G$, I'm only assuming it is a group action by its complex points, $G(\mathbb{C})$. Although see below for where the action originates.

My question is: suppose that $\mathfrak{g}$ stabilizes some finite-dimensional subspace $W\subseteq V$. Then must $G$ also stabilize $W$?

Actual motivation and setting of question: I have an action of $G$ on an algebraic variety $X$, and there is a finite-dimensional subspace $W$ of the field of rational functions on $X$, that the action of $\mathfrak{g}$ preserves, and I want to know if that implies $G$ must preserve it as well.

  • $\begingroup$ Your definition of compatibility is a bit weird. It says that the action of the Lie algebra on V is G-equivariant, and that is quite not good enough to get what you want. $\endgroup$ – Mariano Suárez-Álvarez Mar 29 '18 at 5:06
  • $\begingroup$ In your situation, you really have a lot more than you are using. Perhaps you should state precisely the situation you are actually interested in, instead of trying to axiomatize it? $\endgroup$ – Stephen Mar 29 '18 at 17:33
  • $\begingroup$ Thanks for your responses. I was away from technology for several days and so I'm only seeing this now. I think you are both right, and Dunay's answer below is correct as well. Clearly more hypotheses are needed... $\endgroup$ – freeRmodule Apr 2 '18 at 16:06

I see no one has answered your question, so I will try.

I think the answer is no. Just consider an any Lie group $G$ action on a vector space and have its lie algebra $\mathfrak{g}$ act trivially on it. Then the actions will be compatible. But, any subspace will be stabilized by $\mathfrak{g}$, although some may not be stabilized by $G$.

I think you need more hypothesis to get the result.

  • $\begingroup$ Thanks. I will continue trying to think about what else I need (and perhaps I should just roll with the actual hypotheses I have instead of attempting to abstract them). $\endgroup$ – freeRmodule Apr 2 '18 at 16:07

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.