0
$\begingroup$

This question is on my course of Lie Groups. I am struggling on finding a better proof than the one I suggest below:

Let $V$ be a $\mathbb{C}$ vector space of dimension $n$. Let $u,v \in End(V,\mathbb{C})$ such that $[u,v] = \alpha u + \beta v$ where $\alpha , \beta \in \mathbb{C}$. Show that there exists a basis in which the matrices of u and v are triangular.

My Attempt:

There exists a basis B such that the matrix of $[u,v]$ is triangular. Then,

$[u,[u,v]] = u(\alpha u + \beta v) - (\alpha u + \beta v)u = \beta uv - \beta vu = \beta [u,v]$

$[v,[u,v]] = -\alpha [u,v]$

are also triangular. Let $w = [u,v]$, then all the elements below the diagonal in $uw$ are equal to those in $wu$. However, if $w$ is triangular then $uw$ must also be. Which implies that $wu$ is triangular (Idem for $v$). I am not sure on how to proceed from here. Does this imply that $u$ is triangular?

Do we have to go through each element: Let $(e_{n1})$ be the $n1$ entry of $wu$ which is equal to the dot product of the $n^{th}$ row of $w$ and the $1^{st}$ column of $u$. But, the $n^{th}$ row of $w$ can only have one entry which is non zero (i.e. $(w_{nn}$)), hence, ($u_{n1}$) must also be zero. We can apply the same argument for the whole $n^{th}$ row of $u$. Then we use recursion for the other rows to show that $u$ is triangular.

Is there something wrong with my proof? Is there another proof with is more elegant? I find my method a bit clumsy. Thank you for your help.

$\endgroup$
0
$\begingroup$

The vectors $u$ and $v$ span a $2$-dimensional solvable complex Lie algebra. By passing to the (faithful) adjoint representation we may assume that $u$ and $v$ are matrices of size $2$. Actually, this is assumed here already anyway. By Lie's theorem, there is a basis such that $u$ and $v$ are simultaneously upper-triangular.

$\endgroup$

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.