0
$\begingroup$

I am trying to understand the proof of the following claim (see A.L. Onishchik, E.B. Vinberg (Eds.) Lie Groups and Lie Algebras III, p.50, Theorem 3.1).

Theorem 3.1 (ii) If the Lie group $G$ is solvable, then it has a connected normal Lie subgroup of codimension 1.

Proof.

Consider the Lie subgroup $G'_1:= \overline{[G,G]}$, the closure of the commutator group $[G,G]$. The solvability of the Lie group $G$ implies that $G_1' \ne G$. The group $G/G_1'$ is abelian, and evidently has a connected Lie subgroup $A'$ of codimension $1$. If $\pi: G \to G/G_1'$ is the natural epimorphism, then $G_1: = \pi^{-1}(A')$ is the desired normal Lie subgroup of $G$. Q.E.D.

So, I don't understand why any abelian Lie group has a connected Lie subgroup of codimension $1$. I tried to think as follows: maybe there is a nice Lie homomorphism $f:A \to A$ that has a constant rank $1$ and thus its kernel is a subgroup of the desired type but I cannot see this homomorphism in general.

$\endgroup$

1 Answer 1

1
$\begingroup$

I think I've got it. As I understood any $n$-dimensional connected abelian Lie group $A$ has a form $\mathbb{T}^k \times \mathbb{R}^{n-k}$ and thus we get set $B:= \mathbb{T}^k \times \mathbb{R}^{n-k-1}$ if $n-k >0$ and $B:=\mathbb{T}^{n-1}\times\{e\}$ in otherwise and we are done, i.e,. $B$ is a subgroup of codimension $1$.

$\endgroup$
4
  • 2
    $\begingroup$ A perhaps slightly more ‘conceptual’ explanation is that the exponential map is a submersive homomorphism, and the image of a codimension-$1$ subspace of the Lie algebra (which is also a subalgebra in this case) will therefore be a codimension-$1$ subgroup. $\endgroup$
    – LSpice
    Commented Mar 13, 2022 at 18:09
  • $\begingroup$ Dear LSpice, thank you very much. As I undertand, you suggest to take any vector subspace of codimension 1 (being an abelian subalgebra and consider the corresponding exp image.) If so, what about irrational winding of a torus? Can we get it also by this way? $\endgroup$
    – Viktor
    Commented Mar 13, 2022 at 18:32
  • $\begingroup$ The irrational winding of a torus is practically written this way: by definition, it is $\exp(L)$, where $L$ is a line of irrational slope. $\endgroup$
    – LSpice
    Commented Mar 13, 2022 at 19:09
  • $\begingroup$ I see. Thank you! $\endgroup$
    – Viktor
    Commented Mar 13, 2022 at 20:16

Not the answer you're looking for? Browse other questions tagged .