4
$\begingroup$

Ok so I'm confused about the relation between these two concepts. If I have a Lie Group $G$ I can associate a Lie algebra $\mathfrak{g}$ by taking his tangent space in the identity, with the appropriate bracket.

Now I have three questions (it's actually one):

1) Say I have $\mathfrak{g}$ and $\mathfrak{h}$ Lie algebras of two Lie groups $G$ and $H$. Does $\mathfrak{g} \simeq \mathfrak{h}$ imply $G \simeq H$ (as Lie groups, obviously) ? (I think no)

2) Now the opposite: if I have $G \simeq H$, can I say $\mathfrak{g} \simeq \mathfrak{h}$? (I think yes)

3) What can I say in case one, if the answer is no?

Let me give you an example: talking about Spin groups, wikipedia states that the accidental isomorphisms between low dimensional Spin groups and classic Lie groups are due to root system isomorphisms. But those are a thing of algebras, I don't understand. Can someone give me a brief overview of those concepts? Thanks

$\endgroup$
  • 1
    $\begingroup$ Question 1 can be answered in the negative with the following example: let $G$ be $\mathbb{R}$ under addition and let $H$ be the unit circle in $\mathbb{C}$ under complex multiplication. We have $\mathfrak{g}\cong\mathbb{R}\cong\mathfrak{h}$, but certainly $G$ is not isomorphic to $H$. $\endgroup$ – 211792 Feb 3 '15 at 18:49
  • 2
    $\begingroup$ Another example to how question 1 is negative is that $SO(n)$ and $O(n)$ have the same Lie algebra, but they are not diffeomorphic as Lie groups since one is connected (in fact a subspace of the other) and the other isn't. $\endgroup$ – Mnifldz Feb 3 '15 at 18:54
  • $\begingroup$ Thanks! Now about this last thing... is there some particular kind of Lie groups that is 1:1 with Lie algebras? Like, I don't know, simply connected ones? $\endgroup$ – AnalysisStudent0414 Feb 3 '15 at 18:55
3
$\begingroup$

1) was answered by AustinC and Mnifldz in the comments: there are many examples showing the answer is "no."

2) This is true. The key idea of the proof is the following lemma: Any group homomorphism $f:G\rightarrow H$ induces a map $f_\ast:\mathfrak{g}\rightarrow \mathfrak{h}$ in a functorial fashion. That is, if the group homomorphism is the identity map, then so is the induced map. Further, the induced map of a composition is the composition of the induced maps.

Assuming this, suppose $f:G\rightarrow H$ is an isomorphism with inverse $g:H\rightarrow G$. Then $g\circ f$ is the identity on $G$, so the induced map is the identity on $\mathfrak{g}$. On the other hand, the induced map is the composition $g_\ast \circ f_\ast$, so $g_\ast \circ f_\ast = Id_{\mathfrak{g}}$. On the other hand, applying the same argument to $f\circ g$ shows that $f_\ast \circ g_\ast = Id_{\mathfrak{h}}$. This implies that $f_\ast$ is an isomorphism of Lie algebras with inverse $g_\ast$.

3) You can say a lot, but not everything. First, since $\mathfrak{g}$ and $\mathfrak{h}$ are computed using the tangent space as the identity, they are indifferent to components other than the identity component of $H$ and $G$. (This is how Mnifldz's example arises). But even if $G$ and $H$ are connected, you can still have $\mathfrak{g}$ and $\mathfrak{h}$ isomorphic without $G$ and $H$ being isomorphic.

You can still save something: If $G$ and $H$ have isomorphic Lie algebras, then there is another (connected) Lie group $K$ with is simultaneously covers both $G$ and $H$. So, up to covers, $G$ and $H$ are the same. (This is how AustinC's example arises: $\mathbb{R}$ covers $S^1$.)

As a simple corollary to this if $G$ and $H$ are simply connected, then an isomorphism on the algebra level does imply that $G$ and $H$ are isomorphic.

There are a few other instances where you can recover something like this. For example, if $G$ and $H$ are compact Lie groups which are centerless, then they have isomorphic Lie algebras iff they are isomorphic Lie groups.

$\endgroup$

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.