I'm trying to understand the complexification of Lie groups from page $207$ here and I'm having trouble with a computation.

Assume $A, B$ are hermitian metrices, and $k$ is a unitary matrix. I want to show that the two paths $\alpha(t):= \exp(A)k\exp(tB)$ and $\beta(t):=\exp\left(A + tAd(k)B\right)k$ have the same tangent vector at $t=0$ (i.e. when you differentiate the above two matrix paths at $0$, you get the same matrix).

Now one can see that $\alpha'(0)= \exp(A)(Ad(k)B)k$.

I'm having trouble with the other curve. I got $$\beta'(0) = \exp(A)\left(\frac{1-\exp(-ad_{A})}{ad_{A}}(Ad(k)B)\right)k$$ using the formula for the differential of the exponential map. Or more directly,

$$\beta(t) = \left[I + \left(A+tAd(k)B\right) + \frac{1}{2!}\left(A+tAd(k)B\right)^2 + \dots \right]k$$ so

$$\beta'(0)= \left[\sum\limits_{n=0}^{\infty}\frac{1}{(n+1)!}\left(\sum\limits_{m=0}^{n}A^m\left(Ad(k)B\right)A^{n-m}\right)\right]k.$$

At this point, it looks like $\beta'(0) = \alpha'(0)$ if and only if $[A,Ad(k)B]=0$. What am I missing?

Edit: What I'm trying to show is that if $K\subset U(n)$ is a closed connected subgroup with Lie algebra $\mathfrak{l}$, then the set $\left\lbrace \exp(iX)k : X \in \mathfrak{l}, k \in K \right\rbrace$ is a subgroup of $GL(n,\mathbb{C})$ with Lie algebra $\mathfrak{l}\otimes \mathbb{C}. $


This question has an open bounty worth +50 reputation from Sir Wilfred Lucas-Dockery ending in 3 days.

This question has not received enough attention.

Please help this clerk balance his books.

  • $\begingroup$ The finite products of $\exp(t (A+iB)), A+iB \in Lie(K) \otimes \Bbb{C}$ form a connected real Lie group $K_\Bbb{C}$ with Lie algebra $Lie(K) \otimes \Bbb{C}$ (when considering smooth functions of the form $f(t) = \prod_{j=1}^J \exp( h_j(t) (A_j+iB_j)), t,h_j(t) \in \Bbb{R}$ then clearly $f'(0) \in Lie(K) \otimes \Bbb{C}$, so the only difficulty is to justify why it suffices to look at the case $J$ finite). Let a basis $Lie(K) = \sum_{m=1}^d \Bbb{R} A_m$, then $K_\Bbb{C}$ has a complex Lie group structure from the chart $z \in \Bbb{C}^d \mapsto \prod_{m=1}^d \exp(z_m A_m)$. $\endgroup$ – reuns 20 hours ago
  • $\begingroup$ Thanks. So this is like constructing the connected Lie subgroup associated to the Lie algebra $\mathfrak{l}\otimes \mathbb{C}$? Is there a way to show the fundamental group of this constructed subgroup is the same as that of $K$? I think something like this is required to prove that complex representations of $K$ extend to that of $K_{\mathbb{C}}$. $\endgroup$ – Sir Wilfred Lucas-Dockery 20 hours ago
  • $\begingroup$ I'm sorry, I'm quite lost. My previous comment agreed that it was possible to construct some connected complex group G with $K \subset G \subset GL(n,\mathbb{C})$ and with $Lie(G)= Lie(K)\otimes \mathbb{C}$. However, to prove that $G$ enjoys some universal property, I pointed out that we need that the inclusion induces $\pi_1(K)\simeq\pi_1(G)$. I think this is exactly why Bump tries to construct $G$ with typical element $\exp(iX)k$ (the isomorphism on fundamental groups follows easily from this construction of $G$). Does this make sense? I might be misunderstanding what's written in the text. $\endgroup$ – Sir Wilfred Lucas-Dockery 18 hours ago
  • $\begingroup$ Let $G = \bigcup_l \gamma_l K$ with $K$ the connected component of identity. I think you can look at how each $\gamma_l$ commutes with each $\exp(t A)$ (ie. $\forall tA \in Lie(K),\gamma_l \exp(tA) = \exp(f_l(tA)) \gamma_l$ with $f_l(A) = \gamma_l A \gamma_l^{-1} \in GL(Lie(K))$) and extend $f_l$ naturally to $f_l \in GL(Lie(K) \otimes \Bbb{C})$ to obtain the group law on $G_\Bbb{C} = \bigcup_l \gamma_l K_\Bbb{C}$. This way the complexification of $\Bbb{R+iZ}$ will be $\Bbb{C} \times \Bbb{Z}$ not $\Bbb{C}$. $\endgroup$ – reuns 18 hours ago

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.