3
$\begingroup$

I am reading some books about Lie group and Lie algebra. Denote the set of all the left invariant vector fields as $\mathfrak{X}_L$, and the tangent space at $e$ of $G$ as $T_eG$. They say that the $\mathfrak{X}_L$ and $T_eG$ are isomorphics. So we can extend a vector $\xi\in T_eG$ to the vector field on $G$ by this way: $$X(g)=dL_g(\xi),\mbox{for any }g\in G$$

where $X(g)$ is a vecotr at $g$, $L_g:G\rightarrow G$ is the left translation, and $dL_g:T_eG\rightarrow T_gG$ is the pushforward. My question is what is $dL_g(\xi)$ exactly?

Thank you.

$\endgroup$
2
  • 1
    $\begingroup$ Just checking: You're aware that not every vector field on a Lie group is left-invariant...? (This is something of a comical understatement; the space of left-invariant fields in your example is finite-dimensional. The point is, neither of your examples is left-invariant.) $\endgroup$ Commented Feb 11, 2015 at 15:20
  • $\begingroup$ @user86418 thanks, I understood wrongly the left invariant field. $\endgroup$ Commented Feb 11, 2015 at 15:56

1 Answer 1

3
$\begingroup$

Your first example may not be so illustrative. Since $$ \begin{pmatrix} 1 & 0 & a_1 \\ 0 & 1 & a_2 \\ 0 & 0 & 1 \end{pmatrix} \begin{pmatrix} 1 & 0 & x_1 \\ 0 & 1 & x_2 \\ 0 & 0 & 1 \end{pmatrix}= \begin{pmatrix} 1 & 0 & a_1+x_1 \\ 0 & 1 & a_2+x_2 \\ 0 & 0 & 1 \end{pmatrix} $$ your Lie group is isomorphic to $\mathbb{R}^2$ under addition. But if you think of those two positions as coordinates $y_1$ and $y_2$, then \begin{align*} \frac{\partial y_1}{x_1} &= 1 &\frac{\partial y_1}{x_2} &= 0\\ \frac{\partial y_2}{x_1} &= 0 &\frac{\partial y_1}{x_2} &= 1 \end{align*} So for each $(a_1,a_2)$, the map $dL_{(a_1,a_2)}$ is the identity map.

A better example might be \begin{align*} G &= SU(2) = \left\{\begin{bmatrix} \alpha & -\bar\beta \\ \beta & \bar\alpha \end{bmatrix} : \alpha,\beta\in \mathbb{C},\ |\alpha^2| + |\beta^2| = 1\right\} \\ \mathfrak{g} &= \mathfrak{su}(2) = \left\{\begin{bmatrix} ix & -\bar\beta \\ \beta & -ix \end{bmatrix} : x\in\mathbb{R},\beta\in \mathbb{C}\right\} \end{align*} Let $\xi = \begin{pmatrix} i & 0 \\ 0 &-i\end{pmatrix} \in \mathfrak{g}$. Then $\xi$ is the derivative at $t=0$ of the path in $SU(2)$ $$ \gamma(t) = \exp\begin{pmatrix} it & 0 \\ 0 &-it\end{pmatrix} = \begin{pmatrix} e^{it} & 0 \\ 0 &e^{-it}\end{pmatrix} $$ So for $g = \begin{pmatrix} \alpha & -\bar\beta \\ \beta & \bar\alpha \end{pmatrix} \in G$, we have \begin{align*} L_g(\xi) &= \left.\frac{d}{dt} \begin{pmatrix} \alpha & -\bar\beta \\ \beta & \bar\alpha \end{pmatrix} \begin{pmatrix} e^{it} & 0 \\ 0 & e^{-it} \end{pmatrix}\right|_{t=0} \\ &= \left.\frac{d}{dt} \begin{pmatrix} \alpha e^{it} & -\bar\beta e^{-it} \\ \beta e^{it} & \bar\alpha e^{-it} \end{pmatrix}\right|_{t=0} \\ &= \left.\begin{pmatrix} \alpha e^{it}(i) & -\bar\beta e^{-it}(-i) \\ \beta e^{it}(i) & \bar\alpha e^{-it}(-i) \end{pmatrix}\right|_{t=0} \\ &= \begin{pmatrix} \alpha i & \bar\beta i \\ \beta i & -\bar\alpha i\end{pmatrix}\\ &= \begin{pmatrix} \alpha & -\bar\beta \\ \beta & \bar\alpha \end{pmatrix} \begin{pmatrix} i & 0 \\ 0 & -i\end{pmatrix}=g\xi\\ \end{align*}

Also, I'm not sure what your pictures are supposed to represent.

$\endgroup$
7
  • $\begingroup$ Thank you. your answer is very clear. My question is: when $G$ and a vector $\xi$, which is at $e$ of $G$, are given, I say: "we can identify a unique left invariant vector field on $G$ from this vector $\xi$." am I right? In my opinion, this is a conclusion of the isomorphic property between $T_eG$ and $\mathfrak{X}_L$. But in the two pictures, $G_1=G_2$ (both are planes), we have the same $\xi$ at $e=(0,0)$, but they have two types of vector field. I'm new to this, so I am very confused. $\endgroup$ Commented Feb 11, 2015 at 15:06
  • $\begingroup$ "But in the two pictures, $G_1=G_2$ (both are planes)" ... I don't understand this statement. What are the multiplications on $G_1$ and $G_2$? The pictures you've posted don't seem to have anything to do with Lie groups. Can you edit the question with this information? $\endgroup$ Commented Feb 11, 2015 at 15:10
  • $\begingroup$ I m aware that I have some mistakes of understanding the definition of left invariant vector field, so those two pictures are meaningless. left invariant field $X$ is: $dL_g(X)=X(L_g)$. What I dont understand is $dL_g$. In the books, they all say that $dL_g$ is a tangent map from $T_eG$ to $T_gG$, but it is too abstract. They dont explain What form it has. How is it defined? $\endgroup$ Commented Feb 11, 2015 at 16:27
  • $\begingroup$ @Martial: A tangent vector at $e$ can be realized as the derivative of a path in $G$ through $e$. Multiply every point on that path by $g$, and you have a path through $g$. Take the derivative and you have a tangent vector to $e$. That's what I did in my $SU(2)$ example. $\endgroup$ Commented Feb 11, 2015 at 17:33
  • $\begingroup$ thank you very much. your answer is very helpful. $\endgroup$ Commented Feb 11, 2015 at 17:50

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.