Let $G$ be a Lie group, $L_g$ the left-translation on this group with differential $d L_g$. A vector field $X$ on $G$ is called left-invariant if

$$ X \circ L_g = d L_g \circ X \quad \forall g \in G$$


$$ X_{gh} = (d L_g)_h (X_h) \quad \forall g,h \in G. $$

Now, this definition seems so natural to me that I cannot come up with a non-trivial counterexample for a vector field that is $\textit{not}$ left-invariant. In my mind, pushing forward on the tangent space is basically always the same as the group action...

Could you provide me with such a counterexample that helps understand the notion of left-invariance?

  • 4
    $\begingroup$ Take any non-zero vector in any tangent space. Show that it extends uniquely to a left-invariant vector field. Conclude that the space of left-invariant vector fields has dimension $\dim{G}$, whereas the dimension of the space of vector fields is infinite. Take any non-zero vector field, multiply it by a smooth bump function and get a non-invariant vector field (unless you get zero). Now if you think about how to construct a non-zero vector field on a manifold you should be able to find plenty of non-invariant examples... $\endgroup$ – t.b. Aug 6 '12 at 10:24

You could start taking the Lie group $(\mathbb R^n,+),$ and considering what does it means for a vector field $X$ on $\mathbb R^n$ to be left-invariant.

  • $\begingroup$ Thanks! Do you know another example, where the differential $(dL_g)_h$ is not the identity? $\endgroup$ – madison54 Aug 6 '12 at 10:11

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.