In the Book "Morse Theory" by J. Milnor in chapter 21 they study the structure of Riemannian Lie Groups with bi-invariant metrics.

For Left-Invariant vector fields $X,Y,Z,W$ and a bi-invariant riemannian metric $g$ it is quite straightforward to show that


Now he uses this relation to prove that the sectional curvature

$K(X,Y)=g([X,Y],[X,Y])=1/4g([X,Y],[X,Y]) \geq 0 $.

What i don't understand at this point, is why it is sufficient to look only at left-invariant vectorfields X,Y to prove that the sectional curvature is generally positive.

My guess is that is has to do with the bi-invariance of the metric and the fact that the space of left-invariant vector fields is isomorphic to the tangent Space at the identity. But i don't see how exactly i can make sense of that.


1 Answer 1


For any $g\in G$ and any two planes $\Pi $ in $T_gG$, let $v, w$ be an orthonormal basis for $\Pi$. Let $V, W$ be two left-invariant vector fields on $G$ so that $V(g) = v$ and $W(g) = w$. Then the sectional curvature with respect to $\Pi$ is given by

$$ S_\Pi = Rm(v, w, w, v) = Rm (V, W, W, V) = \frac 14 g ([V, W], [V, W]) \ge 0.$$

We are using that $Rm$ is a tensor, so it suffices to use any vector fields $\tilde V, \tilde W$ with $\tilde V (g) = v, \tilde W(g) = w$ to calculate $Rm(v, w, w, v)$. We also use that for all $v\in T_gG$, there is a left-invariant vector fields $V$ with $V(g) = v$.


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.