MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

The Grassmann manifold $\widetilde{Gr}(k,\Bbb{R}^n)$ of oriented $k$-planes in $\Bbb{R}^n$ is a double cover of the Grassmann manifold $Gr(k,\Bbb{R}^n)$ of non-oriented $k$-planes. We can give $\widetilde{Gr}(k,\Bbb{R}^n)$ the covering metric making the covering a local isometry.

Is it possible to describe the geodesics in $\widetilde{Gr}(k,\Bbb{R}^n)$ simply in terms of the geodesics in the non-oriented $Gr(k,\Bbb{R}^n)$?

share|cite|improve this question
Yes of course. Not only is the map from the oriented Grassmannian to the unoriented ones a covering map but it's a local isometry. So there's a direct correspondence between the geodesics given by lifting and taking the composite with the covering map, respectively. Does this answer your question? – Ryan Budney Dec 27 '12 at 4:30
@Ryan: I'm not sure that it does. I'm having trouble picturing the geodesics in the oriented Grassmannian in terms of the geodesics in the unoriented Grassmannian. For example, consider a plane with its two orientations, which corresponds to two distinct points in the oriented Grassmannian. Under the covering map, these points get map to a single point. The geodesic through these two points in the oriented Grassmannian get map presumably to a closed geodesic in the unoriented Grassmannian. But how do we describe the geodesic through these two oriented planes in terms of the closed geodesic? – Oliver Jones Dec 27 '12 at 6:36
@Oliver: I think the answer to your comment is already in Ryan's comment: lift. In more detail: Imagine your unoriented plane P, moving along a closed geodesic back to its starting position. Now attach to the initial P one of its orientations and, as P moves continuously along the geodesic, keep it oriented in a continuous way; in other words, drag the orientation along with the plane. When P returns to its initial position, the orientation might return to the initial orientation or to the opposite one. The latter possibility is the situation in your comment. – Andreas Blass Dec 27 '12 at 15:43
@Andreas: You're right, I didn't fully appreciate that covering spaces have the lifting property. Thanks for clarifying. This brings me to a related question. There are two ways in which to define a metric on the Grassmnnian of oriented planes; one is to treat it as a homogeneous space and the other is to pull back the metric from the Grassmannian of unoriented planes under the covering map. Do these metrics coincide? – Oliver Jones Dec 28 '12 at 21:39
@Oliver: I assume the metric on the unoriented version is obtained by treating it as a homogeneous space. Then the two metrics in your comment should coincide, provided you've normalized them the same way. The reason is that the 2-to-1 projection from the oriented to the unoriented version is invariant under the actions (on both versions) of the orthogonal group, so pulling back an invariant metric from the unoriented version will produce an invariant metric on the oriented version. – Andreas Blass Dec 30 '12 at 2:12

(CW answer just to get this off the unanswered list. See the comments above by Ryna Budney ans Andreas Blass for more details.)

Yes. Let $f:(M,g) \to (N,h)$ a covering map that is a local isometry, and let $p\in M$. If $\gamma:[0,1]\to N$ is a geodesic such that $\gamma(0) = f(p)$, then we can lift $\gamma$ to a geodesic $\tilde{\gamma}:[0,1]\to M$ a geodesic with $\tilde{\gamma}(0) = p$.

Of course, it is possible that $\gamma$ is a closed geodesic while $\tilde{\gamma}$ is not: consider the simple case of the covering $\mathbb{R}^1\to \mathbb{T}^1$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.