nLab geodesic convexity




The definition of geodesic convexity is like that of convexity, but with straight lines in an affine space generalized to geodesics in a Riemannian manifold or metric space.


In a Riemannian manifold

Let (X,g)(X,g) be a Riemannian manifold and CXC \subset X a subset. We say that CC is

  • weakly geodesically convex if for any two points from CC there exists exactly one minimizing geodesic in CC connecting them;

  • geodesically convex if for any two points in CC there exists exactly one minimizing geodesic in XX connecting them, and that geodesic arc lies completely in CC

  • strongly geodesically convex if for any two points in C¯\overline{C} there exists exactly one minimizing geodesic in XX connecting them, and that geodesic arc lies completely in CC, except possibly the endpoints; and furthermore there exists no nonminimizing geodesic inside CC connecting the two points.


The convexity radius at a point pXp \in X is the supremum (which may be ++ \infty) of rr \in \mathbb{R} such that for all η<r\eta \lt r the geodesic ball B p(r)B_p(r) is strongly geodesically convex.

The convexity radius of (X,g)(X,g) is the infimum over the points pXp \in X of the convexity radii at these points.

In a metric space

For XX a metric space, a distance-preserving path is a function x:[a,b]Xx \colon [a,b]\to X which is an isometry. This is a metric-space analogue of an “arc-length-parametrized minimizing geodesic” on a Riemannian manifold. In particular, the existence of such a path implies that d(x(a),x(b))=bad(x(a),x(b)) = b-a. We then say that XX is geodesic (or geodesically convex) if any two points can be connected by a distance-preserving path.

We say that XX is a length space if for any xx and yy, the distance d(x,y)d(x,y) is the infimum of the lengths of all continuous paths from xx to yy. The Hopf-Rinow theorem? says that for a metric space in which every closed bounded subset is compact, being a length space is equivalent to being geodesic.



At any point pXp \in X of a Riemannian manifold, the convexity radius is positive.

This is due to (Whitehead).


If XX is a compact space then the convexity radius of (X,g)(X,g) is positive.

This is reproduced for instance as proposition 95 in (Berger)


Every paracompact manifold admits a complete Riemannian metric with bounded absolute sectional curvature and positive injectivity radius.

This is shown in (Greene).


Original literature includes

  • J. H. C. Whitehead, Convex regions in the geometry of paths Quart. J. Math. 3, 33–42 (1932).
  • R. Greene, Complete metrics of bounded curvature on noncompact manifolds Archiv der Mathematik Volume 31, Number 1 (1978)

A review of geodesic convexity in Riemannian manifolds is in

  • Isaac Chavel, Riemannian geometry – A modern introduction Cambridge University Press (1993)
  • Marcel Berger, A panoramic view of Riemannian geometry

A categorical perspective on geodesic convexity for metric spaces can be found in

  • Taking categories seriously, Reprints in Theory and Applications of Categories, No. 8, 2005, pp. 1–24. (pdf)

Much of the issues on geodesic convexity becomes more complicated when trying to generalize from Riemannian to Lorentzian manifolds, as discussed at length at

  • John K. Beem, Paul E. Ehrlich, Kevin L. Easley, Global Lorentzian geometry, Marcel Dekker, 1996, 635 pages
  • John K. Beem, Lorentzian geometry in the large, Math. of gravitation I, Lorentzian geometry and Einstein equations, Banach Center Publications 41, Inst. of Math. Polish Acad. of Sci. Warszawa 1997 pdf

In particular, the conclusions of the Hopf-Rinow Theorem fail to hold for complete Lorentzian manifolds.

Last revised on January 21, 2014 at 15:01:05. See the history of this page for a list of all contributions to it.