Skip to Main Content

On the Geometry of Metric Spaces

Manuel Ritoré

Communicated by Notices Associate Editor Chikako Mese

Article cover

1. Introduction

The measurement of distances because of practical reasons, as the delimitation of parcels of farmland after floodings or those related to construction problems, lies behind the origins of mathematics in ancient civilizations. It was Euclid, in Proposition 20 of the first book of the Elements, who proved that the sum of two sides of a triangle is greater than the remaining one. This fact was probably used by Archimedes to state as assumption, in his work On the sphere and cylinder, that the straight line is the shortest between two points. Both achievements are now proven in basic courses on Linear Algebra.

Presently, the notion of distance on a nonempty set is formalized as a function satisfying


if and only if ,


for all ,


for all .

These conditions imply . The second property is the symmetry one. In Gromov’s words this assumption “[…] unpleaseantly limits many applications.” Removing this symmetry condition gives rise to the concept of asymmetric distance, see Busemann’s Local metric geometry (1958). The third property is the well-known triangle inequality.

The pair , where is a nonempty set and a distance on , is called a metric space.

Metric spaces were introduced by Fréchet in 1906 in the context of functional analysis. Shortly after, a comprehensive treatment of the theory was presented by Hausdorff in the second part of his monograph Hau14 published in 1914, where the notion of topological space was introduced.

Metric spaces appear everywhere in modern mathematics. They are an invaluable tool in functional analysis, geometry of manifolds, graph theory, artificial intelligence, and many other fields. The aim of this note is to give a glance of the role of metric spaces in several fields of mathematics without any intention of being exhaustive. We focus on different notions of curvature in metric spaces, compactness results for Riemannian manifolds, and an application of metric graph theory to group theory.

2. Length Spaces

Given a subset of a Euclidean space we may consider the restriction of the Euclidean distance to the set. For instance, the distance between two antipodal points on a unit sphere is . However, this distance is not intrinsic in the sense that cannot be measured inside the subset.

Given a continuous curve on a metric space, its length can be defined as the supremum of the quantities

over all partitions , , of the interval . When is finite the curve is said to be rectifiable. Hence the length of curves, sometimes infinite, can be measured in metric spaces.

A metric space is a length space if the distance between any pair of points can be computed as

where is any rectifiable curve connecting and (i.e., such that , ). A length space is geodesic if for every pair of points , there exists a rectifiable curve connecting , such that . A geodesic in is a rectifiable curve such that for all . The interested reader is referred to Burago, Burago and Ivanov’s BBI01 for a complete presentation of length spaces.

Examples of length spaces are Riemannian, Finsler and sub-Riemannian manifolds. A Riemannian manifold is a manifold together with an scalar product on the fiber bundle (a continuously varying scalar product on each tangent space). In a Finsler manifold the scalar product is replaced by a smooth norm. A sub-Riemannian manifold is a manifold together with a nonintegrable horizontal distribution and a positive definite scalar product on . The classical example of a sub-Riemannian manifold is the Heisenberg group : the -dimensional Euclidean space together with a noncommutative product and the left-invariant vector fields

induce a left-invariant Riemannian metric making an orthonormal basis. The vectors fields generate a nonintegrable distribution that, together with the restriction of to , provide a sub-Riemannian structure on . On a connected sub-Riemannian manifold , Chow–Rashevskii theorem, see §0.4 and §1.1, §1.2 in Gromov Gro96, implies that every pair of points can be connected by a smooth horizontal curve on (i.e., everywhere tangent to the horizontal distribution).

To define a distance we consider, in the Riemannian and Finsler cases, the class of piecewise smooth curves and, in the sub-Riemannian case, the class of piecewise smooth horizontal curves. In all cases, the length of a curve can be computed as

where is the norm associated to in the Riemannian case, the Finsler norm, or the one associated to in the sub-Riemannian case. Associated to this length we can define a distance between two points as the infimum of the length of the curves joining both points. This way we obtain the Riemannian distance, the Finsler distance, or the Carnot-Carathéodory distance. Geodesics are curves of minimum distance connecting two given points.

Figure 1.

Geodesics in the Riemannian and sub-Riemannian Heisenberg group connecting two points in the same vertical line. The vertical line is a Riemannian geodesic. The helicoidal curve is a sub-Riemannian geodesic, forced to be tangent to the horizontal distribution.

Graphic without alt text

3. Metric Spaces with Bounded Curvature

Curvature in Riemannian geometry was introduced as a measure of how a Riemannian space deviates from Euclidean space, see Gromov Gro91 for an enlightening discussion. For surfaces embedded in Euclidean space, the curvature coincides with the product of the principal curvatures of the surface and it is referred to as the Gauß curvature. The classical Gauß–Bonnet theorem provides a first hint on how the curvature affects the behavior of geodesic triangles on a surface. If is a triangle delimited by geodesics making inner angles and then

Hence the sum of the inner angles of a triangle on a surface depends on the sign of the Gauß curvature: it is equal to on surfaces with , larger than on surfaces with and smaller than when .

Figure 2.

Behavior of geodesics triangles according to Gauß–Bonnet theorem.

\renewcommand{\arraystretch}{1} \setlength{\unitlength}{1.0pt} \begin{tikzpicture} \coordinate(A) at (0,0); \begin{scope}[shift={(6,5.5)}] \draw(0,0) node{$K>0$}; \end{scope} \begin{scope}[shift={(6,3)}] \draw(0,0) node{$K=0$}; \end{scope} \begin{scope}[shift={(6,0.5)}] \draw(0,0) node{$K<0$}; \end{scope} \begin{scope}[shift={(1,5)}] \draw(0,0) to[out=-10, in=210] (3,0.5); \draw(0,0) to[out=90,in=235] (0.5,1.75); \draw(0.5,1.75) to[out=5,in=130] (3,0.5); \draw[fill=black] (0,0) circle (0.05); \draw[fill=black] (3,0.5) circle (0.05); \draw[fill=black] (0.5,1.75) circle (0.05); \end{scope} \begin{scope}[shift={(1,2.5)}] \draw(0,0) -- (3,0.5); \draw(0,0) -- (0.5,1.75); \draw(0.5,1.75) -- (3,0.5); \draw[fill=black] (0,0) circle (0.05); \draw[fill=black] (3,0.5) circle (0.05); \draw[fill=black] (0.5,1.75) circle (0.05); \end{scope} \begin{scope}[shift={(1,0)}] \draw(0,0) to[out=35,in=175] (3,0.5); \draw(0,0) to[out=65,in=270] (0.5,1.75); \draw(0.5,1.75) to[out=-35,in=165] (3,0.5); \draw[fill=black] (0,0) circle (0.05); \draw[fill=black] (3,0.5) circle (0.05); \draw[fill=black] (0.5,1.75) circle (0.05); \end{scope} \end{tikzpicture}

According to the historical account in the preface of Alexander, Kapovitch, and Petrunin’s monograph, see arXiv:1903.08539, the first synthetic description of curvature is due to A. Wald in a paper published in 1936, followed by a paper by Alexandrov in 1941. It is generally considered that Rauch’s comparison results for Jacobi fields in the early 1950s can be thought of as an infinitesimal triangle comparison result which was soon followed by the celebrated Toponogov’s comparison theorem in the late 1950s.

A version of triangle comparison not involving angles can be stated as follows. Assume we have a point in a Riemannian manifold and a unit-speed geodesic in a small neighborhood of . We consider a second unit-speed geodesic in the plane and a point such that , . This way we construct a comparison triangle whose vertices , , lie at the same distance as the ones of the original triangle . We define the functions , . Then we have the following.


Under the conditions above


If has nonnegative sectional curvatures then .


If has nonpositive curvature then .

This result is typically proven by means of Jacobi fields, see Cheeger and Ebin (1975), but perhaps the reader could find useful the following alternative argument. Assume that the sectional curvatures of are no larger than . Take , a normal neighborhood of where the exponential map is a diffeomorphism, and a geodesic parameterized by arc length. A basic result on Jacobi fields, similar to the one used in the proof of Bishop’s volume comparison, implies that the Hessian of the function , where is the distance to , satisfies

for any , , and . In Euclidean space we would have equality in 1 replacing by the Euclidean distance to . Then the functions and satisfy

Hence on and, since the function vanishes at the endpoints of the interval, we get , which implies .

Figure 3.

Illustration of triangle comparison. The one to the right lies in a Riemannian manifold . The one to the left is a comparison triangle in the Euclidean plane. If the sectional curvature of is nonnegative then . The opposite inequality holds when has nonpositive sectional curvature.

\renewcommand{\arraystretch}{1} \setlength{\unitlength}{1.0pt} \begin{tikzpicture} \begin{scope}[shift={(4.5,3)}] \draw(0,1) -- (3,0); \draw[fill=black] (1,-2) circle (0.05) node[below]{$\bar{p}$}; \draw[fill=black] (0,1) circle (0.05) node[above]{$\bar{\gamma}(0)$}; \draw[fill=black] (3,0) circle (0.05) node[above]{$\bar{\gamma}(T)$}; \draw[dashed] (0,1) -- (1,-2); \draw[dashed] (3,0) -- (1,-2); \draw[fill=black] (1.35,0.55) circle (0.05) node[above]{$\bar{\gamma}(t)$}; \draw(1,-2) -- (1.35,0.55); \end{scope} \begin{scope}[shift={(.5,3)}] \draw(0,1) to[out=5, in=135] (3,0); \draw[fill=black] (1,-2) circle (0.05) node[below]{$p$}; \draw[fill=black] (0,1) circle (0.05) node[above]{$\gamma(0)$}; \draw[fill=black] (3,0) circle (0.05) node[xshift=5,yshift=12]{$\gamma(T)$}; \draw[dashed] (0,1) to[out=-90, in=130] (1,-2); \draw[dashed] (3,0) to[out=-120, in=20] (1,-2); \draw[fill=black] (1.45,0.93) circle (0.05) node[above]{$\gamma(t)$}; \draw(1,-2) to[out=90,in=250] (1.45,0.93); \end{scope} \end{tikzpicture}

This comparison can be stated also in terms of angles. Consider a small geodesic triangle in a Riemannian manifold . If has nonnegative curvature then, at every vertex, the angle made by the sides of the triangle is larger than or equal to the comparison angle at the vertex (i.e., the Euclidean angle of a comparison triangle). For nonpositive curvature we have the opposite inequality. Similar comparisons with a simply connected surface of constant curvature hold when the sectional curvature of is no smaller or no larger than .

The triangle comparison for Riemannian manifolds allows to introduce a synthetic notion of metric space with curvature bounded, above or below, in terms of properties of geodesic triangles on the metric space.

Note that angles can be defined in a metric space using the law of cosines. Given three different points in a metric space , we can build a comparison Euclidean triangle of vertices by requiring that the distances equal , , . The comparison angle is defined as the Euclidean angle , that is,

If we have two curves emanating from the same point , the angle is

when the limit exists.

The term CAT spaces is used for metric spaces with as an upper curvature bound (CAT stands for Cartan–Alexandrov–Topogonov) and the term Alexandrov spaces for spaces with lower curvature bounds. The reader should be aware that sometimes a different terminology is used depending on the direction of the bound.

The theory of CAT spaces and the one of Alexandrov spaces differ in many respects. In the case of curvature bounded above, Hadamard manifolds, complete simply connected length spaces of nonpositive curvature, play a central role. Almost all classical results for Riemannian manifolds have been extended to Hadamard manifolds, like uniqueness of geodesics connecting two given points, the validity of triangle comparison for large triangles, the fact that the squared distance function is globally concave. Essential references on CAT spaces are Bridson and Haeffliger BH99, and Ballman, Gromov, and Schroeder BGS85. As for Alexandrov spaces, there are two main points that make the theory quite different from the one of CAT spaces. The first is that triangle comparison holds for arbitrarily large triangles without additional assumptions. This is the content of Toponogov’s theorem. The second one concerns the local structure: in an Alexandrov space the Hausdorff dimension coincides with the topological dimension, which is an integer or infinite. An Alexandrov space is a manifold except on a small set of points. Examples of Alexandrov spaces are the boundaries of convex sets in Euclidean spaces. Numerous results valid for Riemannian manifolds also hold in the class of Alexandrov spaces with curvature bounded from below by . To cite just a few, an upper bound on the diameter when , a splitting theorem, Gromov–Bishop inequalities for the volume of balls. Petrunin also proved a Levy–Gromov type isoperimetric inequality (see the author’s monograph Rit23 for the Riemannian case). The reader is referred to Burago, Burago and Ivanov BBI01, the already mentioned monograph by Alexander, Kapovitch, and Petrunin, and the references cited therein, for excellent introductions to the theory of metric spaces with bounds on curvature.

4. Convergence of Metric Spaces

The introduction of curvature bounds on metric spaces might seem an abstract construction with no special interest. However, it is related to the behavior of sequences of Riemannian manifolds. One can consider the space of all Riemannian manifolds with some topology and asks whether some given subset is compact or precompact with this topology. In many cases the limit of the sequence is not a Riemannian manifold. This is one of the situations where metric spaces with bounds on curvature play a role.

A standard way of measuring the distance between two sets in a metric space is the Hausdorff distance , defined by

where for any set , is the open tubular neighborhood of of radius .

Gromov, see Gro81Gro07 introduced a way of measuring the distance between two metric spaces , nowadays known as the Gromov–Hausdorff distance, denoted by . To compute it, we take all metric spaces which contain isometric copies of and of , compute their Hausdorff distance on , and take the infimum over all possible metric spaces .

Gromov–Hausdorff distance is really a distance on the space of isometric classes of metric spaces since two isometric metric spaces have -distance equal to . To measure the Gromov–Hausdorff distance using its definition is not very practical. However, for compact metric spaces , , should we have two -nets on , and on such that

then we would have .

Important classes of Riemannian manifolds which are precompact in the Gromov–Hausdorff distance are the following:

For and , the class of -dimensional Riemannian manifolds with volume and injectivity radius .

For any and , , the class of -dimensional Riemannian manifolds with and sectional curvature . The same result holds assuming instead that the Ricci curvature is .

While the Gromov–Hausdorff distance is useful when considering compact metric spaces, it is usually a very strong convergence for unbounded metric spaces. In this case it should be replaced by the pointed Gromov–Hausdorff convergence.

Natural measures which can be considered on metric spaces are the Hausdorff measures. Given , the -dimensional Hausdorff measure associated to a set is given, up to a constant, by

where is the supremum of taken over all coverings of by sets satisfying . For a metric space there exists a constant such that for and for all . The quantity is called the Hausdorff dimension of and is denoted by .

It is also important to note that Gromov–Hausdorff limits (see next section) of Alexandrov spaces of curvature are themselves Alexandrov spaces of curvature . There are also compactness results for Alexandrov spaces. Gromov himself proved that the space composed of the Alexandov spaces with curvature , diameter and Hausdorff (topological) dimension is compact in the Gromov–Hausdorff topology.

Although the Hausdorff measure on a metric space , with , is a natural measure to consider, sometimes it is more interesting to take a different one, see Appendix 2 in CC97. This leads to the notion of metric measure space , which is nothing but a metric space together with a Borel measure on . If is a length (geodesic) space we refer to as a length (geodesic) measure space.

A sequence of metric measure spaces converges in measured Gromov–Hausdorff topology to a metric measure space if there is a sequence of measurable maps and a sequence of real numbers converging to such that the maps are -quasi-isometries:

the tubular neighborhood of radius of is , and the push-forward measures converge in weak topology to , cf. Fukaya (1987).

5. Metric Spaces with a Lower Bound on Ricci Curvature

That the curvature of a Riemannian manifold implies metric properties of triangles was the key to introduce a synthetic notion of metric spaces with bounded curvature. An analogous approach should be possible for the Ricci curvature on a Riemannian manifold.

It looks like the first ones to discuss the possibility of such generalization were Cheeger and Colding in Appendix 2 in CC97, who suggested to use lower bounds on the volume growth of sectors. See also §5.44 in Gromov Gro07. In a Riemannian manifold , the classical Bishop–Gromov theorem implies that if is a model Riemannian manifold (complete simply connected with constant sectional curvature ) of the same dimension as and satisfying , we have

for arbitrary , where is the annulus , and the corresponding annulus in , whose volume is independent of the base point. The symbol denotes the Riemannian volumes. This inequality is also valid if we consider a geodesic sector with focal point , which is nothing but a set such that, for each , there is a geodesic segment connecting and . Let . Then we have

for arbitrary .

Figure 4.

Comparison of volume of geodesic sectors.

\renewcommand{\arraystretch}{1} \setlength{\unitlength}{1.0pt} \begin{tikzpicture} \begin{scope}[shift={(4,0)},rotate=-30] \draw[rotate=63.2] (0,0) -- (2.65,0); \draw[rotate=116.8] (0,0) --(2.65,0); \draw[ rotate=76] (0,0) -- (2.72,0); \draw[ rotate=104] (0,0) -- (2.72,0); \begin{scope}[scale=0.8,shift={(1.5,3)}] \draw(0,0) arc [start angle=0, end angle=180,x radius=1.5,y radius=0.35]; \draw[dashed] (0,0) arc [start angle=360, end angle=180,x radius=1.5,y radius=0.35]; \end{scope} \begin{scope}[scale=0.5,shift={(1.5,3)}] \draw(0,0) arc [start angle=0, end angle=180,x radius=1.5,y radius=0.35]; \draw[dashed] (0,0) arc [start angle=360, end angle=180,x radius=1.5,y radius=0.35]; \end{scope} \begin{scope}[scale=0.3,shift={(1.5,3)}] \draw(0,0) arc [start angle=0, end angle=180,x radius=1.5,y radius=0.35]; \draw[dashed] (0,0) arc [start angle=360, end angle=180,x radius=1.5,y radius=0.35]; \end{scope} \draw(0.8,0.8) node{$r_1$}; \draw(1.1,1.3) node{$r_2$}; \draw(1.6,2.2) node{$r_3$}; \end{scope} \begin{scope}[shift={(0,0)}, rotate=-30] \draw(0,0) to[out=140,in=255] (-1.055,2.7); \draw(0,0) to[out=115, in=260] (-0.4,2.98); \begin{scope}[xscale=-1,yscale=1] \draw(0,0) to[out=140,in=255] (-1.055,2.7); \draw(0,0) to[out=115, in=260] (-0.4,2.98); \end{scope} \begin{scope}[scale=1,shift={(1.05,2.7)}] \draw(0,0) arc [start angle=0, end angle=180,x radius=1.05,y radius=0.3]; \draw[dashed] (0,0) arc [start angle=360, end angle=180,x radius=1.05,y radius=0.3]; \end{scope} \begin{scope}[scale=1,shift={(1.09,1.7)}] \draw(0,0) arc [start angle=0, end angle=180,x radius=1.09,y radius=0.3]; \draw[dashed] (0,0) arc [start angle=360, end angle=180,x radius=1.09,y radius=0.3]; \end{scope} \begin{scope}[shift={(0,0.8)},scale=0.66,shift={(1.05,0)}] \draw(0,0) arc [start angle=0, end angle=180,x radius=1.05,y radius=0.3]; \draw[dashed] (0,0) arc [start angle=360, end angle=180,x radius=1.05,y radius=0.3]; \end{scope} \draw(1,0.7) node{$r_1$}; \draw(1.4,1.6) node{$r_2$}; \draw(1.35,2.65) node{$r_3$}; \end{scope} \end{tikzpicture}

Another synthetic approach to Ricci curvature mentioned in Cheeger and Colding CC97 is the use of the Laplacian of the distance function. The classical Bishop volume comparison for balls in an -dimensional manifold with follows from comparison of the mean curvatures of geodesic spheres with the same radii in and the space form of constant sectional curvature . The mean curvature of a geodesic sphere is the Laplacian of the distance function to a fixed point. So, for instance, when , this translates into the equation

at least for small radius. Note that is the mean curvature of a ball of radius in the Euclidean space . Equation 2 was shown to hold in a weak sense in Riemannian manifolds by Calabi (1958). In order for this approach to work we need the extension of the notion of Laplacian to a metric space. Observe that for a Lipschitz function on a metric measure space we can define

and the Cheeger energy

By approximation of measurable functions by Lipschitz functions, this energy can be extended to wider classes. In general, is not a quadratic form. We say that a metric measure space is infinitesimally Hilbertian if the Cheeger energy is a quadratic form in . On such spaces it is possible to define a weak notion of Laplacian, thus giving sense to inequality 2. This extension of differential calculus to metric spaces has been carried out by Gigli Gig15. See also Ambrosio’s lecture at ICM2018 Amb18. A recent monograph by Heinonen, Koskela, Shanmugalingam, and Tyson (2015) is a very good introduction to analysis on metric spaces.

Yet another generalization of the notion of Ricci curvature comes from the well-known Bochner’s formula. Recall that Bochner’s formula in a Riemannian manifold reads

If we assume and use the estimate we arrive at Bochner’s inequality

Here and, of course, can be replaced by a larger number. All the operators appearing in this formula can be defined weakly on a metric measure space assuming some extra hypotheses, thus providing a notion of Ricci curvature bounded below.

5.1. An approach based on mass transport

In the first decade of the 21st century, Lott and Villani LV09 and Sturm Stu06aStu06b introduced independently equivalent notions of metric measure spaces with Ricci curvature bounded below. These notions were based on mass transportation properties on Riemannian manifolds with Ricci curvature bounded below.

Let us introduce first a few concepts, see Villani Vil09 for an excellent exposition. Given a metric space we consider the space of probability measures on and the subset of those satisfying

for some (all) . Given , a transport plan between and is a probability measure on with marginals (i.e., the push-forward measures , by the projections , , are and ). We denote the set of transport plans between and by Plan. The Wasserstein distance on is defined by

A minimizer of this problem is called an optimal transport plan between and . The metric space inherits many properties of . The equivalence between geodesics in Wasserstein space and certain optimal transport plans was established by Lott and Villani LV09, §2.3 and Sturm Stu06a, §2.3.

Given , we define the Shannon and Rényi entropy functionals for a measure by