A path-connected space is a topological space in which any two points can be joined by a continuous path. It formalizes the idea that movement between any two locations is possible without leaving the space or making a discontinuous jump. Path connectedness is stronger than connectedness in general, but the two properties coincide for many familiar spaces, including connected open subsets of Euclidean space and connected manifolds.
Definition
A path in a space is a continuous function
where has its usual topology. The points and are its initial and terminal points.
The space is path connected if, for every , there is a path satisfying
A subset is called path connected when it has this property with its subspace topology. Equivalently, the joining paths must take all their values in , not merely in the surrounding space .
No differentiability, finite length, or injectivity is required. A path may pause, retrace part of its route, or intersect itself. Thus path connectedness is a topological condition, not a statement about smoothness or distance.
Under the definition above, the empty space is path connected vacuously. Some authors impose nonemptiness as an additional requirement. Every one-point space is path connected.
Reversal, concatenation, and path components
Paths can be reversed and joined. If runs from to , its reverse
runs from to .
If runs from to and runs from to , their concatenation is
The two expressions agree at , and the resulting map is continuous.
Consequently, the relation
is an equivalence relation: constant paths establish reflexivity, reversal establishes symmetry, and concatenation establishes transitivity.
Its equivalence classes are the path components of . Each is a maximal path-connected subset, and every path-connected subset lies in exactly one path component. For a nonempty space, path connectedness is equivalent to having exactly one path component.
Relationship with connectedness
Every path-connected space is connected. Suppose, to the contrary, that is the union of two disjoint nonempty open sets and . A path beginning in and ending in would give, by inverse images, a separation of into two disjoint nonempty open subsets. This contradicts the connectedness of the interval.
The converse fails. Connectedness rules out separation by open sets; it does not guarantee the existence of continuous routes between points.
The topologist’s sine curve
A standard counterexample is the closed topologist’s sine curve
The oscillating graph is path connected, since it is the continuous image of the interval . Its closure is , so is connected.
Nevertheless, there is no path in joining the vertical segment to the oscillating graph. Any attempted continuous departure from the segment into the graph would force the second coordinate to oscillate between and arbitrarily close to the departure time, contradicting continuity.
Thus has exactly two path components: the graph and the vertical segment. It also demonstrates that the closure of a path-connected set need not be path connected.
Every path component is contained in a connected component. Unlike connected components, however, path components need not be closed. In this example, the graph is a path component whose closure is all of .
Local path connectedness
A space is locally path connected if every point has a neighborhood basis of path-connected sets. Explicitly, whenever belongs to an open set , there is a path-connected open set such that
This is a local condition, distinct from global path connectedness. In a locally path-connected space:
- Every path component is open.
- Path components coincide with connected components.
- Every connected open subset is path connected.
In particular, a connected, locally path-connected space is path connected.
To see why path components are open, take a point in a path component and a path-connected open neighborhood of that point. The whole neighborhood lies in the same path component. Each component is therefore a union of open sets. If a connected component contained more than one path component, these open path components would separate it.
Every topological manifold is locally path connected, because sufficiently small coordinate neighborhoods correspond to open balls in Euclidean space. Hence a nonempty connected manifold is path connected.
Global path connectedness does not imply local path connectedness. For example, adjoining the line segment from to to the closed topologist’s sine curve makes the resulting space path connected, but it still fails to be locally path connected near the upper part of its vertical segment.
Examples and geometric criteria
Convex and star-shaped sets
Every nonempty convex subset is path connected. Given , the straight-line path
lies entirely in .
More generally, a set is star-shaped with respect to if the line segment from to every point of the set remains in the set. Any two points can then be joined by traveling through . Convexity is therefore sufficient but not necessary for path connectedness.
Open subsets of Euclidean space
For an open subset , connectedness and path connectedness are equivalent. A stronger statement holds: any two points in a connected open set can be joined by a polygonal path, consisting of finitely many straight-line segments contained in .
Openness matters. The closed topologist’s sine curve is a connected subset of that is not path connected.
Spheres and punctured spaces
The circle and, more generally, every sphere with are path connected. By contrast, consists of two isolated points.
Similarly,
is path connected for , whereas has two path components. In higher dimensions a path can go around the missing point; on the real line it cannot.
Preservation properties
Path connectedness behaves well under several standard constructions.
Continuous images. If is path connected and is continuous, then is path connected: composing a path in with joins the corresponding image points. In particular, path connectedness is invariant under homeomorphisms.
Products. A product of nonempty path-connected spaces, equipped with the product topology, is path connected. Given paths joining the coordinates of two points, the coordinatewise formula
defines a continuous path in the product. Conversely, if a nonempty product is path connected, each factor is path connected by projection.
Unions with a common point. If a family of path-connected subsets has a common point, its union is path connected. Paths can be joined through that point. In particular, two intersecting path-connected subsets have path-connected union.
Quotients. A quotient space of a path-connected space is path connected, because the quotient map is continuous and surjective.
Arbitrary subsets and intersections need not preserve the property. For instance, an upper semicircle and a lower semicircle are individually path connected, but their intersection consists of two distinct endpoints and is not path connected.
Role in algebraic topology
Path connectedness is a basic organizing condition in algebraic topology. The set of path components is commonly denoted
A continuous map induces a map between these sets by sending each path component into a path component.
The fundamental group records loops based at a chosen point, up to homotopy fixing their endpoints. It is defined even when is not path connected, but it only detects the path component containing . A path between two basepoints induces an isomorphism between their fundamental groups; this isomorphism generally depends on the chosen path.
Path connectedness alone places no restriction on whether loops can be contracted. The circle is path connected but has nontrivial fundamental group. A simply connected space, under the usual convention, is path connected and has every loop homotopic to a constant loop. Thus the existence of paths between points and the contractibility of loops are separate topological properties.