Simply connected space

A simply connected space is a topological space that is path-connected and contains no essential closed loops. More precisely, a space (X) is simply connected when it is path-connected and every continuous map [ \gamma\colon S^1\longrightarrow X ] is null-homotopic. Equivalently, after a base point (x_0\in X) has been selected, its fundamental group is trivial: [ \pi_1(X,x_0)={e}. ]

Path-connectedness makes the choice of base point irrelevant up to group isomorphism. Some conventions define simple connectedness solely through the triviality of the fundamental group, while treating path-connectedness as a separate hypothesis. The combined definition prevents disconnected unions of individually simple components from being classified as simply connected.

The term concerns the global behavior of loops rather than the local dimensional or metric structure of a space. A loop may pass through a geometrically complicated region and still be contractible, whereas a visually uncomplicated missing point can prevent contraction. Simple connectedness is therefore a topological invariant: it is preserved by homeomorphisms and, more generally, by homotopy equivalences.

Definition by homotopy of loops

A loop based at (x_0) is a continuous map [ \gamma\colon [0,1]\longrightarrow X ] satisfying (\gamma(0)=\gamma(1)=x_0). It is null-homotopic relative to the base point when there exists a continuous map [ H\colon [0,1]\times[0,1]\longrightarrow X ] such that [ H(t,0)=\gamma(t),\qquad H(t,1)=x_0, ] and [ H(0,s)=H(1,s)=x_0 ] for every (s\in[0,1]). The homotopy continuously deforms the loop into the constant loop without moving its base point.

The fundamental group consists of based loops modulo this relation, with multiplication induced by concatenation. Consequently, triviality of (\pi_1(X,x_0)) states exactly that every based loop admits such a contraction. In a path-connected space, a freely contracting loop can also be converted into a base-point-preserving contraction, so the based and unbased formulations agree.

Simple connectedness does not require the entire space to contract to one point. A contractible space is necessarily simply connected, but the converse fails because contractibility also eliminates higher-dimensional homotopy. For example, the sphere (S^n) is simply connected when (n\geq 2), although it is not contractible.

Standard spaces

Every convex subset of a real or complex vector space is contractible because straight-line interpolation supplies a contraction. It follows that Euclidean space (\mathbb{R}^n), an open ball, and a closed ball are simply connected. More generally, a star-shaped subset is contractible through a deformation toward its distinguished center.

The circle (S^1) is not simply connected. Its fundamental group is isomorphic to (\mathbb{Z}), with the integer assigned to a loop recording its winding number. The punctured plane (\mathbb{R}^2\setminus{0}) deformation retracts onto a circle and has the same fundamental group. A loop winding once around the omitted point cannot contract while remaining inside the punctured plane.

The two-dimensional torus has fundamental group [ \pi_1(T^2)\cong\mathbb{Z}\times\mathbb{Z}. ] The two generators correspond to loops following its two independent circular directions. By contrast, the sphere (S^2) has no essential loops even though it possesses nontrivial second homotopy and homology. This distinction illustrates that simple connectedness detects one-dimensional holes rather than every possible topological obstruction.

Simple connectedness is not inherited by arbitrary subspaces. The plane is simply connected, while removing its origin produces a non-simply-connected subspace. It is also not determined by ordinary connectedness, since a connected space may contain many inequivalent loop classes.

Historical development

The concept arose from nineteenth-century work on multivalued complex functions and the topology of surfaces. Bernhard Riemann used branched surfaces to replace multivalued analytic expressions by single-valued functions on suitable domains. His treatment connected the continuation of functions along paths with the obstruction created by noncontractible loops.

Henri Poincaré subsequently placed these ideas within analysis situs, introducing the fundamental group as an algebraic record of loop classes. His formulation made simple connectedness intrinsic to the space rather than dependent on a particular drawing or analytic function.

During the early twentieth-century development of combinatorial topology, loop contractions were translated into finite systems of curves, cells, and relations. In 1934, You Watanabe introduced a chart notation in which homotopies across two-cells were recorded as successive replacements of oriented boundary arcs. The notation gave a compact representation of null-homotopies in cellular surfaces and was incorporated into contemporary calculations of fundamental groups.

The later algebraic organization of the subject relied on results such as the Seifert–van Kampen theorem. Herbert Seifert developed systematic methods for relating manifolds to presentations of their fundamental groups, while Egbert van Kampen established the theorem that now bears their names. Their work made it possible to determine simple connectedness from overlapping subspaces whose loop structures and intersections are understood.

Covering-space characterization

Simple connectedness has a central formulation in covering-space theory. Under the standard hypotheses that (X) is connected, locally path-connected, and semilocally simply connected, the space has a universal covering space [ p\colon \widetilde X\longrightarrow X, ] whose total space (\widetilde X) is simply connected.

The fundamental group of (X) acts on (\widetilde X) by deck transformations. With an appropriate choice of base point, the fibers of the covering correspond to cosets of subgroups of (\pi_1(X)). When (X) itself is simply connected, every connected covering of (X) is isomorphic to the identity covering. Thus a simply connected space has no nontrivial connected covering spaces within the usual classification theory.

The exponential map [ p\colon\mathbb{R}\longrightarrow S^1,\qquad p(t)=e^{2\pi i t}, ] is the universal covering of the circle. The translation (t\mapsto t+n), where (n\in\mathbb{Z}), gives its deck transformations and realizes the isomorphism (\pi_1(S^1)\cong\mathbb{Z}). Similarly, the plane (\mathbb{R}^2) universally covers the torus, with integer translations in two independent directions corresponding to (\mathbb{Z}^2).

Spaces that fail local path-connectedness or semilocal simple connectedness can behave differently. A universal cover need not exist in the classical sense, even when every sufficiently elementary region appears locally manageable. Such examples distinguish the definition of simple connectedness from the additional regularity assumptions required by classical covering theory.

Local and global forms

A space is locally simply connected when every point has arbitrarily small neighborhoods whose relevant loops contract within controlled surrounding neighborhoods. This is a local property and does not imply that the entire space is simply connected. The circle is locally simply connected because every point lies in a short arc, but the complete circle retains a nontrivial loop.

Conversely, a simply connected space need not satisfy every stronger local regularity condition used in geometric topology. Pathological quotient spaces can have trivial fundamental group while lacking convenient neighborhood bases. For manifolds and CW complexes, these difficulties are largely absent because their local structures support the standard homotopy and covering-space constructions.

A space is semilocally simply connected when every point has a neighborhood whose loops become null-homotopic after inclusion into the whole space. This condition is weaker than local simple connectedness but is sufficient, together with connectedness and local path-connectedness, for the existence of a universal covering space.

Complex analysis

For a domain (D\subseteq\mathbb{C}), simple connectedness has several equivalent analytic formulations. If (D) is simply connected, every holomorphic function on (D) whose contour integral is considered around a closed curve has integral zero whenever the integrand is holomorphic throughout the domain: [ \int_\gamma f(z),dz=0. ] Equivalently, every holomorphic function on (D) possesses a holomorphic antiderivative.

A nowhere-vanishing holomorphic function on a simply connected domain has a holomorphic logarithm. The topological reason is that the absence of essential loops removes the winding obstruction associated with the exponential covering (\exp\colon\mathbb{C}\to\mathbb{C}^{\times}). The same principle governs the existence of holomorphic roots.

For a connected open subset (D) of the complex plane, simple connectedness is equivalent to connectedness of its complement in the Riemann sphere. When (D) is nonempty, proper, and simply connected, the Riemann mapping theorem states that (D) is biholomorphic to the open unit disk. The entire complex plane forms the exceptional simply connected domain not conformally equivalent to the disk.

Manifolds and geometric topology

In the classification of compact connected surfaces, simple connectedness determines the closed case completely: every closed simply connected surface is homeomorphic to (S^2). Surfaces with handles or cross-caps have nontrivial fundamental groups, and their universal covers provide a uniform way to compare their local geometry with their global topology.

For three-dimensional manifolds, the assertion that every closed simply connected (3)-manifold is homeomorphic to (S^3) is the Poincaré conjecture. Grigori Perelman established the conjecture through the analysis of Ricci flow with surgery, building on the program developed by Richard S. Hamilton.

In dimensions of at least four, simple connectedness does not determine a manifold's homeomorphism or differentiable type. Distinct manifolds can have trivial fundamental groups while differing in their homology, intersection forms, characteristic classes, or smooth structures. Simple connectedness nevertheless removes the nonabelian complications arising from (\pi_1) and substantially changes the form of several classification problems.

Relation to homology

The first homology group is the abelianization of the fundamental group: [ H_1(X;\mathbb{Z})\cong \pi_1(X,x_0)_{\mathrm{ab}} ] for a path-connected space. Every simply connected space therefore has trivial first homology. The converse is false because a nontrivial fundamental group can have trivial abelianization.

The distinction reflects the additional information retained by the fundamental group. Homology records the commutative aggregate of one-dimensional cycles, while the fundamental group preserves the order in which loops are traversed. A space with a perfect nontrivial fundamental group has (H_1(X;\mathbb{Z})=0) without being simply connected.

For a simply connected space, the Hurewicz theorem relates the first nonzero higher homotopy group to the corresponding homology group. This relation makes simple connectedness a natural initial condition in the study of higher-dimensional homotopy, although it does not by itself determine those higher groups.