ConceptComplete

Connected and Disconnected Spaces

Connectedness captures the intuitive notion that a space is "in one piece." It is one of the fundamental topological properties, preserved by continuous maps, and forms the basis for the intermediate value theorem and many existence results.


Definition of Connectedness

Definition4.1Connected Space

A topological space XX is connected if it cannot be written as a union of two nonempty disjoint open sets. Equivalently, XX is connected if the only subsets of XX that are both open and closed (clopen) are \emptyset and XX itself.

A space that is not connected is called disconnected.

Definition4.2Separation

A separation (or disconnection) of a topological space XX is a pair of nonempty open sets U,VXU, V \subseteq X such that X=UVX = U \cup V and UV=U \cap V = \emptyset. A space is connected if and only if it admits no separation.

Equivalently, {U,V}\{U, V\} is a separation of XX if and only if UU and VV are nonempty, UV=XU \cup V = X, and neither UU nor VV contains a limit point of the other.

ExampleBasic Examples
  1. R\mathbb{R} is connected (proved below as a theorem).
  2. Q\mathbb{Q} is disconnected: Q=(Q(,2))(Q(2,))\mathbb{Q} = (\mathbb{Q} \cap (-\infty, \sqrt{2})) \cup (\mathbb{Q} \cap (\sqrt{2}, \infty)) is a separation.
  3. The discrete topology on a set with more than one point is disconnected: every singleton is clopen.
  4. The indiscrete topology on any set is connected (the only clopen sets are \emptyset and XX).
  5. {0}{1}\{0\} \cup \{1\} with the subspace topology from R\mathbb{R} is disconnected.

Connected Subsets

Definition4.3Connected Subset

A subset AA of a topological space XX is connected if AA is connected in the subspace topology. Equivalently, there do not exist open sets U,VU, V in XX with AUVA \subseteq U \cup V, AUA \cap U \neq \emptyset, AVA \cap V \neq \emptyset, and AUV=A \cap U \cap V = \emptyset.

Theorem4.1Properties of Connected Sets

Let XX be a topological space.

  1. If CC is connected and CACC \subseteq A \subseteq \overline{C}, then AA is connected.
  2. If {Cα}\{C_\alpha\} is a family of connected subsets with αCα\bigcap_\alpha C_\alpha \neq \emptyset, then αCα\bigcup_\alpha C_\alpha is connected.
  3. The continuous image of a connected space is connected.
Proof

(1): Suppose A=UVA = U \cup V is a separation. Since CAC \subseteq A is connected, CC lies entirely in UU or entirely in VV; say CUC \subseteq U. Then CU\overline{C} \subseteq \overline{U}. Since UU and VV are disjoint and open in AA, UV=\overline{U} \cap V = \emptyset in AA. Thus ACUA \subseteq \overline{C} \subseteq \overline{U}, giving V=AUUUV = A \setminus U \subseteq \overline{U} \setminus U. But VV is open and UU\overline{U} \setminus U contains no open subset of AA, so V=V = \emptyset, contradicting the assumption.

(2): Fix pαCαp \in \bigcap_\alpha C_\alpha. If αCα=UV\bigcup_\alpha C_\alpha = U \cup V is a separation, then pp belongs to exactly one, say pUp \in U. Each CαC_\alpha is connected and contains pUp \in U, so CαUC_\alpha \subseteq U. Hence αCαU\bigcup_\alpha C_\alpha \subseteq U, giving V=V = \emptyset, a contradiction.


Connectedness of Intervals

Theorem4.2Connected Subsets of $\mathbb{R}$

A subset ARA \subseteq \mathbb{R} is connected if and only if AA is an interval (including rays and R\mathbb{R} itself).

This is a deep result that relies on the completeness of R\mathbb{R}. We prove the forward direction here; the converse is proved in the theorems section.

Proof

(\Leftarrow, sketch): Suppose AA is an interval and A=UVA = U \cup V is a separation with aUa \in U and bVb \in V, a<ba < b. Let c=sup(U[a,b])c = \sup(U \cap [a, b]). Since UU is closed in AA (it equals AVA \setminus V), cUc \in U. Since UU is open in AA, there exists ϵ>0\epsilon > 0 with (cϵ,c+ϵ)AU(c - \epsilon, c + \epsilon) \cap A \subseteq U. If c<bc < b, then c+ϵUc + \epsilon' \in U for small ϵ\epsilon', contradicting c=sup(U[a,b])c = \sup(U \cap [a,b]). If c=bc = b, then bUV=b \in U \cap V = \emptyset, a contradiction.


Operations on Connected Spaces

ExampleProducts of Connected Spaces

If XX and YY are connected, then X×YX \times Y is connected. More generally, an arbitrary product αXα\prod_\alpha X_\alpha is connected (in the product topology) if and only if each XαX_\alpha is connected.

For the finite case: fix (a,b)X×Y(a, b) \in X \times Y. For any (x,y)X×Y(x, y) \in X \times Y, the sets X×{b}X \times \{b\} and {x}×Y\{x\} \times Y are connected (homeomorphic to XX and YY respectively) and share the point (x,b)(x, b). So X×{b}{x}×YX \times \{b\} \cup \{x\} \times Y is connected. Since all these sets share the point (a,b)(a, b), their union X×YX \times Y is connected.

RemarkTopologist's Sine Curve

The topologist's sine curve S={(x,sin(1/x)):x>0}{(0,y):1y1}S = \{(x, \sin(1/x)) : x > 0\} \cup \{(0, y) : -1 \leq y \leq 1\} is connected but not path-connected. It demonstrates that these two notions of connectedness are genuinely different.

SS is connected because it is the closure of the graph {(x,sin(1/x)):x>0}\{(x, \sin(1/x)) : x > 0\}, which is a continuous image of (0,)(0, \infty) (hence connected), and the closure of a connected set is connected.