Connectedness and Path-Connectedness

A topological space $X$ is connected if it cannot be written as the union of two disjoint, nonempty open sets; equivalently, the only subsets that are simultaneously open and closed ("clopen") are $\varnothing$ and $X$ itself. A space is path-connected if for any two points $a,b\in X$ there is a continuous map $\gamma :[0,1]\to X$ with $\gamma (0)=a$ and $\gamma (1)=b$. These notions formalize the intuitive idea of a space being "in one piece." They matter because connectedness is a topological invariant — preserved by homeomorphisms — and underwrites foundational results like the Intermediate Value Theorem. The surprising fact: connectedness and path-connectedness are genuinely different; there exist connected spaces in which most pairs of points cannot be joined by any path at all.

The Core Insight

Connectedness is a negative, global property — defined by the impossibility of a certain decomposition. Path-connectedness is a positive, constructive property — defined by the existence of joining maps. The constructive notion is stronger (it implies the negative one), but the negative notion can hold "by accident" in pathological spaces where no actual paths exist. The bridge between them is local path-connectedness: where a space is locally nice, the two notions coincide.

Formal Statement

Separation and Connectedness

Let $(X,\tau )$ be a topological space. A separation of $X$ is a pair of open sets $U,V\in \tau$ satisfying

The space $X$ is connected if no separation exists. A subset $A\subseteq X$ is connected if it is connected in the subspace topology.

Symbols: $\tau$ is the collection of open sets; $U,V$ are candidate clopen blocks; $\varnothing$ is the empty set.

Path-Connectedness

A path in $X$ from $a$ to $b$ is a continuous map

Clopen Characterization

Connectedness via Continuous Maps to a Two-Point Set

Let $\{0,1\}$ carry the discrete topology. Then $X$ is connected if and only if every continuous map $f:X\to \{0,1\}$ is constant:

Path-Connected ⟹ Connected

Every path-connected space is connected.

Path-connected implies connected

1. Suppose not. Let $X$ be path-connected but admit a separation $(U,V)$ as in equation [1].
2. Since $U,V$ are nonempty, pick $a\in U$ and $b\in V$.
3. By path-connectedness there is a continuous $\gamma :[0,1]\to X$ with $\gamma (0)=a$, $\gamma (1)=b$.
4. Consider the preimages ${\gamma }^{-1}(U)$ and ${\gamma }^{-1}(V)$. Each is open in $[0,1]$ (continuity), they are disjoint, their union is all of $[0,1]$, and both are nonempty ($0\in {\gamma }^{-1}(U)$, $1\in {\gamma }^{-1}(V)$).
5. Hence $({\gamma }^{-1}(U),{\gamma }^{-1}(V))$ is a separation of $[0,1]$.
6. But $[0,1]$ is connected (a standard fact: any clopen subset of an interval containing $0$ is, by the supremum argument, all of it). Contradiction.
7. Therefore no separation of $X$ exists; $X$ is connected. $■$

:::proof[Interval $[0,1]$ is connected — the keystone lemma]

1. Suppose $[0,1]=U\bigsqcup V$ with $U,V$ open, disjoint, nonempty; say $0\in U$.
2. Define $s=\sup \{x:[0,x]\subseteq U\}$. The set is nonempty (contains $0$) and bounded, so $s$ exists.
3. Claim $s\in U$ is impossible if $s<1$, and $s\in V$ is impossible too — examine cases. If $s\in U$: openness gives an interval $(s-\delta ,s+\delta )\subseteq U$, so $[0,s+\delta /2]\subseteq U$, contradicting that $s$ is the supremum (unless $s=1$).
4. If $s\in V$: openness gives $(s-\delta ,s+\delta )\subseteq V$, but then points just below $s$ lie in $V$, contradicting $[0,s^{\prime })\subseteq U$ for $s^{\prime }$ near $s$.
5. The only consistent outcome is $s=1$ and $[0,1]\subseteq U$, forcing $V=\varnothing$ — contradiction.
6. Thus $[0,1]$ has no separation and is connected. $■$

Felix Hausdorff | 1868–1942

Hausdorff's 1914 Grundzüge der Mengenlehre gave topology its modern axiomatic footing and clean treatments of connectedness within general (now "Hausdorff") spaces. Building on Cantor's point-set work and Fréchet's metric spaces, Hausdorff separated the intuitive idea of a connected continuum from the set-theoretic machinery needed to state it rigorously. His framework let later workers (Knaster, Kuratowski, Sierpiński) construct the bizarre connected-but-not-path-connected and totally-disconnected examples that pin down exactly how subtle the concept is. A polymath who also wrote poetry under the pen name "Paul Mongré," he died tragically in 1942 to escape Nazi deportation.

Building Intuition

Picture connectedness as the question: "Can I draw a fence that completely separates the space into two camps, with no point left on the fence?" If every attempted fence must either pass through the space (leaving boundary points unassigned to an open camp) or leave one camp empty, the space is connected. Path-connectedness is the stronger, more tactile idea: "Can I physically walk from any point to any other without leaving the space?" Walking is constructive — you produce an actual continuous trajectory. Connectedness only forbids a clean split; it doesn't promise you can travel. The topologist's sine curve is the cautionary tale: it is "one piece" (no clean fence exists), yet you cannot walk from the wiggling tail to the vertical segment because the oscillations get infinitely dense — any walking path would have to "jump."

The topologist's sine curve: connected but not path-connected

<svg viewBox="0 0 460 280" xmlns="http://www.w3.org/2000/svg">
<rect width="460" height="280" fill="#0a0e17"/>
<line x1="60" y1="140" x2="430" y2="140" stroke="#94a3b8" stroke-width="1"/>
<line x1="80" y1="40" x2="80" y2="240" stroke="#94a3b8" stroke-width="1"/>
<text x="40" y="40" fill="#e2e8f0" font-size="12">y</text>
<text x="430" y="160" fill="#e2e8f0" font-size="12">x</text>
<!-- vertical segment (limit set) -->
<line x1="80" y1="50" x2="80" y2="230" stroke="#94a3b8" stroke-width="3"/>
<text x="20" y="145" fill="#e2e8f0" font-size="11">segment {0}×[-1,1]</text>
<!-- sine wiggle key curve -->
<path d="M82,140 Q90,55 100,140 Q108,225 120,140 Q126,60 138,140 Q145,222 160,140 Q170,70 188,140 Q200,210 225,140 Q245,90 280,140 Q310,180 360,140 Q400,125 430,140"
 stroke="#d4a853" stroke-width="2" fill="none"/>
<text x="250" y="60" fill="#d4a853" font-size="12">y = sin(1/x), x &gt; 0</text>
<text x="180" y="265" fill="#e2e8f0" font-size="11">oscillations infinitely dense as x→0⁺ — no path crosses the gap</text>
</svg>

Worked Examples

:::example[Showing $[0,3]\cup \{5\}$ is disconnected — with explicit separation] Given: $X=[0,3]\cup \{5\}\subset \mathbb{R}$ with the subspace topology.

1. Take $U=(-1,4)\cap X=[0,3]$ and $V=(4,6)\cap X=\{5\}$.
2. Both are open in $X$ (intersections of open reals with $X$).
3. Check $U\cap V=\varnothing$, $U\cup V=X$, and both nonempty.
4. This is a valid separation per equation [1].

The set $GL_n(\mathbb{R})$ is disconnected; counting components

1. The determinant $\det :G{L}_n(\mathbb{R})\to \mathbb{R}\setminus \{0\}$ is continuous (polynomial in entries).
2. $\mathbb{R}\setminus \{0\}=(-\infty ,0)\bigsqcup (0,\infty )$ is disconnected with 2 components.
3. The preimages ${\det }^{-1}(0,\infty )$ and ${\det }^{-1}(-\infty ,0)$ are disjoint nonempty open sets covering $G{L}_n$.
4. Each piece is path-connected (provable via row operations / polar decomposition), so there are exactly two components.

Common Misconceptions

The topologist's sine curve

Let

Connected, locally connected nowhere — and totally disconnected $\mathbb{Q}$

The rationals $\mathbb{Q}\subset \mathbb{R}$ are totally disconnected: the only connected subsets are singletons. Between any two rationals $p<q$ lies an irrational $\alpha$, giving the separation $((-\infty ,\alpha )\cap \mathbb{Q},(\alpha ,\infty )\cap \mathbb{Q})$. Yet $\mathbb{Q}$ is not discrete — a vivid reminder that "totally disconnected" ≠ "discrete."

Connections

Connectedness threads through nearly all of topology and analysis. It is the engine behind the Intermediate Value Theorem (a continuous image of a connected set is connected, and connected subsets of $\mathbb{R}$ are exactly intervals). It refines into Connected Components (maximal connected subsets, which partition any space) and Path Components. The local versions — Local Connectedness and Local Path-Connectedness — control when the two notions coincide. In algebraic topology, path components are precisely the elements of ${\pi }_0(X)$, the Fundamental Group's zeroth-order shadow. Connectedness also interacts with Compactness (continua = compact connected Hausdorff spaces) and with Quotient Spaces (quotients of connected spaces are connected).

Real-World Applications

1. Complex analysis / domains: Holomorphic functions are studied on connected open sets ("domains"); the Identity Theorem requires connectedness so that local agreement forces global agreement.
2. Image processing: Connected-component labeling segments digital images into distinct objects (foreground blobs) using discrete connectivity (4- or 8-neighborhoods).
3. Robotics / motion planning: A robot can move between two configurations iff they lie in the same path component of the free configuration space; path-connectedness is literally solvability of navigation.
4. Electrical networks: A circuit's nodes form a connected graph iff current can flow between any two terminals; connectivity analysis underlies fault detection.
5. Phase transitions / percolation theory: Physicists study when a random medium develops a connected "spanning cluster," a connectedness phase transition central to materials science.
6. GIS and road networks: Determining whether all cities are reachable is a connectivity query on a graph embedded in a topological space.
7. Data analysis (TDA): Persistent $H_0$ homology tracks connected components across scales to extract cluster structure from point clouds.

Historical Development

• 1872 — Dedekind/Cantor: Rigorous construction of $\mathbb{R}$ and the line continuum sets the stage for analyzing "continuity" of the space itself.
• 1883 — Cantor: Introduces perfect and connected point sets in his work on the continuum; early definitions of "connectedness" via point accumulation.
• 1893 — Camille Jordan: The Jordan Curve Theorem implicitly relies on separating connected regions of the plane, motivating careful topology of connectedness.
• 1906 — Maurice Fréchet: Introduces metric spaces ("L-spaces"), giving a setting where connectedness can be defined and tested abstractly.
• 1914 — Felix Hausdorff: Grundzüge der Mengenlehre axiomatizes neighborhood-based topology; connectedness gets its modern general-space formulation.
• 1921 — Knaster & Kuratowski: Construct exotic connected sets (the Knaster–Kuratowski "fan"), showing how pathological connectedness can be.
• 1922 — Kuratowski: Closure-axiom foundation for topology cleanly encodes clopen sets and the connectedness criterion.
• 1920s — R. L. Moore school: Develops the theory of continua (compact connected metric spaces), classifying indecomposable continua and refining path vs. arc-connectedness.
• 1930s — Bourbaki: Systematizes connectedness, components, and the path-connected/connected distinction in the canonical Topologie générale, fixing today's terminology.

Subtleties & Limits

While connectedness for "nice" spaces (manifolds, CW complexes, open subsets of ${\mathbb{R}}^n$) is fully understood — there the two notions coincide — the structure of indecomposable continua and the classification of all connected, totally path-disconnected planar sets remains subtle and combinatorially rich. Questions about homogeneous continua (e.g., the pseudo-arc) and how connectedness behaves under exotic quotients and inverse limits continue to generate research in continuum theory. There is no finite algorithm to decide connectedness of an arbitrary semianalytic set in full generality, though for semialgebraic sets cylindrical algebraic decomposition gives a (doubly-exponential) procedure — the precise complexity frontier is still studied.

See Also