Connectedness in Metric Spaces

Understanding when a space is "in one piece" is a fundamental concern across analysis, topology, and their applications. Connectedness formalizes this intuitive idea, separating spaces that are intrinsically unified from those that can be decomposed into isolated parts. This concept is crucial for extending core results like the Intermediate Value Theorem, analyzing the structure of spaces, and underpinning arguments in calculus, differential equations, and geometry.

The Formal Definition of a Connected Set

In a metric space $(X,d)$, we are not interested in physical separation but in separation by open sets. A subset $A\subseteq X$ is connected if it cannot be expressed as the union of two non-empty, disjoint sets that are *open in the subspace topology of $A$*. Formally, $A$ is disconnected if there exist open sets $U$ and $V$ in $X$ such that:

1. $A\subseteq U\cup V$
2. $A\cap U\ne \varnothing$ and $A\cap V\ne \varnothing$
3. $A\cap U\cap V=\varnothing$

If no such $U$ and $V$ exist, then $A$ is connected. The classic example is an interval in $\mathbb{R}$. The interval $[0,1]$ is connected. In contrast, the set $[0,1]\cup [2,3]$ is disconnected, as you can choose $U=(-1,1.5)$ and $V=(1.5,4)$ to satisfy the conditions above. A more subtle example is the set of rational numbers $\mathbb{Q}$. It is disconnected because you can separate it into, for instance, rationals less than $\sqrt{2}$ and rationals greater than $\sqrt{2}$ using open intervals in $\mathbb{R}$.

A key theorem states that the continuous image of a connected set is connected. If $f:X\to Y$ is a continuous function and $A\subseteq X$ is connected, then $f(A)$ is connected in $Y$. This property is powerful for proving global results from local assumptions.

Connectedness and the Intermediate Value Theorem

The Intermediate Value Theorem (IVT) from calculus is a quintessential connectedness result. Its generalization can be stated as follows: Let $f:X\to \mathbb{R}$ be a continuous function from a connected metric space $X$ to $\mathbb{R}$. If $f$ takes two distinct values $a$ and $b$, then it takes every value between $a$ and $b$.

The proof leverages the connectedness property. Consider two points $p,q\in X$ with $f(p)=a$ and $f(q)=b$. The image $f(X)$ must be a connected subset of $\mathbb{R}$. The connected subsets of $\mathbb{R}$ are precisely the intervals. Since $a,b\in f(X)$, the entire interval $[a,b]$ (or $[b,a]$) must be contained in $f(X)$. This not only proves the theorem but also characterizes connected subsets of $\mathbb{R}$: a set in $\mathbb{R}$ is connected if and only if it is an interval. This equivalence makes $\mathbb{R}$ a primary arena for developing intuition about connectedness.

Path-Connectedness: A Stronger Form of Unity

A more intuitive, but often stronger, concept is path-connectedness. A metric space $X$ is path-connected if for every pair of points $x,y\in X$, there exists a continuous function $\gamma :[0,1]\to X$ such that $\gamma (0)=x$ and $\gamma (1)=y$. This function $\gamma$ is called a path from $x$ to $y$.

Path-connectedness is generally more restrictive than connectedness. Every path-connected space is connected. The reasoning is straightforward: if a path-connected space could be split into two disjoint open sets $U$ and $V$, a path from a point in $U$ to a point in $V$ would have to cross the gap. Since the image of the connected interval $[0,1]$ under the continuous path must be connected, this is impossible.

The converse, however, is false. A connected space need not be path-connected. The standard counterexample is the topologist's sine curve, defined as: $$S=\{(x,\sin (1/x))\mid x\in (0,1]\}\cup \{(0,y)\mid y\in [-1,1]\}.$$ This set $S$ in ${\mathbb{R}}^2$ is connected but not path-connected. There is no continuous path connecting a point on the oscillating curve to a point on the vertical segment at $x=0$. This example is vital for distinguishing between these two forms of connectedness in analysis.

Connected Components and Topological Applications

For spaces that are not connected, we can partition them into maximal connected pieces. The connected component of a point $x$ in a metric space $X$ is the union of all connected subsets of $X$ that contain $x$. This component is itself connected and is the largest connected set containing $x$. Two points are in the same connected component if and only if there exists a connected subset containing both.

Key properties of components include:

• Every space is the disjoint union of its connected components.
• Connected components are always closed subsets of the space (though they are not necessarily open, as seen in $\mathbb{Q}$).
• Continuous functions map components into components.

In path-connected spaces, we have the analogous concept of path-components. The path-component of a point is the set of all points that can be joined to it by a path. Path-components can be finer than connected components, as shown by the topologist's sine curve, which is one connected component but two distinct path-components.

These concepts have direct topological applications. They are topological invariants, meaning that if two spaces are homeomorphic, they must have the same number of connected components and path-components. This makes connectedness a primary tool for distinguishing between spaces. For instance, a circle is connected and path-connected, while two disjoint circles are not connected, proving they are not homeomorphic to a single circle.

Common Pitfalls

1. Assuming path-connectedness and connectedness are equivalent. This is perhaps the most frequent error. Always remember: path-connected $⟹$ connected, but not vice-versa. The topologist's sine curve is the canonical counterexample to keep in mind to avoid this trap.
2. Misidentifying connected components. A connected component is a maximal connected subset. It's not enough for a subset to be connected; it must not be properly contained in a larger connected set. Furthermore, components are always closed in the space, but students often mistakenly try to prove they are open (they are only open if the space has finitely many components).
3. Confusing "open in $X$" with "open in the subspace $A$" in the definition. When testing if $A\subseteq X$ is connected, the separating sets $U$ and $V$ must be open in $X$, but the conditions $A\cap U$ and $A\cap V$ must be non-empty and disjoint. These intersections are open *in the subspace topology of $A$*, which is the correct framework for the definition.
4. Overlooking the role of the codomain in the IVT generalization. The generalized IVT requires the codomain to be $\mathbb{R}$. The theorem relies on the fact that connected subsets of $\mathbb{R}$ are intervals. This property does not hold for ${\mathbb{R}}^2$ or other spaces, so a direct analog of the IVT does not exist there.

Summary

• A metric space is connected if it cannot be partitioned into two non-empty, disjoint subsets that are open in its subspace topology. This is the foundational topological notion of being "in one piece."
• The Intermediate Value Theorem is a direct consequence of connectedness: the continuous image of a connected set is connected, and in $\mathbb{R}$, connected sets are intervals.
• A space is path-connected if any two points can be joined by a continuous path. This implies connectedness but is a strictly stronger condition, as demonstrated by counterexamples like the topologist's sine curve.
• Any space can be decomposed into its maximal connected parts, called connected components. These components are closed, disjoint, and collectively partition the space. Path-components provide a similar, sometimes finer, partition.
• Connectedness and its related concepts are fundamental topological invariants used to classify and distinguish between different metric and topological spaces.