Compactness and Tychonoff's Theorem

💡 The Core Insight

Compactness captures the essence of "finiteness" in an infinite setting: every attempt to cover a compact space by infinitely many open sets can be reduced to a finite subcollection. Tychonoff's theorem reveals that this property is preserved under arbitrary products, making compact spaces the "atoms" from which complex topological structures are built.

Definition

Compactness: A topological space $(X, \tau)$ is compact if every open cover of $X$ has a finite subcover. That is, if $\{U_\alpha\}_{\alpha \in A}$ is any collection of open sets with $X = \bigcup_{\alpha \in A} U_\alpha$ , then there exist finitely many indices $\alpha_1, \ldots, \alpha_n \in A$ such that $X = U_{\alpha_1} \cup \cdots \cup U_{\alpha_n}$ .

Tychonoff's Theorem: Let $\{(X_\alpha, \tau_\alpha)\}_{\alpha \in A}$ be a family of topological spaces indexed by an arbitrary set $A$ . If each $X_\alpha$ is compact, then the product space $\prod_{\alpha \in A} X_\alpha$ equipped with the Product Topology is compact.

TheoremMain Result

Theorem 1 (Tychonoff's Theorem): Let $\{X_\alpha\}_{\alpha \in A}$ be a family of compact topological spaces. Then $X = \prod_{\alpha \in A} X_\alpha$ is compact in the product topology.

Theorem 2 (Characterization by the Finite Intersection Property): A topological space $X$ is compact if and only if every family of closed sets with the finite intersection property has nonempty intersection. (A family $\mathcal{F}$ of sets has the finite intersection property if every finite subfamily has nonempty intersection.)

Theorem 3 (Compactness in Hausdorff Spaces): In a Hausdorff Space, every compact subset is closed, and every closed subset of a compact space is compact.

Proof

Proof of Theorem 1 (Tychonoff's Theorem using Alexander Subbase Theorem): We employ the Alexander Subbase Theorem, which states that a space is compact if every cover by subbasic open sets has a finite subcover.

Step 1: Recall that the product topology on $X = \prod_{\alpha \in A} X_\alpha$ has as a subbase all sets of the form $\pi_\beta^{-1}(U)$ where $\beta \in A$ and $U$ is open in $X_\beta$ , and $\pi_\beta: X \to X_\beta$ is the projection map.

Step 2: Let $\mathcal{S}$ be a cover of $X$ by subbasic open sets. We must show $\mathcal{S}$ has a finite subcover. We may write

\mathcal{S} = \bigcup_{\alpha \in A} \mathcal{S}_\alpha

where $\mathcal{S}_\alpha = \{\pi_\alpha^{-1}(U) : \pi_\alpha^{-1}(U) \in \mathcal{S}\}$ consists of those subbasic sets coming from the $\alpha$ -th coordinate.

Step 3: Suppose by contradiction that no finite subfamily of $\mathcal{S}$ covers $X$ . We claim that for each $\alpha \in A$ , the collection $\mathcal{S}_\alpha$ does not cover $X$ .

Step 4: Proof of claim: If $\mathcal{S}_\alpha$ covered $X$ , then $\pi_\alpha(X) = X_\alpha$ would be covered by $\{U : \pi_\alpha^{-1}(U) \in \mathcal{S}_\alpha\}$ . Since $X_\alpha$ is compact, finitely many sets $U_1, \ldots, U_n$ would cover $X_\alpha$ . Then $\{\pi_\alpha^{-1}(U_1), \ldots, \pi_\alpha^{-1}(U_n)\}$ would be a finite subcover of $X$ from $\mathcal{S}$ , contradicting our assumption.

Step 5: Since $\mathcal{S}_\alpha$ does not cover $X$ , there exists a point $x^{(\alpha)} \in X$ not in any element of $\mathcal{S}_\alpha$ . Define $y = (y_\alpha)_{\alpha \in A}$ where $y_\alpha = \pi_\alpha(x^{(\alpha)})$ .

Step 6: The point $y \in X$ must lie in some element of $\mathcal{S}$ , say $y \in \pi_\beta^{-1}(U)$ for some $\beta \in A$ and open $U \subseteq X_\beta$ . This means $\pi_\beta(y) = y_\beta \in U$ .

Step 7: By construction, $y_\beta = \pi_\beta(x^{(\beta)})$ , so $x^{(\beta)} \in \pi_\beta^{-1}(U) \in \mathcal{S}_\beta$ . This contradicts the choice of $x^{(\beta)}$ as not being in any element of $\mathcal{S}_\beta$ . This contradiction shows that $\mathcal{S}$ must have a finite subcover. By the Alexander Subbase Theorem, $X$ is compact. □

∎

⚠️ Common mistake

Common Misconception: Students often believe that a product of compact spaces is compact in the Box Topology (where basic open sets are arbitrary products of open sets, not just those with finitely many non-whole factors). This is false! The classic counterexample is $\prod_{n=1}^\infty [0,1]$ with the box topology, which is not compact. The product topology is strictly coarser than the box topology, and this difference is essential for Tychonoff's theorem. The proof fundamentally uses that basic open sets in the product topology have only finitely many "interesting" coordinates.

mathematician

Andrey Nikolayevich Tychonoff (1930) — Proved that arbitrary products of compact spaces are compact, establishing one of the most powerful tools in general topology and demonstrating the deep connection between compactness and the axiom of choice.

visualization

Example1

The Hilbert Cube is Compact Consider the Hilbert cube $H = \prod_{n=1}^\infty [0, 1/n]$ with the product topology inherited from $\mathbb{R}^\infty$ . Solution: Each factor $[0, 1/n]$ is a closed and bounded interval in $\mathbb{R}$ , hence compact by the Heine-Borel Theorem. By Tychonoff's Theorem, the product

H = [0,1] \times [0, 1/2] \times [0, 1/3] \times \cdots

is compact in the product topology. Note that $H$ is a subspace of the metric space $\ell^2$ (under the natural embedding), and the product topology on $H$ coincides with the subspace topology from $\ell^2$ . This gives an alternative proof: $H$ is a closed subset of the unit ball in $\ell^2$ , and while the unit ball itself is not compact, $H$ is totally bounded and complete in the inherited metric, hence compact by the metric space characterization of compactness. The Hilbert cube is homeomorphic to $[0,1]^\mathbb{N}$ and serves as a universal compact metric space: every compact metric space embeds in the Hilbert cube.

Example2

Stone-Čech Compactification Application Let $X$ be a completely regular space and $\beta X$ its Stone-Čech Compactification. We can construct $\beta X$ as a subspace of a product of compact spaces. Construction: For each bounded continuous function $f: X \to \mathbb{R}$ , let $I_f = [\inf f, \sup f]$ (a compact interval). Consider the evaluation map

e: X \to \prod_{f \in C^*(X)} I_f, \quad e(x) = (f(x))_{f \in C^*(X)}

where $C^*(X)$ denotes all bounded continuous real-valued functions on $X$ . By Tychonoff's theorem, $Y = \prod_{f \in C^*(X)} I_f$ is compact. The image $e(X)$ is dense in its closure $\overline{e(X)} \subseteq Y$ . This closure is compact (being a closed subset of a compact space) and can be shown to be $\beta X$ up to homeomorphism. The compactness of the product is essential—without Tychonoff's theorem, this construction of the Stone-Čech compactification would fail.

Counterexample

Failure Without Compactness: $\mathbb{R}^\mathbb{N}$ is Not Compact Consider $X = \prod_{n=1}^\infty \mathbb{R}$ with the product topology. Each factor $\mathbb{R}$ is not compact (the open cover $\{(-n, n)\}_{n \in \mathbb{N}}$ has no finite subcover). Tychonoff's theorem does not apply. Indeed, $X$ is not compact: Consider the open cover

\mathcal{U} = \left\{\prod_{n=1}^\infty (-n, n)\right\} \cup \left\{\pi_k^{-1}(\mathbb{R} \setminus [-k, k]) : k \in \mathbb{N}\right\}

For any finite subcollection, if we omit the set $\prod_{n=1}^\infty (-n, n)$ , we can only exclude points where some coordinate lies outside finitely many bounded intervals. But the point $(1, 2, 3, 4, \ldots)$ (with $k$ -th coordinate equal to $k$ ) is not covered by any finite subcollection. This illustrates that all factors must be compact for Tychonoff's theorem to apply. Compactness of a product is not preserved when even one factor is non-compact (in infinite products).

🧠 Intuition

Mental Model: Think of compactness as "inability to escape to infinity." In $\mathbb{R}^n$ , the Heine-Borel theorem says compact = closed + bounded, capturing both "no holes" and "no infinitely distant points." For products, imagine each factor $X_\alpha$ as a dimension. A point in the product specifies one coordinate in each dimension. An open cover tries to "chase" all these infinite coordinates simultaneously. But here's the key: in the product topology, a basic open set restricts only finitely many coordinates—all others can be arbitrary. Tychonoff's theorem works because we can "coordinate" the compactness: when trying to build a finite subcover, we handle each dimension's compactness separately. The product topology's weak structure (only finitely many coordinates matter in each basic open set) allows the compactness of individual factors to collectively control the entire infinite-dimensional space. Visualize it as: each factor $X_\alpha$ says "I can be covered by finitely many sets," and the product topology lets us combine these finite collections across all coordinates without losing control, even though there are infinitely many coordinates.

connection

Functional Analysis and Weak- Compactness</strong>: Tychonoff's theorem directly implies the <span class="wiki-link-missing">Banach-Alaoglu Theorem</span>, one of functional analysis's fundamental results. The theorem states that the closed unit ball in the dual space $X^*$ of a normed vector space is compact in the weak- topology. The proof embeds $X^*$ into $\prod_{x \in X} \mathbb{C}$ (or $\mathbb{R}$ for real spaces) via $\phi \mapsto (\phi(x))_{x \in X}$ . The unit ball embeds into $\prod_{x \in X} \overline{B}(0, \|x\|)$ where $\overline{B}(0, r)$ is the closed ball of radius $r$ in $\mathbb{C}$ . Each factor is compact, so Tychonoff gives compactness of the product, and the unit ball (being a closed subset) is compact. This result underpins reflexivity theory, existence of minimizers in optimization, and the theory of distributions.

Compactness and Tychonoff's Theorem

Overview

Core Theory

Examples and Applications

Historical Development

Connections to Other Areas

What You Need First

What This Opens Up

See Also