Topological Spaces, Continuity, and Compactness

⚡ Key Takeaways

Topological spaces generalize the notion of "nearness" without reference to distance: a topology is a collection of "open sets" satisfying three axioms. Continuity becomes the requirement that preimages of open sets are open. Compactness is the abstraction of "closed and bounded" — every open cover admits a finite subcover. Key result: a continuous image of a compact space is compact, generalizing the Extreme Value Theorem. Used throughout analysis, geometry, and algebra. Foundations laid by Hausdorff (1914) building on Cantor, Fréchet, and Riesz.

💡 The Core Insight

Forget distance. Keep only the question: which sets are "open"? Everything analytic — limits, continuity, compactness — can be rephrased purely in terms of open sets. Continuity is not about "small changes producing small changes" but about open sets pulling back to open sets. Compactness is not "closed and bounded" (a metric accident) but every open cover has a finite subcover — a finiteness principle that turns local information into global conclusions.

DefinitionTopological Space

A topological space is a pair $(X, \tau)$ where $X$ is a set and $\tau \subseteq \mathcal{P}(X)$ is a collection of subsets called open sets, satisfying:

1

\quad \varnothing \in \tau \quad\text{and}\quad X \in \tau

Derivation hint: The empty set and the whole space are always open — these are axioms, the "boundary cases" needed for the union/intersection rules to be coherent.

2

\quad \{U_i\}_{i \in I} \subseteq \tau \;\Longrightarrow\; \bigcup_{i \in I} U_i \in \tau \quad (I \text{ arbitrary})

Derivation hint: Arbitrary (even uncountable) unions of open sets are open; this asymmetry with intersections is the heart of topology.

3

\quad U_1, \dots, U_n \in \tau \;\Longrightarrow\; \bigcap_{k=1}^{n} U_k \in \tau \quad (n < \infty)

Derivation hint: Only finite intersections are guaranteed open; allowing infinite intersections would collapse the theory (see the warning below). Here $\mathcal P(X)$ is the power set, $I$ an arbitrary index set. A set $C \subseteq X$ is closed iff its complement $X \setminus C$ is open. The interior $\operatorname{int}(A)$ is the largest open set contained in $A$ ; the closure $\overline{A}$ is the smallest closed set containing $A$ ; the boundary is $\partial A = \overline A \setminus \operatorname{int}(A)$ .

DefinitionBasis, Subbasis, and Generated Topologies

A collection $\mathcal{B} \subseteq \tau$ is a basis for $\tau$ if every open set is a union of members of $\mathcal{B}$ . A family $\mathcal{B} \subseteq \mathcal{P}(X)$ is a basis for some topology iff it covers $X$ and:

4

\quad \forall\, B_1, B_2 \in \mathcal B,\ \forall\, x \in B_1 \cap B_2,\ \exists\, B_3 \in \mathcal B :\ x \in B_3 \subseteq B_1 \cap B_2

Derivation hint: This is exactly the condition allowing $B_1 \cap B_2$ to be re-expressed as a union of basis elements, ensuring closure under finite intersection. The metric topology on a metric space $(X,d)$ takes as basis the open balls $B(x,r) = \{y : d(x,y) < r\}$ .

DefinitionContinuity

Let $f : (X,\tau_X) \to (Y,\tau_Y)$ . Then $f$ is continuous iff

5

\quad \forall\, V \in \tau_Y :\ f^{-1}(V) \in \tau_X

Derivation hint: "Preimage of open is open." Note we pull back along $f$ , not push forward — images of open sets need not be open. Equivalently, $f$ is continuous iff $f^{-1}(C)$ is closed for every closed $C \subseteq Y$ , iff $f(\overline A) \subseteq \overline{f(A)}$ for all $A \subseteq X$ , iff $f$ is continuous at every point $x$ :

6

\quad \forall\, V \ni f(x)\ \text{open},\ \exists\, U \ni x\ \text{open} :\ f(U) \subseteq V

Derivation hint: The local $\varepsilon$ – $\delta$ statement with neighbourhoods replacing balls; collecting over all $x$ recovers the global preimage condition. A bijection $f$ with $f$ and $f^{-1}$ both continuous is a homeomorphism; homeomorphic spaces are topologically identical.

DefinitionCompactness

$X$ is compact iff every open cover has a finite subcover:

7

\quad X = \bigcup_{i \in I} U_i,\ U_i \in \tau \;\Longrightarrow\; \exists\, \{i_1,\dots,i_n\} \subseteq I :\ X = \bigcup_{k=1}^n U_{i_k}

Derivation hint: A finiteness extraction: any cover, no matter how redundant or infinite, can be trimmed to finitely many sets that still cover. A subset $K \subseteq X$ is compact iff it is compact as a subspace, equivalently every cover of $K$ by sets open in $X$ has a finite subcover.

TheoremContinuous Image of Compact is Compact

If $f : X \to Y$ is continuous and $X$ is compact, then $f(X) \subseteq Y$ is compact.

TheoremHeine–Borel

A subset $K \subseteq \mathbb{R}^n$ (standard topology) is compact iff it is closed and bounded.

TheoremExtreme Value Theorem (topological form)

If $X$ is compact and $f : X \to \mathbb{R}$ continuous, then $f$ attains a maximum and a minimum on $X$ .

ProofContinuous Image of Compact is Compact

Step 1. Let $\{V_i\}_{i\in I}$ be an open cover of $f(X)$ in $Y$ , so $f(X) \subseteq \bigcup_i V_i$ .

Step 2. By continuity, each $f^{-1}(V_i)$ is open in $X$ (equation [5]).

Step 3. These preimages cover $X$ : for any $x \in X$ , $f(x) \in f(X)$ lies in some $V_i$ , hence $x \in f^{-1}(V_i)$ . So $\{f^{-1}(V_i)\}_{i\in I}$ is an open cover of $X$ .

Step 4. By compactness of $X$ , extract a finite subcover $X = \bigcup_{k=1}^n f^{-1}(V_{i_k})$ .

Step 5. Apply $f$ : for any $y = f(x) \in f(X)$ , $x \in f^{-1}(V_{i_k})$ for some $k$ , so $y = f(x) \in V_{i_k}$ . Thus $f(X) \subseteq \bigcup_{k=1}^n V_{i_k}$ .

Step 6. We have produced a finite subcover of $f(X)$ . Hence $f(X)$ is compact. $\blacksquare$

∎

ProofExtreme Value Theorem from the above

Step 1. $f(X) \subseteq \mathbb{R}$ is compact by the previous theorem.

Step 2. By Heine–Borel, $f(X)$ is closed and bounded.

Step 3. Bounded $\Rightarrow$ $M = \sup f(X)$ exists in $\mathbb{R}$ .

Step 4. $M \in \overline{f(X)} = f(X)$ (closed), so $M = f(x_0)$ for some $x_0$ : the maximum is attained. The minimum follows by applying the argument to $-f$ . $\blacksquare$

∎

ProofClosed subset of a compact space is compact

Step 1. Let $C \subseteq X$ be closed, $X$ compact. Take an open cover $\{U_i\}$ of $C$ by sets open in $X$ .

Step 2. Add the open set $X \setminus C$ . Now $\{U_i\} \cup \{X \setminus C\}$ covers all of $X$ .

Step 3. Extract a finite subcover of $X$ : finitely many $U_{i_k}$ together with possibly $X \setminus C$ .

Step 4. Discard $X \setminus C$ (it contains no point of $C$ ). The remaining $U_{i_1}, \dots, U_{i_n}$ cover $C$ . $\blacksquare$

∎

mathematician · Felix Hausdorff | 1868–1942

Hausdorff, a German mathematician (and, under the pseudonym Paul Mongré, an essayist and playwright), gave topology its modern axiomatic form in Grundzüge der Mengenlehre (1914). He defined a space via neighbourhood systems satisfying four axioms — the fourth being the separation property now called the Hausdorff condition ( $T_2$ ): distinct points have disjoint neighbourhoods. His framework distilled the scattered ideas of Cantor, Fréchet, and Riesz into a usable theory. Persecuted under Nazi rule, Hausdorff took his own life with his wife and sister-in-law in 1942 to avoid deportation to a concentration camp. His name survives in Hausdorff spaces, Hausdorff dimension, and the Hausdorff metric.

🧠 Intuition

Think of open sets as the collection of "wiggle room" regions: an open set around $x$ is somewhere you can move $x$ a little and stay inside. The axioms encode common sense: the whole space and nothing are wiggle-rooms; merging any number of wiggle-rooms gives a wiggle-room (union); but you can only safely intersect finitely many before the room shrinks to a point.

Continuity = "no tearing": if you can find wiggle room around $f(x)$ in the target, you can guarantee wiggle room around $x$ in the source that maps inside it. Preimages stay open because nearby outputs come from nearby inputs.

Compactness is subtler. The slogan "compact = closed and bounded" is only true in $\mathbb{R}^n$ . The real meaning is a finiteness magic: any attempt to cover the space with open patches, however cleverly, can be reduced to finitely many. This converts "for each point, locally..." into "globally, with finitely many cases." That is why compactness makes maxima exist, makes continuous functions uniformly continuous, and makes products well-behaved.

Figure

Open cover of a compact interval reduced to a finite subcover

```svg

```

ExampleVerifying compactness of a finite union

Given: In $\mathbb{R}$ with the standard topology, consider $K = [0,1] \cup [2,3]$ . Show $K$ is compact and find $\max f$ for $f(x) = x^2 - 2x$ on $K$ .

Step 1 (compactness). Each $[0,1]$ and $[2,3]$ is closed and bounded, hence compact by Heine–Borel. A finite union of compact sets is compact: given a cover of $K$ , restrict it to each piece, extract finite subcovers $F_1, F_2$ , and $F_1 \cup F_2$ is a finite cover of $K$ .

Step 2 (continuity + EVT). $f(x) = x^2 - 2x = (x-1)^2 - 1$ is continuous; $K$ compact $\Rightarrow$ $f$ attains its max.

Step 3 (evaluate). Critical point $x=1$ gives $f(1) = -1$ (a minimum). Check endpoints: $f(0) = 0$ , $f(1) = -1$ , $f(2) = 0$ , $f(3) = 9 - 6 = 3$ .

Step 4. Compare values $\{0, -1, 0, 3\}$ .

\boxed{\max_{x \in K} f(x) = 3 \text{ at } x = 3,\qquad \min_{x\in K} f(x) = -1 \text{ at } x=1}

(dimensionless; pure numbers.)

ExampleA continuous bijection that is NOT a homeomorphism

Given: Let $X = [0, 2\pi)$ with the subspace topology from $\mathbb{R}$ , and $S^1 = \{(\cos t, \sin t)\}$ the unit circle. Define $f : X \to S^1$ , $f(t) = (\cos t, \sin t)$ .

Step 1. $f$ is a continuous bijection (each angle in $[0,2\pi)$ hits each circle point once).

Step 2. Is $f^{-1}$ continuous? Consider the set $U = [0, 1)$ , which is open in $X$ (it equals $(-1,1) \cap X$ ).

Step 3. $f(U)$ is an arc including the point $(1,0)$ but missing its left-hand neighbourhood on the circle — it is not open in $S^1$ . So $(f^{-1})^{-1}(U) = f(U)$ is not open, hence $f^{-1}$ is not continuous.

Step 4. Conclusion:

\boxed{f \text{ is a continuous bijection but NOT a homeomorphism}}

Why it matters: $X = [0,2\pi)$ is not compact (it's bounded but not closed). The theorem "a continuous bijection from a compact space to a Hausdorff space is a homeomorphism" fails here precisely because compactness fails. Had we used $X = [0,2\pi]$ (compact) with a quotient identification, the map would be closed and the result would hold.

⚠️ Common mistake

"Compact means closed and bounded." This is the Heine–Borel theorem and holds only in $\mathbb{R}^n$ (or finite-dimensional normed spaces). In infinite-dimensional spaces, the closed unit ball is bounded and closed but not compact. In a general topological space, "bounded" may not even be definable.

"Infinite intersections of open sets are open." False — equation [3] permits only finite intersections. In $\mathbb{R}$ , $\bigcap_{n=1}^\infty (-\tfrac1n, \tfrac1n) = \{0\}$ , which is closed, not open.

"Continuity preserves open sets." No — preimages of open sets are open, but images need not be. $f(x) = x^2$ maps the open set $(-1,1)$ to $[0,1)$ , which is not open. (Maps that send open sets to open sets are called open maps, a separate notion.)

"Sequences detect everything." In metric spaces, closure and continuity can be tested with sequences. In general topological spaces this fails; one needs nets or filters. Sequential continuity is strictly weaker than continuity outside first-countable spaces.

CounterexampleThe closed unit ball in

\ell^2

is bounded, closed, but not compact

Let $\ell^2$ be the space of square-summable sequences with $\|x\|^2 = \sum x_n^2$ . The closed unit ball $B = \{x : \|x\| \le 1\}$ is closed and bounded. Consider the standard basis vectors $e_n = (0,\dots,0,1,0,\dots)$ (the $1$ in slot $n$ ). Each $\|e_n\| = 1$ , so $e_n \in B$ . But for $m \neq n$ ,

\|e_m - e_n\| = \sqrt{1 + 1} = \sqrt 2.

No subsequence of $(e_n)$ is Cauchy, hence none converges. So $B$ is not sequentially compact, and (being a metric space) not compact. Heine–Borel breaks down in infinite dimensions: compactness is a genuinely stronger condition than closed-and-bounded.

connection

Topology sits at the crossroads of analysis, geometry, and algebra:

<a class="wiki-link" href="/wiki/math-topgeo-metric-spaces">Metric Spaces</a> are topological spaces whose topology comes from a distance function; they are the historical and intuitive prototype, but topology strips away the metric to find what is essential.
Connectedness (a space not splitting into two disjoint open pieces) and the Intermediate Value Theorem are topological cousins of compactness — both convert local data into global theorems.
Continuous Functions in real analysis are the metric-space special case of equation [5]; uniform continuity on compact sets is a topological consequence.
Product Topology and Tychonoff's Theorem (arbitrary products of compact spaces are compact) reveal compactness as the "right" finiteness condition that survives infinite products.
Fundamental Group and Homology in algebraic topology classify spaces up to homeomorphism using algebraic invariants, building directly on the notions defined here.
Hausdorff Spaces ( $T_2$ separation) provide the setting where limits are unique and compact subsets are automatically closed.

🔬 Where this appears

⏳ History

- 1872 — Cantor: While studying Fourier series, Georg Cantor introduces accumulation (limit) points, derived sets, and the notion of closed sets in

\mathbb{R}^n

— the embryo of point-set topology.

- 1894 — Borel: Émile Borel proves that a countable open cover of a closed bounded interval has a finite subcover (the Borel covering theorem).

- 1895–1904 — Poincaré: Henri Poincaré founds Analysis Situs, the combinatorial/algebraic study of shape, introducing homology and the fundamental group.

- 1906 — Fréchet: Maurice Fréchet introduces metric spaces ("classes (E)") and abstract notions of limit, in his thesis — the first general framework for convergence.

- 1906–1908 — Riesz, Lebesgue: Lebesgue extends Borel's covering result to arbitrary (uncountable) covers, giving the full Heine–Borel theorem; Riesz explores neighbourhood-based abstractions.

- 1914 — Hausdorff: Grundzüge der Mengenlehre gives the modern definition of topological space via neighbourhood axioms, including the Hausdorff separation axiom; this is the founding text.

- 1922 — Kuratowski: Kazimierz Kuratowski axiomatizes topology via the closure operator, an elegant equivalent formulation.

- 1923–1929 — Alexandroff, Urysohn, Tietze: Compactness gets its modern open-cover definition (Alexandroff–Urysohn); the Urysohn lemma and Tietze extension theorem develop separation theory.

- 1930 — Tychonoff: Andrey Tychonoff proves that arbitrary products of compact spaces are compact — later shown equivalent to the Axiom of Choice (Kelley, 1950).

open-problem

Where the theory gets hard and what remains genuinely subtle:

Compactness vs. its cousins. In general spaces, compact (finite subcovers), sequentially compact (every sequence has a convergent subsequence), and limit-point compact are inequivalent. They coincide for metric spaces, but examples like the ordinal space $[0, \omega_1)$ (sequentially compact, not compact) and $\{0,1\}^{[0,1]}$ (compact, not sequentially compact) show the landscape is intricate.

Set-theoretic independence. Many topological statements are independent of ZFC. The existence of a normal Moore space that is not metrizable depends on axioms beyond ZFC; questions about the Stone–Čech compactification $\beta\mathbb{N}$ (e.g., whether all its points away from $\mathbb{N}$ are alike) hinge on the Continuum Hypothesis and forcing.

Tychonoff and Choice. Tychonoff's theorem in full generality is equivalent to the Axiom of Choice — a deep entanglement of compactness with the foundations of set theory.

Continuum-theory problems. Whether certain "homogeneous indecomposable continua" exist with prescribed properties remains an active research area; the structure of compact connected metric spaces (continua) holds open questions.

Computability of topology. In computable/constructive mathematics, compactness behaves differently (e.g., the constructive Heine–Borel theorem is subtle), an active frontier in theoretical computer science and reverse mathematics.

Topological Spaces, Continuity, and Compactness

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also