Skip to content
Tutorials Center
My learning

Topics

Lesson 18 of 36 · Continuity and Homeomorphism

Hausdorff and Separation

Learn the separation axioms T0, T1 and Hausdorff (T2), why Hausdorff spaces have unique limits, and how the cofinite topology breaks that rule, with code checks.

Some topologies are so coarse that two different points cannot be told apart by open sets. Others separate points as cleanly as the real line does. Separation axioms grade topologies by how well their open sets distinguish points. The best-known level, the Hausdorff property, is the one that makes limits behave and that most of analysis quietly assumes.

The first three levels

Let (X, τ) be a topological space and x ≠ y two points.

Name Condition
T₀ (Kolmogorov) Some open set contains exactly one of x, y.
T₁ (Fréchet) There is an open set containing x but not y, and another containing y but not x.
T₂ (Hausdorff) There are disjoint open sets U ∋ x and V ∋ y.

Each level implies the one before: T₂ ⟹ T₁ ⟹ T₀. None of the arrows reverse.

  • The Sierpiński space {0, 1} with open sets ∅, {1}, {0, 1} is T₀: the open set separates the points. It is not T₁, since every open set containing 0 also contains 1.
  • The cofinite topology on an infinite set is T₁ but not T₂, as we show below.
  • The indiscrete topology on two or more points is not even T₀.
  • Every metric space is Hausdorff: for x ≠ y, let r = d(x, y)/2. The balls B(x, r) and B(y, r) are disjoint, since a common point z would give d(x, y) ≤ d(x, z) + d(z, y) < 2r = d(x, y), a contradiction.

A useful reformulation: a space is T₁ if and only if every singleton is closed. Since is closed when its complement is open, we need an open set around each other point y that avoids x, which is exactly the T₁ condition.

Some books number the axioms differently or define "Hausdorff" with extra conditions. Always check the definition if you read another text. Here, "Hausdorff" means T₂ as in the table.

Checking the axioms by code

On a finite space, each axiom is a finite check over pairs of points. The code tests the discrete, indiscrete and Sierpiński topologies on two points and the particular point topology on three points.

Python T0, T1 and Hausdorff checks on finite spaces

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The discrete space satisfies all three; the Sierpiński space and the particular point space are only T₀; the indiscrete space satisfies none.

Unique limits

Hausdorff spaces are exactly the spaces where limits of sequences (and, more generally, of nets and filters) are unique.

Theorem. In a Hausdorff space, a sequence converges to at most one point.

Proof. Suppose xₙ → p and xₙ → q with p ≠ q. Pick disjoint open sets U ∋ p and V ∋ q. Eventually xₙ ∈ U and eventually xₙ ∈ V, so some xₙ lies in U ∩ V = ∅, a contradiction. ∎

The failure of this for non-Hausdorff spaces is not just theory. In the indiscrete topology every sequence converges to every point. Subtler is the cofinite topology on ℕ, where the open sets are ∅ and the sets whose complement is finite. Consider the sequence xₙ = n. For any point p ∈ ℕ and any open U ∋ p, the complement of U is a finite set F, and all xₙ with n > max F lie in U. So xₙ → p for every p. Limits are as far from unique as they can be.

Python The sequence 1, 2, 3, ... in the cofinite topology

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The test uses random open sets as a sample; the argument in the text covers all of them, since each has a finite complement.

Why Hausdorff is the standard assumption

Most spaces used in analysis, geometry and applications are Hausdorff, and many theorems need it:

  • limits are unique, so "the limit" makes sense;
  • a compact subset is closed (shown in the next chapter);
  • the set where two continuous functions agree is closed, so two continuous functions that agree on a dense set agree everywhere;
  • the diagonal {(x, x)} is closed in the product X × X.

Without Hausdorff, spaces like the line with a doubled origin appear, where two different points are never separated; they are important in algebraic geometry but they break many familiar facts.

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. Which implication always holds between separation axioms?

    Disjoint open sets around x and y give an open set around each that avoids the other, so T2 implies T1 (and T1 implies T0).
  2. 2. Which space is T0 but not T1?

    Every open set containing 0 also contains 1, so T1 fails, but {1} is open and separates the points, so T0 holds.
  3. 3. Why is every metric space Hausdorff?

    A common point would make d(x, y) < 2r = d(x, y) by the triangle inequality.
  4. 4. In the cofinite topology on the natural numbers, to which points does 1, 2, 3, ... converge?

    Every open set around p has a finite complement, so the sequence is eventually inside it, for every p.

Summary

  • T₀, T₁ and T₂ (Hausdorff) say how well open sets separate points; T₂ ⟹ T₁ ⟹ T₀.
  • A space is T₁ exactly when all singletons are closed; every metric space is Hausdorff.
  • In Hausdorff spaces, limits of sequences are unique.
  • The cofinite topology on ℕ is T₁, not Hausdorff, and 1, 2, 3, … converges to every point.
  • Finite T₁ spaces are discrete; many standard theorems assume Hausdorff.

Next: Compact Spaces and Open Covers