Skip to content
Tutorials Center
My learning

Topics

Lesson 10 of 36 Β· Topological Spaces

Definition of a Topology

Learn the three axioms of a topology, test collections of sets with a checker, and count all topologies on 1, 2, 3 and 4 points by brute force.

In a metric space we defined open sets using balls. That works, but distance carries more information than we need. The key step of topology is to throw away the metric and keep only the list of open sets, requiring that this list obey the rules we proved in the last chapter. Anything that obeys them counts as a topology.

The definition

Let X be a set. A topology on X is a collection Ο„ of subsets of X, called the open sets, such that:

  1. βˆ… ∈ Ο„ and X ∈ Ο„.
  2. Unions: the union of any collection of members of Ο„ is in Ο„ (finite or infinite).
  3. Intersections: the intersection of any finitely many members of Ο„ is in Ο„.

The pair (X, Ο„) is a topological space. Its elements are called points. A subset is closed if its complement is open.

Axioms 2 and 3 can be checked in a simpler way when X is finite: it is enough that the union and the intersection of any two members are in Ο„, because repeating the step covers longer lists.

Why "arbitrary unions but only finite intersections"? Because that is exactly what metric spaces satisfy. The intersection of the infinitely many open intervals (βˆ’1/n, 1/n) is , which is not open in ℝ, so demanding infinite intersections would exclude the real line.

Every metric space is a topological space

Given a metric d on X, declare U open when every point of U has a ball inside U. The previous chapter showed that this collection satisfies the three axioms. It is called the metric topology. So all the examples of Chapter 2, including ℝ, ℝⁿ and function spaces, are topological spaces. The converse is false: some topologies cannot come from any metric. We will see small ones in a moment.

Checking the axioms

For a finite set we can write a complete checker. Test it on three collections on X = {1, 2, 3}.

Python A topology checker for finite sets

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

Program input Β· 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.

Collections B and C each break one axiom, and the checker names the missing set.

Counting topologies by brute force

How many different topologies exist on a set of n points? Every topology contains βˆ… and X, and the other 2ⁿ βˆ’ 2 subsets are each either in Ο„ or not, so there are at most 2^(2ⁿ βˆ’ 2) candidates. For n = 4 that is 2¹⁴ = 16,384, small enough to test all of them. The count of labeled topologies (where the points have names) is known to be 1, 4, 29, 355, 6942, … for n = 1, 2, 3, 4, 5. The code recovers the first four.

Python Counting all topologies on 1 to 4 points

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

Program input Β· 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 numbers grow quickly, and no simple formula is known. Many of these topologies look different only because the points carry different labels; in Homeomorphisms we count them up to renaming.

Coarser and finer

If τ₁ βŠ† Ο„β‚‚ are two topologies on the same set, then τ₁ is coarser (weaker, smaller) and Ο„β‚‚ is finer (stronger, larger). The indiscrete topology {βˆ…, X} is the coarsest topology and the discrete topology (every subset open) is the finest. A finer topology has more open sets, so it can separate points more sharply, and it makes it harder for a sequence to converge.

Is there a metric behind every topology?

No. On the two-point set {a, b} take Ο„ = {βˆ…, {a}, {a, b}}. This is called the SierpiΕ„ski space. In a metric space, two distinct points can be separated by balls: B(a, r) and B(b, r) with r = d(a, b)/2 are disjoint open sets. In the SierpiΕ„ski space the only open set containing b is the whole set, which also contains a. No metric can produce this topology, so topology is strictly more general than metric geometry.

Check your understanding

Quick check Β· 4 questions

of 4 answered You scored / 4 Β· perfect!

  1. 1. Which axiom allows unions of infinitely many open sets?

    A topology must be closed under arbitrary unions, finite or infinite.
  2. 2. X = {1, 2, 3} and tau = { empty set, {1}, {2}, {1,2,3} }. Why is this not a topology?

    A topology must contain the union of any two members, and {1, 2} is not in the list.
  3. 3. How many topologies exist on a 2-point set {a, b}?

    They are the indiscrete one, the discrete one, and the two Sierpinski topologies ({a} open or {b} open): 4 in total.
  4. 4. Which topology is the coarsest on any set X?

    Every topology contains the empty set and X, so {empty set, X} is contained in all of them.

Summary

  • A topology is a collection of open sets containing βˆ… and X, closed under arbitrary unions and finite intersections.
  • Metric spaces give topologies, but not every topology comes from a metric (the SierpiΕ„ski space does not).
  • On a finite set, pairwise unions and intersections are enough to check the axioms.
  • There are 1, 4, 29 and 355 labeled topologies on 1, 2, 3 and 4 points.
  • Coarser topologies have fewer open sets; the indiscrete topology is coarsest and the discrete one is finest.

Next: Examples of Topologies