Skip to content
Tutorials Center
My learning

Topics

Lesson 27 of 36 · Product and Quotient Spaces

The Product Topology

Build the product topology from open rectangles, see why projections are continuous, how products inherit properties, and compute one for finite spaces.

How do you put a topology on a pair of spaces? The plane ℝ² is the product ℝ × ℝ; a torus is the product of two circles; a cylinder is a circle times an interval. In each case we want a topology on X × Y built from the topologies of X and Y in the most natural way. This is the product topology, and it is the tool for building new spaces and for proving that properties like compactness and connectedness survive products.

The definition

Let X and Y be topological spaces. A basic open set in X × Y is a set U × V where U is open in X and V is open in Y. The product topology consists of all unions of such rectangles.

The basic open sets are closed under finite intersection, because (U₁ × V₁) ∩ (U₂ × V₂) = (U₁ ∩ U₂) × (V₁ ∩ V₂). So they form a basis, and unions of them satisfy the axioms of a topology.

Not every open set of the product is a rectangle. An open disk in ℝ² is not a product U × V, but it is a union of small open rectangles.

Projections

The projections are π₁(x, y) = x and π₂(x, y) = y. In the product topology they are continuous, since π₁⁻¹(U) = U × Y is a basic open set. In fact the product topology is the coarsest (smallest) topology that makes both projections continuous. That is the reason it is the right choice: it is the topology you get if you demand "a map into X × Y is continuous if and only if both coordinate maps are".

Universal property. A function f: Z → X × Y is continuous exactly when π₁ ∘ f and π₂ ∘ f are continuous. This is how you show that a curve such as t ↦ (cos t, sin t) is continuous: each coordinate is.

Products of metric spaces

If X and Y are metric spaces there are several natural metrics on X × Y:

  • d_max = max(d_X, d_Y), whose balls are rectangles;
  • d_sum = d_X + d_Y, whose balls are "diamonds" in ℝ²;
  • d_euc = √(d_X² + d_Y²), whose balls are round disks in ℝ².

They look different but give the same topology, the product topology. The reason is the inequality d_max ≤ d_euc ≤ d_sum ≤ 2·d_max: every ball of one kind contains a ball of another kind with the same center.

Python Three product metrics, one 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 chain of inequalities in the assertion is what makes the balls nest. A second experiment follows, on a finite product where you can list every open set.

A finite product, computed

Take X = {a, b} with open sets ∅, {a}, {a, b} (the Sierpiński space) and Y = {0, 1} with the discrete topology. Build the product topology by taking all unions of rectangles.

Python Building the product topology of two 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 point ("a", 0) is open because is open in the Sierpiński space and is open in the discrete space, but ("b", 0) is not open, because is not open in X. The product topology reflects both factors.

What products preserve

  • Hausdorff: a product of Hausdorff spaces is Hausdorff.
  • Connected: a product of connected spaces is connected. (Slice × Y and use that connected sets sharing a point have a connected union.)
  • Path-connected: a product of path-connected spaces is path-connected.
  • Compact: a product of two compact spaces is compact. This is the finite case of Tychonoff's theorem: an arbitrary product of compact spaces is compact. (For finite products the proof uses the tube lemma.)
  • Second countable / metrizable: finite or countable products of such spaces keep the property.

Examples: [0, 1]ⁿ is compact; the torus S¹ × S¹ is compact and connected; the cylinder S¹ × [0, 1] is compact and connected. A closed bounded subset of ℝⁿ is a closed subset of a box [−M, M]ⁿ, which is how Heine–Borel is finished.

Infinite products

For infinitely many factors the basis is slightly different: a basic open set is a product ∏ U_i in which U_i = X_i for all but finitely many indices. That is the product topology. Allowing every U_i to be arbitrary gives the box topology, which is finer and loses Tychonoff's theorem. The product topology is the one that makes pointwise convergence the notion of convergence.

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. What do basic open sets of the product topology on X × Y look like?

    Open rectangles U × V form a basis; the open sets are unions of them.
  2. 2. What is special about the product topology compared with other topologies on X × Y?

    Both projections are continuous, and any topology with that property must contain all the sets π₁⁻¹(U) = U × Y and π₂⁻¹(V) = X × V, hence their intersections U × V.
  3. 3. Which statement is true?

    Connectedness (like compactness and the Hausdorff property) passes to products. ℝ² is ℝ × ℝ with the product topology, which equals the Euclidean topology.
  4. 4. Why do d_max, d_sum and d_euc on ℝ² give the same topology?

    From d_max ≤ d_euc ≤ d_sum ≤ 2 d_max, every ball in one metric contains a ball in the other with the same center.

Summary

  • The product topology has basis U × V (open rectangles).
  • Projections are continuous, and a map into a product is continuous iff each coordinate is.
  • Several natural product metrics give the same topology.
  • Hausdorff, connectedness, path-connectedness and compactness pass to products.
  • For infinite products only finitely many factors may be restricted in a basic open set.

Next: The Quotient Topology