Ma · Mathematics

Why Is Every Singleton Set Open in a Finite Metric Space?

Published By the 9to5Science editorial team

At a glance

  1. 1. Recall the definition

    {x} is open if some ball B(x, r) with r > 0 contains only x.

  2. 2. Take the minimum distance

    Let r be the smallest distance from x to any other point.

  3. 3. Check that r is positive

    Finitely many positive distances have a positive minimum.

  4. 4. Conclude

    B(x, r) = {x}, so every singleton, and hence every subset, is open.

Recall the definitions. In a metric space (X,d)(X, d), the open ball of radius r>0r > 0 around xx is

B(x,r)={ y∈X:d(x,y)<r }.B(x, r) = \{\, y \in X : d(x, y) < r \,\}.

A set U⊆XU \subseteq X is open if every point of UU has some open ball around it that stays inside UU.

For U={x}U = \{x\} this means: we need a radius r>0r > 0 with B(x,r)={x}B(x, r) = \{x\}, so that the ball contains no other point.

The proof

Let XX be a finite metric space and x∈Xx \in X.

Case 1: X={x}X = \{x\}. Then {x}=X\{x\} = X, and the whole space is always open.

Case 2: XX has other points. Define

r=min⁡{ d(x,y):y∈X, y≠x }.r = \min\{\, d(x, y) : y \in X,\ y \neq x \,\}.

This minimum exists because it is taken over a finite, non-empty set of numbers. Each of those numbers is positive, because d(x,y)>0d(x, y) > 0 whenever y≠xy \neq x. So r>0r > 0.

Now if y≠xy \neq x, then d(x,y)≥rd(x, y) \ge r, so y∉B(x,r)y \notin B(x, r). Hence

B(x,r)={x}⊆{x},B(x, r) = \{x\} \subseteq \{x\},

and {x}\{x\} is open. ■\blacksquare

Consequence: every subset is open

Any subset A⊆XA \subseteq X is the union of its singletons, A=⋃a∈A{a}A = \bigcup_{a \in A} \{a\}. A union of open sets is open, so AA is open. Taking complements, every subset is also closed.

So a finite metric space always carries the discrete topology. In particular, a finite metric space with at least two points is never connected.

Where finiteness is used

The only step that needs finiteness is “the minimum exists and is positive”. In an infinite space the distances from xx to the other points can have infimum 00, and then no ball isolates xx:

  • In R\mathbb{R}, the distances from 00 to the points 1/n1/n are 1/n→01/n \to 0. Every ball around 00 contains some 1/n1/n, so {0}\{0\} is not open.
  • Some infinite spaces still have open singletons. In Z\mathbb{Z} with d(m,n)=∣m−n∣d(m, n) = |m - n|, the ball B(n,1/2)={n}B(n, 1/2) = \{n\}, so Z\mathbb{Z} is discrete too.

What matters is whether xx is an isolated point: whether inf⁡y≠xd(x,y)>0\inf_{y \neq x} d(x, y) > 0. Finiteness guarantees this for every point.

Common mistakes

  • Writing “let r=min⁡d(x,y)r = \min d(x,y)” without saying why it is positive. The proof depends on y≠x⇒d(x,y)>0y \neq x \Rightarrow d(x,y) > 0, which is one of the metric axioms.
  • Using a closed ball. The closed ball {y:d(x,y)≤r}\{y : d(x,y) \le r\} with this rr contains the nearest neighbour. Either use the open ball or a smaller radius such as r/2r/2.
  • Forgetting the one-point space, where the minimum is over an empty set.

References

  1. Rudin, W. (1976). Principles of Mathematical Analysis (3rd ed.), Chapter 2. McGraw-Hill.
  2. Munkres, J. R. (2000). Topology (2nd ed.), Section 20. Prentice Hall.

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.