I am having some problems understanding when Cauchy sequence don't converge, I know that if the metric space ($X$) is not complete the sequence don't converge IN $X$, every example I can think is embedded in $\mathbb{R}^n$, so it converge in the "big" space. For example: $$ a_n = \left(1+\frac{1}{n}\right)^n $$ does not converge in $\mathbb{Q}$, but it converges.

But I don't know if it is true or trivial that this happens always. The intuition tells me that a Cauchy sequence allways converge (even if it doesnt do it in the subspace that I am considering).


1 Answer 1


This is true. Call a metric space complete if all Cauchy sequnces converge (with limit in the given space).

For each metric space $(X,d)$ there exists a metric space $(X',d')$ such that

  1. $X \subset X'$ and $d'(x,y) = d(x,y)$ for $x,y \in X$, i.e. the inclusion map $i : X \hookrightarrow X'$ is an isometry.

  2. $X$ is dense in $X'$.

  3. Each Cauchy sequence in $X'$ converges in $X'$, i.e. $(X',d')$ is complete.

This is called the completion of $(X,d)$ and it is essentially unique which means that if $(X'',d'')$ has the same properties 1. - 3., then there exists a unique isometry $\phi : (X',d') \to (X'',d'')$ such that $\phi(x) = x$ for all $x \in X$. Note that this universal property implies that $\phi$ is a bijection (take the analogous isometry $\psi : (X'',d'') \to (X',d')$ and observe that $\psi \circ \phi$ and $\phi \circ \psi$ must be the identity maps).

I am not going into details of the construction, you should consult a textbook. Also see here.

Note that if $(X,d)$ is already complete, then it is its own completion. The completion of $\mathbb Q$ is $\mathbb R$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.