In topological space $(X, ‎\tau )$ every compact subspace of $X$ is closed, so no infinite subspace of $X$ can have the cofinite topology.

Is it right to say:

Each infinite subspace of $X$ contains an infinite discrete subspace.

How can I prove it?

Thank you.

  • $\begingroup$ Isn't $[0,1]\subseteq \Bbb{R}$ a counterexample? It's compact, but the only discrete subspaces are the finite sets. $\endgroup$
    – user149890
    Jun 6, 2014 at 12:40
  • 2
    $\begingroup$ @mathmax No. $\{1/n:n=1,2,\ldots\}$ is discrete. $\endgroup$ Jun 6, 2014 at 12:49
  • $\begingroup$ Leila means: suppose $X$ is a KC-space (every compact subset is closed). Does it follow that every infinite subspace of $X$ contains an infinite discrete subspace? This is well-known to be true for Hausdorff spaces, which are a special class of KC spaces. $\endgroup$ Jun 6, 2014 at 15:20

1 Answer 1


A nice "folklore" topology theorem that is often useful, see the paper: Minimal Infinite Topological Spaces, John Ginsburg and Bill Sands, The American Mathematical Monthly Vol. 86, No. 7 (Aug. - Sep., 1979), pp. 574-576.

Suppose $X$ is any infinite topological space. Then there exists a countably infinite subspace $A$ of $X$ such that $A$ is homeomorphic to one of the following five spaces:

  1. $\mathbb{N}$ in the indiscrete topology.
  2. $\mathbb{N}$ in the cofinite topology.
  3. $\mathbb{N}$ in the upper topology (all non-trivial open sets are of the form $n^\uparrow = \{m \in \mathbb{N}: m \ge n\}$).
  4. $\mathbb{N}$ in the lower topology (all non-trivial open sets are of the form $n^\downarrow = \{m \in \mathbb{N}: m \le n\}$).
  5. $\mathbb{N}$ in the discrete topology.

Now, if $X$ is a KC-space (i.e. all compact subsets of $X$ are closed in $X$), then every subspace $Y$ of $X$ is also a KC-space. As the first 4 spaces are all non-KC (in the first 3 spaces all subsets are compact, and in the lower topology exactly all finite sets are compact, but not all of these are closed, e.g. $\{2\}$ is not closed as $3$ is in its closure) this means that every infinite (subset of a ) KC-space contains an infinite discrete subspace.

(Added) Note that this question and its answer give a (detailed explanation of a) direct proof for this. This uses recursion to construct the set, using at every stage that a countably infinite subspace $A$ of $X$ cannot have the cofinite topology. The latter is clear, as I remarked above: suppose $X$ is a KC-space, then $A \subset X$ is also a KC-space (a compact $K \subset A$ is also compact in $X$ so closed in $X$, so $K$ is closed in $A$ as well) and the cofinite countable space is not a KC-space, as all subsets are compact but only the finite subsets are closed (and the whole subspace).


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .