Let $(X,\tau)$ be a topological space. Let us call $x,y\in X$ swappable if there is $f:X\to X$ continuous such that $f(x)=y$ and $f(y)=x$. This relation is obviously reflexive and symmetric, but not necessarily transitive.

Moreover, we call $(X,\tau)$ rigid if the identity is the only homeomorphism from $X$ to itself.

Is there a rigid space $(X,\tau)$ with $|X| > \aleph_0$ such that for all $x,y\in X$ we have that $x,y $ are swappable?


The answer is yes by a 1951 result of Miroslav Katětov, who proved that there is an (uncountable) rigid totally disconnected compact space. Any such space $X$ is rigid and has the property that any two points are swappable by continuous functions, since if $x\neq y$, we can find an open separation $X=U\sqcup V$ with $x\in U$ and $y\in V$, and then let $f$ be constant on each piece with value $y$ or $x$, respectively. This is a continuous function swapping $x$ and $y$.

  • $\begingroup$ I'm not sure if there would be an easier example or not. $\endgroup$ – Joel David Hamkins Apr 10 '17 at 16:53
  • $\begingroup$ If you don't insist on compactness, then it is relatively easy to show that there are zero-dimensional uncountable rigid spaces (for example, you can use transfinite recursion to construct a rigid subset of the reals with cardinality of the continuum). Your argument still works for such spaces. $\endgroup$ – Will Brian Apr 10 '17 at 17:37
  • $\begingroup$ Thanks! Indeed, any rigid subspace of the reals will be totally swappable, since rigidity ensures that it will be totally disconnected, and then the locally constant maps I mention will perform the desired swaps. $\endgroup$ – Joel David Hamkins Apr 10 '17 at 20:08
  • $\begingroup$ Also every path connected metrizable rigid space is swappable. Such examples are not hard to construct. $\endgroup$ – Moishe Kohan Apr 12 '17 at 23:34

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.