Could you help me with this problem?

Let $d$ be a metric on $[0,1]$ consistent with the standard topology. Prove that the metric space: $([0,1], d)$ has at most one isometry (except for identity).

I would really appreciate your help.

Thank you.


Notice that it is enough to show that an isometry which fixes the endpoints is the identity.

Suppose then $f:[0,1]\to[0,1]$ is an isometry such that $f(0)=0$, and let $x\in[0,1]$. Notice that $f$ is strictly increaing —because it is continuous for the usual topology and injective.

It is easy to check that the sequence $(f^n(x))_{n\geq1}$ is monotone, so that it converges with respect to the usual topology of $[0,1]$. The hypothesis then implies that it also converges with respect to $d$ and, in particular, it satisfies the condition of Cauchy. Since $f$ is an isometry, $d(f^n(x),f^{n+1}(x))$ does not depend on $n$ because $f$ is an isometry, and then the Cauchy condition implies that this distance is zero: in particular, $d(x,f(x))=0$, that is $x=f(x)$.

  • $\begingroup$ It may be a stupid question, but how do we know that the sequence $(f^n(x))_{n\geq1}$ is monotone? $\endgroup$
    – Andrew
    Apr 22 '13 at 17:19
  • $\begingroup$ Think about it for a while. I'll add the details in a few hours if you did not see it on your own. $\endgroup$ Apr 22 '13 at 17:25
  • $\begingroup$ I still don't see that. I don't think it somehow results from the fact that isometry preserves distance, does it? I've tried various combinations with distance between $f^{n}, \ f^{n-1}, \ f(x), \ x$, but it doesn't seem to be the right approach. $\endgroup$
    – Andrew
    Apr 22 '13 at 18:03
  • $\begingroup$ Could you post the details now, please? $\endgroup$
    – Andrew
    Apr 23 '13 at 5:03

Hint: Continuity is a topological property.

Spoiler: The following is the shape of a proof.

  1. Isometries are continuous functions.
  2. Continuity is a topological property.
  3. Isometries are bijections.
  4. Continuous bijections $([0,1],d_\textrm{Eucl.}) \to ([0,1],d_\textrm{Eucl.})$ are strictly increasing or strictly decreasing.
  5. A strictly increasing bijective function $[0,1]\to[0,1]$ which is an isometry w.r.t. the metric $d$ is unique.
  6. Similarly for a strictly decreasing bijective function.

Edit: For point 5, suppose $f,g$ both have this property. Then consider $h=f^{-1}\circ g$. This is a continuous bijection (check), so is strictly increasing/decreasing. Suppose $h(x)\neq x$. Then $d(h^n(x),h^{n-1}(x)) = d(h(x),x)\neq 0$. But let $n\to\infty$; by monotonicity, $h^n(x)\to x_0$, and by continuity, $d(h^n(x),h^{n-1}(x)) \to 0$. Contradiction.

  • $\begingroup$ Could you tell me a bit more about point 5 ? $\endgroup$
    – Andrew
    Apr 22 '13 at 17:20
  • $\begingroup$ Whoops, just realized this is more or less equivalent to the other answer. $\endgroup$ Apr 22 '13 at 17:47

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.