Let $f: \mathbb R \to \mathbb R$ be a strictly monotone functions. Consider the following metric on $\mathbb R$: $$ d(x,y) = |f(x) - f(y)|, \qquad x,y \in \mathbb R. $$ Whether that metric topology is the same as the standard one?

It is the same if $f$ is continuous or has finite number of points of discontinuity. But whether it's the same for arbitrary $f$, e.g. if it discontinuous on $\mathbb Q$?


Actually, I wanted to ask, whether that metric topology always contain the standard one? It is true if there are finite number of points of discontinuity: then we have the standard topology generated by the open intervals plus we have points and semiopen (in usual topology) intervals that are open.

  • $\begingroup$ I believe a strictly monotone function has to be continuous. $\endgroup$ – Yanko Oct 7 '17 at 22:07
  • $\begingroup$ @yanko $f(x)=x\lfloor 1+x^2\rfloor$ is strictly increasing, yet not continuous. $\endgroup$ – zwim Oct 7 '17 at 22:11
  • $\begingroup$ @zwim right nice example! $\endgroup$ – Yanko Oct 7 '17 at 22:13
  • $\begingroup$ @yanko moreover, it can have a countable everywhere dense set of points of discontinuity: math.stackexchange.com/questions/172753/… $\endgroup$ – vanger Oct 7 '17 at 22:22

Consider the strictly monotonic function $f$ given by $$ f(x) = \begin{cases} x & x \le 0\\ x +1 & x > 0\\ \end{cases} $$

It's continuous except at $0$, but in the topology induced by it the sequence $x_n = 1/n$ no longer converges to the point $0$. So your statement about giving the same topology isn't correct.

I haven't proved it, but I think the topology induced by this function is a "separated union" of $(-\infty, 0]$ and $(0, +\infty)$, which you can think of taking regular $\mathbb{R}$ and breaking it at $0$ so that the point $0$ stays attached to the negatives. You can generalize this to multiple points of discontinuity, and think about what the difference would be if the function was right-continuous, or neither left- nor right- continuous, at the points of discontinuity.

  • $\begingroup$ Yeah, of course, I was wrong about the topologies being the same. We could take $f(x) = -1 + x$ for $x < 0$, $f(0) = 0$, $f(x) = 1 + x$ for $x > 0$ and see that $\{0\}$ is open which is not in the usual topology. I wanted to ask, whether that metric topology contains the usual one. Topologies of your and my example contain all sets that open in the usual topology plus some more sets. I updated the question. $\endgroup$ – vanger Oct 7 '17 at 23:51

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.