1
$\begingroup$

As a continuation to the fully answered question:

Injective and Integrable Mapping from $\mathbb R^3$ to $\mathbb R$

Does there exist an injective mapping $f:\mathbb R^n\rightarrow[0,1]$ that has only a finite number of discontinuities? Or maybe one can come with some topological claim showing it is impossible?

$\endgroup$
1
  • 1
    $\begingroup$ I extended my answer to injective maps $f:\mathbb{R}^n\to\mathbb{R}^{n-1}$. $\endgroup$
    – GH from MO
    Jan 2, 2013 at 7:30

1 Answer 1

8
$\begingroup$

Assume $f:\mathbb{R}^n\rightarrow[0,1]$ is an injective map. Take any circle $C\subset \mathbb{R}^n$. If $f$ is continuous on $C$, then the restriction of $f$ to $C$ is a homeomorphism $C\to f(C)$ (because $C$ is compact), which is a contradiction, because $C$ minus any point is connected, while $f(C)$ minus any point in $f(C)\cap(\inf f(C),\sup f(C))$ is disconnected. This shows that $f$ has a discontinuity on any circle in $\mathbb{R}^n$. In particular, when $n>1$, the cardinality of the discontinuities is continuum.

Added: By a similar argument combined with the Borsuk–Ulam theorem, the number of discontinuities of any injective map $f:\mathbb{R}^n\to\mathbb{R}^{n-1}$ is continuum.

$\endgroup$
6
  • $\begingroup$ Thanks! Please note that in the mentioned question (in that link), there is an example with only countably many discontinuities. $\endgroup$
    – Ohad Asor
    Jan 2, 2013 at 7:12
  • 1
    $\begingroup$ No, in the answer to the mentioned question the discontinuities lie on the union of boundaries of cubes, which is a two-dimensional object of cardinality continuum (and measure zero). As I proved above, there is no injective map with less than continuum many discontinuities. $\endgroup$
    – GH from MO
    Jan 2, 2013 at 7:20
  • $\begingroup$ In fact Gerald Edgar commented on your original question (in that link): "Cannot be done with countably many discontinuities." And you responded: "Right. and thanks for your answer. if this is indeed the problem, so make it zero measure of discontinuities." $\endgroup$
    – GH from MO
    Jan 2, 2013 at 7:24
  • $\begingroup$ i agree i dont have a full understanding, but this will come with time... i'm doing my best $\endgroup$
    – Ohad Asor
    Jan 2, 2013 at 7:25
  • 1
    $\begingroup$ So, can we say this... if $A \subset \mathbb R^n$ has Hausdorff dimension ${}\lt n-1$, then $\mathbb R^n \setminus A$ is connected? $\endgroup$ Jan 2, 2013 at 14:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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