5

For non-arbitary spaces we can discuss for such case, like how many components are there or other properties. But is it true for any space? It seems if we have a homeomorphism $f$ from $S$ to $S' = S - \{p\}$, $f(p) = q$, but since a space is homeomorphic to itself, there is some $g(r)=q$. Then there is no inverse if $f$ and $g$ coincide. However they don't have to and maybe $f$ is somehow the homeomorphism since I can't deduce more information.

If it is not true, a counter-example will be super helpful! Thank you.

MonkeyKing
  • 3,178
  • 1
    Even two different puncturings of a space needn’t be homeomorphic; it may matter which point is removed. Consider removing a point from a figure “X”. The result might have 1, 2, or 4 components, depending on which point is removed. – MPW Dec 24 '19 at 21:32

3 Answers3

8

Another example: $\mathbb C \setminus \mathbb Z$. This is connected and homeomorphic to the punctured version of itself.

GEdgar
  • 111,679
6

Any infinite discrete space is homeomorphic to itself minus any point. For example, the map $n\mapsto (n+1)$ is a homeomorphism $\mathbb{N}\to \mathbb{N}\setminus\{0\}$ (where $\mathbb{N}$ has the discrete topology).

The same example works if you give $\mathbb{N}$ the trivial topology or the Alexandrov topology (where open sets are upwards-closed sets).

Alex Kruckman
  • 76,357
3

$\mathbb{Q}$ is homeomorphic to its punctured version (all countable metric spaces without isolated points are). Same for the irrationals $\mathbb{P}$.

Henno Brandsma
  • 242,131