3
$\begingroup$

Let $A$ be a contractible subset of a topological space $X$. A set is said to be contractible if it has the same homotopy type as a point. Alternately, a space is contractible if the identity map of the space is homotopic to a constant map.

The first thing that I was wondering is if $A$ is contractible then what could we say about $A^\circ$ and $\overline A$; are they also contractible? The answer to both is a clear no.

  1. Let $X=\mathbb R^2$. Take $A$ to be the union of two closed discs, touching at a single point. Then $A^\circ$ is the union of the interiors of each disc and this space is not even path connected.

  2. Let $X=\mathbb R^2$. Take $A$ to be the circle $(S^1)$ minus a point. This is homeomorphic to $\mathbb R$ and is therefore contractible. The closure is $S^1$ which is not.

The first answer gave rise to the following question: If $A$ is contractible and $A^\circ$ is path connected can we say that $A^\circ$ is contractible?

So far I have neither been able to come up with a proof or a counter example. The only thing I can say is that we may assume without loss of generality that $A$ contracts to a point in $A^\circ$.

$\endgroup$

1 Answer 1

2
$\begingroup$

Take the standard open cone in $\mathbb{R}^3$ (i.e. it does not contain the tip of the cone), about each point on the cone include an (open) line segment normal to that point of length $f(x)$, were $f$ is some function that tends to 0 fast enough so that the resulting space is homeomorphic (in the obvious way) to the open cone crossed with $\mathbb{R}$. Finally, add the cone point.

The space we have described is contractible to the cone point, but its interior is homeomorphic to the open cone crossed with $\mathbb{R}$, hence, noncontractible.

$\endgroup$
2
  • $\begingroup$ Wow! I didn't realise how simple the answer would be! Infact if $D$ is the open unit disc in $\mathbb R^2$ then we can simply take $A=(D^2-\{(0,0)\}\times R)\bigcup \{(0,0,0)\}$ and this would work right? $\endgroup$
    – R_D
    Commented Dec 12, 2021 at 8:49
  • $\begingroup$ @R_D Yeah it would! $\endgroup$ Commented Dec 12, 2021 at 15:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .