1
$\begingroup$

Suppose $d\in \{3,4,\dotsc\}$ and $A\subseteq \mathbb{R}^d$ is non-empty, open and connected with its complement $A^c$ connected too and $\text{int}(A^c)\neq \emptyset$. Its boundary $S:=\partial A$ is then connected too. Let $A^{(2)}$ be the set of 2-dimensional affine planes within the affine space $\mathbb{R}^d$. For any $x\in S$, define \begin{gather*} S_x:=\bigcup_{n=1}^\infty S_x^{(n)} \\ S_x^{(1)}:=\{y\in S\mid\exists H\in A^{(2)}\mathrel:\text{ $x$ and $y$ are in the same connected component of }H\cap S\},\\ S_x^{(n+1)}:=\{y\in S|\,\exists z\in S_x^{(n)}:\,y\in S_z^{(1)}\}. \end{gather*} Is $S_x=S$ for all $x \in S$? If not, is $\{S_x\mid x\in S\}$ a countable partition of $S$?

$\endgroup$
1

1 Answer 1

2
$\begingroup$

Consider the set $S=\{\,(t,t^2,t^3)\in\mathbb R^3\mid\,t\in\mathbb R\}$. Clearly $S$ is connected, and so is its complement $A=\mathbb R^3\setminus S$.

Note that each plane has at most 3 points of intersection with $S$. It follows that $S_x=\{x\}$ for any $x\in S$.

The only problem is the set $A^c=S=\partial A$ has empty interior, but it is easy to fix by removing from $A$ a closed ball centered at a point in $S$.

$\endgroup$
5
  • $\begingroup$ Thx! Of course, this $S$ is not the boundary of an open set. It is not a 'surface'. $\endgroup$
    – 5th decile
    Commented Apr 27, 2022 at 8:34
  • $\begingroup$ TeX note: $\mathbb R^3 \setminus S$ \mathbb R^3 \setminus S spaces better than $\mathbb R^3 \backslash S$ \mathbb R^3 \backslash S. I edited accordingly. $\endgroup$
    – LSpice
    Commented Apr 27, 2022 at 22:53
  • $\begingroup$ @ThibautDemaerel fixed now. [Sorry I read $\text{int}(A^c)= \emptyset$ insted of $\text{int}(A^c)\neq \emptyset$.] $\endgroup$ Commented Apr 28, 2022 at 0:01
  • $\begingroup$ Sorry, I didn't see your updated answer for a while. I realize I should better have asked a slightly different question, namely the same one except that we take $S=\partial\left(\overline{A}\right)$ ($A$ still open). $\endgroup$
    – 5th decile
    Commented May 16, 2022 at 17:30
  • $\begingroup$ @ThibautDemaerel this version seems to be open --- mathoverflow.net/questions/363950/curves-on-potatoes $\endgroup$ Commented May 18, 2022 at 3:13

You must log in to answer this question.

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