Sign up ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

This is yet another question of mine related to this other stack overflow question.

Let $A$ be a path connected, dense subset of ${\mathbb R}^2$. Let $B$ denote the complement of $A$, and $C$ denote the set of cut points in $B$, i.e. the set of all $b\in B$ such that $B\setminus \lbrace b \rbrace$ is not path connected. Assume that $B$ contains at least three points and has no isolated points.

Prove or find a counterexample : $C$ is dense in $B$.

share|cite|improve this question
B is the closed interval and b is one of the endpoint, then B\{b} is path connected. And if b is a midpoint of the interval, then it is not path connected. – lee Dec 9 '12 at 15:02
In fact, it may be the case that $B\setminus\{b\}$ is path connected, but $B$ itself is not: Consider a two-point set $B$. – Hagen von Eitzen Dec 9 '12 at 15:13
@lee : question edited so as to take your example into account. – Ewan Delanoy Dec 9 '12 at 16:15
@HagenvonEitzen : question edited so as to take your example into account. – Ewan Delanoy Dec 9 '12 at 16:15
Can a single comment contain notifications to several users? – Ewan Delanoy Dec 9 '12 at 16:17

1 Answer 1

up vote 2 down vote accepted

Consider $b\in B$. Let $a$ be another point of $B$. Assume that $C$ does not intersect an open neighbourhood $U$ of $b$. Select $b_0\in U\cap B\setminus \{a,b\}$ (using the fact that $b$ is not an isolated point). Then $b_0\notin C$, hence there is a path $\gamma_0\colon[0,1]\to B$ from $a$ to $b$ avoiding $b_0$. Select $b_1=\gamma_0(t_0)$ on $\gamma_0([0,1])\cap U\setminus\{a,b\}$. Then $b_1\notin C$, hence there is a path $\gamma_1\colon[0,1]\to B$ from $a$ to $b$ avoiding $b_1$. The set $T_>:=\{t\in[t_0,1]\mid \gamma_0(t)\in\gamma_1([0,1])\}$ is compact, we have $1\in T_>$, $t_0\notin T_>$. Let $t_1=\min T_>$. The set $T_<:=\{t\in[0,t_0]\mid \gamma_0(t)\in\gamma_1([0,1])\}$ is compact, we have $0\in T_<$, $t_0\notin T_<$. Let $t_2=\max T_<$. There exists $s_1\in [0,1]$ with $\gamma_1(s_1)=\gamma_0(t_1)$. The set $S := \{s\in[0,1]\mid \gamma_1(s)=\gamma_0(t_0)\}$ is compact. Let $s_0$ be an element that minimizes the continuous function $s\mapsto |s-s_1|$. Then we can glue together a simple closed curve $\gamma\colon[0,1]\to B$ per $$\gamma(t)=\begin{cases}\gamma_0(t_0+2t(t_1-t_0))&\text{if }t<\frac12 \\ \gamma_1(s_1+(1-2t)(s_0-s_1))&\text{otherwise}\end{cases}$$ By the Jordan curve theorem, we obtain two open sets, the interior and te exterior of the curve. Since $A$ is dense, there are interior and exterioir points of $A$. Any path between such points intersects our Jordan curve $\subseteq B$, contradicting the path connectedness of $A$. Therefore $C$ intersects every openneighbourhoodof $B$, i.e. $b\in\overline C$. Since $b\in B$ was arbitrary, $C$ is dense in $B$.

share|cite|improve this answer
The definition of your set $S$ looks strange to me. Since $\gamma_1$ avoids $b_1=\gamma_0(t_0)$, must not $S$ be empty ? – Ewan Delanoy Dec 9 '12 at 17:42
In general it is not at all simple to extract a simple path from an arbitrary path (which can be fractal, Peano-like, nowhere differentiable, you name it). – Ewan Delanoy Dec 9 '12 at 17:45
Since any path connected subset of the plane is also arc connected, however, one can assume that $\gamma_0$ and $\gamma_1$ are arcs. – Ewan Delanoy Dec 9 '12 at 19:27

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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