1

Q. Prove that: If we remove 12 points from the Euclidean plane $\mathbb{R}^2$ to get set $A$, then $A$ is connected.

I don't really understand how to proceed with the question. I thought of going via contradiction but I am stuck. Should I approach it via path connectedness?

  • Seems like this should be true if you replace $12$ by any finite number. Maybe a proof by induction? – Robert Shore Nov 11 '21 at 05:37
  • Complement of any countable set is connected. – Kavi Rama Murthy Nov 11 '21 at 05:40
  • notice that at any point in the plane, there are uncountably many lines extending radially from that point. given two distinct points in the plane missing finitely many points, at least one pair of lines from the two points must intersect in the plane and miss all of the points removed from the plane. use this as your path between the points. – C Squared Nov 11 '21 at 06:51

2 Answers2

1

We know that if $X$ is connected by paths then $X$ is connected. If we think of a polygonal path that dribbles the n points, the test would end, but I think this way does not show what may be happening behind the scenes. I will try to give a fairly visual demonstration: So suppose that $A=\mathbb R^2\setminus\{x_1,x_2,\dots,x_k\}$ where $x_i\neq x_j$ if $i\neq j$. Let $y,z\in A$, if the line that joins $y$ with $z$ does not contain any of the points $\{x_1,x_2,\dots,x_k\}$, finish the test. Otherwise, let $p$ be the midpoint of $y$ and $z$. Let $\mathscr L=\{p+tv:t\in\mathbb R\}$ the equidistant line of points $y$ and $z$.

enter image description here

Then for each $n\in\mathbb N$ there are two lines $\alpha_n$ and $\beta_n$ that join the points $y$ with $p+nv$ and $p+nv$ with $z$, respectively. Let $\gamma_n=\alpha_n*\beta_n$ a path that $y$ with $z$. Then we notice that $\gamma_n\cap\gamma_m=\{y,z\}$ for $n\neq m$. Let's suppose $\forall n\in\mathbb N,\exists r_n\in\{1,2,\dots,k\}$ such that $x_{r_n}\in\gamma_n$, which is false since we can list the $n + 1$ first steps and have: $$ \color{red}{k+1\text{ elements}}\left\{ \begin{array}{l} \color{red}{x_{r_1}}\in\gamma_1\\ \color{red}{x_{r_2}}\in\gamma_2\\ \vdots\\ \color{red}{x_{r_k}}\in\gamma_k\\ \color{red}{x_{r_{k+1}}}\in\gamma_{n+1}\\ \end{array} \right. $$ thus there exists a point $x_r$ that belongs to two paths $\gamma's$, which is a contradiction. Therefore $\exists n_0\in\mathbb N$ such that $\{x_1,x_2,\dots,x_k\}\not\in\gamma_{n_0}$. Thus there is a path $\gamma_{n_0}$ in $\mathbb R^2\setminus\{x_1,x_2,\dots,x_k\}$ that joins point $y$ with $z$. Then $\mathbb R^2\setminus\{x_1,x_2,\dots,x_k\}$ is connected by paths, then $\mathbb R^2\setminus\{x_1,x_2,\dots,x_k\}$ is connected.

On the other hand if you want you can work with portions of circles instead of working with line segments, like this:

enter image description here

Zaragosa
  • 1,679
0

If you remove the whole $\mathbb{Q}^2$ (i.e. the set $\mathbb{R}^2 - \mathbb{Q}^2$), you still get a path connected set. You can find the proof of this on this website.