Existence of a limit point implies existence of inifintely many limit points?

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
8 replies · 3K views
julypraise
Messages
104
Reaction score
0

Homework Statement


Prove the following statement is true or not:

the statement:
Let [itex](X,d)[/itex] be a non-empty metric space and [itex]A[/itex] is a non-empty subset of [itex]X[/itex]. Then if [itex]A'[/itex] is not empty, then [itex]A'[/itex] is infinte.


Homework Equations


Definition of limit point and its negation.


The Attempt at a Solution


I tried to prove by contradiction in this way: (to prove the statement is true)
Suppose [itex]p_{1},\dots,p_{N}[/itex] are limit points of [itex]A[/itex]. Since [itex]A[/itex] has a limit point, it is infinte. Now observe that if [itex]p'\in A,p\notin\{p_{1},\dots,p_{N}\}[/itex] then [itex]p'[/itex] is not a limit point of [itex]A[/itex], i.e., [itex]A\cap B(p';r_{p'})=\{p'\}[/itex] for some [itex]r_{p'}>0[/itex]. Thus [itex]A[/itex] has a kind of gap inside it.

But from here, I cannot go further. It almost seems that [itex]A'[/itex] finite is consistent.
 
Physics news on Phys.org
julypraise said:
But from here, I cannot go further. It almost seems that [itex]A'[/itex] finite is consistent.

Go with that. I can think of a metric space with one limit point. Can you?
 
Last edited:
Dick said:
Go with that. I can think of a metric space with one limit point. Can you?

Sorry, I tried but couldn't think of one. Could you give me at least a hint? Like, what kind of metric space is it with what kind metric??
 
Well, one obvious way to avoid "an infinite number of limit points" is to start with a space that only contains a finite number of points!
 
HallsofIvy said:
Well, one obvious way to avoid "an infinite number of limit points" is to start with a space that only contains a finite number of points!

Humm.. But, you know, if a set has only finite points then there is no limit point of this set for whatever metric space is talked about. Isn't it?
 
julypraise said:
Sorry, I tried but couldn't think of one. Could you give me at least a hint? Like, what kind of metric space is it with what kind metric??

Nothing fancy. Pick a subset of the real numbers. They should be discrete but approach a single limit point. Like 0.
 
Wow, this one right? [itex]\bigcup_{n=1}^{\infty}\{-\frac{1}{n},\frac{1}{n}\}[/itex].
 
julypraise said:
Wow, this one right? [itex]\bigcup_{n=1}^{\infty}\{-\frac{1}{n},\frac{1}{n}\}[/itex].

Looks generally ok to me. But I think you want to include 0 in your subset. What do you think?
 
Last edited:
Dick said:
Looks generally ok to me. But I think you want to include 0 in your subset. What do you think?

Ah... I think it's okay to disprove the statement. I showed that A' is not empty but finite anyway. Thanks!