Connectedness: Show Y U A & Y U B are Connected

  • Thread starter Thread starter radou
  • Start date Start date
Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
8 replies · 4K views
Messages
3,149
Reaction score
8

Homework Statement



This one is giving me serious trouble.

Let Y be a subset of X. Let both X and Y be connected. SHow that if A and B form a separation of X\Y, then Y U A and Y U B are connected.

The Attempt at a Solution



I know all the basic definitions and theorems from the chapter about connectedness preceding this exercise section, but any way I try it, I don't seem to get anywhere.

Any ideas?
 
Physics news on Phys.org
When you say "[tex]A[/tex] and [tex]B[/tex] form a separation of [tex]X \setminus Y[/tex]", what does this mean, exactly? My guess would be what I usually call "[tex]\{A, B\}[/tex] disconnect [tex]X \setminus Y[/tex]", that is, [tex]A[/tex] and [tex]B[/tex] are disjoint (relatively) open subsets of [tex]X \setminus Y[/tex], whose union is [tex]X \setminus Y[/tex]. Is this correct?
 
Well, partially. According to Munkres, there's a subtle difference:

If X is a topological space, a separation of X is a pair of non empty, disjoint and open subsets U and V of X whose union is X. (This is a definition.)

If Y is a subspace of X, a separation of Y is a pair of non empty disjoint sets U, V whose union if Y. (This is a Lemma)

Note that in the Lemma it is not required for the separation sets to be open.

So, I assume if we're talking about X\Y, which is a subspace of X, the sets in the separation are not required to be open.
 
That won't work. You need some kind of relative openness hypothesis, otherwise the result is false. Take [tex]X = \mathbb{R}[/tex], [tex]Y = [0, 1][/tex], [tex]A = (-\infty, -1] \cup (1, 2)[/tex], [tex]B = (-1, 0) \cup [2, \infty)[/tex]. Then [tex]A[/tex] and [tex]B[/tex] form a separation (according to the conditions you gave) of [tex]X \setminus Y = (-\infty, 0) \cup (1, \infty)[/tex], but [tex]Y \cup A = (-\infty, -1] \cup [0, 2)[/tex] and [tex]Y \cup B = (-1, 1] \cup [2, \infty)[/tex] are both disconnected.

Also, with the definition of "separation" you cite for subspaces, you can give a separation for a connected subspace, such as [tex]X = \mathbb{R}, Y = [0, 1], U = [0, \textstyle\frac12), V = [\textstyle\frac12, 1][/tex]. That doesn't make sense.
 
ystael said:
That won't work. You need some kind of relative openness hypothesis, otherwise the result is false. Take [tex]X = \mathbb{R}[/tex], [tex]Y = [0, 1][/tex], [tex]A = (-\infty, -1] \cup (1, 2)[/tex], [tex]B = (-1, 0) \cup [2, \infty)[/tex]. Then [tex]A[/tex] and [tex]B[/tex] form a separation (according to the conditions you gave) of [tex]X \setminus Y = (-\infty, 0) \cup (1, \infty)[/tex], but [tex]Y \cup A = (-\infty, -1] \cup [0, 2)[/tex] and [tex]Y \cup B = (-1, 1] \cup [2, \infty)[/tex] are both disconnected.

Also, with the definition of "separation" you cite for subspaces, you can give a separation for a connected subspace, such as [tex]X = \mathbb{R}, Y = [0, 1], U = [0, \textstyle\frac12), V = [\textstyle\frac12, 1][/tex]. That doesn't make sense.

Interesting, it's exactly what says in the book. So, in the "separation lemma" for subspaces, the sets should be open too?
 
At least relatively open in the subspace, otherwise they are useless for establishing disconnectedness. I don't have a copy of the new edition of Munkres, so I can't guess what he might be thinking there.
 
ystael said:
At least relatively open in the subspace, otherwise they are useless for establishing disconnectedness. I don't have a copy of the new edition of Munkres, so I can't guess what he might be thinking there.

OK, I looked it up. The problem is that you missed an important hypothesis in the lemma 23.1, which changes things entirely.

Lemma 23.1 of Munkres says that a separation of [tex]Y \subset X[/tex] is a disjoint pair of nonempty sets [tex]A, B[/tex] whose union is [tex]Y[/tex], which satisfy the condition that neither of [tex]A, B[/tex] contains a limit point of the other. This latter condition is exactly what you need for [tex]A[/tex] and [tex]B[/tex] to be relatively open (in fact, clopen) in [tex]Y[/tex]. The examples I gave above do not satisfy the latter condition.
 
ystael said:
OK, I looked it up. The problem is that you missed an important hypothesis in the lemma 23.1, which changes things entirely.

Lemma 23.1 of Munkres says that a separation of [tex]Y \subset X[/tex] is a disjoint pair of nonempty sets [tex]A, B[/tex] whose union is [tex]Y[/tex], which satisfy the condition that neither of [tex]A, B[/tex] contains a limit point of the other. This latter condition is exactly what you need for [tex]A[/tex] and [tex]B[/tex] to be relatively open (in fact, clopen) in [tex]Y[/tex]. The examples I gave above do not satisfy the latter condition.

You're right, for some reason I forgot to mention this.