The fact that a subspace of a normal topological space is normal is known to be wrong.
I'm trying to find the wrong step in the following proof:
Let $X$ be a normal topological space and $Y$ be a subspace. Let $C_1, C_2$ be closed sets in $Y$.
So there exist $C_1', C_2'$ closed in $X$ such that $C_i = Y \cap C_i' $ for $i=1,2$.
Since $X$ is normal, there are open disjoint sets $U_1, U_2$ such that $C_i' \subset U_i$.
Thus $C_i=C_i' \cap Y \subset U_i \cap Y$.
So if we let $V_i = U_i \cap Y$, we get that $V_1, V_2$ are open in $Y$, disjoint, and satisfy $C_i \cap V_i$.
Therefore $Y$ is a normal topological space.
Where is the mistake?
Asked
Active
Viewed 104 times
5
diracdeltafunk
- 14,911
Omer
- 2,490
-
1Scrutinize your proof with a counterexample in hand. – Anne Bauval Dec 30 '23 at 16:15
-
1$C_1,C_2$ disjoint does not imply $C'_1,C'_2$ disjoint. – Anne Bauval Dec 30 '23 at 16:19
-
I think that (C′i∩Y)⊂(Ui∩Y) would be clearer; I assume that's what you mean. – Acccumulation Dec 31 '23 at 01:30
1 Answers
5
Since $X$ is normal, there are open disjoint subsets $U_1, U_2$ such that $C_i' \subset U_i$
No, this only follows if you also know that $C_1' \cap C_2' = \varnothing$!
This is actually the source of two separate problems in your proof. Can you identify them?
diracdeltafunk
- 14,911
-
2Hard hour. I didn't even assume the original C1, C2 were disjoint apparently. Thanks, I'll accept the answer when the time limit expires – Omer Dec 30 '23 at 16:19
-
2