0

Let $ f: X \rightarrow Y $ be a continuous function. And let $ E $ be a connected subset of $ X $.

I want to show that $ f(E) $ is also connected.

I am able to show that if $ X $ is connect if follows that $ f(X) $ is also connect.

I noticed that there are proofs on this site, but they all use a different definition of connected. I want to prove it using the following definition: $ A \subset X $ is connected if there are no non-empty disjoint open sets $ U $ , $ V $ such that $ A = U \cup V $.

My work: For $ f: X \rightarrow Y$ continuous. Suppose $ A \subset X $ is connect. For sake of contradiction assume $ f(A) $ is not connected. Thus there are open sets $ U, V $ such that $ U_1 := U \cap f(A) $ and $ V_1 := V \cap f(A) $, so that $ U_1 $ and $ V_1$ are disjoint and nonempty as well as $ f(A) = U_1 \cup V_1 $.

It follows that $ f^{-1} (U_1) = f^{-1} (U) \cap f^{-1}f(A)) $. However, if $ A \neq X $ I am unable to show that $ f^{-1} (U_1) $ needs to be open or that $ f^{-1} (U_1) \cup f^{-1} (V_1) = f(A) $.

Any help would be much appreciated.

user007
  • 159

1 Answers1

1

I think you meant "suppose $A\subseteq X$ is connected" (not $A\cup X$).

Hint: instead of focusing on $U_1$, focus on $f^{-1}(U)$. This is open because $f$ is continuous.

Elaboration: Since $U\cup V\supseteq f(A)$, it follows $f^{-1}(U)\cup f^{-1}(V)\supseteq A$ with both terms on the left hand side being open. It remains to show that $f^{-1}(U)\cap A$ and $f^{-1}(V)\cap A$ are disjoint.

FShrike
  • 40,125
  • I see that $ f^{-1} (U) $ is open since $ U $ is open. However, I don't see how this helps me show that $ A $ wouldn't be connected since the only condition $ U $ has to follow is that it is open. – user007 Sep 18 '23 at 19:51
  • @user007 I elaborated slightly – FShrike Sep 18 '23 at 19:55
  • Thanks for the hints. We want to prove $(f^{-1}(U) \cap A) \cap (f^{-1}(V) \cap A) = \emptyset$. That's equivalent to $(f^{-1}(U) \cap f^{-1}(V) \cap A = \emptyset$. Suppose that $ (f^{-1}(U) \cap f^{-1}(V) \cap A \neq \emptyset $. We can follow that $U \cap V \cap f(A) \neq \emptyset$. We also have $U \cap V_1 \neq \emptyset$. But this implies that there's an element $a$ such that $ a \in V_1$ thus in $ f(A)$ and $ V $ but also in $U$. Therefore it is in $U_1$. But that is a contradiction. Therefore, $(f^{-1}(U) \cap A) \cap (f^{-1}(V) \cap A) = \emptyset$. But that is a contradiction. – user007 Sep 18 '23 at 20:18
  • Is that correct? – user007 Sep 18 '23 at 20:18
  • 1
    @user007 That looks good. I might just say, suppose $f^{-1}(U)\cap f^{-1}(V)\cap A$ is nonempty. Then there is $x$ in all of these. $f(x)\in U$ and $f(x)\in V$ and $f(x)\in f(A)$ so $f(x)\in(U\cap f(A))\cap(V\cap f(A))$ but by assumption these are disjoint – FShrike Sep 18 '23 at 20:42