Let $f: X \rightarrow Y$, and $A$ is subset of $X$.
Then is it right $f(A^c) = f(A)^c$ if $f$ is injective?
What about the condition "injective" is not involved?
Let $f: X \rightarrow Y$, and $A$ is subset of $X$.
Then is it right $f(A^c) = f(A)^c$ if $f$ is injective?
What about the condition "injective" is not involved?
On
Why is injectivity needed? Consider $f: \{1, 2\} \to \{0\}$, with $f(1) = f(2) = 0$, and $A = \{1\}$. $f(A) = \{0\}$ and $f(A^c) = \{0\}$ as well.
Why is injectivity sufficient? Say $x \in f(A^c)$, then there is $w \in A^c$ such that $f(w) = x$. But $f$ is injective. So no other thing $w' \in X$ will have $f(w') = x$. In particular, for any $w' \in A$, $w' \not= w$, and thus $f(w') \not= x$. This means $x \not\in f(A)$ and thus $x \in f(A)^c$. Since $x$ is chosen arbitrary, this shows that $f(A^c) \subseteq f(A)^c$.
The other direction is similar.
First, $B$ should be a subset of X, not of Y. Yes, injectivity is required but also surjectivity is required for the equality to hold. With injectivity we only have, $ f(B^c) \subset (f(B))^c$ A counterexample would be
$f(x) = tan^{-1}x$
Since taking $ B = \mathbb{R}\setminus{0} $
we have, $f(B^c) = f(0) = 0$
But, $ (f(B))^c = [π/2, \infty) \cup (-\infty, -π/2] \cup {0}$