My book has two adjacent exercises, in which $\partial A$ is the set of boundary points of $A$ (a subset of some metric space $X$) and $\overline{A}$ is the closure of $A$. The first asks me to prove that $\overline{A} = A \cup \partial A$, which was simple and is discussed in this question.

The second exercise asks me to prove that $\partial A = \overline{A} \cap \overline{A^c}$. I proved this one way that matches the book's solution, but I also have another proof that uses the first exercise, the fact that $\partial A = \partial A^c$, and DeMorgan's Laws.

\begin{align*} \overline{A} \cap \overline{A^c} &= \left(A \cup \partial A\right) \cap \left(A^c \cup \partial A^c\right) \\ &= \left(A \cup \partial A\right) \cap \left(A^c \cup \partial A\right) \\ &= \partial A \cup \left(A \cap A^c\right) \\ &= \partial A \cup \varnothing \\ &= \partial A \end{align*}

Is this a correct proof?


I don't quite see how you go from $(A \cup \partial A) \cap (A^c \cup \partial A)$ to the line below it in one step.

I would expand to 4 terms: $(A \cap A^c) \cup (\partial A \cap A^c) \cup (A \cap \partial A) \cup (\partial A \cap \partial A)$, where the first term equals $\emptyset$, the final one $\partial A$ and the middle terms $(\partial A \cap A) \cup (\partial A \cap A^c)$ can be regrouped as $\partial A \cap (A \cup A^c) = \partial A \cap X = \partial A$ as well, so the total group becomes indeed $\partial A$.

Of course the definition of the boundary already gives the formula almost directly: $x \in \partial A$ iff every open neighbourhood of $x$ intersects $A$ (i.e. $x \in \overline{A})$ and every open neighbourhood of $x$ intersects $X \setminus A = A^c$ (i.e. $x \in \overline{A^c}$. So $\partial A = \overline{A} \cap \overline{A^c}$, almost by definitions of boundary and closure.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.