I'm having trouble of thinking of such a set (real or complex) because:

Let X $\subset \mathbb{R}$ or X $\subset \mathbb{C} $, then by definition the closure is:

$\bar{X}$:=X $\cup \partial X$

While by definition $\partial X$ := $\bar{X}$ \ int(X),

where int(X) is the interior of X.

My initial thought was to use the Cantor set, but this has no interior.

  • $\begingroup$ You're correct that the definition of the boundary is $\partial X := \bar X \setminus \mathop{\rm int}(X)$; but the definition of the closure is not $\bar X := X \cup \partial X$ (that would make the definitions circular). The definition of closure is the usual fundamental definition; $\bar X = X \cup \partial X$ is a consequence of the definition of $\partial X$. $\endgroup$ Feb 9, 2014 at 2:41
  • $\begingroup$ @Daniel, you're right. The original title was wrong, it is now corrected. $\endgroup$
    – BBaire
    Feb 9, 2014 at 3:47
  • $\begingroup$ Look at a dense set in $[0, 1]$. $\endgroup$ Feb 9, 2014 at 3:48
  • 1
    $\begingroup$ You're thinking way too complicated: $X = (0,1) \cup (1,2)$. $\endgroup$ Feb 9, 2014 at 9:18

1 Answer 1


Let $X = \mathbb R \setminus \{0\}$, in $\mathbb R$. Its boundary is $\{0\}$, closure is $\mathbb R$, and so boundary of its closure is empty.


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.