Let $A$ be an uncountable set of real numbers. Then by using bolzano-weierstrass theorem, it can be easily shown that $A$ has at least one limit point. But my question is how many limit point? Can an uncountable set has countable number of limit points? Please suggest me. Thanks in advance.

  • $\begingroup$ Let $U = \bigl\{ x \in \mathbb{R} : \bigl(\exists \varepsilon > 0\bigr)\bigl(A \cap (x-\varepsilon, x+\varepsilon) \text{ is countable}\bigr)\bigr\}$. What do you know about $U$? $\endgroup$ – Daniel Fischer Sep 21 '15 at 11:23
  • $\begingroup$ then what you have shown? $\endgroup$ – neelkanth Sep 21 '15 at 11:25
  • $\begingroup$ @neela: See this answer. $\endgroup$ – Brian M. Scott Sep 21 '15 at 11:32

Without loss of generality, we can assume $A$ is closed, since the limit points of $\overline{A}$ are exactly the limit points of $A$. The Cantor-Bendixson Theorem says that every uncountable, closed subset of the reals contains a nonempty perfect subset. (A perfect set is a set with no isolated points, i.e., a set where every point is a limit point.) It is a fact that every (nonempty) perfect subset of $\mathbb{R}$ has cardinality $2^{\aleph_0}$ (proof idea: build a tree of closed sets, and use Bolzanno-Weierstrass to show that every path along the tree contains a distinct point), $A$ has at least $2^{\aleph_0}$ limit points. Since there are only $2^{\aleph_0}$ real numbers to begin with, this answer is exact.

  • $\begingroup$ how assuming closed do not loss generality? $\endgroup$ – neelkanth Sep 23 '15 at 7:33
  • $\begingroup$ @neela In the more general case, the proof I wrote can show that $\overline{A}$ has $2^{\aleph_0}$ limit points. Since those limit points will all also be limit points of $A$, $A$ has $2^{\aleph_0}$ limit points, too. $\endgroup$ – Mike Haskel Sep 23 '15 at 10:14

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.