Does anyone know a proof that the Cantor set, $\{0,1\}^{\mathbb{N}}$, has a dense subset homeomorphic to the Baire space, $\mathbb{N}^{\mathbb{N}}$? Thank you.


2 Answers 2


Define a mapping $f: \mathbb{N}^\mathbb{N} \to \{ 0,1 \}^\mathbb{N}$ by $$f : \langle x_n \rangle_{n \in \mathbb{N}} \mapsto 0^{x_0} 1 0^{x_1} 1 \cdots$$ where by $0^k$ we mean the length $k$ sequence consisting of only zeroes. It is fairly straightforward to show that this is a homeomorphic embedding of $\mathbb{N}^\mathbb{N}$ into $\{ 0,1 \}^\mathbb{N}$. It is also not too difficult to show that the range of $f$ is $$\{ \langle y_n \rangle_{n \in \mathbb{N}} \in \{ 0,1 \}^\mathbb{N} : y_n = 1\text{ for infinitely many } n \},$$ which is dense in $\{ 0,1 \}^\mathbb{N}$. (Use the fact that basic open sets in $\{ 0,1 \}^\mathbb{N}$ are determined by finite binary sequences, and then extend that finite binary sequence in any way to have infinitely many $1$s.)


If you happen to have it handy, there’s a pile-driver that takes care of the problem in short order: a characterization of the irrationals due originally to Aleksandrov, if I’m not mistaken. The space of irrationals is (up to homeomorphism) the unique zero-dimensional, separable, Čech-complete metrizable space that is nowhere locally compact. A Tikhonov space is Čech-complete iff it’s a $G_\delta$ in some (and in fact in any) compactification. Let $$X=\left\{x\in\{0,1\}^{\Bbb N}:x\text{ is not eventually constant}\right\}\;.$$ ($X$ corresponds to the points of the middle-thirds Cantor set that are not endpoints of removed intervals.)

  • $X$ and $\{0,1\}^{\Bbb N}\setminus X$ are both dense in $\{0,1\}^{\Bbb N}$, which is therefore a compactification of $X$.
  • $\{0,1\}^{\Bbb N}\setminus X$ is countable, so $X$ is a $G_\delta$ in its compactification $\{0,1\}^{\Bbb N}$ and is therefore Čech-complete.
  • $\{0,1\}^{\Bbb N}$ is a zero-dimensional, separable metrizable space, so $X$ is as well.
  • That $X$ is nowhere locally compact follows easily from the fact that its complement is dense in $\{0,1\}^{\Bbb N}$.
  • 1
    $\begingroup$ I must admit it seemed a bit obfuscated to me until I realised that "separable + Čech-complete + metrizable" = Polish. $\endgroup$
    – user642796
    Dec 9, 2012 at 15:22

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.