-1

Consider the map $\varphi$ from $A = \prod_{n = 1}^{\infty}\{ x \in \mathbb{Z} \mid 0 \leq x \leq 9 \}$ to $[0, 1]$ such that $$ \varphi : (a_{1} , a_{2}, \dots) \mapsto 0.a_{1}a_{2}\dots $$ Since $A$ is countable, and $\varphi$ is surjective (every real number in $[0, 1]$ has a decimal representation), $[0, 1]$ is also countable, which is not the case, what's wrong with the reasoning?

Edit: $A$ is uncountable. I thought that countable union of countable sets is countable, however in this case it's the countable product of countable set rather than union, it's the assumption that's wrong, not the conclusion.

1 Answers1

1

Since $A$ is countable

Unfortunately it is not. And this reasoning here can serve as a proof of that, assuming you already know that $[0,1]$ is uncountable.

If you don't, then the standard Cantor's diagonal argument applies. Assume that there is a sequence $f_{n}$ that covers whole $A$. So it is a sequence of sequences and thus $f_{n,m}$ makes sense.

Define $$a_k=(f_{k,k}+1)\text{ mod }10$$ There is nothing special about this formula. The point is that $a_k\neq f_{k,k}$. I've added mod only to stay in digits set. Any such choice will work. I leave as a simple exercise that $(a_k)$ cannot be in the image of $f_n$.

freakish
  • 42,851