6

I know and I have proved more than once that the set of irrational numbers ($\mathbb{I}$) is uncountable, but now I'm given to solve this problem:

Show that $|\mathbb{I}|=|\mathbb{R}|$,

How can I do that?

Do I need to assume the Continuum Hypothesis in order to that statement to be true?

Thanks

  • 6
    The two differ by a countable set. – David Mitra Mar 01 '14 at 21:39
  • 2
    @David: That's not sufficient (without the axiom of choice) – Asaf Karagila Mar 01 '14 at 21:43
  • Why do people still care about AC? – user2345215 Mar 01 '14 at 21:44
  • 3
    @user2345215: Why do people still care about sets? Or mathematics? – Asaf Karagila Mar 01 '14 at 21:44
  • 2
    @Asaf Karagila: But why would you point that out? It's not like I go to questions and say: "hey but this requires the axiom of replacement!" – user2345215 Mar 01 '14 at 21:46
  • @AsafKaragila Is it difficult to do the standard trick in finding a bijection between the two without using AC? (By standard trick I mean what you suggested in your answer.) – David Mitra Mar 01 '14 at 21:47
  • 2
    @user2345215: Because things requiring choice are usually things that you can't just write down in a formula. Also, there are way more things that you need to use choice for than things you need replacement for (in ordinary mathematics, anyway). But it's an interesting experiment, go ahead and try that! :-) – Asaf Karagila Mar 01 '14 at 21:49
  • @David: If $\Bbb I$ was a set which didn't have a countably infinite subset, how would you do it? – Asaf Karagila Mar 01 '14 at 21:49
  • @AsafKaragila I suppose I should have been more explicit in my first comment: "construct a bijection using the fact that the two differ by a countable set". – David Mitra Mar 01 '14 at 21:52
  • @David: Well, that's still a bit ambiguous. If anything, it's by the fact that $|\Bbb I|>\aleph_0$. – Asaf Karagila Mar 01 '14 at 21:53

4 Answers4

8

Fix an enumeration of $\Bbb Q$, $q_n$, and find a countably infinite subset of $\Bbb I$, $r_n$.

Now find a map which fixes all the points which are not $r_n$'s, and maps the union $\{q_n,r_n\mid n\in\Bbb N\}$ into $\{r_n\mid n\in\Bbb N\}$.


The above can be translated quite neatly to cardinal arithmetic. Write $\Bbb I$ as $A\cup B$ where $A$ is countably infinite, and $B\cap A=\varnothing$. Then we have:

$$|\Bbb R|=|\Bbb Q|+|\Bbb I|=\aleph_0+(|A|+|B|)=\aleph_0+(\aleph_0+|B|)=\aleph_0+|B|=|\Bbb I|.$$

Asaf Karagila
  • 393,674
3

Here is how to construct an explicit bijection. First we construct a bijection $\alpha$ from $\mathbb{N}$ to $\mathbb{Q}$. (For definiteness I will take $\mathbb{N}$ to include $0$, this makes no real difference.) For brevity, call $\alpha(i)$ by the name $r_i$.

Now we proceed semi-formally. Map all numbers which are not rational or of the form $r+\sqrt{2}$, where $r$ is rational, to themselves.

Map rational $r_i$ to $r_{2i+1}+\sqrt{2}$. Map $r_i+\sqrt{2}$ to $r_{2i}+\sqrt{2}$.

André Nicolas
  • 507,029
3

In what follows, I write $\mathfrak c$ for $|\mathbb R|$, as customary. That $\mathfrak c=2^{\aleph_0}$ can be proved in several ways. For example, the Cantor set $C$, by construction, has size $2^{\aleph_0}$. This shows that $2^{\aleph_0}\le\mathfrak c$. On the other hand, it is easy to see that $\mathbb R$ has the same size as any open interval, and $|(0,1)|\le2^{\aleph_0}$, as we can identify the binary expansion of a number with an infinite sequence of zeros and ones. (I am using that if $A$ and $B$ inject into each other, then there is a bijection between them. This is the Bernstein-Schröder theorem.)

To see the relevance of this, your question is asking for a very particular case of a result discussed here, namely (without assuming the axiom of choice) if $\mathfrak m$ and $\mathfrak n$ are cardinalities, $\mathfrak m+\mathfrak m=\mathfrak m$, and $\mathfrak m+\mathfrak n=2^{\mathfrak m}$, then $\mathfrak n=2^{\mathfrak m}$.

Your question is the case where $\mathfrak m=|\mathbb Q|=\aleph_0$, and $\mathfrak n=|\mathbb I|$, precisely because $|\mathbb R|=2^{\aleph_0}$.

There are other approaches, of course. As discussed here and here, there are reals $r$ such that the set $C+r=\{x+r\mid x\in C\}$ is contained in $\mathbb I$. This shows that $\mathfrak c\le|\mathbb I|$. Since clearly $|\mathbb I|\le\mathfrak c$, we again gave equality.

For yet another variant of this idea, note that ($|\mathbb I|\le\mathfrak c$ and) obviously, $2^{\aleph_0}\le{\aleph_0}^{\aleph_0}$, and the latter is the size of $\mathbb I$, as discussed for example here, where it is shown that not only there is a bijection betwen the space $\mathbb I$ and Baire space $\mathcal N=\mathbb N^{\mathbb N}$ but, in fact, the two spaces are homeomorphic (where the latter has the product topology of countably many copies of the discrete set $\mathbb N$).

0

$\mathbb{R} = \mathbb{Q} \cup \mathbb{I}$, so $|\mathbb{R}| = |\mathbb{Q}| + |\mathbb{I}| = \max(|\mathbb{Q}|, |\mathbb{I}|) = |\mathbb{I}|$, as the latter is uncountable.

Henno Brandsma
  • 242,131