8

$\aleph_0^\mathfrak c=2^\mathfrak c$

This is a question on my homework, and I am not sure how to show they equal each other, I tried Cantor-Schroeder-Bernstein but got stuck, any help would be great.

Asaf Karagila
  • 393,674
  • 1
    Please try to write your questions clearer and better titles, also try and be correct about the theorems you quote, especially their names. It should not be very hard to find the correct name on Wikipedia and copy-paste it here. – Asaf Karagila May 04 '11 at 19:13

3 Answers3

8

For all infinite cardinals $\kappa$, you have $2^{\kappa}=\lambda^{\kappa} \leq(\kappa^+)^{\kappa}$ for all cardinals $\lambda$, $2\leq \lambda\leq \kappa^+$, where $\kappa^+$ is the successor cardinal of $\kappa$.

Certainly, we know that since $2\leq \lambda\leq\kappa^+$ and $\kappa$ is infinite, then $2^{\kappa}\leq \lambda^{\kappa}\leq (\kappa^+)^{\kappa}$, since the set of all functions from $\kappa$ to $2$ naturally embeds into the set of all functions from $\kappa$ to $\lambda$, which naturally embeds into the set of all functions from $\kappa$ to $\kappa^+$.

Now, since $\kappa\lt 2^{\kappa}$ by Cantor's Theorem, we also know that $\kappa^+\leq 2^{\kappa}$, so $(\kappa^+)^{\kappa}\leq (2^{\kappa})^{\kappa} = 2^{\kappa\kappa} = 2^{\kappa}$, since $\kappa\kappa = \kappa$ (assuming the Axiom of Choice). So we have the chain of inequalities $$2^{\kappa}\leq \lambda^{\kappa} \leq (\kappa^+)^{\kappa} \leq 2^{\kappa},\qquad \text{if }2\leq \lambda\leq\kappa^+;$$ so all inequalities are equalities.

Since $\aleph_0\lt \mathfrak{c}$, it follows that $$\aleph_0^{\mathfrak{c}} = (\mathfrak{c}^+)^{\mathfrak{c}} = 2^{\mathfrak{c}}.$$

Note. The equality $\kappa\kappa=\kappa$ for arbitrary infinite cardinals/sets is equivalent to the Axiom of Choice; but one can prove that for alephs we have $\aleph_{\alpha}\aleph_{\alpha}=\aleph_{\alpha}$ without using AC, and from there we get that $\aleph_{\alpha}+\aleph_{\beta} = \aleph_{\alpha}\aleph_{\beta} = \max\{\aleph_{\alpha},\aleph_{\beta}\}$ without having to invoke AC. For the result you want, though, you don't need it because $\mathfrak{c}\mathfrak{c}= 2^{\aleph_0}2^{\aleph_0} = 2^{\aleph_0+\aleph_0} = 2^{\aleph_0} = \mathfrak{c}$ which does not require AC, as pointed out by Asaf in comments.

Arturo Magidin
  • 398,050
  • @Arturo: I think in the second paragraph, you wanted to write $2^{\kappa}\leq \lambda^{\kappa}\leq (\kappa^+)^{\kappa}$ instead of $2^{\kappa}\leq \lambda^{\kappa}\leq (\kappa)^{\kappa}$. – Martin Sleziak May 04 '11 at 18:35
  • @Martin: Oops; yes. I had some extra successors that I took out, took out one too many. Thanks! – Arturo Magidin May 04 '11 at 18:37
  • Do we not need the axiom of choice to show that $\kappa \kappa = \kappa$? – Zhen Lin May 04 '11 at 20:17
  • 2
    @Zhen Lin: I believe so; certainly, $|X\times X|=|X|$ for all infinite cardinals is equivalent to AC (Tarski's Theorem). For alephs you don't need it, though; so if we assume only that $\mathfrak{c}$ is an aleph, the desired result holds for $\lambda=\aleph_0$. – Arturo Magidin May 04 '11 at 20:26
  • @Asaf: Do you mean $2^{\kappa}\cdot 2^{\kappa}=2^{\kappa+\kappa}$, and $\aleph_0+\aleph_0$, no? I know that $\kappa^{\lambda}\cdot\kappa^{\mu}=\kappa^{\lambda+\mu}$ holds even in the absence of AC, but how do you go from that to $\kappa^{\lambda\mu}$? – Arturo Magidin May 04 '11 at 20:40
  • @Arturo: You are correct. I will give out the details I meant to give later. In the meantime I will delete the previous comment. – Asaf Karagila May 04 '11 at 21:02
  • 1
    @Arturo: Yes, what you wrote was exactly what I meant. Regarding your comment about multiplication in the power of the exponent, well... even in the absence of choice $(\lambda^\kappa)^\mu=\lambda^{\kappa\cdot\mu}$; the thing I wrote before was wrong. In the case of $\aleph_0$ in the power, then $\aleph_0+\aleph_0=\aleph_0\cdot\aleph_0=\aleph_0$. :-) – Asaf Karagila May 04 '11 at 22:21
  • @Asaf: Right: I knew the usual exponentiation rules hold even without AC (Jech has them before he introduces AC, which is how I know (-:) It just wasn't clear to me how you were going from $2^{\kappa+\kappa}$ to $2^{\kappa\kappa}$, since $\kappa+\kappa=\kappa\times\kappa$ for infinite $\kappa$ seems like something that might require (or be equivalent) to Choice... But now that you've clarified, all is fine. – Arturo Magidin May 05 '11 at 01:44
  • @Arturo, (or anyone reading this) do you have a reference for why $\aleph_{\alpha}\aleph_{\alpha}=\aleph_{\alpha}$ without AC where I could read up on this? – yunone May 09 '11 at 04:39
  • @yunone: I know it's in Jech's "Set Theory". I can look it up tomorrow morning. – Arturo Magidin May 09 '11 at 04:41
  • @Arturo, thanks, the title is enough. I found it as Theorem 3.5 on page 30 for anyone else interested. – yunone May 09 '11 at 04:44
7

$2^{\mathfrak c} \le \aleph_0^{\mathfrak c}$ since $2\le \aleph_0$

If you know, that for any cardinal numbers $a^b \le 2^{ab}$ and that $\aleph_0\cdot\mathfrak c=\mathfrak c$ then you get

$$\aleph_0^{\mathfrak c} \le 2^{\aleph_0\cdot\mathfrak c} = 2^{\mathfrak c}$$

Now you can use Cantor-Bernstein (you have both inequalities).

EDIT:

The inequality $a^b \le 2^{ab}$ follows from the fact that $A^B \subseteq \mathcal P(B\times A)$. (Here $A^B$ denotes the set of all maps from $B$ to $A$; every such function is a subset of $B\times A$.)

EDIT2:

Arturo's post reminded me of another possibility of proving the inequality $a^b \le 2^{ab}$. Since we know $a<2^a$ from Cantor theorem, we get $a^b \le (2^a)^b = 2^{ab}$.

  • 1
    About Edit 2: Yes, but the proof of the last equality then requires a little argument (any function $f:B\to(A^C)$ can be identified with a function $g_f:B\times C\to A$). – Andrés E. Caicedo May 04 '11 at 19:16
3

Here is an alternate "more elementary" proof that $\aleph_0^\mathfrak c \leq 2^\mathfrak c$:

We use that $c$ is the cardinality of both $\mathbb{R}$ and $(0,1)$.

Let $f: (0,1) \rightarrow \mathbb{N}$ be any function.

We define $F_f := \{ x+ f(x) | x \in (0,1) \}$.

Then it is easy to check that any function $f$ is uniquely identified by $F_f$ and thus

$f \to F_f$ is a 1-1 function from $\mathbb{N}^{(0,1)}$ to ${\mathcal P}(\mathbb{R})$.

Arturo Magidin
  • 398,050
N. S.
  • 132,525
  • @user9176: The common shortcuts \N and \R for blackboard bold are not part of the standard LaTeX packages; you have to write \mathbb{N} and \mathbb{R}. – Arturo Magidin May 05 '11 at 01:42