Is the cardinality of an infinite product easily expressed in simpler terms?

The specific problem is this. Suppose $\alpha$ is a countable limit ordinal. The set I want to count has size equal to $\Pi_{\beta<\alpha} \beth_\beta$.

I would like the output to be $\beth_\alpha$ but honestly I'm not sure how this works out. Finite cardinal multiplication is very easy, but I don't know anything about infinite products like this. Certainly I can bound it above; $\beth_\beta<\beth_\alpha$, so the product has size at most $\beth_\alpha^\omega$, but by Konig's theorem this is strictly larger than $\beth_\alpha$ and doesn't sit well with me.

  • $\begingroup$ Nothing much can be said about infinite product of cardinals. Sure we can say this and that. But something concrete? Not very often. $\endgroup$ – Asaf Karagila Sep 10 '15 at 19:20
  • $\begingroup$ So in this case, is it likely to say "it's at most $\beth_\alpha^\omega$" and leave it there? $\endgroup$ – Richard Rast Sep 10 '15 at 20:23
  • $\begingroup$ I have to disagree with Asaf, though I figure it is really that we interpret the quantifier "nothing much" in different ways. For a general result in particular containing the computation you are asking, see theorem 13 in this blog post (which refers to this one). $\endgroup$ – Andrés E. Caicedo Sep 23 '15 at 15:25

It turns out this does have a nice answer. I originally had a much more complicated proof, and almost immediately afterward, a simple one was given to be by Douglas Ulrich. The answer is that for any limit ordinal $\alpha$, $\prod_{\beta<\alpha} \beth_\beta=\beth_{\alpha+1}$. What follows is a proof in more detail than is maybe necessary.

The essential reason is the $\prod_i 2^{\kappa_i}=2^{\sum_i \kappa_i}$, which is easily verified. This gives us the following:

$\prod_{\beta<\alpha}\beth_\beta=\prod_{\beta<\alpha}2^{\beth_\beta}$; this is seen by seeing both sides inject naturally into the other. Left to right is that $\beth_\beta\leq 2^{\beth_\beta}$, while right to left is because this is a product of a smaller set (since $\alpha$ is a limit ordinal).

Then $\prod_{\beta<\alpha}2^{\beth_\beta}=2^{\sum_\beta\beth_\beta}$. Clearly $\sum_\beta\beth_\beta=\beth_\alpha$, so this last is $2^{\beth_\alpha}=\beth_{\alpha+1}$, as desired.


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.