MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In a paper placed on the arXiv today Shelah references theorem 0.9 from this paper (also Shelah) that uses $\aleph_{736}$ as an upper bound. This strikes me as analogous to Skewes' number. Are there any other examples where explicit mentions of 'large alephs' are used in proofs or theorems? Large in this instance is $\aleph_n$ for $n > 3$, say.

Edit: The first version of this question asks for $n$ a natural number, because clearly there are uses of $\aleph_\omega$ that are not all that uncommon, I imagine. But if there are uses of alephs - in isolation, not as part of a general transfinite induction scheme - that have $n$ an infinite ordinal like $\omega^2\cdot 5 + 45$ or some specific polynomial in $\omega$ which is 'not boring' (e.g. $\omega, \omega+1$), then I'd like to hear those as well.

Bonus points, even though this is a separate question: where does the 736 come from? Is it due to some sort of Goedelian numbering scheme that encodes the statement “G is a free abelian group”, which is part of the theorem?

share|cite|improve this question
You probably mean $n$ to be a natural number? – Joel David Hamkins Nov 17 '10 at 1:50
Well, probably. But if there was an instance of $\aleph_{25\omega^2 + 1001}$ that was genuinely needed - in isolation, and not like in Gentzen's proof of the consistency of PA or similar, I would accept that. In fact I'll edit the question accordingly. – David Roberts Nov 17 '10 at 2:00
Isn't it usually spelled 'Shelah'? I hardly know anything about transliteration schemes from Semitic languages, but I've never seen it spelled like this. – Todd Trimble Nov 17 '10 at 2:01
Opps!1 fixed. Got confused with – David Roberts Nov 17 '10 at 2:10
Since the arxiv has Shelah as the author, I would use Shelah if for no other reason than to find the paper in the arxiv. Also, the author seems to use the spelling "Shelah" in most, if not all, of his publications. Gerhard "Ask Me About System Design" Paseman, 2010.11.16 – Gerhard Paseman Nov 17 '10 at 2:12


  1. As for the "736" in Shelah's paper, this is just his way of saying "for every $n>1$, there is a group as in the theorem which is free iff the continuum is below $\aleph_n$", there is nothing privileged or special about $\aleph_{736}$ in this result, other than it is ``random''.

  2. Of course, what should be by far the most famous occurrence of a "large" $\aleph_n$ is $\aleph_4$, in a result due to Shelah, which is a cornerstone of modern cardinal arithmetic. The result is usually stated about exponentiation, although it is really about the pcf structure. For example:

If $2^{\aleph_0}<\aleph_\omega$, then $\aleph_\omega^{\aleph_0}<\aleph_{\aleph_4}$.

This is quite strange, really. We do not know at the moment that $\aleph_4$ is needed. The best results (due, once again, to Shelah) show that it cannot be anything below $\aleph_1$. Some people suspect $\aleph_2$ should be the right number, but at the moment $\aleph_4$ is what we get. And we do not understand it well. When Shelah found the result, he told Leo Harrington about it, and Leo asked, quite naturally "Why the HELL is 4?" So, of course, when Shelah wrote his "Cardinal Arithmetic" book, that was the title of the relevant section.

3. Though they are not strictly $\aleph_n$s, you get an interesting collection of possible cardinalities as the uncountable spectrum of a countable theory. Given a consistent complete first-order theory $T$ in a countable language which does not admit finite models, let $I(T,\kappa)$ denote the number of isomorphism classes of models of $T$ of size $\kappa$.

The possible values of $I(T,\aleph_0)$ are not known (Vaught's conjecture). But for $\kappa$ uncountable, the possibilities for $I(T,\kappa)$ are known. This is essentially due to Shelah, but the last pieces of the result were obtained in Bradd Hart, Ehud Hrushovski, Michael C. Laskowski, "The uncountable spectra of countable theories", Ann. of Math. (2) 152 (2000), no. 1, 207--257: alt text

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.