Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I'm trying to prove that there is an uncountable ordinal all which members are countable ordinal. This is fairly easy if I can state that the class of all countable ordinals is a set and then take the union on of that set (which I already know is an ordinal). However I'm not sure how to justify the fact that it is indeed a set and not a proper class.

Secondly, I was wondering if a union of a proper class of ordinals is necessarily an ordinal in the same way a union of a set of ordinals is?

Looking forward for your answers. Thanks!

share|cite|improve this question
Related. – Cameron Buie Jan 26 '13 at 18:16
up vote 5 down vote accepted

First note that every countable ordinal is in bijection with $\omega$, therefore we can view it simply as a relation on $\omega$ which well-orders it by this order type.

Therefore we may define an equivalence relation on $\mathcal P(\omega\times\omega)$ in which every two well-orders are equivalent if and only if they are isomorphic (and everything which is not a well-ordering of $\omega$ is lumped into one big equivalence class).

Now you can show that this too makes a set, it is a definable collection of $\mathcal P(\mathcal P(\omega\times\omega))$, and since every well-ordered relation is isomorphic to a unique ordinal there is a definable map from that set onto the class of countable ordinals. Therefore countable ordinals make a set.

The union of this set is again a set, by the axiom of union, and it is an ordinal by showing it is transitive and $\in$ well-orders it.

As for the second question the answer is obviously no. If the union of a proper class of ordinals would be an ordinal then it would have to be an ordinal which is a set, which has a proper class of members. The union of any proper class of ordinals, if so, is the entire class of ordinals.

Also related: How do we know an $ \aleph_1 $ exists at all?

share|cite|improve this answer
Thanks Assaf, I had a feeling I could expect a swift reply from you. As always you've been most helpful! Could you elaborate please on how to formally show that a union of every proper class of ordinal is the entire class of ordinals? – Serpahimz Jan 12 '13 at 16:19
@Serpahimz: Every set of ordinals is a subset of an ordinal. So if we have a proper class of ordinals it has to be unbounded in the ordinals themselves, and so every ordinal is an element of a member of this class. – Asaf Karagila Jan 12 '13 at 16:23
@Serpahimz: Also, I should point out that "Assaf" and "Asaf" both write the same in Hebrew, but in the myriad of set theorists named like this (and I know at least three) I am "Asaf" with one "s". :-) – Asaf Karagila Jan 12 '13 at 16:43
Heh, I will remember it for next time :) It's also rather surprising there are so many set theorists that share the name, maybe you should consider forming a society :) – Serpahimz Jan 12 '13 at 20:39
There is one, but I already said too much... Watch out for any cluster of three or more people walking with cellphones at the same time. They know you know now. – Asaf Karagila Jan 12 '13 at 20:44

The union of any proper class of ordinals is the class of all ordinals, actually, which is not, itself, an ordinal. If this were not so, then there would be a least ordinal $\alpha$ not in the union. Then $\alpha$ would necessarily be an upper bound on the ordinals in the class, so the class is contained in the set $\alpha\cup\{\alpha\}$, and so not proper, at all.

For your first question, consider the set $$\mathcal{W}=\bigl\{\langle X,R\rangle\in\mathcal{P}(\omega)\times\mathcal{P}(\omega\times\omega):R\text{ well-orders }X\bigr\}.$$ Every countable well-order type is represented by at least one of the members of this set, so by replacement, the collection of countable well-order types (ordinals) is a set.

share|cite|improve this answer
Thanks a lot :) – Serpahimz Jan 12 '13 at 16:30
You're quite welcome. – Cameron Buie Jan 12 '13 at 17:26

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.