Let $\kappa$ be a singular cardinal with $\operatorname{cf}\kappa = \lambda > \omega$.

Let $C = \{ \alpha_{\zeta} \mid \zeta < \lambda \}$ be a strictly increasing continuous sequence of cardinals with limit $\kappa$. I want to show this is a closed unbounded set in $\kappa$.

Obviously it is unbounded, because $\sup C = \kappa$, but how do I show it is closed? Is it to do with continuity? Also I am used to dealing with clubs in regular cardinals, are there any differences for clubs in singular cardinals?

Thanks very much.


Yes, you are correct. The fact that $C$ is closed has everything to do with the sequence $\{ \alpha_\zeta \}_{\zeta < \lambda}$ being continuous.

The main idea is that since the sequence is strictly increasing, then for any limit ordinal $\gamma < \lambda$ we have that $\lim_{\zeta < \gamma} \alpha_\zeta = \sup ( \{ \alpha_\zeta : \zeta < \gamma \} )$. Therefore, if $\alpha < \kappa$ is a limit ordinal such that $\sup ( C \cap \alpha ) = \alpha$, let $\gamma = \min \{ \zeta < \lambda : \alpha_\zeta \geq \alpha \}$. We can show that $\gamma$ must be a limit ordinal, and that $C \cap \alpha = \{ \alpha_\zeta : \zeta < \gamma \}$. Using continuity we have that $$\alpha_\gamma = \lim_{\zeta < \gamma} \alpha_\zeta = \sup ( \{ \alpha_\zeta : \zeta < \gamma \} ) = \sup ( C \cap \alpha ) = \alpha,$$ and so $\alpha \in C$.

| cite | improve this answer | |
  • $\begingroup$ Thanks, but why is $\gamma$ a limit ordinal? I can't find a contradiction if I suppose it is a successor. Even it if it's not a limit ordinal, surely $\alpha_{\gamma} = \lim_{\zeta \to \gamma} \alpha_{\zeta}$ still holds? $\endgroup$ – Paul Slevin Feb 27 '12 at 15:12
  • $\begingroup$ No sorry, that is wrong. If $\gamma = \beta + 1$, then $\alpha_\beta = \lim_{\zeta \to \gamma}\alpha_{\zeta}$. I think the answer to my question follows after showing that $C \cap \alpha = \{ \alpha_{\zeta} \mid \zeta < \gamma \}$, and then we get a contradiction if we assume $\gamma$ is a successor. $\endgroup$ – Paul Slevin Feb 27 '12 at 15:22
  • $\begingroup$ @Paul: You're correct that the contradiction will follow once you show that $C \cap \alpha = \{ \alpha_\zeta : \zeta < \gamma \}$. This little observation stems from the definition of $\gamma$ and the fact that the original sequence is strictly increasing. $\endgroup$ – user642796 Feb 27 '12 at 16:24

Recall the definition of a closed and unbounded set in $\kappa$:

  1. For every $\alpha<\kappa$ there is some $\alpha_\zeta\in C$ such that $\alpha<\alpha_\zeta$ (unboundedness);
  2. For every limit ordinal $\delta<\lambda$ we have $\sup\{\alpha_\zeta\mid\zeta<\delta\}=\alpha_\delta$ (continuity).

Now if $\sup C=\kappa$ then it is clearly unbounded in $\kappa$, and continuity is exactly the demand that the sequence is closed. As you may recall, a continuous sequence is the range of a continuous function with respect to the order topology.

As for clubs of singular cardinals, those are indeed a bit different. In regular cardinals being a club means being a very large set in every possible way, while clubs in singular cardinals can be quite small in terms of cardinality.

| cite | improve this answer | |
  • $\begingroup$ I forget, Paul, did we discuss the continuity in the previous answer or on the chat? $\endgroup$ – Asaf Karagila Feb 24 '12 at 16:35
  • $\begingroup$ you defined a new sequence in your previous answer, which was continuous $\endgroup$ – Paul Slevin Feb 27 '12 at 14:12
  • $\begingroup$ Or if you mean in terms of continuity in continuous functions, we talked about it in chat a little bit $\endgroup$ – Paul Slevin Feb 27 '12 at 14:28

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.