Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

Let $\kappa$ be a regular uncountable cardinal, and let $\lambda < \kappa$ be a regular cardinal. Then the set $$E_{\lambda}^{\kappa} := \{ \alpha < \kappa \mid \operatorname{cf}\alpha = \lambda \}$$ is stationary in $\kappa$. It is a result that every stationary subset of $E_{\lambda}^{\kappa}$ is the disjoint union of $\kappa$ stationary sets. Now define $$S = \{\alpha < \kappa \mid \operatorname{cf}\alpha<\alpha\}.$$

In Chapter 8, Jech makes the following claim: "Every stationary subset $W \subseteq S$ is the disjoint union of $\kappa$ stationary sets: By Fodor’s Theorem, there exists some $λ < κ$ such that $W ∩ E_{\lambda}^{\kappa}$ is stationary."

I'm a bit confused by this. Here is what I managed to extract from this problem: Is it true that S is always stationary? In other words, is it always possible for it to contain stationary subsets $W$?

If so, then consider the function $\operatorname{cf}:W \to \kappa$ which gives the cofinality of $\alpha \in S$. By definition this is regressive (i.e. $\operatorname{cf}\alpha < \alpha$) so by Fodor's Theorem, there exists a $\lambda < \kappa$ and a stationary set $T \subseteq W$ such that $\operatorname{cf}\alpha = \lambda$ on $T$. If $\lambda$ is a regular cardinal then I can see that $T \subseteq E_{\lambda}^{\kappa} \cap W$ and so $W\cap E_{\lambda}^{\kappa}$ is stationary as a superset, and we're done. But what if $\lambda$ is not a regular cardinal, and then $E_{\lambda}^{\kappa}$ doesn't make sense?

Any help with this would be appreciated.

share|cite|improve this question
$\bf{NOTE}$: I am not allowed to use Solovay's theorem where I can just partition any stationary set, since I use this to prove it. – Paul Slevin Mar 3 '12 at 3:35
You have $\operatorname{cf}\alpha=\lambda$ for $\alpha\in T$. Clearly $\lambda$ is regular, since it’s a cofinality. – Brian M. Scott Mar 3 '12 at 3:46
Thanks, but what if $\lambda$ is not a limit ordinal? – Paul Slevin Mar 3 '12 at 3:55
It has to be: if it’s not, $T$ consists entirely of successor ordinals and must therefore be non-stationary. – Brian M. Scott Mar 3 '12 at 3:57
up vote 1 down vote accepted

Suppose that $W\subseteq S$ is stationary. Then as you say, $\operatorname{cf}$ is a pressing-down function on $W$, so it is constant on a stationary subset of $W$, i.e., there are a stationary $T\subseteq W$ and a $\lambda\in\kappa$ such that $\operatorname{cf}\alpha=\lambda$ for each $\alpha\in T$. Cofinalities are always regular, so $\lambda$ is regular, $E_\kappa^\lambda$ is defined, and clearly $T\subseteq W\cap E_\kappa^\lambda$. Since $T$ is stationary, the possibly larger set $W\cap E_\kappa^\lambda$ is also stationary, and all’s well.

Added: Note that $\lambda$ cannot be $1$: that would mean that every $\alpha\in T$ was a successor ordinal, and $T$ would be disjoint from the cub of limit ordinals and hence not stationary after all.

share|cite|improve this answer
What happens if the $\lambda$ that Fodor's theorem gives is not a limit? – Paul Slevin Mar 3 '12 at 3:58
@Paul: It has to be a limit. The only non-limit cofinality is $1$, and since $T$ is stationary, it must contain limit ordinals, and therefore $\lambda$ must be at least $\omega$. – Brian M. Scott Mar 3 '12 at 4:00
Of course, thank you. – Paul Slevin Mar 3 '12 at 12:22

Let $\Lambda$ denote the set of limit ordinals less than $\kappa$. Since $\Lambda$ is a club $W\cap \Lambda$ is stlationary. Now your argument holds applied to this stationary.

share|cite|improve this answer

@Paul Slevin: I don't follow your argument. First of all, if $\operatorname{cf}(\alpha)=\lambda$, then $\lambda$ must be regular. The set of all $\alpha\in W$ with $\operatorname{cf}(\alpha)=\lambda$ is just $W\cap E_\lambda^\kappa$. By Fodor's lemma, this is stationary for some $\lambda<\kappa$. Also, as was pointed out before, the set of successor ordinals is not stationary and hence the $\lambda$ that you get from Fodor's lemma is not $1$. So, we know that for some regular $\lambda>1$, $W\cap E_\lambda^\kappa$ is stationary. But we don't know whether $E_\lambda^\kappa\subset W$, which you seem to claim.

The following hasn't been addressed yet, it seems: whether the set $S$ contains a stationary set.

Since $\kappa$ is uncountable, there is a regular (infinite) cardinal $\lambda<\kappa$. Now the set of all $\alpha<\kappa$ with $\operatorname{cf}(\alpha)=\lambda$ and $\alpha>\lambda$ is a stationary subset of $S$. Why? Because it is the intersection of a stationary set, namely $E_\lambda^\kappa$, and a club, namely $\kappa\setminus(\lambda+1)$.

share|cite|improve this answer
I made a small mistake in my proof - I was writing it at 4am. I should have said that by Fodor's theorem we find a $T \subseteq W$, which is stationary, and $T \subseteq E_{\lambda}^{\kappa} \cap W$. – Paul Slevin Mar 3 '12 at 12:18

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.