So I've read up a little on Mahlo cardinals and found a definition along the lines of of $\kappa$ is Mahlo iff it is inaccessible and regular cardinals below it form a stationary set.

Using this definition, I can see why $\aleph_0$ and $\aleph_1$ are not Mahlo, in that the first is not inaccessible because it is countable and that the second is a successor cardinal (and therefore not inaccessible).

But then I get to $\aleph_\omega$... Which is inaccessible, because it is uncountable and a weak limit cardinal. The stationary subset part is tripping me up though. The cardinals below it are $\{\aleph_0, \aleph_1, ..., \aleph_n, ... | n < \omega\}$.

So I guess, is $\aleph_\omega$ Mahlo? If yes, why? And if no, why not? Further, what is the "smallest" Mahlo cardinal?

  • $\begingroup$ $\aleph_\omega$ is not even inaccessible. Note that it is not a regular cardinal since it has cofinality $\aleph_0$. $\endgroup$ – Simply Beautiful Art Jul 14 '17 at 14:43
  • 1
    $\begingroup$ Why did this get downvoted? It's a basic issue (cofinality), but a basic issue that trips practically everyone up when first learning this stuff. Seems like a good question to me. $\endgroup$ – Noah Schweber Jul 14 '17 at 17:28
  • $\begingroup$ An inaccessible, whether weak or strong, must also be regular, which $\aleph_{\omega}$ certainly is not. $\endgroup$ – BrianO Jul 14 '17 at 19:16

A Mahlo cardinal has to be regular, which $\aleph_\omega$ is not. $\aleph_\omega =\bigcup\aleph_n$, so $\operatorname{cf}(\aleph_\omega)=\aleph_0$. Every strong inaccessible $\kappa$ satisfies $\kappa=\aleph_\kappa$, but even that is not enough as the lowest $\kappa$ satisfying that has $\operatorname{cf}(\kappa)=\aleph_0$. As we can't prove even that strong inaccessibles exist, we can't say where they are in the $\aleph$ heirarchy.

  • $\begingroup$ Your last sentence is not quite accurate. We can easily show inaccessible cardinals are $\aleph$ fixed points. $\endgroup$ – Asaf Karagila Jul 15 '17 at 7:38
  • $\begingroup$ @AsafKaragila assuming that strong inaccessibles exist, let's say we take the first aleph fixed point (where $\kappa = \aleph_\kappa$). It is inaccessible for sure, but also regular right? Why wouldn't this be Mahlo? Does $cf(\kappa)$ not equal $\kappa$ in this case? $\endgroup$ – user3684314 Jul 17 '17 at 3:27
  • 2
    $\begingroup$ @user3684314: No, the first fixed point is not inaccessible. The first fixed point has cofinality $\omega$. And inaccessible cardinal is not just any fixed point, if $\kappa$ is inaccessible, then there is a club of order type $\kappa$ of fixed points before it. $\endgroup$ – Asaf Karagila Jul 17 '17 at 4:57
  • $\begingroup$ It isn't even inaccessible. It is the limit of a countable sequence. $\endgroup$ – Ross Millikan Jul 17 '17 at 4:59
  • 3
    $\begingroup$ Yes, every inaccessible is a fixed point but not every fixed point is an inaccessible. An inaccessible has to be regular, with $cf(\kappa)=\kappa$ but the first fixed point has cofinality $\aleph_0$ $\endgroup$ – Ross Millikan Jul 17 '17 at 17:46

If $k$ is an uncountable cardinal and $cf(k)<k$ then k has a closed unbounded subset of singular members:

If $\omega=cf(k)<k$ let $f:\omega \to k$ \ $\omega $ be strictly increasing with $\cup_{n\in \omega}f(n)=k.$ Let $s=\{f(n)+1: n\in \omega\}.$ (Ordinal addition.) Then $s$ is club in $k$ and does not even contain any cardinals.

If $\omega<cf(k)<k$ let $t\subset k$ \ $cf(k)$ with $|t|=cf(k)$ and $\cup t=k.$ Let $s=\{x\in k: x=\cup (x\cap t)\}.$ I will leave it as an exercise that $s$ is club in $k$ and that no member of $s$ is regular.

It is a matter of convenience of definition that $\omega$ (i.e. $\aleph_0$) is excluded from being inaccessible and from being Mahlo.


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.