Let $\lambda$ be an ordinal. A cardinal $\kappa$ is $\lambda$-strong iff there is some inner model $M$ and an elementary embedding $$ j \colon V \rightarrow M $$ s.t. $crit(j) = \kappa$ and $V_\lambda \subseteq M$.

I am given the following exercise:

Let $\alpha < \beta < \kappa$ be ordinals, where $\kappa$ is $(\kappa+\beta)$-strong. Then $$ \{ \mu < \kappa \mid \mu \text{ is } (\kappa+\alpha) \text{-strong} \} $$ has size $\kappa$.

I'd like to show that $$ M \models \kappa \text{ is } (j(\kappa)+\alpha) \text{-strong}, $$ but I'm not sure where to begin...

(It seems that I have to "stretch" a $(\kappa, \kappa+\alpha)$-extender $E \in V$ to a $(\kappa, j(\kappa)+\alpha)$-extender $\tilde E \in M$, but how? Is this even the right approach?)


Consider the case $\alpha =0 $ and recall that a cardinal $\mu$ is $\mu$-strong iff it is measurable.

Now, if $\kappa$ is $\kappa+2$-strong, we may fix a witnessing measure $U$ on $\kappa$. Then $U \in V_{\kappa+2} \subseteq M$ and some basic observations yield $$ M \models \kappa \text{ is measurable} $$ Thus for any fixed $\xi < \kappa$ we get $$ M \models \exists \mu \colon j(\xi) < \mu < j(\kappa) \wedge \mu \text{ is measurable} $$ (take $\mu = \kappa$) and elementarity yields that there is some measurable $\mu \in (\xi, \kappa)$. This proves that there are $\kappa$ many $\mu$-strong cardinals $\mu < \kappa$ and also suggest that we should be more careful and require something like $\alpha+n < \beta < \kappa$ for some $0 < n < \omega$. (Given the kind of proof I have for the edited question below, the exact value of $n$ is then easily calculated.)

I think that this exercise contains a typo and should read $$ \{ \mu < \kappa \mid \mu \text{ is } (\mu+\alpha) \text{-strong} \} $$ instead. I will ask my professor and in the case that this is what he was asking for, I will tidy up my post and provide an answer for future reference.

  • $\begingroup$ At which level $V_\alpha$ can we find an extender witnessing that there is an embedding $j:V\to M$ with critical point $\kappa$ and such that $V_{\kappa+\tau}\subset M$? $\endgroup$ Jul 4, 2015 at 19:23
  • $\begingroup$ @Andres Such an extender is an element of $V_{\kappa + \tau +1}$ and an argument similar to the one above yields that $\{ \mu < \kappa \mid \mu \text{ is } (\mu+\alpha) \text{-strong} \}$ has size $\kappa$. I still fail to see how I can improve this result to $\kappa$ many $(\kappa+\alpha)$-strong cardinals below $\kappa$. $\endgroup$ Jul 4, 2015 at 19:38
  • $\begingroup$ Actually, it may be $V_{\kappa+\tau+3}$... (using flat pairing functions it seems that we could get away with $+2$ at the end)? $\endgroup$ Jul 4, 2015 at 19:42
  • 1
    $\begingroup$ This is false, notice that when $\kappa$ is measurable, then it is $(\kappa+1)$-strong as we can get a transitive class $M$ and an elementary embedding $j:V\rightarrow M$ with $crit(j)=\kappa$ and $M^\kappa\subseteq M$. If what you're trying to prove was true, you would always get $\kappa$ many measurables below any measurable $\kappa$. $\endgroup$ Jul 5, 2015 at 14:29
  • $\begingroup$ @Camilo Yeah, I already figured that, but I somehow managed to delete the relevant part of a previous edit where I said that for this reason we should be more carefully and require $\alpha + 1 < \beta < \kappa$ (or maybe even a little more, but this will fall out of a careful observation of a proof). $\endgroup$ Jul 5, 2015 at 14:36

1 Answer 1


We can show that if $\kappa$ is $(\kappa+\alpha+1)$-strong, then the set of $(\kappa+\alpha)$-strong cardinals below $\kappa$ has size $\kappa$. First off, as is easy to verify, $M\vDash \kappa\text{ is }(\kappa+\alpha)\text{-strong}$. Let $k: M\prec N$ witness this. Let $e=k\circ j$. Then $e: V\prec N$ witnesses that $\kappa$ is $(\kappa+\alpha)$-strong with target $\lambda\gt j(\kappa)+\alpha$, and let $E$ be the $(\kappa,e(\kappa))$-extender defined by $X\in E_a$ if and only if $X\in Y$ and $a\in e(X)$, where $Y=\{X\in M|X\subseteq [\kappa]^{|a|}\}$. Then $j_E^M: M\prec \text{Ult}_E(M)$ witnesses, in $M$, the $(j(\kappa)+\alpha)$-strongness of $\kappa$, because $e(\kappa)\gt |V_{(j(\kappa)+\alpha)}|^M$ if and only if $k(\kappa)\gt |V_{\kappa+\alpha}|$. To see this, note that $k(\kappa)\gt |V_{\kappa+\alpha}|^M$, because $k(\kappa)\gt |V_{\kappa+\alpha}|^N$, and since $M\vDash k(\kappa)\times V_{\kappa+\alpha}\subseteq N$, $|V_{\kappa+\alpha}|^M=|V_{\kappa+\alpha}|^N$, and since $k(\kappa)\times V_{\kappa+\alpha}\subseteq M$, $|V_{\kappa+\alpha}|^M=|V_{\kappa+\alpha}|$. To establish $E$ is an extender, condition 1. is easy. Coherence follows from the fact that all the functions, if they exist in $M$, are $\Delta_0$. Normality is also simple. Finally, it is easy to establish that the direct limit is well-founded. The rest follows by standard reflection arguments.

All of this aside, I do think it was probably a typo. This process is rather involved (For an exercise), and the result that $\{\mu\lt\kappa|\mu\text{ is }(\mu+\alpha)\text{-strong}\}$ is much simpler and easier. Plus, I didn't even realize you wrote $\kappa+\alpha$ instead of $\mu+\alpha$, which indicates the author might have as well.

  • $\begingroup$ Why should $M$ see $j_E^M$? $E$ is defined using $j$. If your argument worked, there could be no least measurable cardinal, as pointed out by Camilo Arosemena-Serrato, and thus no measurable cardinal. $\endgroup$ Nov 8, 2021 at 21:15
  • $\begingroup$ Also, $k(\kappa) = j(\kappa)$ since $M \vDash \text{$j(\kappa)$ is inaccessible and greater than $\kappa+\alpha$}$, so $k(j(\kappa)) = j(\kappa)$ $\endgroup$ Nov 8, 2021 at 21:23
  • $\begingroup$ I meant $e(\kappa)=j(\kappa)$. $\endgroup$ Nov 8, 2021 at 21:36

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.