8
$\begingroup$

On page 14 of the introduction to Vol. II of Gödel's collected works:

By a slightly more difficult argument one can show that GCH continues to hold if $V=L[a]$ and $a\subseteq\aleph_1$.

Does anyone know this argument?

$\endgroup$
2
  • $\begingroup$ This is also an exercise from Jech's Set Theory, 13.26 (p.198) $\endgroup$
    – Asaf Karagila
    Mar 11, 2013 at 8:24
  • $\begingroup$ (It is a natural question. It appears as an exercise in a few books.) $\endgroup$ Mar 11, 2013 at 22:44

1 Answer 1

8
$\begingroup$

Suppose $A\subseteq\omega_1$ and $V=L[A]$.

Trying to imitate the usual argument (which I assume you have seen), we should start with an appropriate version of the condensation lemma. What we get is that if $X\prec L_\alpha[A]$ for some infinite limit $\alpha$, then $X\cong L_\beta[B]$ for some $B$ and, in fact, if $\pi:X\to L_\beta[B]$ is the transitive collapse, then $B$ is simply $\pi''(X\cap A)$.

Suppose in particular that $X$ is countable, $X\prec L_{\eta}[A]$, $\eta$ large enough. We use the standard fact that $X\cap\omega_1$ is an ordinal $\gamma$. (This is because for any ordinal $\tau\in X\cap\omega_1$, $X\models$"$\tau$ is countable", so there is in $X$ a surjection $f:\omega\to\tau$. But $\omega\subset X$, so $\tau=f''\omega\subset X\cap\omega_1$.)

We have then that $\pi''(X\cap A)=\pi''(A\cap\gamma)=A\cap\gamma$. Any "real" $r\subset\omega$ belongs to some such $X$, and so $\pi(r)=r$, which means that $r\in L_\beta[A\cap\gamma]$ for some countable ordinals $\beta$ and $\gamma$.

We conclude by noting that any such structure $L_\beta[A\cap\gamma]$ belongs to $L_{\omega_1}[A]$, which has size $\omega_1$, and therefore there are only $\omega_1$ many reals, that is, $\mathsf{CH}$ holds.

For $\kappa>\omega$, we can just implement the usual argument directly, assuming $X$ has size $\kappa$ which is at least $\omega_1$, and $\kappa\subset X$, and $A\in X$. Now the collapse of $X$ has the form $L_\beta[A]$ for some $\beta<\kappa^+$. Given $Y\subset\kappa$, we may assume $Y\in X$, which gives us $\pi(Y)=Y\in L_\beta[A]$, so each subset of $\kappa$ belongs to $L_{\kappa^+}[A]$, which has size $\kappa^+$. The conclusion is that $2^\kappa=\kappa^+$ as well.

$\endgroup$
1
  • $\begingroup$ No harassment, :-) But the argument is right. $\endgroup$ Jun 1, 2020 at 20:44

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.