Under suitable large-cardinal assumptions, in the inner model $L(\mathbb R)$ one can have $\omega_1$ and $\omega_2$ measurable (this follows from determinacy).

I was wondering if it is possible to reconcile measurability of $\omega_1$ (or other cardinals) with the model $L( \mathbb R^{\omega_1} )$ or objects alike?

Here $\mathbb R^{\omega_1}$ is the product space of $\omega_1$-many copies of the real line.

  • 3
    $\begingroup$ The notations $L[A]$ and $L(A)$ switched roles at some point. $L[\Bbb R]$ is a model generated by adding the predicate for the set of real numbers into your language. It is a model of ZFC, whereas you seem to mean $L(\Bbb R)$, which is the smallest model containing all the reals... $\endgroup$
    – Asaf Karagila
    Sep 10, 2022 at 12:47
  • 1
    $\begingroup$ In fact, $L[\mathbb{R}]$ is just $L$ (since "is a real" is a sufficiently low-complexity property). $\endgroup$ Sep 10, 2022 at 17:17

1 Answer 1


For each $\alpha<\omega_1$, choose a real $r_\alpha$ that codes the ordertype $\alpha$. This sequence $\langle r_\alpha : \alpha < \omega_1 \rangle$ then codes a sequence of surjections from $\omega$ to each countable ordinal. From this, you can then run the Ulam matrix construction and show that there is no countably complete ultrafilter on $\omega_1$.

Let $F$ be a countably complete filter on $\omega_1$ in $L(\mathbb R^{\omega_1})$, and let $\sigma_\alpha : \omega \to \alpha$ be the surjection coded by $r_\alpha$. For $n<\omega$ and $\beta<\omega_1$, let $S^\beta_n = \{ \alpha > \beta : \sigma_\alpha(n) = \beta \}$. Since $\bigcup_{n<\omega} S_n^\beta$ is the tail interval $(\beta,\omega_1)$, we can take $n_\beta$ to be the least $n$ such that $S^\beta_n$ is $F$-positive, by countable completeness. Let $T_n = \{ \beta : n_\beta = n \}$. Since $\omega_1$ is uncountable, there is $n$ such that $|T_n| > 1$. So for such $T_n$, if $\beta<\gamma$ are in $T_n$ and $\alpha \in S_n^\beta \cap S_n^\gamma$, then $\sigma_\alpha(n) = \beta = \gamma$, which is impossible. Thus $S_n^\beta$ and $S_n^\gamma$ are disjoint $F$-positive sets, and so $F$ is not an ultrafilter.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.