Let $T$ be a countable first-order theory, and assume that $T$ has exactly one atomic model up to isomorphism in every uncountable cardinality. (By "atomic" I mean a model which omits the non-principal types).

Now let $\mathfrak{M}$ be a countable transitive model of set theory, and assume that $T$ is also (countable) in $\mathfrak{M}$.

Is the above property preserved in $\mathfrak{M}$? i.e. does it hold in $\mathfrak{M}$ that there exists only one atomic model of $T$ in every uncountable cardinality in $\mathfrak{M}$?


This is really a comment, but I need a bit more space.

If $\phi$ is a sentence of ${\cal L}_{\omega_1,\omega}$ that is $\aleph_0$-categorial, there is a complete first order theory $T$ in an expanded vocabulary such that the models of $\phi$ are exactly the reducts atomic models of $\phi$. The expansion is done in such a way that two structures will be isomorphic in the original language if and only if they are isomorphic in the expanded language.

So your question is really the same as: For $\phi$ a sentence of ${\cal}L_{\omega_1,\omega}$ is ``$\phi$ is $\kappa$-categorical for all infinite $\kappa$" absolute?

This is, as far as I know, still an open question. It is also open if ``$\phi$ is $\aleph_1$-categorical" is absolute.

John Baldwin in his paper ``Amalgamation, Absoluteness, and Categoricity" addresses some issues around this. Here is a link http://homepages.math.uic.edu/~jbaldwin/pub/singsep2010rev.pdf

  • 7
    $\begingroup$ I think this is absolutely valuable as an answer, not a mere comment. $\endgroup$ – Todd Trimble Dec 30 '13 at 19:00
  • 1
    $\begingroup$ Shelah showed that for an atomic class $K$ under the assumption $\forall n\in\omega,\; 2^{\aleph_n}<2^{\aleph_{n+1}}$ (called very weak GCH (VWGCH)), "categoricity in all uncountable powers" is equivalent to "excellence plus categoricity in $\aleph_1$". Excellence in an amalgamation property of countable models. Therefore, it is absolute by Shoefield absoluteness. If $\aleph_1$-categoricity is absolute (which is still open), then your question has a positive answer. $\endgroup$ – Ioannis Souldatos Oct 17 '16 at 17:15

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.