Shelah's Main Gap Theorem states that for all first-order, complete theories, T, in a countable language, we have that either $$I(T,\aleph_\alpha)=2^{\aleph_\alpha}$$ or $$I(T,\aleph_\alpha)<\beth_{\omega_1}(\alpha)$$ This result, while beautiful, leaves some questions. Can analogues of this result be extended to larger languages? Do other theorems exists for logics like $L_{\omega_1 \omega}$ or $L_{\omega_1 \omega}(Q)$? Do there exist any logics without the 'gap phenomenon'?

Any resources would be helpful. Thanks!

  • $\begingroup$ It seems that the problem for $L_{\omega_1,\omega}$ is still open, as well as even more modest problems (see the discussion on the end of page 43 here): shelah.logic.at/files/702.pdf $\endgroup$
    – Haim
    Commented Jul 13, 2014 at 3:27
  • $\begingroup$ @Haim: Ahh, your right. On page 25, he remarks that it is still a problem to prove the main gap for uncountable theories! $\endgroup$ Commented Jul 13, 2014 at 4:04

1 Answer 1


Extending Shelah's main gap to non first-order or even to first-order theories in an uncountable language is a major hard open problem.

I am familar with couple of extensions to non f.o. (both papers are available from my web page):

  1. Rami Grossberg and Bradd Hart. The classification theory of excellent classes, Journal of Symbolic Logic, 54, (1989) 1359--1381.
  2. Rami Grossberg and Olivier Lessmann. Abstract decomposition theorem and applications, Contemporary Mathematics, Vol 380, (2005), AMS, pp. 73--108.

For full $L_{\omega_1,\omega}$ this seems to be super hard as it may require proving Shelah's categoricity conjecture for $L_{\omega_1,\omega}$ which despite of many hundreds of pages of approximation is still open.


Your Answer

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

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