I have an issue with a proof given "Théorie axiomatique des ensembles" by J.L Krivine.

Ordinal definition given of $\alpha$:

  1. $"x \in y"$ defines a strict, total, well-foundedness relation R in $\alpha$
  2. $x \in \alpha \Rightarrow x \subset \alpha$

We are trying to prove:

Given two ordinal sets $\alpha$ and $\beta$, one of these cases is true: $\alpha \in \beta$ or $\beta \in \alpha$ or $\alpha = \beta$

The proof starts by considering $\xi = \alpha \cap \beta$

The sketch of the proof is:

$\xi \: initial \: segment \: of \: \alpha \Rightarrow \xi = \alpha \: or \: \xi \in \alpha$

$\xi \: initial \: segment \: of \: \beta \Rightarrow \xi = \beta \: or \: \xi \in \beta$

4 cases are studied. The last case $\xi \in \alpha$ and $\xi \in \beta$ is excluded because it will lead to $\xi \in \xi$ (impossible for ordinals)

My question is: how do we know $\xi \ne \emptyset$ ? Why two random ordinal sets have a common element ? Or does the proof holds if $\xi = \emptyset$ ?

  • 1
    $\begingroup$ Where does the proof uses that $\xi\neq\varnothing$? $\endgroup$ – Asaf Karagila Jun 13 at 8:12
  • $\begingroup$ If $\xi = \emptyset$ then would it still be an initial segment of $\alpha$ ? $\endgroup$ – g.lahlou Jun 13 at 8:17
  • 1
    $\begingroup$ Is the empty set a subset of $\alpha$? $\endgroup$ – Asaf Karagila Jun 13 at 8:17
  • $\begingroup$ The empty set is a subset of $\alpha$ but it does not prove it is an element of $\alpha$, does it ? Every initial segment of $\alpha$ should be an element of $\alpha$, but is $\emptyset$ an element of every ordinal? $\endgroup$ – g.lahlou Jun 13 at 8:22
  • 1
    $\begingroup$ Is $\varnothing$ a transitive set? Is $\in$ a well-order of it? $\endgroup$ – Asaf Karagila Jun 13 at 8:23

Okay, I think I may have an answer to my question. Thanks @Asaf for pointing me to the right direction. According to a comment in this thread, if an ordinal is not the emptyset, it contains the emptyset. So, if $\xi$ is the emptyset, it is still an initial segment of $\alpha$ and $\beta$ considering none of them is an emptyset. If $\alpha$ or $\beta$ is the emptyset, the theorem we are trying to prove is trivial.

| cite | improve this answer | |

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.