Definition: An ordinal number $\alpha$ is called a limit ordinal number if there is no ordinal number immediately preceding $\alpha$.

Now my lecture notes say that $\omega, 2\omega, \omega^2, \omega^\omega$ are limit ordinal numbers whereas $\omega+3,2^\omega+5$ are not which is intuitively clear. But is there a characterization of limit ordinals that may work when proving the case with some rigor? Or is it "just look and identify" sort of a thing using the ordering of the family of ordinal numbers?

  • 1
    $\begingroup$ Generally, the sum, product, or power of two limit ordinals tends to be limit ordinals, which might be helpful to prove. Also, you probably meant $\omega2$ instead of $2\omega$ at the beginning of the second paragraph. $\endgroup$ – Simply Beautiful Art Jun 10 '17 at 17:00
  • $\begingroup$ What I meant was $\omega+\omega$. $\endgroup$ – Janitha357 Jun 10 '17 at 17:06
  • 2
    $\begingroup$ Then I believe that should be $\omega2$. See here for more details. $\endgroup$ – Simply Beautiful Art Jun 10 '17 at 17:07

You already have a characterization of limit ordinals. In any specific case you just have to verify that this characterization is fulfilled. Consider for example $\omega^2$:

$$ \omega^2 = \omega \cdot \omega = \sup \{ \omega \cdot n \mid n < \omega \}. $$

Hence, if $\alpha < \omega^2$, there is some $n < \omega$ such that $\alpha < \omega \cdot n$. But then $$ \alpha + 1 < \omega \cdot n + 1 < \omega \cdot (n+1) \le \omega^2. $$

Thus $\omega^2$ is a limit ordinal.

| cite | improve this answer | |

There can be several answers, depending on what you mean by "identify".

The simplest answer would be to look at the Cantor normal form of $\alpha$, and see if it has any finite ordinal there. If the answer is no, then $\alpha$ is a limit ordinal (or $0$, which may or may not be a limit ordinal depending on you convention) and otherwise it is a successor ordinal.

Another answer would be that $\alpha$ is a limit ordinal if and only if for every $\beta<\alpha$, $\beta+1<\alpha$ (with the same caveat about $0$ as before). Although it seems not to be exactly what you are looking for.

| cite | improve this answer | |
  • $\begingroup$ The last paragraph seems to be what I was looking for. $\endgroup$ – Janitha357 Jun 10 '17 at 17:03

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.