One definition of a finite set is that it can be injected into an initial segment of $ \Bbb N$, thus any $n$ in $\Bbb N$ is finite.

Accordingly, if it's legitmate to define every element in $^* \Bbb N$ as its own intial segment, then every element $^* \Bbb N$ is finite.

What about Dedekind-infinity? Can we define an Dedekind-infinite element in $^* \Bbb N$ by constructing an injection into its proper subset?

Added: I'm not sure whether the construction of $^* \Bbb N$ is compatible with $\bf ZFC$ in the same way as $\Bbb N$ is, which is probably not, I guess. But how?

  • $\begingroup$ how exactly are you considering an element in $^*\mathbb N$ to be a set? $\endgroup$ – Ittay Weiss May 17 '13 at 7:56
  • $\begingroup$ @IttayWeiss: Perhaps I got it wrong. Can't we define an element in $^{\ast} \Bbb N$ as its own initial segment? $\endgroup$ – Metta World Peace May 17 '13 at 8:00
  • $\begingroup$ I'm not sure how you intend to do that. $\endgroup$ – Ittay Weiss May 17 '13 at 8:03
  • $\begingroup$ You say that something is finite, and then ask whether it can be Dedekind infinite. But "finite" implies "Dedekind-finite" in ZF. $\endgroup$ – Trevor Wilson May 17 '13 at 15:52

There is a problem with this idea because we use $\in$ to construct $\omega$ which is a surrogate for $\Bbb N$ in $\sf ZFC$. This is fine because both $\in$ and $\Bbb N$ are well-founded.

But $^*\Bbb N$ is not well-founded. The non-standard integers must have a decreasing sequence. So we cannot use $\in$ to model them in a model of $\sf ZFC$. Of course, we can have non-standard models of $\sf ZFC$ whose $\omega$ is actually $^*\Bbb N$, but that's besides the point. Inside these models, the set which they think is $\omega$ is well-founded (internally!).

You could perhaps use $\subseteq$ to define the order $^*\Bbb N$ somehow and use $\omega$ for the initial segment of the standard integers. In that case it shouldn't be hard to note that the standard integers correspond to Dedekind-finiteness, and the non-standard integers are Dedekind-infinite sets.

Why? Because any non-standard integer would have to have infinitely many smaller integers, which correspond to infinitely many subsets, which means that it is a Dedekind-infinite set (remember we assume $\sf AC$).

To be fair, though, I don't see what good comes from that definition.


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.