0
$\begingroup$

In the book "Numbers" by Ebbinghaus et. al, the Natural numbers are defined as:

The natural numbers form a set $\mathbb N$, containing a distinguished element $0$, called zero, together with a successor function $S:\mathbb N \to \mathbb N$, of $\mathbb N$ into itself, which satisfies the following axioms:

(S1) $S$ is injective,

(S2) $0 \notin S(\mathbb N)$,

(S3) If a subset $M \subset \mathbb N$ contains zero and is mapped into itself by $ S $, then $M = \mathbb N$.

Does this definition puts any strict requirement that only $0$ can be starting element? In other words, does (S2) excludes any possibility of another starting element, say $w_0 \notin S(\mathbb N)$? If it was strict requirement, I think the author would have explicitly mentioned something like :(S2) only $0 \notin \mathbb N$

Also, as (S1) says $S$ is just injective (and not surjective), so there can be some other elements outside of range of $S(\mathbb N)$ similar to $0$.

The reason I ask, as if there exists such element $w_0 \notin S(\mathbb N)$, we can have a chain of elements in sequence $ w_0 \to w_1 \to w_2 $ and so on with starting element $w_0$ independent of the one with the $0$ (zero). Then, there can be two different sets, $M_1$ and $M_2$ for S(3), where $M_1$ include $w_0$ and $M_2$ doesn't, but both still include $0$.

then, $M_1 = \mathbb N$

and also, $M_2 = \mathbb N$

giving two different sets for $\mathbb N$ with different sizes, which seems wrong to me.

Can you please let me know what I understood wrong here?

$\endgroup$
4
  • $\begingroup$ What do you mean by $w_0$, $M_1$, et cetera? Do you mean subscripts? $\endgroup$
    – user1134696
    Commented Aug 15, 2023 at 22:22
  • $\begingroup$ Also your $(S3)$ condition doesn't make sense. Please edit it. $\endgroup$
    – user1134696
    Commented Aug 15, 2023 at 22:28
  • $\begingroup$ @ShyamalSayak Thanks and apologies from my side. Fixed both now. $\endgroup$ Commented Aug 15, 2023 at 22:37
  • $\begingroup$ ($S3$) is still not fixed properly. $\endgroup$
    – user1134696
    Commented Aug 16, 2023 at 15:22

1 Answer 1

3
$\begingroup$

Suppose that $\mathbb N$ contains an element $w_0$ like you describe, that is, $w_0 \neq 0$ yet $w_0$ is not the successor of any element. Then consider the set $\mathbb N \setminus \{w_0\}$. Clearly this set contains 0. And it is also mapped into itself by $S$, since by assumption $w_0$ is not the successor of any element. It follows by (S3) that $\mathbb N \setminus \{w_0\} = \mathbb N$. But this is a contradiction. So no such element $w_0$ exists.

More informally, consider the set $$ \{0, S(0), S(S(0)), S(S(S(0))), \ldots\}. $$ This set satisfies the requirements of (S3), and therefore it is precisely this set that is $\mathbb N$, and there cannot be any other elements.

$\endgroup$
4
  • $\begingroup$ Thanks. Yes, this is the exact contradiction I was observing. So the sole purpose of (S3) is to just exclude any such element $w_0$ from $\mathbb N$? Also, I'm curious still, wouldn't then adding only to (S2) i.e. $S(\mathbb N)$ does not contain only $0$ (or rewording it as: $S(\mathbb N)$ contains all elements in $\mathbb N$, except 0) , would have avoided the need for S(3) and made definition simpler? Or does (S3) serves some other purpose too? $\endgroup$ Commented Aug 15, 2023 at 23:04
  • 1
    $\begingroup$ @dheerajSuthar, (S3) serves lots of purposes! For instance, it also rules out that there's a loop: the "regular" natural numbers plus elements $a, b, c$ with $S(a) = b, S(b) = c, S(c) = a$. And it rules out that, besides the "regular" natural numbers, there are infinite chains: $\ldots, x_{-2}, x_{-1}, x_0, x_1, x_2, \ldots$ where $S(x_i) = x_{i+1}$. $\endgroup$ Commented Aug 15, 2023 at 23:22
  • $\begingroup$ Thanks @mees-de-vries. The loop example was quite helpfu. To confirm, if I understood the infinite chain argument correctly, (S3) would enforce we have a starting point, here 0? In summary, the purpose of (S3) is to give us an intersection of all such possible valid sub sets of $\mathbb N$, thereby ensuring we get a minimum set fulfilling properties expected out of Natural numbers (start with 0, in sequence without loop, infinite in one direction etc.), and nothing else. $\endgroup$ Commented Aug 15, 2023 at 23:52
  • 1
    $\begingroup$ (S3) is exactly the principle of mathematical induction, so it is good for everything induction is good for, which is quite a lot of things. $\endgroup$ Commented Aug 16, 2023 at 6:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .