0
$\begingroup$

Let $N$ be a set for which $$\forall a\left(a\in N\iff a=\{\}\lor\exists b\left(b\in N\land a=\left\{b\right\}\right)\right)$$ Prove that $$\forall P\left(\left\{\right\}\in P\land \forall a\left(a\in N\implies\left(a\in P\implies\{a\}\in P\right)\right)\implies\forall a\left(a\in N\implies a\in P\right)\right)$$

I'm having troubles with this problem. The first thing I tried to do is to expand the definition of set $N$. It seems that there is exactly one set $N$ for which the first statement is true. The second statement should be true for all $P$ and $a$ (we need to prove it).

My first attempt was to replace $a\in N$ in the second statement with $a=\{\}\lor\exists \left(b\in N\land a=\left\{b\right\}\right)$. The part $\{\}\in P\implies\{\}\in N$ is true, according to the definition of $N$. Now, I'm not sure how to deal with $\forall a\left(a\in N\implies\left(a\in P\implies\{a\}\in P\right)\right)$. Becuase there is an implication after this statement, the first idea which came to my mind is to try contradiction approach. Suppose that $\forall a\left(a\in N\implies\left(a\in P\implies\{a\}\in P\right)\right)$ is not true. So, I need to prove that then exists some $a$ from $N$ which is not in $P$. But, how to do that? I ran out of ideas.

One thing I noticed is that both in the definition of $N$ and the second statement, which we need to prove, appear $\{b\}$ (and in the second $\{a\}$). I suppose the trick is to somehow apply this part of the definition of set $N$ (or a set $N$, if there may be multiple sets $N$) such that these two parts cancel out using some logical implications.

I'm looking for an axiom-level proof. I mean, applying logical and ZFC axioms to this statement until we obtain an equivalence where on the left side is the statement we need to prove and on the right side is something which is true by definition.

Simply, what would be the easiest way to prove this? Thank you in advance if you decide to help :)

$\endgroup$
  • 2
    $\begingroup$ You're going to need to use the axiom of regularity; the result is not true without it. $\endgroup$ – Eric Wofsey Nov 16 '17 at 20:01
  • $\begingroup$ @EricWofsey. The result is not true without it. - did I anywhere mention that the axiom of regularity is rejected? No, I'm just wondering how exactly to apply that axiom. $\endgroup$ – user503399 Nov 16 '17 at 20:03
  • 4
    $\begingroup$ Well, my point is that most arguments in set theory don't use the axiom of regularity, and if you're just applying your general intuition about sets, you are likely to be overlooking the extra power the axiom of regularity gives you in this context. $\endgroup$ – Eric Wofsey Nov 16 '17 at 20:09
  • 1
    $\begingroup$ @Eric Wofsey . There is a missing bracket in the first display-line of the Q. Is this the problem? $\endgroup$ – DanielWainfleet Nov 27 '17 at 1:49
  • 1
    $\begingroup$ @DanielWainfleet: Ah, yes, there should be. I now realize what you are saying is correct if you parenthesize it the other way (but it did not even occur to me that this might be intended, since it is a very unnatural way to state the condition in that case!). $\endgroup$ – Eric Wofsey Nov 27 '17 at 1:54
5
$\begingroup$

The key idea you're missing is to use the axiom of regularity in order to make a stronger assumption about $a$. Specifically, suppose there does exist $a\in N$ such that $a\not\in P$. Let $S=\{a:a\in N\wedge a\not\in P\}$ (this set exists by Separation from $N$). By assumption, $S$ is nonempty, so by the axiom of regularity there exists some $a\in S$ such that for all $b\in S$, $b\not\in a$. Now try applying your line of thinking to this particular special $a$.

(Incidentally, I would very strongly advise against trying to think about things like this in purely formal "axiom"-level terms. First come up with an informal argument based on your intuitive understanding of how sets and logic work and what the axioms mean, and then go back and try to make it fully formal once you have the informal argument.)

$\endgroup$
  • $\begingroup$ Seems valid. As soon as I completely understand your solution, I'll accept your answer. BTW, as a side question, is there exactly one set $N$ with such properties? $\endgroup$ – user503399 Nov 16 '17 at 20:58
  • 2
    $\begingroup$ @Math_974: Yes. $N$ has as elements all iterated singletons of $\emptyset$, and if there was anything else besides in it, that would violate regularity. $\endgroup$ – tomasz Nov 16 '17 at 21:23
  • $\begingroup$ I still didn't understand your answer. But in meanwhile, please tell me what do you think about this? $\endgroup$ – user503399 Nov 23 '17 at 20:37
  • $\begingroup$ Sorry for bothering you again, but can you please take another look at my previous comment and tell me is the thing DanielWainfleet noticed correct and your answer wrong, or is your answer correct and his comment wrong? Can you just spend a minute to explain what is happening here? Thanks. $\endgroup$ – user503399 Nov 26 '17 at 6:11
  • $\begingroup$ @Math_974: His comments are wrong, but in any case are irrelevant to my answer. $\endgroup$ – Eric Wofsey Nov 26 '17 at 6:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy