3

I'm reading the first chapter of "Elements of Set Theory" by Enderton. I found a solution of the book and saw it after doing exercises on my own. But I think there are 3 errors among 7 exercises, which is frequent.

Exercise 1.1. Which of the following become true when $\in$ is inserted in place of the blank? Which become true when $\subseteq$ is inserted?
(a) $\{\emptyset\}$__$\{\emptyset,\{\emptyset\}\}$
(b) ...
(c) ...
(d) ...
(e) ...
Solution. Choices (a) and (d) become true when $\in$ is inserted in place of the blank. Choices (b) and (c) become true when $\subseteq$ is inserted in place of the blank. Choice (e) is not true in either case.
My solution. Choice (a) becomes true when $\in$ or $\subseteq$ is inserted in place of the blank. ...

Assume $V_{a+1} = \mathscr P\,V_a$ and $V_0 = \emptyset$

Exercise 1.5 Define the rank of a set $c$ to be the least $a$ such that $c\subseteq V_a$. Compute the rank of $\{\{\emptyset \}\}$. Compute the rank of $\{\emptyset, \{\emptyset\}, \{\emptyset, \{\emptyset \}\}\}$.
Solution. Observe that $V_{a+1}=\mathscr P\,V_a$. Taking $V_0 = A = \emptyset$, it follows that $$V_1 = \mathscr P\,V_0 = \{\emptyset, \{\emptyset\}\}$$
Hence the rank of $\{\{\emptyset\}\}$ is 1. ...
My solution. $V_1 = \mathscr P\,V_0 = \{\emptyset\}$. $V_2 = \mathscr P\,V_1 = \{\emptyset, \{\emptyset\}\}$ so the rank of $\{\{\emptyset\}\}$ is 2. ...

Exercise 1.7 List all the members of $V_3$. List all the members of $V_4$.
Solution. Without listing them, $V_3$ has 16 members, and $V_4$ has 32 members.
My solution.$V_3$ has 4 elements and $V_4$ has 16 elements.

I think my corrections are all true. So I think a writer of this solution didn't take a lot care in it. But I have no answer to compare without this. When I start proof from axiom and exactly proved theorem, that answer won't be likely to be wrong. But how do I know I'm not mistaken when a process of proving gets long? Is it inevitable being wrong to some degree when studying mathematics at any level?

op ol
  • 465
  • 2
  • 7
  • 1
    The title of this post and the context are totally disconnected. – NicNic8 Jul 09 '21 at 03:59
  • @NicNic8 Thanks for pointing. I edited to fit. Is it good now? – op ol Jul 09 '21 at 04:03
  • 1
    I think the verb "misleading" isn't being used correctly here. Maybe you meant just "mistaken"? Like, how do you know you're not mistaken when you come up with your own solution to a problem? – littleO Jul 09 '21 at 04:06
  • @littleO Thanks, I edited again. – op ol Jul 09 '21 at 04:08
  • 1
    Your second answer is wrong. $\mathcal P(\emptyset)={\emptyset}.$ – Thomas Andrews Jul 09 '21 at 04:08
  • 1
    You are correct in the third case. $$|V_{n+1}|=2^{|V_n|}.$$ There is no way to get $2^5=32$ elements. $|V_0|=0,$ followed by $1,2,4,16,2^{16}.$ – Thomas Andrews Jul 09 '21 at 04:14
  • 2
    How do we know that we are not mistaken? One solution is to use a machine-aided verifier. If your solution can pass the verifier, then it is correct - that is, assuming that the machine works correctly. Unfortunately, this is not widely accepted in the mathematics society, and that's the reason why many research papers are wrong, either slightly or unsavably. – WhatsUp Jul 09 '21 at 04:15
  • 1
    If you can somehow find an expert to check your carefully written solutions, that will help a lot as you're developing mathematically. There is a stackexchange called "Code review" where people post their code for review. I've wondered if we should have a "proof review" stackexchange, where people post their proofs to be checked for correctness and also for stylistic improvements. – littleO Jul 09 '21 at 04:17
  • 5
    You can use the MSE tag "solution-verification"... – Martin Hansen Jul 09 '21 at 04:50
  • @WhatsUp: use a machine-aided verifier $\ldots$ this is not widely accepted in the mathematics society, and that's the reason why many research papers are wrong, either slightly or unsavably --- I seriously doubt many papers can presently be machine verified. For example, to take a random issue of one of the best known journals, I doubt any of these papers can presently be machine verified. – Dave L. Renfro Jul 09 '21 at 04:55
  • @DaveLRenfro The only reason that a proof cannot be machine verified is that the proof is wrong. Whether people are willing to do it is another question ... – WhatsUp Jul 09 '21 at 05:04
  • @WhatsUp: I used the word "presently" (twice) for a reason. And I suspect that by the time the papers I gave examples of can be machine verified, there will be no essential distinction between what we would now consider as "machine brain" and "organic brain" (although perhaps by then these will be distinguished from some other concept of "brain" that is beyond our present ability to even conceptualize). – Dave L. Renfro Jul 09 '21 at 13:03

0 Answers0