0

It's been a while since I studied math, but I am looking back at a mathematical logic book, and found this problem I am currently unable to solve. I assume I am missing something obvious.

It's problem 1.47 (d) from An Introduction to Mathematical Logic by Mendelson.

Prove $\vdash_L (\neg C \Rightarrow \neg B) \Rightarrow (B \Rightarrow C)$

He gives the following axioms for the theory

(A1) $B \Rightarrow (C \Rightarrow B)$
(A2) $((B \Rightarrow (C \Rightarrow D)) \Rightarrow ((B \Rightarrow C) \Rightarrow (B \Rightarrow D)))$
(A3) $(((\neg C) \Rightarrow (\neg B)) \Rightarrow (((\neg C) \Rightarrow B) \Rightarrow C))$

It seems likely, that I want to start off with A3 since it has the premise, and applying (A2) gives me something looking pretty promising.

  1. $(\neg C \Rightarrow \neg B) \Rightarrow ((\neg C \Rightarrow B) \Rightarrow C)$ (A3)
  2. $[(\neg C \Rightarrow \neg B) \Rightarrow ((\neg C \Rightarrow B) \Rightarrow C)] \Rightarrow [((\neg C \Rightarrow \neg B) \Rightarrow (\neg C \Rightarrow B)) \Rightarrow ((\neg C \Rightarrow \neg B) \Rightarrow C))]$ (A2)
  3. $((\neg C \Rightarrow \neg B) \Rightarrow (\neg C \Rightarrow B)) \Rightarrow ((\neg C \Rightarrow \neg B) \Rightarrow C))$ (Modus Ponens, 1, 2)
  4. $C \Rightarrow (B \Rightarrow C)$ (A1)

The two things I see from here:

One is that if I could get something like $((\neg C \Rightarrow B) \Rightarrow C) \Rightarrow (B \Rightarrow C)$, then I could use transitivity of $\Rightarrow$ (proven in 1.47 (b)) along with statement (1). This statement is a tautology, so in theory should be provable, but I haven't been able to figure it out.

Two is that with (4) and the consequent of (3), I could also use transitivity to get my result. Although I am not sure what hack I need to employ to get there since that consequent is obviously not a tautology.

In general I know that $\neg C \Rightarrow \neg B \Leftrightarrow C \vee \neg B \Leftrightarrow \neg B \vee C \Leftrightarrow B \Rightarrow C$, which should imply the above. However, it seems the exercise here is to only use the axioms, statements preceding the $\vdash_L$ of which there are none, and Modus Ponus.

Anyone have any hints on this?

=== SOLUTION ===

I figured out the solution after someone showed me the problem was solved with the deduction theorem, and playing around with the proof of that theorem a bit.

Since I don't have a the deduction theorem, I use something basically just as good, which is the ability to "Export" a formula. That is, if we were to build a statement, $S_i$ with rule $\text{(Rule)}$, we could immediately apply $S_i \Rightarrow (R \Rightarrow S_i) \ \text{(A1)}$ for any formula $R$, and $\text{(MP)}$ to get $R \Rightarrow S_i$. We do this in one line, and mark it as $\text{(Rule,Export)}$.

Now for the proof.

Let $X := \neg C \Rightarrow \neg B$, and $Y := B \Rightarrow C$. We want to show $X \Rightarrow Y$.

Also let $Z := \neg C \Rightarrow B$, $A_1 := B \Rightarrow (Z \Rightarrow C)$, and $A_2 := ((B \Rightarrow Z) \Rightarrow Y)$.

\begin{align*} S_{1} &:= X \Rightarrow ((Z \Rightarrow C) \Rightarrow A_1) &\text{(A1, Export)} \\ S_{2} &:= S_{1} \Rightarrow ((X \Rightarrow (Z \Rightarrow C)) \Rightarrow (X \Rightarrow A_1)) &\text{(A2)} \\ S_{3} &:= (X \Rightarrow (Z \Rightarrow C)) \Rightarrow (X \Rightarrow A_1) &\text{(MP, $S_{1}$, $S_{2}$)} \\ S_{4} &:= X \Rightarrow (Z \Rightarrow C) &\quad\text{(A3)} \\ S_{5} &:= (X \Rightarrow (A_1 \Rightarrow A_2))\Rightarrow ((X \Rightarrow A_1) \Rightarrow (X \Rightarrow A_2)) &\text{(A2)} \\ S_{6} &:= X \Rightarrow (A_1 \Rightarrow A_2) &\quad\text{(A2, Export)} \\ S_{7} &:= (X \Rightarrow A_1) \Rightarrow (X \Rightarrow A_2) &\text{(MP, $S_{6}$, $S_{5}$)} \\ S_{8} &:= X \Rightarrow A_1 &\text{(MP, $S_{4}$, $S_{3}$)} \\ S_{9} &:= (X \Rightarrow A_2) \Rightarrow ((X \Rightarrow (B \Rightarrow Z)) \Rightarrow (X \Rightarrow Y)) &\text{(A2)} \\ S_{10} &:= X \Rightarrow A_2 &\text{(MP, $S_{8}$, $S_{7}$)} \\ S_{11} &:= (X \Rightarrow (B \Rightarrow Z)) \Rightarrow (X \Rightarrow Y) &\text{(MP, $S_{10}$, $S_{9}$)} \\ S_{12} &:= X \Rightarrow (B \Rightarrow Z) &\text{(A1,Export)} \\ S_{13} &:= X \Rightarrow Y &\text{(MP, $S_{12}$, $S_{11}$)} \\ \end{align*}

zrbecker
  • 4,136
  • Are you allowed to use the deduction theorem? – Taroccoesbrocco Feb 17 '25 at 14:27
  • I think I figured out how to do it without the deduction theorem, by inspecting page 32, and comparing to the proof of the deduction theorem. Writing it up now. – zrbecker Feb 17 '25 at 14:31
  • It seems to me that such a system of axioms is of interest only for historical purposes, not unlike Aristotle's Square of Opposition in this regard. Modern forms of natural deduction are much more intuitive and, well... "natural." IMHO, it would be no great loss if you skipped this section of the text. You might come back to it you are interested the history of logic in addition to its modern applications. Note that each of thes above axioms are theorems of natural deduction. – Dan Christensen Feb 17 '25 at 15:30
  • @DanChristensen I assume I am going about what I want to do in a very slow manner, but chatted with some friends about foundational mathematics and felt uncomfortable discussing how logic is built up without logic to start with. I know we talk about a "meta-theory", but it's been a long time since I touched the subject, so decided to go see how it's explained at the undergrad level.

    Mostly doing the problems for the big fun of symbol manipulation. :)

    – zrbecker Feb 17 '25 at 17:47
  • Just know that logic is often much simpler that textbooks on "mathematical logic" make it seem. – Dan Christensen Feb 17 '25 at 20:55

0 Answers0