0

How to prove ¬P∨¬Q → ¬(P∧Q)

using the following axioms?

a1. A→(B→A)

a2. (A→(B→C))→((A→B)→(A→C))

a3.(A∧B)→A

a3'.(A∧B)→B

a4.A→(B→(A∧B))

a5. A→A∨B

a5'. B→A∨B

a6.(A→C)→((B→C)→(A∨B→C))

a7. ¬¬A→A

and only role is Modus Ponens.

a tried with a6, with A=¬P, B=¬Q, and C=¬(P∧Q), but it doesn't work.

  • 1
  • @MauroALLEGRANZA https://staff.fnwi.uva.nl/b.vandenberg3/Onderwijs/Proof_Theory_2015/handout_3.pdf i use this axiom, and i can not use deductive theorem, i must do it in harder way. I must prove one of formula: De Morgan or Peirc'si law, nd i dont know neither –  Jan 24 '18 at 15:04
  • See an old post: with $\bot$ as primitive and the definition of $\lnot \varphi$ as $\varphi \to \bot$, we have some alternative ways in order to manage negation: EFQ, Peirce's law and Double Negation (your (Ax.7)). – Mauro ALLEGRANZA Jan 24 '18 at 15:26

1 Answers1

1

Quite correct. You have to prove Contraposition: $(A \to B ) \to (\lnot B \to \lnot A)$.

Then apply it to (Ax.3) and (Ax.3') to get, respectively:

$\lnot P \to \lnot (P \land Q)$ and $\lnot Q \to \lnot (P \land Q)$.


We need "auxiliary lemmas" like the first one: $\vdash A \to A$, needed to prove, with (Ax.1) and (Ax.2) the Dedution Theorem.

A second one is the "auxiliary rule" of Hypothetical Syllogism:

$A \to B, B \to C \vdash A \to C$.

Proof

1) $B \to C \vdash A \to (B \to C)$ --- from (Ax.1)

2) $A \to B, B \to C \vdash A \to (B \to C)$ --- from 1)

3) $A \to B, B \to C \vdash A \to C$ --- from (Ax.2) and 2) by Modus Ponens twice.

Using the Deduction Th we can conclude with:

$\vdash (A \to B) \to ((B \to C) \to (A \to C)).$

In the post: proving $(p \to q) \to ((q \to r) \to (p \to r))$, you can find a Deduction Theorem-free derivation of it.


Now for the final step.

As per you Lectures Notes, $\lnot \varphi$ is an abbreviation of $\varphi \to \bot$.

Thus, we have to use HS in the form: $\vdash (A \to B) \to ((B \to \bot) \to (A \to \bot))$ to get, without abbreviation:

$A \to B \vdash \lnot B \to \lnot A$.



Now for the the proof.

1) $\vdash (P \land Q) \to P$ --- (Ax.3)

2) $\vdash \lnot P \to \lnot (P \land Q)$ --- from 1)

3) $\vdash (P \land Q) \to Q$ --- (Ax.3')

4) $\vdash \lnot Q \to \lnot (P \land Q)$ --- from 3)

5) $\vdash (\lnot P → \lnot (P \land Q)) → ((\lnot Q → \lnot (P \land Q)) → ((\lnot P \lor \lnot Q) → \lnot (P \land Q)))$ --- (Ax.6)

6) $\vdash (\lnot P \lor \lnot Q) → \lnot (P \land Q)$ --- from 2), 4) and 5) by MP twice.

  • The literature I use does not have that axiom. its hard to proof something. And only i know to prove in this system is A->A. –  Jan 24 '18 at 13:26
  • @hjasel Mauro is absolutely correct though: you need an additional axiom to introduce negations ... without it, you can't complete this proof! The typical axiom used in a Hilbert system is $(\neg A \rightarrow \neg B) \rightarrow (B \rightarrow A)$, though you can also use $(A \rightarrow B) \rightarrow ((A \rightarrow \neg B) \rightarrow \neg A)$ – Bram28 Jan 24 '18 at 14:04
  • https://staff.fnwi.uva.nl/b.vandenberg3/Onderwijs/Proof_Theory_2015/handout_3.pdf @Bram28 –  Jan 24 '18 at 14:53
  • if it doesn't have that axiom it means i must prove it. But I dont know how. I cant use deductive theorem, so im stuck. –  Jan 24 '18 at 14:59
  • @hjasel Notice that the proof for Theorem 1.4 does not actually show those derivations, and indeed, there is no way that the rules given in that document can prove something like $A \rightarrow \neg \neg A$. So, at this point the 'Conversely, ...' claim is just a false claim: the system as stated is not complete and cannot derive all theorems that are derivable by classical deduction. So forget about trying to prove it: you just can't. Look, the author of the document just left off an axiom. If you ask him or her about it I'm sure they'll readily acknowledge that. – Bram28 Jan 24 '18 at 15:06
  • @Bram28 Axiom 7 as listed above is sufficient to make this a complete system, in particular Axioms 1, 2, 7 suffice. – B. Mehta Feb 06 '18 at 16:43
  • 1
    @B.Mehta - 1,2 and 7 will not suffice if $\land$ and $\lor$ are not defined but assumed as primitive (as per the question above). As said before, the OP overlooked a fundamental point: that the system has $\bot$ and it is used to define $\lnot$. – Mauro ALLEGRANZA Feb 06 '18 at 16:54
  • @MauroALLEGRANZA Fair, but we do not need any additional axioms (on top of the ones in the OP) to work with negation in this system. – B. Mehta Feb 06 '18 at 16:58