1
$\begingroup$

Is it possible to show

${\phi,\phi\to\psi}⊢\neg(\phi\to\neg\psi)$

only with modus ponens,deduction theorem and these three axiom?

A1: $\phi\to(\psi\to\phi)$

A2: $(\phi\to(\psi\to\pi))\to((\phi\to\psi)\to(\phi\to\pi))$

A3: $(\neg\phi\to\neg\psi)\to(\psi\to\phi)$

looks so easy but it's too hard for me to take even a step

$\neg\neg\phi\to\phi$ was quiet devastating too

$\endgroup$
  • $\begingroup$ Welcome to MSE. Please use MathJax. $\endgroup$ – José Carlos Santos May 1 '18 at 13:11
  • $\begingroup$ does MathJax viable for the title too? $\endgroup$ – tolmekia May 1 '18 at 13:56
  • 1
    $\begingroup$ Yes, it is viable. $\endgroup$ – José Carlos Santos May 1 '18 at 14:04
0
$\begingroup$

Hint : the derivation is long and boring ...

Form Axioms above, we have to derive some preliminary results :

A) $\vdash \lnot \lnot \phi \to \phi$

B) $\vdash \phi \to \lnot \lnot \phi$

C) $\vdash (\phi \to \psi) \to (\lnot \psi \to \lnot \phi)$. The derivation needs A), B), D) and Ax.3.

D) $\phi \to \psi, \psi \to \chi \vdash \phi \to \chi$. The derivation is straightforward: it needs modus ponens twice and a final application of the Deduction Th.

Now the derivation :

1) From modus ponens, by Deduction Th twice :

$\vdash \phi \to ((\phi \to \lnot \psi) \to \lnot \psi)$.

2) From C) above :

$\vdash ((\phi \to \lnot \psi) \to \lnot \psi) \to (\lnot \lnot \psi \to \lnot (\phi \to \lnot \psi))$.

3) From 1) and 2) by D) :

$\vdash \phi \to (\lnot \lnot \psi \to \lnot (\phi \to \lnot \psi))$.

4) $\phi$ --- premise

5) $\phi \to \psi$ --- premise

6) $\psi$ --- from 4) and 5) by mp

7) $\lnot \lnot \psi$ --- from 6) and B) by mp

8) $\lnot (\phi \to \lnot \psi)$ --- from 3), 4) and 7) by mp twice.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.