I already asked this but too many typos caused confusion and the editing was confusing me so I re-wrote it here:

If $\theta$ is a sentence in the language of number theory, and we know the following, with the corner braces being the Gödel number: $$N \vdash \theta \leftrightarrow Thm_N(\overline{\ulcorner \neg \theta \urcorner})$$ Note that $N$ is the acioms of Robinsons arithmetic and we have defined $Thm_N(f) \equiv (\exists c)Deduction(c,\,f)$ which basically says $f$ is a Gödel-number of a formula such that $c$ is a Gödel-number for a deduction of $f$.

The question is: is $\theta$ a true sentence (in $\mathfrak{N}$)?

My immediate thought is no, but Im worried I'm interpreting this wrong. Is $Thm_N(\overline{\ulcorner \neg \theta \urcorner})$ saying that there exists a deduction of $\neg \theta$? If so does that mean that $\theta$ must be false? I'm having trouble understanding.

  • 2
    $\begingroup$ $\theta$ is not true, but $\lnot\theta$ is not provable. $\endgroup$ – Andrés E. Caicedo Nov 28 '19 at 22:19
  • $\begingroup$ @AndrésE.Caicedo Thank you for your response! I am trying to understand why this is the case, any suggestions on how to see it? $\endgroup$ – Ben A Nov 28 '19 at 22:26
  • $\begingroup$ $N$ is sound and $\mathfrak N\models N$. Argue from there. $\endgroup$ – Andrés E. Caicedo Nov 28 '19 at 22:28

Below I conflate sentences with their Godel numbers' numerals, and assume the soundness of $N$.

Pretty clearly $\theta$ can't be $N$-decidable, the key point being the very nice property of $\Sigma_1$-completeness - or rather, its particular instance that, for any sentence $\psi$, $N\vdash\psi$ implies $N\vdash Thm_N(\psi)$:

  • If $N$ proved $\theta$, by $\Sigma_1$-completeness $N$ would also prove $Thm_N(\theta)$. By the assumption on $\theta$ this means that $N$ would prove $Thm_N(\theta)\wedge Thm_N(\neg\theta)$ - that is, $N$ would prove that $N$ is inconsistent. This contradicts the soundness of $N$.

  • If $N$ proved $\neg\theta$, by $\Sigma_1$-completeness $N$ would also prove $Thm_N(\neg\theta)$, and so by assumption on $\theta$ $N$ would in fact prove $\theta$. This contradicts the consistency of $N$.

And this addresses the truth value of $\theta$ (in the standard model $\mathfrak{N}$) as well: the second bulletpoint above shows that $\theta$ is in fact false in $\mathfrak{N}$ since for all $\psi$ we have $\mathfrak{N}\models Thm_N(\psi)$ iff $N\vdash\psi$.


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.