Arithmetic analogues of McAloon's unique Rosser sentences (Q1114685)

In what follows, \(T_0,T_1,\ldots,T_n,\ldots\) is a r.e. sequence of theories containing PRA, and (if the \(T_n\)'s form a chain) \(T\) is their union. MPr(z) is the formula \[ \exists\ x[ \Pr_{Tx}(z)\&\forall y\leq x\sim \Pr_{Ty}(neg(z))], \] MPr'(z) is defined similarly with \(\leq\) replaced by \(<\). The McAloon-Rosser sentences are sentences \[ \Phi\Leftrightarrow \sim M\Pr (\ulcorner \Phi \urcorner)\quad\text{and}\quad\Phi \Leftrightarrow \sim M\Pr {}'(\ulcorner \Phi \urcorner). \] It turns out that McAloon-Rosser sentences are unique over PRA but their uniqueness over \(T\) depends on the \(T_n\)'s. If the \(T_n\)'s grow sufficiently fast (like \(T_n= \text{PRA}+\Sigma_{n+1}\)-induction, for instance), then the sentences are unique over \(T\). On the other hand it is shown that for a (very short however) sequence (because consisting only of \(T_0 = \text{PRA}\) and \(T_1=T= \text{PRA}+Con_{\text{PRA}})\), McAloon-Rosser sentences are not unique. Other results include: another uniqueness result under different assumptions about \(T_n\)'s (not satisfied by \(\Sigma_{n+1}\)-induction) -- here also a difference between MPr and MPr' is displayed; results about Henkin sentences corresponding to MPr and MPr'; and applications of the results to constructions of end extensions of models of arithmetic.