On derivability

From MaRDI portal





The notion of derivability here considered may be intuitively explained as follows: If we have given a certain system of elements, a monadic predicate \(f\) (i.e. a class), and an \(n+1\) adic predicate \(g\), \(X\) is derivable on the basis \(f\) by means of \(g\) if and only if there exists a finite series of elements \(Y_1, Y_2, \ldots, Y_m\) such that \(Y_m\) is \(X\) and every \(Y_i\), either belongs to \(f\) or stands in the relation \(g\) to certain \(n\) of its predecessors. As special cases of this notion we have deducibility from the postulates of a formal system with regard to certain rules of procedure, the definition of \(<\) in terms of the successor relation, and -- if we extend the notion to ordered pairs -- the definition of a numerical function by recursion. It is well known that this derivability can be defined by a formula involving quantification with respect to class variables. In this paper the author shows that it can also be formulated in terms of the system of his paper on substitution [Bull. Am. Math. Soc. 42, 561--569 (1936; Zbl 0015.05003)]; this system involves only the functional calculus of the first order with identity, certain constant individuals, and an extra dyadic associative operation of ``concatenation. It follows that ``many of the purposes of non-individual variables and recursive definitions can be met by more elementary machinery. It is claimed -- and the claim is plausible, although Quine gives no proof and the reviewer has not checked it -- that many of \textit{K. Gödel}'s results [Monatsh. Math. Phys. 38, 173--198 (1931; Zbl 0002.00101)] can be obtained on the basis of the author's system.











This page was built for publication: On derivability

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5769238)