Omitting types: application to recursion theory
The starting point of this masterly paper is an old result of \textit{R. O. Gandy} et al. [Bull. Acad. Pol. Sci., Sér. Sci. Math. Astron. Phys. 8, 577--582 (1960); 9, 881--882 (1961; Zbl 0207.30102)] showing that the intersection of all \(\omega\)-models of a \(\Pi_{1}^{1}\)-theory in the language of analysis is included in the collection of all hyperarithmetic sets. A later proof of this result (in severely model theoretic language) is sketched in the reviewer's article [Theory of models, Proc. 1963 Int. Symp. Berkeley, 190--205 (1965; Zbl 0225.02040)] and given in full in Exercise 4 of Chapter 7 on p. 150 of \textit{G. Kreisel} and \textit{J. L. Krivine} [Elements of mathmatical logic (model theory). Amsterdam: North-Holland (1967; Zbl 0155.33801)]. This proof uses certain ``omitting type theorems, proved in Chapter 7 loc. cit. for the special purpose of establishing model theoretic ``invariants, one of the reviewer's pet projects; the general usefulness of omitting types in pure model theory was not stated and, evidently, not realized by Kreisel and Krivine. In contrast, the author takes a proper general view and proves the main Theorems 1 and 2 (pp. 82--84) showing how to omit a continuum of types by two models; the proof is uniform for several important classes of models, including in particular the classes of general and \(\omega\)-models. (The author prefers to use finite formulas and to restrict the class of models in the metalanguage instead of using infinite formulas such as \(\forall x W\left(x=s^{n} 0\right) \) in the object language; cf. top of p. 85.) There are 7 applications to (generalized) recursion theory including most of the gems of this subject, for example on countable admissible sets. Without questioning the general soundness of the paper the reviewer would have preferred a more detailed development of the author's convention (p. 82, 1.6) that ``a model of (a theory) \(T\) must interpret the name of an element of (a given set) ẞ(A\) as that element. For which theories \(T\) can we formulate axioms (concerning those elements of \(A)\) which are adequate to the author's convention? The reviewer needed the concept of structures rigidly contained in (all models of) \(T\) and the explicit definability of their elements of a conceptually clear answer. It would be interesting to see why the particular formal results established in this paper do not need closer analysis of the convention. Reviewer's remarks: On p. 85, bottom, the author describes his Theorem 7 as ``peripheral but needs it, at the very end of the paper, for the (interesting) Application 7 establishing the equivalence between \( \forall \)-hyperarithmeticity and abstract hyperarithmeticity. On pp. 88 and 89 the author quotes a preliminary note by \textit{D. Lacombe} instead of referring to the very elegant detailed later exposition [Logic Colloq. '69, Proc. Summer School Colloq. Math. Logic, Manchester 1969, 3--17 (1971; Zbl 0225.02027)].
This page was built for publication: Omitting types: application to recursion theory
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5655353)