In his long and illuminating paper [1] Joe Barback defined and showed to be non-vacuous a class of infinite regressive isols he has termed “complete y torre” isols. These particular isols a enjoy a property that Barback has since labelled combinatoriality. In [2], he provides a list of properties characterizing the combinatoria isols. In Section 2 of our paper, we extend this list of characterizations to include the fact that an infinite regressive isol X is combinatorial if and only if (...) its associated Dekker semiring D satisfies all those Π2 sentences of the anguage LN for isol theory that are true in the set ω of natural numbers. of the various function and relation symbols of LN via the “lifting ” to D of their Σ1 definitions in ω coincide with their interpretations via isolic extension.) We also note in Section 2 that Π2-correctness, for semirings D, cannot be improved to Π 3-correctness, no matter how many additional properties we succeed in attaching to a combinatoria isol; there is a fixed equation image sentence that blocks such extension. In Section 3, we provide a proof of the existence of combinatorial isols that does not involve verification of the extremely strong properties that characterize Barback's CT isols. (shrink)

Let Λ denote the semiring of isols. We characterize existential completeness for Nerode subsemirings of Λ, by means of a purely isol-theoretic “Σ1 separation property”. Our characterization is purely isol-theoretic in that it is formulated entirely in terms of the extensions to Λ of the Σ1 subsets of the natural numbers. Advantage is taken of a special kind of isol first conjectured to exist by Ellentuck and first proven to exist by Barback . In addition, we strengthen the negative part (...) of [13] by showing that existential completeness is not secured, for a given Nerode semiring, by either a certain “functional closure” property for the extensions of partial recursive functions or the property of “pulling in” some portion of each partial recursive fiber; these latter results are perhaps a little surprising. (shrink)

We construct a recursive ultrapower F/U such that F/U is a tame 1-model in the sense of [6, §3] and FU is existentially incomplete in the models of II2 arithmetic. This enables us to answer in the negative a question about closure with respect to recursive fibers of certain special semirings Γ of isols termed tame models by Barback. Erik Ellentuck had conjuctured that all such semirings enjoy the closure property in question. Our result is that while many do, some (...) do not. (shrink)

Given a Δ1 ultrapower ℱ/[MATHEMATICAL SCRIPT CAPITAL U], let ℒU denote the set of all Π2-correct substructures of ℱ/[MATHEMATICAL SCRIPT CAPITAL U]; i.e., ℒU is the collection of all those subsets of |ℱ/[MATHEMATICAL SCRIPT CAPITAL U]| that are closed under computable functions. Defining in the obvious way the lattice ℒ) with domain ℒU, we obtain some preliminary results about lattice embeddings into – or realization as – an ℒ. The basis for these results, as far as we take the matter, (...) consists of the well-known class of minimal ℱ/[MATHEMATICAL SCRIPT CAPITAL U]'s, which function as atoms, and the class of minimalfree ℱ/[MATHEMATICAL SCRIPT CAPITAL U]'s, to whose nonemptiness a substantial section of the paper is devoted. It is shown that an infinite, convergent monotone sequence together with its limit point is embeddable in an ℒ, and that the initial segment lattices {0, 1,…, n } are not just embeddable in , but in fact realizable as, lattices ℒ. Finally, the diamond is embeddable; and if it is not realizable, then either the 1 - 3 - 1 lattice or the pentagon is at least embeddable. (shrink)