This is an attempt to give a survey of recent results concerning trees. The article is an extended version of our talk in Oberwolfach (Schwarzwald) last May; the forests surrounding the Forschungsinstitut turned out to be a good inspiration.

A tree is a partially ordered set T = (T, ≤) such that for every x ∈ T, the set = {y ∈ T: y < x} is well-ordered. The order type of is called the order of x, o(x), and the length of T is sup {o(x) + 1: x ∈ T}; an α-tree (where α is an ordinal) is a tree of length α. The αth level of T is the set Uα of all elements of T whose order is α. T∣α is the union of all Uβ, β < α; its length is α. A tree (T2, ≤2) is called an extension of (T1, ≤1) if ≤1 = ≤2 ∩ (T1 × T1); T2 is on end-extension of (T1 if T1 = T2∣α for some α. A maximal linearly ordered subset of a tree T is called a branch of T; an α-branch is a branch of length α.

In classical propositional logic it is well known that {7, ⊃ } is a functionally complete set with respect to a two-valued truth function modeling. I.e. all definable logical connectives are definable from 7 and ⊃. Other modelings of classical type propositional logics may have different functionally complete sets; for example, multivalued truth function modelings.

This paper examines the question of a functionally complete set of logical connectives for intuitionistic propositional logic with respect to S. Kripke's modeling for intuitionistic logic.

Since the publication in 1967 of the two papers [1] and [2] by Manuel Blum, the techniques and results of “pure” recursion theory, particularly the recursion theorem and priority methods, have come to play an increasingly important role in studies of computational complexity. This paper gives a typical application of the recursion theorem with a fairly intricate diagonalization to answer a question raised by Blum in [3]. Roughly, we prove the existence of functions which have the property that if we are given any program for computing the function and want to pass to a program which computes the function much more efficiently, then we can only do so at the expense of obtaining a much larger program: the function which describes the necessary increase in the size of the more efficient program must grow more rapidly than any recursive function.

The notion of “support” was introduced by Mostowski in [4] in order to prove that a certain universe satisfied the ordering principle but not the axiom of choice. The notion was refined in [3] and in [1] it was shown to be satisfied in a certain Cohen model of full ZF set theory. This paper is an axiomatic study of universes whose undefined relations are ∈ and a “support structure”, T.

In §2 the general theory is introduced and the universes of [4] and [1] are characterized. §3 examines a more complicated universe which will be used in [5] to show that in many cases a consistency in full ZF set theory may be proven directly by the methods of [4]. The embedding theorems of §4 are crucial to this application.

In spite of the work of Gödel and Cohen, which showed the undecidability of the Generalized Continuum Hypothesis (GCH) from the axioms of set theory, the problem still remains to decide GCH on the basis of new axioms. It is almost 100 years since Cantor first conjectured the Continuum Hypothesis, yet we seem to be no closer to determining its truth (or falsity). Nevertheless, it is a sound methodological principle that given any undecidable set-theoretical statement, we should search for “other (hitherto unknown) axioms of set theory which a more profound understanding of the concepts underlying logic and mathematics would enable us to recognize as implied by these concepts” (see Gödel [7, p. 265]).

Let ε stand for the set of nonnegative integers (numbers), V for the class of all subcollections of ε (sets), Λ for the set of isols, and for the set of mappings from a subset of ε into ε (functions). I f is a function we write δf and ρf for its domain and range respectively. The relation of inclusion is denoted by ⊂ and that of proper inclusion by ⊊. The sets α and β are recursively equivalent [written: α ≃ β], if δf = α and ρf = β for some function f with a one-to-one partial recursive extension f. We denote the recursive equivalence type of α, {σ ∈ V ∣ ≃ α}, by Req(α). Also R stands for Req(ε), while ΛR denotes the collection of all regressive isols. The reader is assumed to be familiar with the contents of [1] and [6].

A pair of sets (A0, A1) forms a minimal pair if A0 and A1 are nonrecursive, and if whenever a set B is recursive both in A0 and in A1 then B is recursive. C. E. M. Yates [8] and independently A. H. Lachlan [4] proved the existence of a minima] pair of recursively enumerable (r.e.) sets thereby establishing a conjecture of G. E. Sacks [6]. We simplify Lachlan's construction, and then generalize this result by constructing two disjoint pairs of r.e. sets (A0, B0) and (A1B1) such that if C0 separates (A0, A1 and C1 separates (B0, B1), then C0 and C1 form a minimal pair. (We say that C separates (A0, A1) if A0 ⊂ C and C ∩ = ∅.) The question arose in our study of (Turing) degrees of members of certain classes, where we proved the weaker result [2, Theorem 4.1] that the above pairs may be chosen so that C0 and C2 are merely Turing incomparable. (There we used a variation of the weaker result to improve a result of Scott and Tennenbaum that no complete extension of Peano arithmetic has minimal degree.)

The purpose of this paper is twofold. In §1 and §2 which are largely expository we develop the known theory of ℵ1-categoricity in terms of strongly minimal sets. In §3 we settle affirmatively Vaught's conjecture that a complete ℵ1-categorical theory has either just one or just ℵ0 countable models, and in §4 we present an example which serves to illustrate the ideas of §3.

As far as we know the only work published on strongly minimal sets is that of Marsh [3]. The present exposition goes beyond [3] in showing that any ℵ-categorical theory has a principal extension in which some formula is strongly minimal.

In Chapter XIII of [4], Prior discusses a system QKt, designed to stand to the “minimal” tense logic Kt as the modal system Q of [3] stands to S5. In this paper I provide semantics for a similar system, slightly weaker than QKt: the weakening is due to the fact that Prior's axioms are slightly too strong for a “minimal” system. An extended post-Henkin style completeness proof for the axiomatization with respect to the semantics provided is then given: the underlying three-valued nature of the semantics requires a twist in the proof which is new to its author at least, and also results in some details being set out which could well be glossed over in the two-valued case.

In this paper we describe generalizations of several approaches to the hyperarithmetic hierarchy, show how they are related to the Kripke-Platek theory of admissible ordinals and sets, and study conditions under which the various approaches remain equivalent.

To put matters in some perspective, let us first review various approaches to the theory of hyperarithmetic sets. For most purposes, it is convenient to first define the semi-hyperarithmetic (semi-HA) subsets of N. A set is then said to be hyperarithmetic (HA) if both it and its complement are semi-HA. A total number-theoretic function is HA if its graph is HA.

In this paper we demonstrate the hereditary undecidability of finite planar graphs. In §2 we introduce the preliminary logical notions used and outline the Rabin–Scott method of semantic embedding. This method is illustrated in §3 by proving the undecidability of the theory of two finite equivalence relations of a special type. In §4 we give a proof of the main theorem by embedding these equivalence relations into finite planar graphs.

The basic idea is first to form a graph which codes a pair of these relations and then to take a representative of it and “squish” it to the plane. This “squishing” requires the introduction of crossings; and edges of the original graph become paths in the new one. To distinguish the original edges we place two different types of “diamonds” about crossing points. We can then uncode our new graphs to recover the equivalence relations by means of simple first-order incidence properties.

For the terminology and motivation of this note, my earlier paper [3] should be consulted. Here I make a slight change by requiring the Intuitionist Propositional Logic H to be given in terms of axiom schemata rather than axioms.

Definition. An axiom schema F is essentially negative iff each schematic letter appearing in F is negated.

Thus the schemata (¬P ∨ ¬¬P), (¬¬(¬P ∨ ¬Q)→(¬P ∨ ¬Q)) are essentially negative, whereas (¬P ∨ P) and (¬¬P→P) are not.

Lemma. Let F be an essentially negative axiom schema. Then F yields at most finitely many intuitionistically nonequivalent axioms whose atoms are chosen from a fixed finite set.

A relational structure is called weakly atomic-compact if and only if every set Σ of atomic formulas (taken from the first-order language of the similarity type of augmented by a possibly uncountable set of additional variables as “unknowns”) is satisfiable in whenever every finite subset of Σ is so satisfiable. This notion (as well as some related ones which will be mentioned in §4) was introduced by J. Mycielski as a generalization to model theory of I. Kaplansky's notion of an algebraically compact Abelian group (cf. [5], [7], [1], [8]).