The relation M 1 ⊰ M 2, where M 1 and M 2 are elementary submodels of a given model M, defines a partial ordering ⊰(M). A natural question is: What are the general properties of this ordering and what special properties can it have for given M? Here we give an example of a prime model M for which ⊰(M) is a dense linear order, in particular ⊰(M) is isomorphic to the natural order of (− ∞, ∞], the reals with a last element (Theorem 1). In fact, this is the only way for ⊰(M) to be a linear order when M is prime (Theorem 3).

We then remark that if N is a model of an ℵ0-categorical theory then ⊰(N) cannot be a linear order, but we give an example where ⊰(N) is a dense partial order, N a model of an ℵ0-categorical theory.

Independently, and about the same time, Benda [1] found nonprime models M where ⊰(M) is isomorphic to any one of a large class of orders, including (− ∞, ∞]. His methods are very similar to ours.

Notation. We assume the reader is familar with Vaught [2] where he will find the definitions and standard results pertaining to prime models, homogeneous models, and ℵ0-categorical theories.

The extreme interest of set theorists in the notion of “closed unbounded set” is epitomized in the following well-known theorem:

Theorem A. For any regular cardinal κ > ω, the intersection of any two closed unbounded subsets of κ is closed and unbounded.

The proof of this theorem is easy and in fact yields a stronger result, namely that for any uncountable regular cardinal κ the intersection of fewer than κ many closed unbounded sets is closed and unbounded. Thus, if, for κ a regular uncountable cardinal, we let denote {A ⊆ κ ∣ A contains a closed unbounded subset}, then, for any such κ, is a κ-additive nonprincipal filter on κ.

Now what about the possibility of being an ultrafilterκ It is routine to see that this is impossible for κ > ℵ1. However, for κ = ℵ1 the situation is different. If were an ultrafilter, ℵ1 would be a measurable cardinal. As is well-known this is impossible if we assume the axiom of choice; however if ZF + “there exists a measurable cardinal” is consistent, then so is ZF + “ℵ1 is a measurable cardinal” [2]. Furthermore, under the assumption of certain set theoretic axioms (such as the axiom of determinateness or various infinite exponent partition relations) can be proven to be an ultrafilter. (See [3] and [5].)

This note presents a practical algorithm for carrying out abstraction on combinatory terms. The well-known abstraction algorithms [1, pp. 188ff.] defining abstracts in terms of the combinator sets { S, K }, { B, C, K, W }, etc. operate on one variable at a time, and result in rather long abstracts when several variables are involved. These algorithms are not practical for the applications of the combinatory logic to the theory of computing which make much use of multi-variable abstraction (e.g., [2], [3]). The present algorithm performs the abstraction with respect to all specified variables in a single step, and yields abstracts in a concise form (with sizes proportional to those of given combinatory terms).

The concept of an α-space was introduced and studied by Dekker in [1] and [2] from which we take our terminology and notation. The reader is assumed to be familiar with their contents. In this paper we are interested in the following conjecture which appears in [2, p. 493]:

(*) S ⊕ C = V and S ∥ C and V an α-space ⇒ both S and C are α-spaces.

It can be simply rephrased as:

(**) The two components of an α-decomposition of an α-space are both α-spaces.

In either form it is relevant to the ideas of α-subspace and α-homomorphism. Dekker [2, T5] has established (*) in the case S (or C) is an isolic α-space. In this paper we will prove (*) in two specific cases, one of which is a generalization of Dekker's result.

An ordinal in a model of set theory is truly countable if its set of predecessors is countable in the real world. We classify the order types of the sets of truly countable ordinals. Models with indiscernibles and other related results are discussed.

In a paper on combinatorial properties and large cardinals [2], Jech extended several combinatorial properties of a cardinal κ to analogous properties of the set of all subsets of λ of cardinality less than κ, denoted by “p κλ”, where λ is any cardinal ≤κ. We shall consider in this paper one of these properties which is historically rooted in a theorem of Ramsey [10] and in work of Rowbottom [12].

As in [2], define [p κλ]2 = {{x, y}: x, y ∈ p κλ and x ≠ y}. An unbounded subset A of p κλ is homogeneous for a function F: [p κλ]2 → 2 if there is a k < 2 so that for all x, y ∈ A with either x ⊊ y or y ⊊ x, F({x, y}) = k. A two-valued measure ü on p κλ is fine if it is κ-complete and if for all α < λ, ü({x ∈ p κλ: α ∈ x}) = 1, and ü is normal if, in addition, for every function f: p κλ → λsuch that ü({x ∈ p κλ: f(x) ∈ x}) = 1, there is an α < λ such that ü({x ∈ p κλ: f(x) = α}) = 1. Finally, a fine measure on p κλ has the partition property if every F: [p κλ]2 → 2 has a homogeneous set of measure one.

For the first Gödel incompleteness theorem, the existence in a formal system of arithmetic L of a sentence which is neither provable nor refutable, all that is required of the formula Th(x) of L used to express the notion ‘x is the g.n. (gödel number) of a theorem of L’ is mere numeralwise correctness, i.e. that for a numeral n, Th(n) is provable in L iff n is the g.n. of a theorem of L. It is well known that much more is needed for the second Gödel incompleteness theorem, the unprovability in L of the formula Con =df ¬(∃y, z)(Th(y) ∧ Th(z) ∧ neg(z, y)), which (if neg expresses negation) expresses the consistency of L. Conditions sufficient for this second theorem, more or less as stated by Hilbert-Bernays [1, p. 286] and elegantly formulated by Löb [2] may with a cavalier disregard for the distinction between use and mention be stated thus: The result of the first incompleteness theorem: there is a sentence G such that ⊢G ↔ ¬Th G), together with, if ⊢A then ⊢Th A, if ⊢(A → B) then ⊢(Th A → Th B), ⊢(Th A → Th Th A). On the other hand Feferman [3], Kreisel [4, p. 154] and Jeroslow [9] have given examples of systems and consistency formulae, based on numeralwise correct formulae Th(x), which are provable within the system.

In [2], [3], J. Grassin characterized those classes of sets which are expressible as a Boolean combination of open sets in the Baire topology on the power set of the natural numbers. From this characterization he deduced that if is a Boolean combination of recursive open sets, then the minimum length for any Boolean combination representing can be realized using recursive open sets.

There is a mistake in [6], discovered by Robert V. Kohn: contrary to the claim on p. 714, the set {◊X i : i ≥ 2} is not independent in S3. As a result three problems thought to be settled are open once more, viz., whether the Post numbers of S2, S3, and K4 are denumerable. In [6] it was claimed they are not. Now we shall prove that that claim is indeed correct for S2 and K4 but, surprisingly enough, not for S3.

One of the most important and distinctive tools in recursion theory has been the priority method whereby a recursively enumerable (r.e.) set A is constructed by stages to satisfy a sequence of conditions {Rn }n∈ω called requirements. If n < m, requirement Rn is given priority over Rm and action taken for Rm at some stage s may at a later stage t > s be undone for the sake of Rn thereby injuring Rm at stage t. The first priority method was invented by Friedberg [2] and Muchnik [11] to solve Post's problem and is characterized by the fact that each requirement is injured at most finitely often.

Shoenfield [20, Lemma 1], and then independently Sacks [17] and Yates [25] invented a much more powerful method in which a requirement may be injured infinitely often, and the method was applied and refined by Sacks [15], [16], [17], [18], [19] and Yates [25], [26] to obtain many deep results on r.e. sets and their degrees. In spite of numerous simplifications and variations this infinite injury method has never been as well understood as the finite injury method because of its apparently greater complexity.

The purpose of this paper is to reduce the Sacks method to two easily understood lemmas whose proofs are very similar to the finite injury case. Using these lemmas we can derive all the results of Sacks on r.e. degrees, and some by Yates and Robinson as well, in a manner accessible to the nonspecialist. The heart of the method is an ingenious observation of Lachlan [7] which is combined with a further simplification of our own.

The notions of recursively saturated and resplendent models grew out of the study of admissible sets with urelements and admissible fragments of L ω1ω, but, when applied to ordinary first order model theory, give us new tools for research and exposition. We will discuss their history in §3.

The notion of saturated model has proven to be important in model theory. Its most important property for applications is that if , are saturated and of the same cardinality then = iff ≅ . See, e.g., Chang-Keisler [3]. The main drawback is that saturated models exist only under unusual assumptions of set theory. For example, if 2κ = κ+ then every theory T of L has a saturated model of power κ+. (Similarly, if κ is strongly inaccessible, then every T has a saturated model of power κ.) On the other hand, a theory T like Peano arithmetic, with types, cannot have a saturated model in any power κ with ω ≤ κ ≤ .

One method for circumventing these problems of existence (or rather non-existence) is the use of “special” models (cf. [3]). If κ = Σλ<κ2λ, κ < ω, then every theory T of L has a special model of power κ. Such cardinals are large and, themselves, rather special. There are definite aesthetic objections to the use of these large, singular models to prove results about first order logic.

Several people have independently been studying Boolean ultrapowers recently; see for example [2], [3], [4], [6]. Boolean ultrapowers are a quite natural generalization of the well-known usual ultrapowers, but it seemed to be unknown whether every Boolean ultrapower is isomorphic to an ultrapower. We give a negative answer to that question. We further show that a Boolean ultrapower by an ℵ1-regular ultrafilter need not be ℵ2-universal, i.e. that Theorem 4.3.12 of [1] does not hold for Boolean ultrapowers.

Let B be a complete Boolean algebra (we identify the algebra with its underlying set), whose operations are denoted by +, ·, −, 0, 1, Σ, Π Let be a structure for some language ℒ.

Let a be an admissible ordinal and let ∧ ≤ α be a limit ordinal. A sequence of a-r.e. degrees is said to be ascending, simultaneous and of length ∧ if (i) there is an α-recursive function t: α × ∧ → α such that, for all ϒ < ∧, A ϒ = {t(σ, ϒ)∣ σ < α} is of degree a ϒ; (ii) if ϒ < ⊤ < ∧, then a ϒ ≤α a τ and (iii) for all ϒ < ∧, there is a ⊤ > ϒ with a ϒ, >α a ϒ. Lerman [4] showed that such an exists for every ∧ ≤ α. An upper bound a of is an α-r.e. degree in which every element of is α-recursive. a is minimal if there is no α-r.e. degree b <α a which is also an upper bound of . Sacks [6] proved that every ascending sequence of simultaneously ω-r.e. degrees of length ω cannot have 0ω′, the complete ω-r.e. degree, as a minimal upper bound. In contrast, Cooper [2] showed that there exists an ascending sequence of simultaneously ω-r.e. degrees of length to having a minimal upper bound which is an ω-r.e. degree. In this paper we investigate the behavior of ascending sequences of simultaneously α-r.e. degrees for admissible ordinals α > ω. Call α Σ∞-admissibIe if it is Σ n -nadmissible for all n. Let Φ(∧) say: No ascending sequence of simultaneously α-r.e. degrees of length ∧ can have 0α′, the complete α-r.e. degree, as a minimal upper bound. Our main result in this paper is:

Let α be either a constructible cardinal with σ2ci(α) < α or Σ∞-admissible. Then σ2cf(α) is the least ordinal ν for which every ∧ ≤ α of cofinality ν (over L α) can satisfy Φ(∧).

Since the late 1940's model theory has found numerous applications to algebra. I would like to indicate some of the points of contact between model theoretic methods and strictly algebraic concerns by means of a few concrete examples and typical applications.

§1. The Lefschetz principle. Algebraic geometry has proved to be a fruitful source of model theoretic ideas. What exactly is algebraic geometry? We consider a field K, and let Kn be the set of n-tuples (a 1 … an ) with coordinates ai in K. Kn is called affine n-space over K. Fix polynomials p 1 …, pk in K[x 1, …, xn ] and define

that is V(p 1 …, pk ) is the locus of common zeroes of the pi in Kn . We call V(Pi …, Pk ) the algebraic variety determined by p 1, …, pk . With this terminology we may say:

Algebraic geometry is the study of algebraic varieties defined over an arbitrary field K. This definition lacks both rigor and accuracy, and we will indicate below how it may be improved.

So far we have placed no restrictions on the base field K. Following Weil [4] it is convenient to start with a so-called “universal domain”; in other words take K to be algebraically closed and of infinite transcendence degree over the prime field. Any particular field can of course be embedded in such a universal domain.