The present paper concerns itself primarily with the decision problem for formal elementary intuitionistic theories and the method is primarily model-theoretic. The chief tool is the Kripke model for which the reader may find sufficient background in Fitting's book Intuitionistic logic model theory and forcing (North-Holland, Amsterdam, 1969). Our notation is basically that of Fitting, the differences being to favor more standard notations in various places.

The author owes a great debt to many people and would particularly like to thank S. Feferman, D. Gabbay, W. Howard, G. Kreisel, G. Mints, and R. Statman for their valuable assistance.

The method of elimination of quantifiers, which has long since proven its use in classical logic, has also been applied to intuitionistic theories (i) to demonstrate decidability ([9], [15], [17]), (ii) to prove the coincidence of an intuitionistic theory with its classical extension ([9], [17]), and (iii), as we will see below, to establish relations between an intuitionistic theory and its classical extension. The most general of these results is to be obtained from the method of Lifshits' quantifier elimination for the intuitionistic theory of decidable equality.

Since the details of Lifshits' proof have not been published, and since the proof yields a more general result than that stated in his abstract [15], we include the proof and several corollaries.

Each of the various “large cardinal” axioms currently studied in set theory owes its inspiration to concrete phenomena in various fields. For example, the statement of the well-known compactness theorem for first-order logic can be generalized in various ways to infinitary languages to yield definitions of compact cardinals, and the reflection principles provable in ZF, when modified in the appropriate way, yields indescribable cardinals.

In this paper we concern ourselves with two kinds of large cardinals which are probably the two best known of those whose origins lie in model theory. They are the Rowbottom cardinals and the Jonsson cardinals.

Let us be more specific. A cardinal κ is said to be a Jonsson cardinal if every structure of cardinality κ has a proper elementary substructure of cardinality κ. (It is routine to see that only uncountable cardinals can be Jonsson. Erdös and Hajnal have shown [2] that for n < ω no ℵn is Jonsson. (In fact, they showed that if κ is not Jonsson then neither is the successor cardinal of κ and that, assuming GCH, no successor cardinal can be Jonsson.) Keisler and Rowbottom first showed that the existence of a Jonsson cardinal contradicts V = L.) The definition of a Rowbottom cardinal is only slightly more intricate. We assume for the moment that our similarity type has a designated one-place relation.

Combinations of capital letters (e.g., G, EVEN) will be used as names of formulas of Peano arithmetic, x, y, z will be variables. Let be a model for Peano arithmetic. The names of formulas with a bar (e.g., G−, EVEN−(x)) will denote the relations defined by the formulas in . The elements of are called numbers.

By a finite sequence of numbers 〈a0, …, an〉 we mean a number . Note that n Pi's and ai's can be nonstandard elements and therefore “finite,” and all concepts which will use it are not absolute but relative to .

If a and b are degrees of unsolvability, a is called (as in [1]) a minimal cover of b if b < a and no degree c satisfies b < c < a. A degree a is called a minimal cover if it is a minimal cover of some degree b. We show in ZF set theory that there is a degree d such that every degree a ≥ d is a minimal cover. In [1] it was remarked that this result follows via a lemma of D. A. Martin [4] from the determinacy of a certain Σ50 Gale-Stewart game (which in turn follows [5] from the existence of a measurable cardinal). The argument here is parallel to that of [1], but the essential new ingredient is the result of Paris [6] that Σ50 determinacy can be proved in ZF. Also it is necessary to replace the Σ50 game of [1] by a related Σ40 game and to use the method, rather than just the statement, of Martin's lemma.

Theorem. There is a degreedsuch that every degreea ≥ dis a minimal cover.

Proof. In a (Gale–Stewart) game on 2ω, players I and II construct a set of integers A as follows: At the nth stage, player I determines whether n ϵ A if n is even and player II if n is odd. Thus player I controls A0 and player II controls A1 where Ai = {n: 2n + i ϵ A}. Let Aij denote (Ai)j and let M (A, B) mean that the degree of A is a minimal cover of that of B. Consider now the particular game in which player I wins iff M(A, A00) holds.

In this paper we prove that the word problem for division rings is recursively unsolvable. Our proof relies on the corresponding result for groups [7], [28], and makes essential use of P. M. Cohn's recent work [11], [13], [15], [16] on division rings.

The word problem for groups is usually formulated in terms of group presentations or finitely presented groups, as in [7], [24], [28], [30]. An equivalent formulation, in terms of the universal Horn sentences of group theory, is mentioned in [32]. This formulation makes sense for arbitrary first-order theories, and it is with respect to this formulation that we show that the word problem for division rings has degree 0′.

Let ω be the nonnegative integers. G. E. Sacks once asked whether there exists an infinite X ⊆ ω such that, for all Y ⊆ X, ω1Yω1 where ω1 is the first nonrecursive ordinal. In this note we negatively answer this question by giving a simple proof that for every infinite set X ⊆ ω there exists Y ⊆ X such that the first recursively inaccessible ordinal. This is accomplished by proving that Hα is hyper-arithmetic in Y whereHα is the αth hyperjump of the empty set ∅, defined in a suitable sense for all ordinals

Background information and undefined notation can be found in Rogers [11]. In particular, we write A ≤hB(A ≤T B) if A is hyperarithmetical (recursive) in B, and A ≡hB if A ≤hB and B ≤hA. We will say that a set A is hyperarithmetically (recursively) encodable if, for every infinite set X ⊆ ω, there exists Y ⊆ X such that A ≤hY (A ≤TY). For any set A (hyperdegree a) let A′ (a′) denote the hyperjump of A (a). Let 0 denote the hyperdegree of ∅. A function f majorizes a function g if f(n) ≥ g(n) for every n. E1 is the representing (type-2) functional of

introduced by Tugué [13] (also Kleene [6]). Let be the smallest ordinal which is not the order type of any well-ordering recursive in E1. Information on can be found in Richter [9] and [10].

Let U be a consistent axiomatic theory containing Robinson's Q [TMRUT, p. 51]. In order for the results below to be of interest, U must be powerful enough to carry out certain arguments involving versions of the “derivability conditions,” DC(i) to DC(iii) below, of [HBGM, p. 285], [F60, Theorem 4.7], or [L55]. Thus it must contain, at least, mathematical induction for formulas whose prenex normal forms contain at most existential quantifiers. For convenience, U is assumed also to contain symbols for primitive recursive functions and relations, and their defining equations. One of these is used to form the standard provability predicate, Prov ˹A˺, “there exists a number which is the Gödel number of a proof of A.” Upper corners denote numerals for Gödel numbers for the enclosed sentences, and parentheses are often omitted in their presence.

This paper contains some results concerning the relation between the sentence A, and the sentence Prov ˹A˺ in the Lindenbaum Sentence Algebra (LSA) for U, the Boolean algebra induced by the pre-order relation A ≤ B ⇔ ⊦A → B. Half of the answer is provided by a theorem of Löb [L55], which states that ⊦Prov ˹A˺ → A ⇔ ⊦A. Hence, in the presence of DC(iii), below, it is never true that Prov ˹A˺ < A in the LSA. However, there is a large and interesting set of sentences, denoted here by Γ, for which A < Prov ⌜A⌝.

The theorem proved in this paper answers some transitivity questions (in the geometric sense) for the type free λ-calculus: Which objects can be mapped on all other objects? How much can an object do by applying it to other objects (see footnote 2)?

The main result is that, for closed terms of the λI-calculus, the following conditions are equivalent:

(a) M has a normal form.

(b) FM = I for some λI-term F.

(c) MN1 … Nn = I for some λI-terms N1 …, Nn.

By the same method it follows that if M is a closed term of the λK-calculus having a normal form, then for some λI-terms (sic) N1, …, Nn, MN1… Nn = I is provable in the λK-calculus.

The theorem of Böhm [2] states that if M1, M2 are terms of the λK-calculus having different βη-normal forms, then ∀A1, A2 ∃N1, …, NnMiN1 … Nn = Ai is provable in the λK-βη-calculus for i = 1, 2. As a consequence of this it was shown (implicitly) in [1, 3.2.20 1/2 (1)] that if M has a normal form, then for some λK-terms N1, …, Nn, MN1 … Nn =I is provable in the λK-calculus.

It was not clear that this also could be proved for the λK-calculus since the proof of the theorem of Böhm essentially made use of λK-terms.

We conjecture that, using the results of this paper, the full theorem of Böhm can be proved for the λI-calculus.

The study of partition relations for cardinal numbers introduced by Erdös and his school in the 1950's has, for the past several years, had a profound impact in logic. Unfortunately, quite early in their development, it was noticed by Rado [1] that the potentially most fruitful class of such relations, infinite exponent partition relations, were always in contradiction with the axiom of choice (AC). As a result, such relations were overlooked. This turned out to be a mistake; for, as has been noticed recently, a close study of infinite exponent partition relations is both interesting and rewarding. For example, there are weakened versions of such relations which are provable in ZF and which have valuable applications in recursion theory and set theory. In addition, the pure theory of these relations, like that of the axiom of determinateness, is fruitful as well as elegant. For more background here one should refer to [3].

At any rate, with a more detailed look at infinite exponent partition relations came a more refined version of Rado's original theorem. Specifically, Rado used the full axiom of choice to carefully construct partitions to violate any desired relation—a more sophisticated look at the actual theory of such relations indicated how one could put together some desired partitions using only well-ordered choice [3].

The distinction between well-ordered choice and full choice is by no means vacuous in this context. For Mathias has shown [4] that the simplest infinite exponent partition relation, ω → (ω)ω, is consistent with countable choice (well-ordered choice of length ℵ0) and, in fact, is consistent with dependent choice.

A standard enumeration of the recursively enumerable (r.e.) sets is an acceptable numbering {Wn}n∈N of the r.e. sets in the sense of Rogers [5, p. 41], together with a 1:1 recursive function f with range In his quest for nonrecursive incomplete r.e. sets Post [4] constructed a hypersimple set Hf, relative to a fixed but unspecified standard enumeration f. Although it was later shown that hyper-simplicity does not guarantee incompleteness, the ironic possibility remained that Post's own particular hypersimple set might be incomplete. We settle the question by proving that H, may be either complete or incomplete depending upon which standard enumeration f is used. In contrast, D. A. Martin has shown [3] that Post's simple set S [4, p. 298] is complete for any standard enumeration. Furthermore, what most modern recursion theorists would regard as the “natural” construction of a hypersimple set (which we give in §1) is also complete for any standard enumeration.

There are two conclusions to be drawn from these results. First, they substantiate the often repeated remark among recursion theorists that Post's hypersimple set construction is a precursor of priority constructions because priorities play a strong role, and because there is a great deal of “restraint” which tends to keep elements out of the set. Secondly, the results warn recursion theorists that more properties than might have been supposed depend upon which standard enumeration is chosen at the beginning of the construction of some r.e. set.

It has long been known that every free associative algebra can be embedded in a skew field [11]; in fact there are many different embeddings, all obtainable by specialization from the ‘universal field of fractions’ of the free algebra (cf. [5, Chapter 7]). This makes it reasonable to call the latter the free field; see §2 for precise definitions. The existence of this free field was first established by Amitsur [1], but his proof is rather indirect and does not provide anything like a normal form for the elements of the field. Actually one cannot expect to find such a normal form, since it does not even exist in the field of fractions of a commutative integral domain, but at least one can raise the word problem for free fields: Does there exist an algorithm for deciding whether a given expression for an element of the free field represents zero?

Now some recent work has revealed a more direct way of constructing free fields ([4], [5], [6]), and it is the object of this note to show how this method can be used to solve the word problem for free fields over infinite ground fields. In this connexion it is of interest to note that A. Macintyre [9] has shown that the word problem for skew fields is recursively unsolvable. Of course, every finitely generated commutative field has a solvable word problem (see e.g. [12]).

The construction of universal fields of fractions in terms of full matrices is briefly recalled in §2, and it is shown quite generally for a ring R with a field of fractions inverting all full matrices, that if the set of full matrices over R is recursive, then the universal field has a solvable word problem. This holds more generally if the precise set of matrices over R inverted over the field is recursive, but it seems difficult to exploit this more general statement.