i) We show for each context-free language L that by considering each word of L as a structure in a natural way, one turns L into a finite union of classes which satisfy a finitary analog of the characteristic properties of complete universal first order classes of structures equipped with elementary embeddings. We show this to hold for a much larger class of languages which we call free local languages, ii) We define local languages, a class of languages between free local and context-sensitive languages. Each local language L has a natural extension L ∞ to infinite words, and we prove a series of “pumping lemmas”, analogs for each local language L of the “uvxyz theorem” of context free languages: they relate the existence of large words in L or L ∞ to the existence of infinite “progressions” of words included in L, and they imply the decidability of various questions about L or L ∞. iii) We show that the pumping lemmas of ii) are independent from strong axioms, ranging from Peano arithmetic to ZF + Mahlo cardinals.

We hope that these results are useful for a model-theoretic approach to the theory of formal languages.

Dowker [1] raised the question of the existence of filters such that for every coloring (partition) of the underlying index set I with two colors there is a relation R on I which (i) is fat (in the sense that sets of the form {y Є I∣xRy} are in the filter) and (ii) has no bichromatic symmetric pairs (i.e., distinct indices x and y such that x R y and y R x). Additionally, he required that the filter have no anti-symmetric fat relation, for such a relation would vacuously satisfy (i) and (ii). The question of the existence of Dowker filters has been studied more recently by Rudin [3], [4], who conjectures [3] that such filters do not exist.

For ZFC the problem remains open. However, Example 2 of this paper shows that one can construct a Dowker filter provided one drops the axiom of choice in favor of the Baire Property (BP) axiom which is known to be incompatible with ZFC but relatively consistent with ZF. In fact, the filter constructed is super-Dowker in the sense that (ii) can be replaced by the requirement that all components of all symmetric pairs have the same color. But, in ZFC the existence of a super-Dowker filter implies the existence of a measurable cardinal.

Let F be a filter on an index set I. A set will be called big, small, or medium depending on whether F contains that set, its compliment, or neither, respectively. We define five cardinals associated with F:

α denotes the smallest cardinal such that there is a family of α big sets whose intersection is not big.

ν denotes the smallest cardinal such that there is a family of ν big sets whose intersection is small.

It follows constructively from weak versions of Markov's principle (MP) and of Church's thesis (WCT) that logical validity for single sentences is not arithmetically definable. This is a direct consequence of the constructive model-theoretic fact that, using MP and WCT, one can prove the categoricity of first-order Heyting arithmetic. By a straightforward refinement of this proof, we then obtain generalizations of the incompleteness theorem of Kreisel [K 2] as presented by van Dalen [Da] and improved by Leivant [L 1]. An analogous result for classical validity in r.e. models appears in [V].

Our methods of proof are new in that they rely not upon variants of the ideas of [K2] but upon a constructive proof idea inspired by the classical result of Tennenbaum ([Te], [E&K], [S]) on recursive models of arithmetic. This line of reasoning was suggested by the applications of readability to recursive mathematics described in [McC1], [McC2] and [McC3].

Markov's principle is more than a convenience in constructive arithmetic and analysis; it is absolutely essential to significant areas of constructive cardinal arithmetic. In turn, logical relations among intuitively appealing principles of constructive cardinal arithmetic parallel relations between MPS and other “problematic axioms” for constructive mathematics, such as the limited principle of omniscience. Finally, simple closure properties on the Dedekind finite sets provide ready examples of statements which are strictly weaker than Markov's principle and yet are independent of extensions of IZF.

MPS, Markov's principle with variables over sets, is equivalent to each of these elementary properties of ⊿, the class of Dedekind finite sets:

1. ∀A(A Є ⊿ ↔ CP(A)).

2. [*]A Є ⊿ ↔ ∀B (A + B is infinite ↔ B is infinite).

3. [**]A Є ⊿ ↔ ∀B((A + 1) x B is infinite ↔ B is infinite).

CP is Tarski's cancellation property. Consequently, MPS is tantamount, in constructive mathematics, to the standard classical characterizations of ⊿ in terms of cardinality.

[*] and [**] imply that ⊿ is closed under addition and multiplication. Closure under addition is, in turn, constructively equivalent to the closure of ⊿ under all combinatorial functions and to the fact that ⊿ is closed under each strict combinatorial function individually. Generally speaking, a function on P(ω) is combinatorial whenever it preserves finiteness, respects cardinality and has a “moduluslike” associate function. It is strict when it is strictly increasing relative to the subset ordering, [cf. §6 for precise definitions.] From this, we see that MPS is foundational for cardinal arithmetic on the constructive Dedekind finite sets.

We combine two techniques of set theory relating to mininal degrees of constructibility. Jensen constructed a minimal real which is additionally a singleton. Groszek built an initial segment of order type 1 + α*, for any ordinal α. This paper shows how to force a singleton such that the c-degrees beneath it, all represented by reals, are of type 1 + α*, for many ordinals α. We also examine the definability α needs to be so represented by a real.

We present a comprehensive study of the axiom schemas (induction and collection schemas for parameter free Σ n formulas) and some closely related schemas.

We prove optimal lower bounds on the computation time for several well-known test problems on a quite realistic computational model: the random access machine. These lower bound arguments may be of special interest for logicians because they rely on finitary analogues of two important concepts from mathematical logic: inaccessible numbers and order indiscernibles.

A recursively enumerable splitting of an r.e. set A is a pair of r.e. sets B and C such that A = B ∪ C and B ∩ C = ⊘. Since for such a splitting deg A = deg B ∪ deg C, r.e. splittings proved to be a quite useful notion for investigations into the structure of the r.e. degrees. Important splitting theorems, like Sacks splitting [S1], Robinson splitting [R1] and Lachlan splitting [L3], use r.e. splittings.

Since each r.e. splitting of a set induces a splitting of its degree, it is natural to study the relation between the degrees of r.e. splittings and the degree splittings of a set. We say a set A has the strong universal splitting property (SUSP) if each splitting of its degree is represented by an r.e. splitting of itself, i.e., if for deg A = b ∪ c there is an r.e. splitting B, C of A such that deg B = b and deg C = c. The goal of this paper is the study of this splitting property.

In the literature some weaker splitting properties have been studied as well as splitting properties which imply failure of the SUSP.

In his Singular points of complex hypersurfaces Milnor proves the following “curve selection lemma” [10, p. 25]:

Let V ⊂ R m be a real algebraic set, and let U ⊂ R m be an open set defined by finitely many polynomial inequalities:

Lemma 3.1. If U ∩ V contains points arbitrarily close to the origin (that is if 0 ∈ Closure (U ∩ V)) then there exists a real analytic curve

with p(0) = 0 and with p(t) ∈ U ∩ V for t > 0.

Of course, this result will also apply to semialgebraic sets (finite unions of sets U ∩ V), and by Tarski's theorem such sets are exactly the sets obtained from real varieties by means of the finite Boolean operations and the projection maps R n+1 → R n . If, in this tiny extension of Milnor's result, we replace ‘R’ everywhere by ‘Q p ’, we obtain a p-adic curve selection lemma, a version of which we will prove in this essay. Semialgebraic sets, in the p-adic context, may be defined just as they are over the reals: namely, as those sets obtained from p-adic varieties by means of the finite Boolean operations and the projection maps . Analytic maps are maps whose coordinate functions are given locally by convergent power series.

Let M be an o-minimal structure or a p-adically closed field. Let be the space of complete n-types over M equipped with the following topology: The basic open sets of are of the form Ũ = {p ∈ Sn (M): U ∈ p} for U an open definable subset of Mn . is a spectral space. (For M = K a real closed field, is precisely the real spectrum of K[X 1, …, Xn ]; see [CR].) We will equip with a sheaf of LM -structures (where LM is a suitable language). Again for M a real closed field this corresponds to the structure sheaf on (see [S]). Our main point is that when Th(M) has definable Skolem functions, then if p ∈ , it follows that M(p), the definable ultrapower of M at p, can be factored through Mp , the stalk at p with respect to the above sheaf. This depends on the observation that if M ≺ N, a ∈ Nn and f is an M-definable (partial) function defined at a, then there is an open M-definable set U ⊂ Nn with a ∈ U, and a continuous M-definable function g:U → N such that g(a) = f(a).

In the case that M is an o-minimal expansion of a real closed field (or M is a p-adically closed field), it turns out that M(p) can be recovered as the unique quotient of Mp which is an elementary extension of M.

We consider countably additive, nonnegative, extended real-valued measures which vanish on singletons. Such a measure is universal on a set X iff it is defined on all subsets of X and is semiregular iff every set of positive measure contains a subset of positive finite measure. We study the problem of existence of a universal semiregular measure on X which is invariant under a given group of bijections of X. Moreover we discuss some properties of universal, semiregular, invariant measures on groups.

Using formally intuitionistic logic coupled with infinitary logic and the completeness theorem for coherent logic, we establish the validity, in Grothendieck toposes, of a number of well-known, classically valid theorems about fields and ordered fields. Classically, these theorems have proofs by contradiction and most involve higher order notions. Here, the theorems are each given a first-order formulation, and this form of the theorem is then deduced using coherent or formally intuitionistic logic. This immediately implies their validity in arbitrary Grothendieck toposes. The main idea throughout is to use coherent theories and, whenever possible, find coherent formulations of formulas which then allow us to call upon the completeness theorem of coherent logic. In one place, the positive model-completeness of the relevant theory is used to find the necessary coherent formulas.

The theorems here deal with polynomials or rational functions (in s indeterminates) over fields. A polynomial over a field can, of course, be represented by a finite string of field elements, and a rational function can be represented by a pair of strings of field elements. We chose the approach whereby results on polynomial rings are reduced to results about the base field, because the theory of polynomial rings in s indeterminates over fields, although coherent, is less desirable from a model-theoretic point of view. Ultimately we are interested in the models.

This research was originally motivated by the works of Saracino and Weispfenning [SW], van den Dries [Dr], and Bunge [Bu], each of whom generalized some theorems from algebraic geometry or ordered fields to (commutative, von Neumann) regular rings (with unity).

We define the notion of Souslin forcing, and we prove that some properties are preserved under iteration. We define a weaker form of Martin's axiom, namely , and using the results on Souslin forcing we show that is consistent with the existence of a Souslin tree and with the splitting number s = ℵ1. We prove that proves the additivity of measure. Also we introduce the notion of proper Souslin forcing, and we prove that this property is preserved under countable support iterated forcing. We use these results to show that ZFC + there is an inaccessible cardinal is equiconsistent with ZFC + the Borel conjecture + -measurability.

A new method is presented for constructing models of set theory, using a technique of forming pseudo-ultrapowers. In the presence of the axiom of choice, the traditional ultrapower construction has proven to be extremely powerful in set theory and model theory; if the axiom of choice is not assumed, the fundamental theorem of ultrapowers may fail, causing the ultrapower to lose almost all of its utility. The pseudo-ultrapower is designed so that the fundamental theorem holds even if choice fails; this is arranged by means of an application of the omitting types theorem. The general theory of pseudo-ultrapowers is developed. Following that, we study supercompactness in the absence of choice, and we analyze pseudo-ultrapowers of models of the axiom of determinateness and various infinite exponent partition relations. Relationships between pseudo-ultrapowers and forcing are also discussed.

This paper is divided into three sections. §1 consists of an argument against the validity of Berry's paradox; §2 consists of supporting arguments for the thesis presented in §1; and §3 examines the possibility of re-establishing the paradox.

Berry's paradox, a semantic antinomy, is described on p. 4 of the textbook [4] as follows:

The aim of this article is to show that a basis for combinatory logic [2] must contain at least one combinator with rank strictly greater than two. We use notation of [1].

Let Q be a primitive combinator given by its reduction rule Qx 1 … xn → C, where C is a pure combination of the variables x 1,…, xn . n is called the rank of the combinator.

A set {Q 1,…,Qn } of combinators is a basis for combinatory logic if for every finite set {x 1,…,xm } of variables and every pure combination C of these variables, there exists a pure combinator Q of Q 1,…,Qn such that Qx 1…xm ↠C.

Property. The Church-Rosser theorem and the (quasi-)normalization theorem are valid for the combinatory reduction system under consideration.

Proof. See [4], [5], and [6].

Any basis must contain at least one combinator with rank strictly greater than two.

Let us assume a basis B with combinators of rank strictly less than 3; then there exists a pure combination X of the combinators in B such that: XABC ↠ CAB.

(*) First of all, we remark that if XABC ↠ M ↠ CAB, M must contain at least one occurrence of each of the variables A, B and C.

Notation. We denote by E[X 1,…,Xn ] expressions that contain at least one occurrence of every term X 1,…,Xn .

Consider k((G)) in the language of valued fields enriched with a unary predicate for the set of constants and another one for the cross-section. For perfect k, this structure is undecidable if it does not satisfy Kaplansky's conditions.

In this note we shall be interested in the following problems.

Problem 1. Can IΔ0 ⊢ ∀x∃y > x(y is prime)?

Here I Δ0 is Peano arithmetic with the induction axiom restricted to bounded (i.e. Δ0) formulae.

Problem 2. Can IΔ0 ⊢ Δ0 PHP?

Here Δ0 PHP (Δ0 pigeonhole principle) is the schema

for θ ∈ Δ0, or equivalently in IΔ0, for a Δ0 formula F(x,y)

written .

By obtaining partial solutions to Problem 2 we shall show that Problem 1 has a positive solution if IΔ0 is replaced by IΔ0 + ∀xx log(x) exists.

Our notation will be entirely standard (see for example [3] and [4]). In particular all logarithms will be to the base 2 and in expressions like log(x), (1 + ε)x, etc. we shall always mean the integer part of these quantities.

Concerning Problem 2 we remark that it is shown in [5] that for k ∈ N and F ∈ Δ0,

As far as we know this is the best result of this form, in that we do not know how to replace log(z) k by anything larger. However, as we shall show in Theorem 1, we can do much better if we increase the difference between the sizes of the domain and range of F.

In what follows let M be a countable nonstandard model of IΔ0, and let be those subsets of M defined by Δ0 formulae with parameters from M.

Theorem 1. For k ∈ N and F ∈ Δ0,

Here log0(x) = x, log k + 1(x) = log(log k (x)).

Proof. To simplify matters, consider first the case k = 1. So assume M ⊨ a log(a) exists and with and a > 1. The idea of the proof is the following.

The price is failure on a class of inductive inference problems that are easily solved, in contrast, by nonBayesian mechanical learners. By “mechanical” is meant “simulable by Turing machine”.

One of the central tenets of Bayesianism, which is common to the heterogeneous collection of views which fall under this rubric, is that hypothesis change proceeds via conditionalization on accumulated evidence, the posterior probability of a given hypothesis on the evidence being computed using Bayes's theorem. We show that this strategy for hypothesis change precludes the solution of certain problems of inductive inference by mechanical means—problems which are solvable by mechanical means when the restriction to this Bayesian strategy is lifted. Our discussion proceeds as follows. After some technical preliminaries, the concept of (formal) learner is introduced along with a criterion of inferential success. Next we specify a class of inductive inference problems, and then define the notion of “Bayesian behavior” on those problems. Finally, we exhibit an inductive inference problem from the specified class such that (a) some nonmechanical Bayesian learner solves the problem, (b) some nonBayesian mechanical learner solves the problem, (c) some mechanical learner manifests Bayesian behavior on the problem, but (d) no mechanical Bayesian learner solves the problem.

Insofar as possible terminology and notation are drawn from Osherson, Stob, and Weinstein [1986].