Linear logic, introduced by Girard, is a refinement of classical logic with a natural, intrinsic accounting of resources. We show that unlike most other propositional (quantifier-free) logics, full propositional linear logic is undecidable. Further, we prove that without the modal storage operator, which indicates unboundedness of resources, the decision problem becomes pspace-complete. We also establish membership in np for the multiplicative fragment, np-completeness for the multiplicative fragment extended with unrestricted weakening, and undecidability for certain fragments of noncommutative propositional linear logic. 1 Introduction Linear logic, introduced by Girard [14, 18, 17], is a refinement of classical logic which may be derived from a Gentzen-style sequent calculus axiomatization of classical logic in three steps. The resulting sequent system Lincoln@CS.Stanford.EDU Department of Computer Science, Stanford University, Stanford, CA 94305, and the Computer Science Labo...

3.672> X with the complex term 1 + 1, not with 2, which, for people unused to Prolog's little ways, tends to come as a bit of a surprise. If we want to carry out the actual arithmetic involved, we have to explicitly force evaluation by making use of the very special inbuilt `operator' is/2. This calls an inbuilt mechanism which carries out the arithmetic evaluation of its second argument, and then unication plays no role here!). On the other hand, \== checks whether its argument are not identical. Arithmetic Prolog contains some built-in operators for handling integer arithmetic. These include *, / +, - (for multiplication, division, addition, and subtraction, respectively) and >, < for comparing numbers. These symbols, however, are just ordinary Prolog operators. That is, they are just a user friendly notation for writing

We investigate semantical normalization proofs for typed combinatory logic and weak -calculus. One builds a model and a function `quote' which inverts the interpretation function. A normalization function is then obtained by composing quote with the interpretation function. Our models are just like the intended model, except that the function space includes a syntactic component as well as a semantic one. We call this a `glued' model because of its similarity with the glueing construction in category theory. Other basic type constructors are interpreted as in the intended model. In this way we can also treat inductively defined types such as natural numbers and Brouwer ordinals. We also discuss how to formalize -terms, and show how one model construction can be used to yield normalization proofs for two different typed -calculi -- one with explicit and one with implicit substitution. The proofs are formalized using Martin-Lof's type theory as a meta language and mechanized using the A...

The semantics of typed lambda calculus is usually described using Henkin models, consisting of functions over some collection of sets, or concrete cartesian closed categories, which are essentially equivalent. We describe a more general class of Kripke-style models. In categorical terms, our Kripke lambda models are cartesian closed subcategories of the presheaves over a poset. To those familiar with Kripke models of modal or intuitionistic logics, Kripke lambda models are likely to seem adequately \semantic." However, when viewed as cartesian closed categories, they do not have the property variously referred to as concreteness, well-pointed-ness, or having enough points. While the traditional lambda calculus proof system is not complete for Henkin models that may have empty types, we prove strong completeness for Kripke models. In fact, every set of equations that is closed under implication is the theory of a single Kripke model. We also develop some properties of logical relations ...

In this article we describe the data model of the MBase system, a webbased,

Abstract not found

We analyze the computational complexity of type inference for untyped -terms in the second-order polymorphic typed -calculus (F 2 ) invented by Girard and Reynolds, as well as higher-order extensions F 3 ; F 4 ; : : : ; F ! proposed by Girard. We prove that recognizing the F 2 - typable terms requires exponential time, and for F ! the problem is nonelementary. We show as well a sequence of lower bounds on recognizing the F k -typable terms, where the bound for F k+1 is exponentially larger than that for F k . The lower bounds are based on generic simulation of Turing Machines, where computation is simulated at the expression and type level simultaneously. Non-accepting computations are mapped to non-normalizing reduction sequences, and hence non-typable terms. The accepting computations are mapped to typable terms, where higher-order types encode reduction sequences, and first-order types encode the entire computation as a circuit, based on a unification simulation of Boolean logic. ...

Even though higher-order calculi for automated theorem proving are rather old, tableau calculi have not been investigated yet. This paper presents two free variable tableau calculi for higher-order logic that use higher-order unification as the key inference procedure. These calculi differ in the treatment of the substitutional properties of equivalences. The first calculus is equivalent in deductive power to the machineoriented higher-order refutation calculi known from the literature, whereas the second is complete with respect to Henkin's general models.

We analyze the inherent complexity of implementing L'evy's notion of optimal evaluation for the -calculus, where similar redexes are contracted in one step via so-called parallel fi-reduction. Optimal evaluation was finally realized by Lamping, who introduced a beautiful graph reduction technology for sharing evaluation contexts dual to the sharing of values. His pioneering insights have been modified and improved in subsequent implementations of optimal reduction. We prove that the cost of parallel fi-reduction is not bounded by any Kalm'ar-elementary recursive function. Not merely do we establish that the parallel fi-step cannot be a unit-cost operation, we demonstrate that the time complexity of implementing a sequence of n parallel fi-steps is not bounded as O(2 n ), O(2 2 n ), O(2 2 2 n ), or in general, O(K ` (n)) where K ` (n) is a fixed stack of ` 2s with an n on top. A key insight, essential to the establishment of this nonelementary lower bound, is that any simply-...

Data Types, though, as Reynolds stresses, is not perfectly suited for higher type or higher order systems and, thus, he proposes a "relational" treatment of invariance: computations do not depend on types in the sense that they are "invariant" w.r.t. arbitrary relations on types and between types. Reynolds's approach set the basis for most of the current work on parametricity, as we will review below (.3). Some twelve years earlier, Girard had given just a simple hint towards another understanding of the properties of "computing with types". In [Gir71], it is shown, as a side remark, that, given a type A, if one defines a term J A such that, for any type B, J A B reduces to 1, if A = B, and reduces to 0, if A ¹ B, then F + J A does not normalize. In particular, then, J A is not definable in F. This remark on how terms may depend on types is inspired by a view of types which is quite different from Reynolds's. System F was born as the theory of proofs of second order intuitionis...