Kit Fine has reawakened a puzzle about variables with a long history in analytic philosophy, labeling it “the antinomy of the variable”. Fine suggests that the antinomy demands a reconceptualization of the role of variables in mathematics, natural language semantics, and first-order logic. The difficulty arises because: (i) the variables ‘x’ and ‘y’ cannot be synonymous, since they make different contributions when they jointly occur within a sentence, but (ii) there is a strong temptation to say that distinct variables ‘x’ (...) and ‘y’ are synonymous, since sentences differing by the total, proper substitution of ‘x’ for ‘y’ always agree in meaning. We offer a precise interpretation of the challenge posed by (i) and (ii). We then develop some neglected passages of Tarski to show that his semantics for variables has the resources to resolve the antinomy without abandoning standard compositional semantics. (shrink)

This paper presents a new take on how argument dependencies in natural language are established and constrained. The paper starts with a rather standard view that (quantificational) argument dependencies are operator-variable dependencies. The interesting twist the paper offers is to eliminate the need for syntax that serves to enforce what the operator-variable dependencies are. Instead the role of ensuring grammatical and generally unambiguous forms is taken up by semantics imposing what are dependency requirements for any interpretation to go through at (...) evaluation time. With this viewpoint there emerges an essential need for localities as the hiding of (embedded) expression material, together with a (limited) form of garbage collection to release grammatical resources used in the current locality for reuse in subsequent localities. (shrink)

This paper describes formalizations of Tait's normalization proof for the simply typed λ-calculus in the proof assistants Minlog, Coq and Isabelle/HOL. From the formal proofs programs are machine-extracted that implement variants of the well-known normalization-by-evaluation algorithm. The case study is used to test and compare the program extraction machineries of the three proof assistants in a non-trivial setting.

The higher order matching problem is the problem of determining whether a term is an instance of another in the simply typed [lgr]-calculus, i.e. to solve the equation a = b where a and b are simply typed [lgr]-terms and b is ground. The decidability of this problem is still open. We prove the decidability of the particular case in which the variables occuring in the problem are at most third order.

Building on previous work by Mints, Buchholz and Schwichtenberg, a simplified version of continuous normalization for the untyped λ-calculus and Gödel’s is presented and analysed in the coalgebraic framework of non-wellfounded terms with so-called repetition constructors.The primitive recursive normalization function is uniformly continuous w.r.t. the natural metric on non-wellfounded terms. Furthermore, the number of necessary repetition constructors is locally related to the number of reduction steps needed to reach the normal form and its size.It is also shown how continuous normal (...) forms relate to derivations of strong normalizability in the typed λ-calculus and how this leads to new bounds for the sum of the height of the reduction tree and the size of the normal form.Finally, the methods are extended to an infinitary λ-calculus with ω-rule and permutative conversions and this is used to derive a strong form of normalization for an iterative version of Gödel’s system , leading to a value table semantics for number-theoretic functions. (shrink)

The topics of structural proof theory and logic programming have influenced each other for more than three decades. Proof theory has contributed the notion of sequent calculus, linear logic, and higher-order quantification. Logic programming has introduced new normal forms of proofs and forced the examination of logic-based approaches to the treatment of bindings. As a result, proof theory has responded by developing an approach to proof search based on focused proof systems in which introduction rules are organized into two alternating (...) phases of rule application. Since the logic programming community can generate many examples and many design goals, the close connections with proof theory have helped to keep proof theory relevant to the general topic of computational logic. (shrink)

The past decade has seen an explosion of work on calculi of explicit substitutions. Numerous works have illustrated the usefulness of these calculi for practical notions like the implementation of typed functional programming languages and higher order proof assistants. It has also been shown that eta-reduction is useful for adapting substitution calculi for practical problems like higher order unification. This paper concentrates on rewrite rules for eta-reduction in three different styles of explicit substitution calculi: λσ, λse and the suspension calculus. (...) Both λσ and λse when extended with eta-reduction rules, have proved useful for solving higher order unification. We enlarge the suspension calculus with an adequate eta-reduction rule which we show to preserve termination and confluence of the associated substitution calculus and to correspond to the eta rules of the other two calculi. We prove that λσ and λse as well as λσ and the suspension calculus are non-comparable while λse is more adequate than the suspension calculus in simulating one-step beta-reduction. After defining the eta-reduction rule in the suspension calculus, and after comparing these three calculi of explicit substitutions , we then concentrate on the implementation of the rewrite rules of eta-reduction in these calculi. We note that it is usual practice when implementing the eta rule for substitution calculi, to mix isolated applications of eta-reduction with the application of other rules of the corresponding substitution calculi. The main disadvantage of this practice is that the eta rewrite rules so obtained are unclean because they have an operational semantics different from that of the eta-reduction of the λ-calculus. For the three calculi in question enlarged with adequate eta rules we show how to implement these eta rules. For the λse we build a clean implementation of the eta rule and we prove that it is not possible to follow the same approach for the λσ and λsusp. (shrink)