This special issue of Mathematical Structures in Computer Science is dedicated to the theory and applications of explicit substitutions, which have attracted a growing community of researchers in the last decade, especially in the study of explicit substitutions as a means of bridging the gap between theory and practice in the implementation of programming languages, as well as theorem provers and proof checkers.

Such implementations typically rely on formal calculi defined using implicit substitution operations that are left at the meta-level, so that they need to turn these meta-level operations into efficient executable code, and this is often fairly intricate and distant from the formal calculi. This causes a significant gap between theory and practice.

Explicit substitutions considerably reduce this gap by bringing the meta-level operations down to the object-level calculus – where they are represented explicitly – allowing us in this way to give formal and robust models for the techniques actually used in implementations, and providing at the same time a more flexible tool for controlling the intermediate steps of evaluation.

All the papers in this issue were invited on the basis of their quality and relevance to the domain, and subjected to the refereeing process of MSCS. Most of them are substantially expanded and revised versions of work originally presented at Westapp'98 and Westapp'99, the first and second ‘Workshop on Explicit Substitutions: Theory and Applications to Programs and Proofs’, which were held in conjunction with RTA'98 in Tsukuba, Japan, and with Floc'99 in Trento, Italy, respectively.

As guest editor, I would like to express my warm thanks both to the authors, for their high-quality contributions to this special issue, and to the referees, whose scientific role was essential in improving the presentation of these contributions.

We define an extension of pure type systems with explicit substitution. We show that the type systems with explicit substitution are strongly normalizing iff their ordinary counterparts are. Subject reduction is shown to fail in general but a weaker, though still useful, subject reduction property is established. A more complicated extension is proposed for which subject reduction does hold in general.

We give a first-order presentation of higher-order logic based on explicit substitutions. This presentation is intentionally equivalent to the usual presentation of higher-order logic based on λ-calculus, that is, a proposition can be proved without the extensionality axioms in one theory if and only if it can be in the other. We show that the Extended Narrowing and Resolution first-order proof-search method can be applied to this theory. In this way we get a step-by-step simulation of higher-order resolution. Hence, expressing higher-order logic as a first-order theory and applying a first-order proof search method is a relevant alternative to a direct implementation. In particular, the well-studied improvements of proof search for first-order logic could be reused at no cost for higher-order automated deduction. Moreover, as we stay in a first-order setting, extensions, such as equational higher-order resolution, may be easier to handle.

We study perpetuality in the calculus of explicit substitutions λx. A reduction is called perpetual if it preserves the possibility of infinite reduction sequences. We then take a look at applications of this study: an inductive characterization of the λx-strongly normalizing terms, two perpetual reduction strategies for λx and finally a proof of strong normalization of a polymorphic lambda calculus with explicit substitutions Fes. To complete the study of Fes, the property of subject reduction is shown to hold by extending type assignments of the typing rules to allow non-pure types (types with possible occurrences of the type substitution operator).

We present a dependent-type system for a λ-calculus with explicit substitutions. In this system, meta-variables, as well as substitutions, are first-class objects. We show that the system enjoys properties like type uniqueness, subject reduction, soundness, confluence, and weak normalization.

We introduce a method to associate calculi of proof terms and rewrite rules with cut elimination procedures for logical deduction systems (i.e., Gentzen-style sequent calculi) in the case of intuitionistic logic. We illustrate this method using two different versions of the cut rule for a variant of the intuitionistic fragment of Kleene's logical deduction system G3.

Our systems are in fact calculi of explicit substitution, where the cut rule introduces an explicit substitution and the left-→ rule introduces a binding of the result of a function application. Cut propagation steps of cut elimination correspond to propagation of explicit substitutions, and propagation of weakening (to eliminate it) corresponds to propagation of index-updating operations. We prove various subject reduction, termination, and confluence properties for our calculi.

Our calculi improve on some earlier calculi for logical deduction systems in a number of ways. By using de Bruijn indices, our calculi qualify as first-order term rewriting systems (TRS's), allowing us to use correctly certain results for TRS's about termination. Unlike in some other calculi, each of our calculi has only one cut rule and we do not need unusual features of sequents.

We show that the substitution and index-updating mechanisms of our calculi work the same way as the substitution and index-updating mechanisms of Kamareddine and Ríos' λs and λt, two well-known systems of explicit substitution for the standard λ-calculus. By a change in the format of sequents, we obtain similar results for a known λ-calculus with variables and explicit substitutions, Rose's λbxgc.

Since Melliès showed that λσ (a calculus of explicit substitutions) does not preserve the strong normalization of the β-reduction, it has become a challenge to find a calculus satisfying the following properties: step-by-step simulation of the β-reduction, confluence on terms with metavariables, strong normalization of the calculus of substitutions and preservation of the strong normalization of the λ-calculus. We present here such a calculus. The main novelty of this calculus (given with de Bruijn indices) is the use of labels that represent updating functions and correspond to explicit weakening. A typed version is also presented.