Recognizability in the Simply Typed Lambda-Calculus

摘要

We define a notion of recognizable sets of simply typed λ-terms that extends the notion of recognizable sets of strings or trees. This definition is based on finite models. Using intersection types, we generalize the notions of automata for strings and trees so as to grasp recognizability for λ-terms. We then expose the closure properties of this notion and present some of its applications.