We study a first-order functional language with the novel combination of the ideas of refinement type (the subset of a type to satisfy a Boolean expression) and type-test (a Boolean expression testing whether a value belongs to a type). Our core calculus can express a rich variety of typing idioms; for example, intersection, union, negation, singleton, nullable, variant, and algebraic types are all derivable. We formulate a semantics in which expressions denote terms, and types are interpreted as first-order logic formulas. Subtyping is defined as valid implication between the semantics of types. The formulas are interpreted in a specific model that we axiomatize using standard first-order theories. On this basis, we present a novel type-checking algorithm able to eliminate many dynamic tests and to detect many errors statically. The key idea is to rely on an SMT solver to compute subtyping efficiently. Moreover, interpreting types as formulas allows us to call the SMT solver at run-time to compute instances of types.

We define a dependent programming language in which programmers can define and compute with domain-specific logics, such as an access-control logic that statically prevents unauthorized access to controlled resources. Our language permits programmers to define logics using the LF logical framework, whose notion of binding and scope facilitates the representation of the consequence relation of a logic, and to compute with logics by writing functional programs over LF terms. These functional programs can be used to compute values at run-time, and also to compute types at compiletime. In previous work, we studied a simply-typed framework for representing and computing with variable binding [LICS 2008]. In this paper, we generalize our previous type theory to account for dependently typed inference rules, which are necessary to adequately represent domain-specific logics, and we present examples of using our type theory for certified software and mechanized metatheory.

Bidirectional typechecking has become popular in advanced type systems because it works in many situations where inference is undecidable. In this paper, I show how to cleanly handle parametric polymorphism in a bidirectional setting. The key contribution is a bidirectional type system for a subset of ML that supports first-class (higher-rank and even impredicative) polymorphism, and is complete for predicative polymorphism (including ML-style polymorphism and higher-rank polymorphism). The system’s power comes from bidirectionality combined with a “greedy ” method of finding polymorphic instances inspired by Cardelli’s early work on System F<:. This work demonstrates that bidirectionality is a good foundation for traditionally vexing features like first-class polymorphism.

This paper is part of a line of work on using the logical techniques of polarity and focusing to design a dependent programming language, with particular emphasis on programming with deductive systems such as programming languages and proof theories. Polarity emphasizes the distinction between positive types, which classify data, and negative types, which classify computation. In previous work, we showed how to use Zeilberger’s higher-order formulation of focusing to integrate a positive function space for representing variable binding, an essential tool for specifying logical systems, with a standard negative computational function space. However, our previous work considers only a simply-typed language. The central technical contribution of the present paper is to extend higher-order focusing with a form of dependency that we call positively dependent types: We allow dependency on positive data, but not negative computation, and we present the syntax of dependent pair and function types using an iterated inductive definition, mapping positive data to types, which gives an account of type-level computation. We construct our language inside the dependently typed programming language Agda 2, making essential use of coinductive types and induction-recursion.

Church’s formulation of the simple theory of types [1] has the pleasing property that every well-formed term has a unique type. The type is thus an intrinsic attribute of a term. Furthermore, we can restrict attention to wellformed terms as the only meaningful ones, because the property of being

The logical framework LF and its implementation as the Twelf metalogic provide both a practical system and a proven methodology for representing deductive systems and their metatheory in a machine-checkable way. An extension of LF with refinement types provides a convenient means for representing certain kinds of judgemental inclusions in an intrinsic manner. I propose to carry out such an extension in full, adapting as much of the Twelf metatheory engine as possible to the new system, and I intend to argue that the extension is both useful and practical. 1

Abstract. Refinement types sharpen systems of simple and dependent types by offering expressive means to more precisely classify well-typed terms. Proof irrelevance provides a mechanism for selectively hiding the identities of terms in type theories. In this paper, we show that refinement types can be interpreted as predicates using proof irrelevance in the context of the logical framework LF, establishing a uniform relationship between two previously studied concepts in type theory. The interpretation and its correctness proof are surprisingly complex, lending credence to the idea that refinement types are a fundamental construct rather than just a convenient surface syntax for certain uses of proof irrelevance. 1

This paper is part of a line of work on using the logical techniques of polarity and focusing to design a dependent programming language, with particular emphasis on programming with deductive systems such as programming languages and proof theories. Polarity emphasizes the distinction between positive types, which classify data, and negative types, which classify computation. In previous work, we showed how to use Zeilberger’s higher-order formulation of focusing to integrate a positive function space for representing variable binding, an essential tool for specifying logical systems, with a standard negative computational function space. However, our previous work considers only a simply-typed language. The central technical contribution of the present paper is to extend higher-order focusing with a form of dependency that we call positively dependent types: We allow dependency on positive data, but not negative computation. Additionally, we present the syntax of dependent pair and function types using an iterated inductive definition, mapping positive data to types, which gives an account of type-level computation. We construct our language inside the dependently typed programming language Agda 2, making essential use of coinductive types and induction-recursion.

Abstract. Refinement types sharpen systems of simple and dependent types by offering expressive means to more precisely classify well-typed terms. We present a system of refinement types for LF in the style of recent formulations where only canonical forms are well-typed. Both the usual LF rules and the rules for type refinements are bidirectional, leading to a straightforward proof of decidability of typechecking even in the presence of intersection types. Because we insist on canonical forms, structural rules for subtyping can now be derived rather than being assumed as primitive. We illustrate the expressive power of our system with examples and validate its design by demonstrating a precise correspondence with traditional presentations of subtyping. Proof irrelevance provides a mechanism for selectively hiding the identities of terms in type theories. We show that LF refinement types can be interpreted as predicates using proof irrelevance, establishing a uniform relationship between two previously studied concepts in type theory. The interpretation and its correctness proof are surprisingly complex, lending support to the claim that refinement types are a fundamental construct rather than just a convenient surface syntax for certain uses of proof irrelevance. 1.