We contribute to the theory of (homotopy) colimits inside homotopy type theory. The heart of our work characterizes the connection between colimits in coslices of a universe, called coslice colimits, and colimits in the universe (i.e., ordinary colimits). To derive this characterization, we find an explicit construction of colimits in coslices that is tailored to reveal the connection. We use the construction to derive properties of colimits. Notably, we prove that the forgetful functor from a coslice creates colimits over trees. We also use the construction to examine how colimits interact with orthogonal factorization systems and with cohomology theories. As a consequence of their interaction with orthogonal factorization systems, all pointed colimits (special kinds of coslice colimits) preserve n-connectedness, which implies that higher groups are closed under colimits on directed graphs. We have formalized our main construction of the coslice colimit functor in Agda.

We present a novel formulation of universe polymorphism in dependent type theory in terms of monads on the category of strict partial orders, and a novel algebraic structure, displacement algebras, on top of which one can implement a generalized form of McBride’s “crude but effective stratification” scheme for lightweight universe polymorphism. We give some examples of exotic but consistent universe hierarchies, and prove that every universe hierarchy in our sense can be embedded in a displacement algebra and hence implemented via our generalization of McBride’s scheme. Many of our technical results are mechanized in Agda, and we have an OCaml library for universe levels based on displacement algebras, for use in proof assistant implementations.

Randomized property-based testing has gained much attention recently, but most frameworks stop short at polymorphic properties. Although Bernardy et al. have developed a theory to reduce a wide range of polymorphic properties to monomorphic ones, it relies upon ad-hoc embedding-projection pairs to massage the types into a particular form. This paper skips the embedding-projection pairs and presents a mechanical monomorphization for a general class of polymorphic functions, a step towards automatic testing for polymorphic properties. The calculation of suitable types for monomorphization turns out to be logarithm .

Abstract We present a cubical type theory based on the Cartesian cube category (faces, degeneracies, symmetries, diagonals, but no connections or reversal) with univalent universes, each containing Π, Σ, path, identity, natural number, boolean, suspension, and glue (equivalence extension) types. The type theory includes a syntactic description of a uniform Kan operation, along with judgmental equality rules defining the Kan operation on each type. The Kan operation uses both a different set of generating trivial cofibrations and a different set of generating cofibrations than the Cohen, Coquand, Huber, and Mörtberg (CCHM) model. Next, we describe a constructive model of this type theory in Cartesian cubical sets. We give a mechanized proof, using Agda as the internal language of cubical sets in the style introduced by Orton and Pitts, that glue, Π, Σ, path, identity, boolean, natural number, suspension types, and the universe itself are Kan in this model, and that the universe is univalent. An advantage of this formal approach is that our construction can also be interpreted in a range of other models, including cubical sets on the connections cube category and the De Morgan cube category, as used in the CCHM model, and bicubical sets, as used in directed type theory.

We present a development of cellular cohomology in homotopy type theory. Cohomology associates to each space a sequence of abelian groups capturing part of its structure, and has the advantage over homotopy groups in that these abelian groups of many common spaces are easier to compute. Cellular cohomology is a special kind of cohomology designed for cell complexes: these are built in stages by attaching spheres of progressively higher dimension, and cellular cohomology defines the groups out of the combinatorial description of how spheres are attached. Our main result is that for finite cell complexes, a wide class of cohomology theories (including the ones defined through Eilenberg-MacLane spaces) can be calculated via cellular cohomology. This result was formalized in the Agda proof assistant.

This is the third in a series of papers extending Martin-Löf's meaning explanations of dependent type theory to a Cartesian cubical realizability framework that accounts for higher-dimensional types. We extend this framework to include a cumulative hierarchy of univalent Kan universes of Kan types, exact equality and other pretypes lacking Kan structure, and a cumulative hierarchy of pretype universes. As in Parts I and II, the main result is a canonicity theorem stating that closed terms of boolean type evaluate to either true or false. This establishes the computational interpretation of Cartesian cubical higher type theory based on cubical programs equipped with a deterministic operational semantics.

We present a development of cellular cohomology in homotopy type theory. Cohomology associates to each space a sequence of abelian groups capturing part of its structure, and has the advantage over homotopy groups in that these abelian groups of many common spaces are easier to compute. Cellular cohomology is a special kind of cohomology designed for cell complexes: these are built in stages by attaching spheres of progressively higher dimension, and cellular cohomology defines the groups out of the combinatorial description of how spheres are attached. Our main result is that for finite cell complexes, a wide class of cohomology theories (including the ones defined through Eilenberg-MacLane spaces) can be calculated via cellular cohomology. This result was formalized in the Agda proof assistant.

Abstract The connection between polymorphic and dynamic typing was originally considered by Curry et al. (1972, Combinatory Logic , vol. ii) in the form of “polymorphic type assignment” for untyped λ-terms. Types are assigned after the fact to what is, in modern terminology, a dynamic language. Interest in type assignment was revitalized by the proposals of Bracha et al. (1998, OOPSLA) and Bank et al. (1997, POPL) to enrich Java with polymorphism (generics), which in turn sparked the development of other languages, such as Scala, with similar combinations of features. In such a setting, where the target language already has a monomorphic type system, it is desirable to compile polymorphism to dynamic typing in such a way that as much static typing as possible is preserved, relying on dynamics only insofar as genericity is actually required. The basic approach is to compile polymorphism using embeddings from each type into a universal “top” type, ${\mathbb{D}}$ , and partial projections that go in the other direction. This scheme is intuitively reasonable, and, indeed, has been used in practice many times. Proving its correctness, however, is non-trivial. This paper studies the compilation of System F to an extension of Moggi's computational meta-language with a dynamic type and shows how the compilation may be proved correct using a logical relation.

Homotopy type theory is a recent research area connecting type theory with homotopy theory by interpreting types as spaces. In particular, one can prove and mechanize type-theoretic analogues of homotopy-theoretic theorems, yielding "synthetic homotopy theory". Here we consider the Seifert-van Kampen theorem, which characterizes the loop structure of spaces obtained by gluing. This is useful in homotopy theory because many spaces are constructed by gluing, and the loop structure helps distinguish distinct spaces. The synthetic proof showcases many new characteristics of synthetic homotopy theory, such as the "encode-decode" method, enforced homotopy-invariance, and lack of underlying sets.

This paper contributes to recent investigations of the use of homotopy type theory to give machine-checked proofs of constructions from homotopy theory. We present a mechanized proof of a result called the Blakers-Massey connectivity theorem, which relates the higher-dimensional loop structures of two spaces sharing a common part (represented by a pushout type, which is a generalization of a disjoint sum type) to those of the common part itself. This theorem gives important information about the pushout type, and has a number of useful corollaries, including the Freudenthal suspension theorem, which was used in previous formalizations. The proof is more direct than existing ones that apply in general category-theoretic settings for homotopy theory, and its mechanization is concise and high-level, due to novel combinations of ideas from homotopy theory and from type theory.