We prove normalization for MTT, a general multimodal dependent type theory capable of expressing modal type theories for guarded recursion, internalized parametricity, and various other prototypical modal situations. We prove that deciding type checking and conversion in MTT can be reduced to deciding the equality of modalities in the underlying modal situation, immediately yielding a type checking algorithm for all instantiations of MTT in the literature. This proof uses a generalization of synthetic Tait computability – an abstract approach to gluing proofs – to account for modalities. This extension is based on MTT itself, so that this proof also constitutes a significant case study of MTT.

We present an intrinsic representation of type theory in the proof assistant Cubical Agda, inspired by Awodey’s natural models of type theory. The initial natural model is defined as quotient inductive-inductive-recursive types, leading us to a syntax accepted by Cubical Agda without using any transports, postulates, or custom rewrite rules. We formalise some meta-properties such as the standard model, normalisation by evaluation for typed terms, and strictification constructions. Since our formalisation is carried out using Cubical Agda’s native support for quotient inductive types, all our constructions compute at a reasonable speed. When we try to develop more sophisticated metatheory, however, the ‘transport hell’ problem reappears. Ultimately, it remains a considerable struggle to develop the metatheory of type theory using an intrinsic representation that lacks strict equations. The effort required is about the same whether or not the notion of natural model is used.

Various mechanisms are available for managing universe levels in proof assistants based on type theory. The Agda proof assistant implements a strong form of universe polymorphism in which universe levels are internalised as a type, making levels first-class objects and permitting higher-rank quantification via ordinary Π-types. We prove normalisation and decidability of equality and type-checking for a type theory with first-class universe levels inspired by Agda. We also show that level primitives can safely be erased in extracted programs. Our development is formalised in Agda itself and builds upon previous work which uses logical relations on extrinsically typed syntax.

Categorical gluing is a powerful technique for proving meta-theorems of type theories such as canonicity and normalization. Synthetic Tait Computability (STC) provides an abstract treatment of the complex gluing models by internalizing the gluing category into a modal dependent type theory with a phase distinction. This work presents a mechanization of STC in the Istari proof assistant. Istari is a Martin-Löf-style extensional type theory with equality reflection, which avoids much of the explicit transport reasoning typically found in intensional proof assistants. This work develops a reusable library for synthetic phase distinction, including modalities, extension types, and strict glue types, and applies it to two case studies: (1) a canonicity model for dependent type theory with dependent products and booleans with large elimination, and (2) a Kripke canonicity model for the cost-aware logical framework. Our results demonstrate that the core STC constructions can be formalized essentially verbatim in Istari, preserving the elegance of the on-paper arguments while ensuring machine-checked correctness.

The correctness of type-checking implementations for proof assistants based on dependent type theory relies on metatheoretical properties that ensure the decidability of typing, some of which require substantial logical strength. Recent mechanizations of such algorithms have highlighted the importance of separating the algorithmic components of the proof - often intricate but requiring relatively low logical strength - from the logical components, which depend on stronger metatheoretical properties, such as normalization or the injectivity of type constructors. In this work, we revisit the logical relations technique and show how it can be used to derive these metatheoretical properties in a direct and uniform way for a core dependent type theory featuring Π-types, N, ⊥ and a universe U. Our presentation yields a compact and conceptually simplified argument that isolates the logically strong reasoning from the algorithmic core. We argue that this approach scales smoothly to richer type theories, and demonstrate this by extending our construction to Exceptional Type Theory (ExcTT), obtaining the first mechanized canonicity proof for this theory.

The metatheory of dependent types has seen a lot of progress in recent years. In particular, the development of categorical gluing finally lets us work with semantic presentations of type theory (such as categories with families) to establish fundamental properties of type theory such as canonicity and normalisation. However, proofs by gluing have yet to reach the stage of computer formalisation: formal proofs for the metatheory of dependent types are still stuck in the age of tedious syntactic proofs. The main reason for this is that semantic presentations of type theory are defined using sophisticated indexed inductive types, which are prone to “transport hell”. In this paper, we introduce a new technique to work with CwFs in intensional type theory without getting stuck in transport hell. More specifically, we construct an alternative presentation of the initial CwF which encodes the substitutions as metatheoretical functions. This has the effect of strictifying all the equations that are involved in the substitution calculus, which greatly reduces the need for transports. As an application, we use our strictified initial CwF to give a short and elegant proof of canonicity for a type theory with dependent products and booleans with large elimination. The resulting proof is fully formalised in Agda.

Traditional category theory is typically based on set-theoretic principles and ideas, which are often non-constructive. An alternative approach to formalizing category theory is to use E-category theory, where hom sets become setoids. Our work reconsiders a third approach - P-category theory - from Čubrić et al. (1998) emphasizing a computational standpoint. We formalize in Rocq a modest library of P-category theory - where homs become subsetoids - and apply it to formalizing algorithms for normalization by evaluation which are purely categorical but, surprisingly, do not use neutral and normal terms. Čubrić et al. (1998) establish only a soundness correctness property by categorical means; here, we extend their work by providing a categorical proof also for a strong completeness property. For this we formalize the full universal property of the free Cartesian-closed category, which is not known to have been performed before. We further formalize a novel universal property of unquotiented simply typed lambda-calculus syntax and apply this to a proof of correctness of a categorical normalization by evaluation algorithm. We pair the overall mathematical development with a formalization in the Rocq proof assistant, following the principle that the formalization exists for practical computation. Indeed, it permits extraction of synthesized normalization programs that compute (long) beta-eta-normal forms of simply typed lambda-terms together with a derivation of beta-eta-conversion.

Metatheorems about type theories are often proven by interpreting the syntax into models constructed using categorical gluing. We propose to use only sconing (gluing along a global section functor) instead of general gluing. The sconing is performed internally to a presheaf category, and we recover the original glued model by externalization.

Our method relies on constructions involving two notions of models: first-order models (with explicit contexts) and higher-order models (without explicit contexts). Sconing turns a displayed higher-order model into a displayed first-order model.

Using these, we derive specialized induction principles for the syntax of type theory. The input of such an induction principle is a boilerplate-free description of its motives and methods, not mentioning contexts. The output is a section with computation rules specified in the same internal language. We illustrate our framework by proofs of canonicity and normalization for type theory.

This paper studies normalisation by evaluation for typed lambda calculus from a categorical and algebraic viewpoint. The first part of the paper analyses the lambda definability result of Jung and Tiuryn via Kripke logical relations and shows how it can be adapted to unify definability and normalisation, yielding an extensional normalisation result. In the second part of the paper, the analysis is refined further by considering intensional Kripke relations (in the form of Artin–Wraith glueing) and shown to provide a function for normalising terms, casting normalisation by evaluation in the context of categorical glueing. The technical development includes an algebraic treatment of the syntax and semantics of the typed lambda calculus that allows the definition of the normalisation function to be given within a simply typed metatheory. A normalisation-by-evaluation program in a dependently typed functional programming language is synthesised.

The implementation and semantics of dependent type theories can be studied in a syntax-independent way: the objective metatheory of dependent type theories exploits the universal properties of their syntactic categories to endow them with computational content, mathematical meaning, and practical implementation (normalization, type checking, elaboration). The semantic methods of the objective metatheory inform the design and implementation of correct-by-construction elaboration algorithms, promising a principled interface between real proof assistants and ideal mathematics. In this dissertation, I add synthetic Tait computability to the arsenal of the objective metatheorist. Synthetic Tait computability is a mathematical machine to reduce difficult problems of type theory and programming languages to trivial theorems of topos theory. First employed by Sterling and Harper to reconstruct the theory of program modules and their phase separated parametricity, synthetic Tait computability is deployed here to resolve the last major open question in the syntactic metatheory of cubical type theory: normalization of open terms.

We prove normalization for (univalent, Cartesian) cubical type theory, closing the last major open problem in the syntactic metatheory of cubical type theory. Our normalization result is reduction-free, in the sense of yielding a bijection between equivalence classes of terms in context and a tractable language of β/η-normal forms. As corollaries we obtain both decidability of judgmental equality and the injectivity of type constructors.

We present two proofs of coherence for cartesian closed bicategories. Precisely, we show that in the free cartesian closed bicategory on a set of objects there is at most one structural 2-cell between any parallel pair of 1-cells. We thereby reduce the difficulty of constructing structure in arbitrary cartesian closed bicategories to the level of 1-dimensional category theory. Our first proof follows a traditional approach using the Yoneda lemma. For the second proof, we adapt Fiore’s categorical analysis of normalisation-by-evaluation for the simply-typed lambda calculus. Modulo the construction of suitable bicategorical structures, the argument is not significantly more complex than its 1-categorical counterpart. It also opens the way for further proofs of coherence using (adaptations of) tools from categorical semantics.

The relationship between categorical gluing and proofs using the logical relation technique is folklore. In this paper we work out this relationship for Martin-Löf type theory and show that parametricity and canonicity arise as special cases of gluing. The input of gluing is two models of type theory and a pseudomorphism between them and the output is a displayed model over the first model. A pseudomorphism preserves the categorical structure strictly, the empty context and context extension up to isomorphism, and there are no conditions on preservation of type formers. We look at three examples of pseudomorphisms: the identity on the syntax, the interpretation into the set model and the global section functor. Gluing along these result in syntactic parametricity, semantic parametricity and canonicity, respectively.

We show how to solve the word problem for simply typed λβη-calculus by using a few well-known facts about categories of presheaves and the Yoneda embedding. The formal setting for these results is 𝒫-category theory, a version of ordinary category theory where each hom-set is equipped with a partial equivalence relation. The part of 𝒫-category theory we develop here is constructive and thus permits extraction of programs from proofs. It is important to stress that in our method we make no use of traditional proof-theoretic or rewriting techniques. To show the robustness of our method, we give an extended treatment for more general λ-theories in the Appendix.

Solves the decision problem for the simply typed lambda calculus with a strong binary sum, or, equivalently, the word problem for free Cartesian closed categories with binary co-products. Our method is based on the semantic technique known as “normalization by evaluation”, and involves inverting the interpretation of the syntax in a suitable sheaf model and, from this, extracting an appropriate unique normal form. There is no rewriting theory involved and the proof is completely constructive, allowing program extraction from the proof.