Intensional Type Theory Research Articles

Interpreting descriptions in intensional type theory

AbstractNatural deduction systems with indefinite and definite descriptions (ε-terms and ι-terms) are presented, and interpreted in Martin-LÖf's intensional type theory. The interpretations are formalizations of ideas which are implicit in the literature of constructive mathematics: if we have proved that an element with a certain property exists, we speak of ‘the element such that the property holds’ and refer by that phrase to the element constructed in the existence proof. In particular, we deviate from the practice of interpreting descriptions by contextual definitions.

State Dependent IO-Monads in Type Theory

We introduce the notion of state dependent interactive programs for Martin-Löf Type Theory. These programs are elements of coalgebras of certain endofunctors on the presheaf category S→Set. We prove the existence of final coalgebras for these functors. This shows as well the consistency of type theory plus rules expressing the existence of weakly final coalgebras for these functors, which represents the type of interactive programs. We define in this type theory the bisimulation relation, and give some simple examples for interactive programs. A generalised monad operation is defined by corecursion on interactive programs with return value, and a generalised version of the monad laws for this operation is proved. The correctness of the monad laws has been verified in the theorem prover Agda which is based on intensional type theory.

Transitivity in coercive subtyping

Coercive subtyping is a general approach to abbreviation and subtyping in dependent type theories with inductive types. Coherence and admissibility of transitivity are important both for understanding of the framework and for its correct implementation. In this paper, we study the issue of transitivity in the context of subtyping for parameterised inductive types. In particular, we propose and study the notion of weak transitivity and show that, for a large class of parameterised inductive types, the natural subtyping rules are coherent and weak transitivity is admissible in an intensional type theory. A possible extension of type theory with certain extensional computation rules is also discussed for achieving admissibility of transitivity in general.

Constructive sets in computable sets

The original interpretation of the constructive set theoryCZF in Martin-Lof’s type theory uses the ‘extensional identity types’. It is generally believed that these ‘types’ do not belong to type theory. In this paper it will be shown that the interpretation goes through without indentity types. This paper will also show that the interpretation can be given in an intensional type theory. This reflects the computational nature of the interpretation. This computational aspect is reinforced by an ω-Set model ofCZF.

The univalence axiom for elegant Reedy presheaves

We show that Voevodsky's univalence axiom for intensional type theory is valid in categories of simplicial presheaves on elegant Reedy categories. In addition to diagrams on inverse categories, as considered in previous work of the author, this includes bisimplicial sets and $\Theta_n$-spaces. This has potential applications to the study of homotopical models for higher categories.