Semantics for Homotopy Type Theory

The main aim of my PhD thesis is to define a semantics for Homotopy type theory based on elementary categorical tools. This led us to extend the study of this system in other directions: we proved a Normalisation theorem, and defined a generic syntax. All those results are obtained for a subset of the whole Homotopy type theory, which we called 1-HoTT theories. A 1-HoTT theory is composed by Martin-Löf type theory with generic inductive types, the axioms of function extensionality and univalence, truncation and generic 1-higher inductive types, which are a subset of the higher inductive types in which the higher constructor of a type T is limited to the type =T . For those theories we obtained some proof theoretic results; the main one is a Normalisation theorem, following Girard's reducibility candidates technique. The semantics is sound and complete, with the completeness result following from the existence of a canonical model, which is also classifying. Our conjecture is that our proof theory and semantics can be extended to every single higher inductive type. The dissertation shows that a very large amount of higher inductive types can be analysed inside our framework: what prevents to extend the results is the lack of a systematic treatment of the syntax of the higher inductive types, which is still an open issue in Homotopy type theory.