I start with a thesis: the natural notion of equality is additional data (paths/morphisms), not a binary relation (the fact that they exist). So, in particular, with such a constructivization (replacing property $\to$ structure): * sets $\to$ $\infty$-groupoids * categories $\to$ $\infty$-categories At the same time, it is somewhat unsatisfactory that the concepts on the right have much more cumbersome, technical definitions. The natural answer to this would be: definitions are given in terms of sets i.e. from a 1-world perspective and one would expect the concept of a $\infty$-category to have a simple and natural definition in the $\infty$-world. I know this is an [important open problem in homotopy type theory](https://mathoverflow.net/questions/145770/how-do-you-define-infinity-1-categories-in-homotopy-type-theory), but homotopy type theory is the internal language of a fairly large class of $\infty$-categories (including all $\infty$-topoi anyway). Thus, it is poorer than the internal $\infty\text{-}\rm{Groupoid}$ language (the most expressive $\infty$-topos?). Questions 1. What is the description of the $\infty\text{-}\rm{Groupoid}$ internal language? 2. Is there a definition of $\infty$-category in this language? P.S. I don't mean that I see specific reasons why moving from HoTT to the internal language of $\infty\text{-}\rm{Groupoid}$ should help (on the contrary: in 1-world the concept of a category is interpreted in any finitely complete category, no advantages from the expressive means of toposes, much less $\rm{Set}$ is not), but I still can't be sure otherwise.