Proposition 1.6.15. Quasicategories provide a model of the axioms for \(\infty \)-categories stated above.
Proof sketch. We describe the interpretation of our axioms in the theory of quasicategories. Functors are maps of simplicial sets, whose composition is strictly unital and associative, and animae are modeled by Kan complexes. For functors \(f,g\colon C\to D\), their identification anima is modeled by the mapping anima from \(f\) to \(g\) in the maximal Kan complex of \(\Fun (C,D)\). Composition and whiskering of natural isomorphisms supply the operations on identifications and their coherence laws. The nerve from Exercise 1.6.12 supplies the classical 1-categories.
The categorical constructions used in our axioms exist in this model and satisfy the stated universal properties, including those on identification animae. In particular, our pullbacks are modeled by homotopy pullbacks of quasicategories. The representability axiom identifies the anima of identifications between two objects with their homotopy pullback, while the Rezk axiom expresses the equivalence between a quasicategory and its \(\infty \)-category of invertible morphisms.
The remaining recognition statements are non-trivial results of quasicategory theory. A quasicategory is a Kan complex if and only if all its morphisms are invertible; see [Land (2021), Corollary 2.1.12]. This result, as well as the criteria for equivalences and invertible natural transformations in Axiom J, Axiom K, is due to Joyal [Joyal (2008)]; for the latter two, see also [Land (2021), Theorems 2.2.1 and 2.3.20]. We refer to these sources for the model-specific results; the rest of the book uses the axioms directly. โก