\(\lambda\)-presentable morphisms, injectivity and (weak) factorization systems (Q857748)

A morphism \(f \colon A \to B\) in a category \({\mathcal C}\) is called \(\lambda\)-presentable if it is a \(\lambda \)-presentable object of the comma category \((A \downarrow {\mathcal C})\). The author shows that every \(\lambda_m\)-injectivity class (i.e., the class of all the objects injective with respect to some class of \(\lambda\)-presentable morphisms) is a weakly reflective subcategory determined by a functorial weak factorization system cofibrantly generated by the class \(K\) of all the \(\lambda\)-presentable morphisms with respect to which it is injective. This means that the left-hand side of the weak factorization system can be obtained as retractions of transfinite compositions of elements of \(K\). This was essentially known for small-injectivity classes since at least \textit{P. Gabriel} and \textit{M. Zisman} [``Calculus of fractions and homotopy theory'', Ergeb. Math. Grenzgeb. 35 (1967; Zbl 0186.56802)], and referred to as the ``small object argument''. An analogous result is obtained for orthogonality classes and factorization systems, where the \(\lambda\)-filtered colimits play the role of the transfinite compositions in the injectivity case. \(\lambda\)-presentable morphisms are also used to organize and clarify some related results (and their proofs), in particular on the existence of enough injectives (resp., pure-injectives) in categories. Finally, locally \(\lambda\)-presentable categories are shown to have all their morphisms as a transfinite composition of \(\lambda\)-presentable ones, which implies that they are cellularly generated (by the set \(C_\lambda\) of their morphisms between \(\lambda\)-presentable objects).