In mathematics, a pseudofunctor F is a mapping from a category to the category Cat of (small) categories that is just like a functor except that F ( f ∘ g ) = F ( f ) ∘ F ( g ) {\displaystyle F(f\circ g)=F(f)\circ F(g)} and F ( 1 ) = 1 {\displaystyle F(1)=1} do not hold as exact equalities but only up to coherent isomorphisms. A typical example is an assignment to each pullback F f = f ∗ {\displaystyle Ff=f^{*}} , which is a contravariant pseudofunctor since, for example for a quasi-coherent sheaf F {\displaystyle {\mathcal {F}}} , we only have:
( g ∘ f ) ∗ F ≃ f ∗ g ∗ F . {\displaystyle (g\circ f)^{*}{\mathcal {F}}\simeq f^{*}g^{*}{\mathcal {F}}.}
Since Cat is a 2-category, more generally, one can also consider a pseudofunctor between 2-categories, where coherent isomorphisms are given as invertible 2-morphisms. The Grothendieck construction associates to a contravariant pseudofunctor a fibered category, and conversely, each fibered category is induced by some contravariant pseudofunctor. Because of this, a contravariant pseudofunctor, which is a category-valued presheaf, is often also called a prestack (a stack minus effective descent).
Definition A pseudofunctor F from a category C to Cat consists of the following data
a category F ( x ) {\displaystyle F(x)} for each object x in C, a functor F f {\displaystyle Ff} for each morphism f in C, a set of coherent isomorphisms for the identities and the compositions; namely, the invertible natural transformations
F ( f ∘ g ) ≃ F f ∘ F g {\displaystyle F(f\circ g)\simeq Ff\circ Fg} ,
F ( id x ) ≃ id F ( x ) {\displaystyle F(\operatorname {id} _{x})\simeq \operatorname {id} _{F(x)}} for each object x such that
F ( f g h ) → ∼ F ( f g ) F h → ∼ F f F g F h {\displaystyle F(fgh){\overset {\sim }{\to }}F(fg)Fh{\overset {\sim }{\to }}FfFgFh} is the same as F ( f g h ) → ∼ F f F ( g h ) → ∼ F f F g F h {\displaystyle F(fgh){\overset {\sim }{\to }}FfF(gh){\overset {\sim }{\to }}FfFgFh} ,
F ( id x ) ∘ F f → ∼ F ( id x ∘ f ) = F f {\displaystyle F(\operatorname {id} _{x})\circ Ff{\overset {\sim }{\to }}F(\operatorname {id} _{x}\circ f)=Ff} is the same as F ( id x ) ∘ F f ≃ id F ( x ) ∘ F f = F f {\displaystyle F(\operatorname {id} _{x})\circ Ff\simeq \operatorname {id} _{F(x)}\circ Ff=Ff} , and similarly for F f ∘ F ( id x ) {\displaystyle Ff\circ F(\operatorname {id} _{x})} .
Higher category interpretation The notion of a pseudofunctor is more efficiently handled in the language of higher category theory. Namely, given an ordinary category C, we have the functor category as the ∞-category
Fct ( C , Cat ) . {\displaystyle {\textbf {Fct}}(C,{\textbf {Cat}}).}
Each pseudofunctor C → Cat {\displaystyle C\to {\textbf {Cat}}} belongs to the above, roughly because in an ∞-category, a composition is only required to hold weakly, and conversely (since a 2-morphism is invertible).
See also Lax functor
References
External links http://ncatlab.org/nlab/show/pseudofunctor
