Limits and colimits preservation under adjoint equivalence of categories

adjoint-functorscategory-theorylimits-colimits

If $F\colon\mathcal{A}\to\mathcal{B}$ is left adjoint to $U\colon\mathcal{B}\to\mathcal{A}$, then $U$ preserves limits and $F$ preserves colimits.

Can we say something more if $F$ is left adjoint to $U$ and they form an adjoint equivalence between $\mathcal{A}$ and $\mathcal{B}$?

My question arises from the following statement on Wikipedia:

The functor $H\colon\mathcal{I}\to\mathcal{C}$ has limit (or colimit)
$\mathcal{I}$ if and only if the functor
$FH\colon\mathcal{I}\to\mathcal{D}$ has limit (or colimit)
$F\mathcal{I}$.

where $F\colon\mathcal{C}\to\mathcal{D}$ is an equivalence. Is $F$ preserving both limits and colimits? (in the Wikipedia case, $F$ is not part of an adjoint equivalence, but, according to this question, every equivalence can be upgraded to an adjoint equivalence).

Best Answer

An adjoint equivalence between categories $\mathcal{A}$ and $\mathcal{B}$ is not a pair $(F,U)$ of functors between them, but a quadruplet $(F,U,\eta,\varepsilon)$, where $\eta$ and $\varepsilon$ are corresponding natural transformations, such that $(F,U,\eta,\varepsilon)$ is an adjunction. Note, that every pair $(F,U)$, where $F$ and $U$ are mutually inverse equivalences of categories, may be completed to such a quadruplet. Therefore, every equivalence of categories is both a left-adjoint and a right-adjoint functor.

Of course, equivalences of categories preserve much more constructions than left-adjoints or right-adjoints. Actually, they preserve "all" categorical information (the things not preserved by equivalences are half-jokingly called "evil"). In particular, equivalences preserve and reflect all limits and colimits (they preserve limits because they are right-adjoints, preserve colimits because they are left-adjoints, reflect limits and colimits because they are fully faithful).

Related Question