I'm currently writing a comprehensive/encyclopedic set of notes on category theory, and one of the things I'm trying to do is gather all sorts of statements of the form

Let $F\colon\mathcal{C}\to\mathcal{D}$ be a functor. The following conditions are equivalent:

  1. For each $X\in\mathrm{Obj}(\mathsf{Cats})$, the precomposition functor $$F^* \colon \mathsf{Fun}(\mathcal{D},\mathcal{X}) \to \mathsf{Fun}(\mathcal{C},\mathcal{X})$$ is X for X some nice property (e.g. conservative).
  2. [Other conditions on $F$ itself]

or with postcomposition $$F_*\colon\mathsf{Fun}(\mathcal{X},\mathcal{C})\to\mathsf{Fun}(\mathcal{X},\mathcal{D})$$ replacing $F^*$.

So far, I've found characterisations for some nice properties, but there are many, many ones I still haven't found anything about.

Thus, I thought it might be a good idea to collect those into a single question here on MO, to keep track of all those:

Known characterisations (I'll link them here once the notes are ready):

  1. $F_*$ is full/faithful/fully faithful.
  2. $F^*$ is full/faithful/fully faithful (though see also MO 468121).
  3. $F^*$ is conservative.

Characterisations which I haven't yet found a reference for or worked out:

  1. $F_*$ is conservative.
  2. $F_*$ is essentially injective, i.e. $F\circ\phi\cong F\circ\psi$ implies $\phi\cong\psi$.
  3. $F^*$ is essentially injective, i.e. $\phi\circ F\cong\psi\circ F$ implies $\phi\cong\psi$.
  4. $F_*$ (resp. $F^*$) is essentially surjective.
  5. $F_*$ (resp. $F^*$) is a monomorphism (resp. an epimorphism).
  6. $F_*$ (resp. $F^*$) is pseudomonic (resp. pseudoepic).
  7. $F_*$ (resp. $F^*$) is dominant.

I'm also interested on the following two natural refinements of Items 2 and 3 above:

  • The functor $$F_*\colon\mathsf{Core}(\mathsf{Fun}(\mathcal{X},\mathcal{C}))\to\mathsf{Core}(\mathsf{Fun}(\mathcal{X},\mathcal{D}))$$ is full, i.e. if $F\circ\phi\cong F\circ\psi$ via a natural isomorphism $\beta$, then $\phi\cong\psi$ via a natural isomorphism $\alpha$ and we have $\beta=\mathrm{id}_{F}\star\alpha$.
  • The functor $$F^*\colon\mathsf{Core}(\mathsf{Fun}(\mathcal{D},\mathcal{X}))\to\mathsf{Core}(\mathsf{Fun}(\mathcal{C},\mathcal{X}))$$ is full, i.e. if $\phi\circ F\cong \psi\circ F$ via $\beta$, then $\phi\cong\psi$ via $\alpha$ and $\beta=\alpha\star\mathrm{id}_{F}$.

A few of these seem hard enough that probably no answer is currently known (or way too hard, like epimorphisms), but I suspect some of these are things that have been worked out already somewhere or are possibly folklore among the correct experts, but so far I haven't had much luck procuring references or statements for them.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.