Let $F:\mathcal{C}\to\mathcal{D}$ be a functor between two categories $\mathcal{C}$ and $\mathcal{D}$ where the notion of $F$ preserving and commuting limits makes sense. I am unable to understand the difference between those two things. So my question is as follows:

What is the difference between a functor that commutes with limits and a functor that preserves limits?

  • $\begingroup$ There is no difference. Both notions always make sense and are the same notion. $\endgroup$ – Qiaochu Yuan Feb 25 '15 at 5:04
  • 1
    $\begingroup$ @QiaochuYuan Oh, OK. Thanks! What about in the quasicategory case? Like what is the difference between "preserve small limits" and "commutes with finite limits"? $\endgroup$ – HiIloveMath Feb 25 '15 at 5:09
  • $\begingroup$ There is still no difference between "preserves" and "commutes with," but small limits are more general than finite limits. $\endgroup$ – Qiaochu Yuan Feb 25 '15 at 5:10
  • 5
    $\begingroup$ Because I wanted to make sure the OP was asking the question they intended to ask. $\endgroup$ – Qiaochu Yuan Feb 25 '15 at 5:47
  • 3
    $\begingroup$ Some people like to say "commutes", some people like to say "preserves". It's like saying a group homomorphism commutes with the group operation vs a group homomorphism preserves the group operation. $\endgroup$ – Zhen Lin Feb 25 '15 at 8:16

In the comments it was claimed that the notions are equivalent. But I claim that there is a subtle difference in the language.

If $F : \mathcal{C} \to \mathcal{D}$ is a functor, then we say that $F$ preserves limits if for every limit cone $(L \to X_i)$ in $\mathcal{C}$ its image $(F(L) \to F(X_i))$ is also a limit cone in $\mathcal{D}$. Notice that this definition makes sense even if $\mathcal{C}$ has not all limits, and it is not assumed a priori that the diagram $(F(X_i))$ in $\mathcal{D}$ has some limit. But this would be necessary to speak of "commutation". Specifically, if $F : \mathcal{C} \to \mathcal{D}$ is a functor, $\mathcal{C}$ has limits of shape $I$ and $\mathcal{D}$ too, then the phrase "$F$ commutes with limits of shape $I$" should mean that the canonical morphism $$F(\lim_i X_i) \to \lim_i F(X_i)$$ is an isomorphism for every $I$-shaped diagram $X$ in $\mathcal{C}$. For $F$ to commute with some operation, we need the existence of such operation.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.