Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Between two categories $\mathcal{C}$ and $\mathcal{D}$ there is a functor category $Fun(\mathcal{C},\mathcal{D})$ with functors as objects and natural transformations as morphisms.

Is there a universal property that can be used to define a functor category (up to equivalence)?

share|cite|improve this question
What do you mean? Inside which category would you consider this universal property? – Berci Jan 18 '13 at 13:00
up vote 3 down vote accepted

Here is a functor category that has a 2-universal property:

Theorem. Let $\mathbb{C}$ be a small category and let $h_\bullet : \mathbb{C} \to [\mathbb{C}^\textrm{op}, \textbf{Set}]$ be the Yoneda embedding. Then, for all locally small and cocomplete categories $\mathcal{E}$, the functor $F \mapsto F h_\bullet$ from the category of cocontinuous functors $[\mathbb{C}^\textrm{op}, \textbf{Set}] \to \mathcal{E}$ to the category of all functors $\mathbb{C} \to \mathcal{E}$ is fully faithful and essentially surjective on objects, and this functor is pseudonatural in $\mathcal{E}$. In other words, $[\mathbb{C}^\textrm{op}, \textbf{Set}]$ is the free cocompletion of $\mathbb{C}$.

But more generally, if $\mathbb{C}$ and $\mathbb{D}$ are both small categories, then $[\mathbb{C}, \mathbb{D}]$ has a 1-universal property:

Theorem. There is a bijection between functors $\mathbb{E} \times \mathbb{C} \to \mathbb{D}$ and functors $\mathbb{E} \to [\mathbb{C}, \mathbb{D}]$ and this bijection is natural in $\mathbb{C}, \mathbb{D}, \mathbb{E}$. In other words, the category of small categories is cartesian closed and the functor category is an exponential object.

share|cite|improve this answer
Thanks, do you have any reference on the second theorem? – Michael Scarn Jan 18 '13 at 13:10
No, but it's a straightforward calculation and analogous to the one for $\textbf{Set}$. – Zhen Lin Jan 18 '13 at 14:16
IIRC this shows an argument for the second point (which really helped me grok why are natural transformations defined that way). Server is not responding at the moment, so I can't check :/ – askyle May 29 '14 at 15:24
OK, the link works now. Check section 7.6 (page 16). – askyle May 29 '14 at 15:49

Any category can be a functor category, because let $\mathcal C:=\{1\}$ the trivial category with one object, then $Fun(\mathcal C,\mathcal D)\simeq \mathcal D$.

share|cite|improve this answer
Sorry, I don't get your point. – Michael Scarn Jan 18 '13 at 13:10
I didn't get yours either :) – Berci Jan 18 '13 at 20:51

Amplifying Zhen's answer. One may show, that $[\mathbb{C}, \mathbb{D}]$ satisfies the 2-universal property, i.e. there exists a 2-natural isomorphism of categories: $$\hom(\mathbb{E} \times \mathbb{C}, \mathbb{D}) \approx \hom(\mathbb{E}, [\mathbb{C}, \mathbb{D}])$$ making $\mathbf{Cat}$ a 2-cartesian closed category. Moreover this structure is the unique 2-closed symmetric monoidal structure on $\mathbf{Cat}$.

share|cite|improve this answer
The amusing thing is that this can be proved using the 1-Yoneda lemma knowing only the 1-universal property. – Zhen Lin Jan 18 '13 at 15:24
Can you elaborate on that a little? – Berci Jan 18 '13 at 20:53
@Berci, every true statement can be proved using any assumptions (responding in line with your answer to the above question). However, I guess there is something more in Zhen's comment. Every cartesian closed category may be thought as enriched over itself. By definition, a 2-category is a category enriched over cartesian structure in $\mathbf{Cat}$. So, particularly, the 2-dimiensional structure on $\mathbf{Cat}$ is induced by its enrichment over itself. (cont) – Michal R. Przybylek Jan 19 '13 at 12:37
This means that $\hom(\mathbb{A}, \mathbb{B})$ is really $[\mathbb{A}, \mathbb{B}]$ and $\hom(\mathbb{E} \times \mathbb{C}, \mathbb{D}) \approx \hom(\mathbb{E}, [\mathbb{C}, \mathbb{D}])$ turns into $[\mathbb{E} \times \mathbb{C}, \mathbb{D}] \approx [\mathbb{E}, [\mathbb{C}, \mathbb{D}]]$. The last condition is trivially true in any cartesian closed category by ordinary (1-dimensional) Yoneda lemma. I guess, by a similar argument, one may also show that in any cartesian closed category these isomorphisms are enriched natural, however, I have not checked all relevant detail. – Michal R. Przybylek Jan 19 '13 at 12:38

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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