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

Let $A\xrightarrow{S}C\xleftarrow{T}B$ be a diagram of categories; the classical definition of the comma category $(S\downarrow T)$ leads to the definition of a functor $$ (S\downarrow-)\colon [B,C]\to\bf Cat $$ sending $T\mapsto (S\downarrow T)$.

Now I wonder if this functor admits a left adjoint: does it commute with limits?

share|cite|improve this question
up vote 3 down vote accepted

In general, no. Here is an example:

  • Let $A = B = \{\ast\}$, the category with one object and one morphism (the identity of that object), $C = \mathbb Z/2$, the category with one object ($\ast'$) and $\hom(\ast', \ast') = \mathbb Z/2$.
  • Check that $[\{\ast\}, \mathbb Z/2] = \mathbb Z/2$ (there is a unique functor that sends $\ast \mapsto \ast'$ and a natural transformation from this functor to itself corresponds to an element of $\mathbb Z/2$ by taking the component at $\ast$).
  • Let $S\colon A \to C$ and $T\colon B \to C$ be that unique functor. Check that $S \downarrow T = \mathrm{Set}(\mathbb Z/2)$, the category with objects $\mathbb Z/2$ and only identity morphisms.

Now if $F\colon\mathbf{Cat} \to [B, C] = \mathbb Z/2$ is a left adjoint that means that for every category $X$ and object $t \in C$ there is a bijection $$\mathbb Z/2 = \hom_{\mathbb Z/2}(F(X), T) \simeq \hom_\mathbf{Cat}(X, \mathrm{Set}(\mathbb Z/2))$$ But if we choose $X = \mathrm{Set}(\{1, 2\})$ then $\hom_\mathbf{Cat}(X, \mathrm{Set}(\mathbb Z/2))$ is just the set maps $\{1, 2\} \to \mathbb Z/2$, and there are $4$ of these, not $2$.

share|cite|improve this answer
Sorry but your proof does not hold. You make 2 mistakes: 1) $[A, C] = [B, C]= C^1$ is not $\mathrm{Set}(\mathrm{ob} \ C)$ , it is exactly the category $C$ itself. 2) $s\downarrow t= Hom(s, t)$ is the discrete category whose objects are the morphisms from s to t and only has the identity morphisms. When you then rewrite the adjoint iso, you see that it actually holds if we set $F(X) = s$. The category $\mathrm{Arr}(s, t))$ that you mention is actually $C\downarrow C$ also known as $C^2$. – magma Sep 18 '13 at 18:36
@magma: Thanks for catching that, I agree I messed up $[A, C]$ and $S \downarrow T$. But $F(X) = s$ does not make the adjoint iso hold because the right hom set still depends on the choice of $X$ whereas the left does not. I've edited my answer to fix those two mistakes. – Jim Sep 19 '13 at 6:59

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.