I am reading Paolo Aluffi's greatly entertaining book "Algebra: Chapter $0$" and I got stuck on some excercises dealing with universal properties.

Let $C$ be a category, and let $A$ and $B$ be two objects in $C$.

Define $C_{A,B}$ to be the category whose objects are morphisms in $C$ with targets $A$ and $B$ and whose morphisms are commutative diagrams in $C$: \begin{equation} Obj(C_{A,B}):=(Z,f,g)\qquad\text{with $Z\in Obj(C), f\in Hom_C(Z\rightarrow A), g\in Hom_C(Z\rightarrow B)$}\\ \sigma\in Hom_{C_{A,B}}((Z_1,f_1,g_1),(Z_2,f_2,g_2))\Longleftrightarrow f_1=f_2\circ\sigma\quad\text{and}\quad g_1=g_2\circ\sigma \end{equation}

In a similar way, define $C^{A,B}$ to be the category whose objects are morphisms in $C$ with sources $A$ and $B$ and whose morphisms are commutative diagrams in $C$: \begin{equation} Obj(C^{A,B}):=(Z,f,g)\qquad\text{with $Z\in Obj(C), f\in Hom_C(A\rightarrow Z), g\in Hom_C(B\rightarrow Z)$}\\ \sigma\in Hom_{C^{A,B}}((Z_1,f_1,g_1),(Z_2,f_2,g_2))\Longleftrightarrow f_2=\sigma\circ f_1\quad\text{and}\quad g_2=\sigma\circ g_1 \end{equation}

In both cases, $\sigma\in Hom_C(Z_1,Z_2)$.

I understand that $C$ has products if there exists a final object in $C_{A,B}$ denoted $A\times B$, and that $C$ has coproducts if there exists an initial object in $C^{A,B}$ denoted $A\amalg B$ (if $C=$ Set, $A\times B$ is the usual product of sets, and $A\amalg B$ is the disjoint union).

My question is: are there initial objects in $C_{A,B}$ and final objects in $C^{A,B}$? I am completely clueless on where to begin my search. Hints will be greatly appreciated!

  • $\begingroup$ There need not be. A category doesn't have to have initial or final objects. $\endgroup$
    – user67803
    Sep 29, 2014 at 6:28
  • $\begingroup$ I know- in fact I was asking if there were any... $\endgroup$ Sep 29, 2014 at 6:33
  • $\begingroup$ Why call having a final object having products, even in scare quotes? $\endgroup$ Sep 29, 2014 at 6:52
  • $\begingroup$ mmm...no reason, really- just following the (non standard, I imagine) wording of the author. $\endgroup$ Sep 29, 2014 at 6:56
  • $\begingroup$ No, that's not the wording Aluffi uses, unless you have an edition with a terrible typo-he says $C$ has products if $C_{A,B}$ has a final object, and similarly. The distinction is important! $\endgroup$ Sep 29, 2014 at 7:00

2 Answers 2


To clarify one mistake in your question: $C_{A,B}$ need not even have a final object-among sets this is the usual cartesian product, so if you take for instance some subcategory of sets not closed under cartesian products then there will be no final object.

You normally should not expect such diagrams to be significant. They exist, for instance, over sets, but they're all just the empty set with its empty maps to $A$ and $B$. The issue is that a $(Z,f,g)$ which factors into every $(Z',f',g')$ must be extremely small, heuristically, since $Z'$ can be arbitrarily small, so small that in reasonable cases it just becomes an initial object from $C$. I can't think off the top of my head of a non-artificial example of anything more interesting. The other question is dual, so equivalent.


The answer is: not necessarily. For example, let $C$ be a category with a final object $P$, and let $A = B = P$. Then $C_{A,B} = C$. Therefore $C_{A,B}$ will have an initial object if and only if $C$ does. For example, if $C$ is the category opposite to the category of nonzero rings, then $C$ has a final object $\mathbf{Z}$ but no initial one.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .