Suppose that $A, B$ are objects of a category with all finite products and exponentials, $\mathbb{C}$. Show that there is a bijection between hom$(A, B)$ and hom$(1, B^A)$ where $1$ is a terminal in $\mathbb{C}$.

My attempt: Consider the isomorphism given by transposition, hom$(1 \times A, B) \cong$ hom$(1, B^A)$. It suffices to show that hom$(A, B) \cong$ hom$(1 \times A, B)$. Since $1$ is terminal, there is unique morphism from any object $Z$ to $1$, $k_Z$. Consider $g \in $ hom$(A, B)$. We can define an isomorphism by $g \mapsto k_A \times g$. This must be an isomorphism since $k_A$ is unique.

This seems correct to me but I'm a little doubtful.


Your idea is right, but I don't think your proposed isomorphism $\newcommand\Hom{\operatorname{Hom}}\Hom(A,B)\simeq \Hom(1\times A,B)$ makes any sense at all.

I.e., you propose sending $g : A\to B$ to $k_A\times g : 1\times A \to A\times B$, which is not an element of $\Hom(1\times A,B)$.

Instead, it suffices to show that $A\simeq 1\times A$, since then $\Hom(A,B)\simeq \Hom(1\times A,B)$.

For this, we use the Yoneda lemma. Since we have $$\Hom(Z,1\times A)\simeq \Hom(Z,1)\times \Hom(Z,A) \simeq \{*\}\times \Hom(Z,A)\simeq \Hom(Z,A),$$ the Yoneda lemma tells us that $1\times A\simeq A$.


If you're not familiar with the Yoneda lemma, Matematleta points out that you can also see $A\simeq 1\times A$ by noticing that $(A,(k_A,\textrm{id}_A))$ is also a product for $1$ and $A$. (If $f:Z\to 1$ and $g:Z\to A$, then $g$ is the unique map from $Z$ to $A$ such that $k_A\circ g = f = k_Z$, and $\textrm{id}_A\circ g = g$).

  • $\begingroup$ Isn't this overkill? $A\simeq 1\times A$ simply because $(A,\langle id_A,!\rangle)$ is also a product. $\endgroup$ – Matematleta Mar 20 '19 at 2:28
  • $\begingroup$ @Matematleta Sure, that also works, I just personally find this more immediately clear. But that's also quite clear and nice. I think I just generally like to think of products as representing functors in the first place. $\endgroup$ – jgon Mar 20 '19 at 2:33
  • 2
    $\begingroup$ I only pointed this out because from the OP's attempt, it seemed to me that he/she hasn't seen Yoneda yet. No problem with your answer (which I was happy to upvote!), I think of products that way too, --- $\endgroup$ – Matematleta Mar 20 '19 at 2:37
  • $\begingroup$ @Matematleta Hm, good point, I'll edit that in as a second viewpoint. $\endgroup$ – jgon Mar 20 '19 at 2:39
  • 1
    $\begingroup$ @Matematleta Thanks! I was trying to read up on Yoneda lemma, but the isomorphism you provided is cleaner. $\endgroup$ – real_father Mar 20 '19 at 2:58

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.