I'm trying to learn some of the basics of category theory, and am stuck on this exercise (from Benjamin Pierce's text):

Show that in a poset considered as a category, the only equalizers are the identity arrows.

Here's an example of where I get confused:

Suppose we have a poset $P =(\{1,2\}, \leq)$ as a category, with arrows $1 \to 1$, $1 \to 2$, and $2 \to 2$. If this claim is true, then $e \colon 1 \to 2$ is not an equalizer of $f,g \colon 2 \to 2$. But how does it not fit the definition of an equalizer? We have $f \circ e = g \circ e$, and $e' \colon 1 \to 2$ is the only arrow such that both $f \circ e' = g \circ e'$ and there is a (unique) arrow $k \colon 1 \to 1$ with $e \circ k = e'$.

I figure I must be misunderstanding something simple, or have setup a bad example, and would really appreciate any help or pointers. Thanks!

(The text gives the following definition of an equalizer, which is where I got the names.)

An arrow $e \colon X \to A$ is an equalizer of a pair of arrows $f \colon A \to B$ and $g \colon A \to B$ if (1) $f \circ e = g \circ e$; and (2) whenever $e' \colon X' \to A$ satisfies $f \circ e' = g \circ e'$, there is a unique arrow $k \colon X' \to X$ such that $e \circ k = e'$.


The equalizer of a pair of identical arrows $f, f : a \to b$ in a category is (isomorphic via a unique isomorphism to) the identity arrow $\text{id}_a : a \to a$. In your situation, the identity arrow equalizes your arrows and it does not factor through $e$ (that is, take $e'$ to be the unique arrow from $2$ to $2$), so $e$ doesn't satisfy the universal property of the equalizer.

  • $\begingroup$ Thanks for such a quick answer! I'm trying to understand: does this mean that $1 \to 2$ is an equalizer but is isomorphic to $2 \to 2$? $\endgroup$ – all-too-human Dec 19 '12 at 7:41
  • $\begingroup$ No, it means that $1 \to 2$ is not an equalizer. I've edited to clarify this. $\endgroup$ – Qiaochu Yuan Dec 19 '12 at 7:44
  • $\begingroup$ Ah, thanks again, I understand now. (I think I had confused myself by just misreading the defn. of equalizer.) $\endgroup$ – all-too-human Dec 19 '12 at 7:50

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.