We work in a category where pullbacks exist. Given a map $f:Y\to Z$, there is a pull-back diagram $$\begin{array}& Y \times_Z Y & \stackrel{g_1}{\longrightarrow} & Y \\ \downarrow{g_2} & & \downarrow{f} \\ Y & \stackrel{f}{\longrightarrow} & Z \end{array}$$ (I am sorry for the awful formatting, is there a better solution?)

Of course, because of symmetry, we should have $g_1=g_2$. But category theory is a formal language, so I would like to have a formal proof of this, without hand-waving. I have not managed to find one so far, although I think that it should not be difficult... The farthest I've got is that there must be a unique map $\phi: Y \times_Z Y \to Y \times_Z Y$ such that $g_1 \phi = g_2 \phi = g_1$, because of the universal property of the pullback. Similarly, there is a unique $\psi$ such that $g_1\psi=g_2\psi=g_2$. But how do I continue from here?


It needn't be true in general that $g_1 = g_2$. In case you want to read more about such things, the morphisms $g_1$ and $g_2$ are called a kernel pair for $f$.

In the category of sets, for example, we can take $$Y \times_Z Y = \{ (a, b) \in Y \times Y \mid f(a) = f(b) \}$$ And then the functions $g_1 : Y \times_Z Y \to Y$ and $g_2 : Y \times_Z Y \to Y$ are given by projection onto the first and second coordinates, respectively. Then typically $g_1 \ne g_2$.

There are some cases where we might have $g_1 = g_2$. For example, if $f$ is a monomorphism, then since $f \circ g_1 = f \circ g_2$ it follows that $g_1 = g_2$. In fact, $f$ is a monomorphism if and only if $(\mathrm{id}_Y, \mathrm{id}_Y)$ is a kernel pair for $f$.


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.