I don't know why the following statement is true:

Theorem : In an abelian category, if the following diagram $\require{AMScd}$ \begin{CD} X' @>{f'}>> Y' \\ @V{g'}VV @VV{g}V\\ X @>>{f}> Y \end{CD} is a pullback and $f$ is an epimorphism, then the diagram is a pushout.

My attempt :

Assume that we are given the following commutative diagram: $\require{AMScd}$ \begin{CD} X' @>{f'}>> Y' \\ @V{g'}VV @VV{b}V\\ X @>>{a}> Z \end{CD} Since the following diagram $\require{AMScd}$ \begin{CD} \mathrm{ker}(f) @>{0}>> Y' \\ @V{i_f}VV @VV{g}V\\ X @>>{f}> Y \end{CD} is commutative, we have a morphism $k:\mathrm{ker}(f)\to X'$ such that $g'k=i_f$ and $f'k=0$.

Then, $ai_f=ag'k=bf'k=0b=0$. And since $f$ is epi, $(Y,f)$ is a cokernel of $i_f$. Then we can make $h:Y\to Z$ such that $hf=a$. But, how can i show that $hg=b$?


A commutative diagram like yours induces a complex $$X'\stackrel{\iota}\to X\oplus Y'\stackrel{\pi}\to Y$$ where $\iota$ is $(g',f')$ and $\pi$ is $(-f,g)$. The commutative square is a pullback iff $\iota=\ker\pi$ and a pushout iff $\pi=\text{coker}\,\iota$.

As you have a pullback, $$0\to X'\stackrel{\iota}\to X\oplus Y'\stackrel{\pi}\to Y$$ is an exact sequence. If $f$ is epi, so is $\pi$ and so we have a short exact sequence $$0\to X'\stackrel{\iota}\to X\oplus Y'\stackrel{\pi}\to Y\to0.$$ Then $\pi=\text{coker}\,\iota$: the commutative square is a pushout.

  • $\begingroup$ I checked all detail. I totally understood your answer. Thank you. $\endgroup$
    – PMJ
    Apr 29 '20 at 21:25

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.