It is known that a chain map $f_\bullet$ between chain complexes $(A_\bullet, d_{A,\bullet})$ and $(B_\bullet, d_{B,\bullet})$ induces a homomorphism $$(f_\bullet)_*: H_\bullet(A_\bullet)\to H_\bullet(B_\bullet).$$


The sources (Hatcher / wikipedia) I read do not explicitly mention how the homomorphism is induced, so I would like to confirm if my idea is correct?

For $\alpha+\text{Im}\,\partial_{A,n+1}\in H_n(A_\bullet)$,


And this works since $f_n$ maps cycles to cycles and boundaries to boundaries?

Thanks for any help.


You are exactly right. To elaborate, the fact that $f_n$ sends cycles to cycles means that we have a map (abusing notation slightly) $$f_n:\ker(d_n^A)\to\ker(d_n^B)$$ and the fact that $f_n$ sends boundaries to boundaries means that $f_n(\operatorname{Im}(d_{n+1}^A))\subseteq \operatorname{Im}(d_{n+1}^B)$. Hence the map $$\pi_n^B\circ f_n:\ker(d_n^A)\to H_n(B_{\bullet})$$ (where $\pi_n^B:\ker(d_n^B)\to H_n(B_{\bullet})$ is the projection) sends every element of $\operatorname{Im}(d_{n+1}^A)\subseteq \ker(d_n^A)$ to the identity in $H_n(B_{\bullet})$. Therefore, by the universal property of quotient groups, $\pi_n^B\circ f_n$ factors uniquely through $H_n(A_{\bullet})$: $$\pi_n^B\circ f_n:\quad \ker(d_n^A)\xrightarrow{\ \pi_n^A \ }H_n(A_{\bullet})\xrightarrow{\ f_* \ }H_n(B_{\bullet}).$$ Since $\pi_n^B\circ f_n = f_*\circ\pi_n^A$, we see that $$f_*(\alpha+\operatorname{Im}(d_{n+1}^A)) = f_n(\alpha)+\operatorname{Im}(d_{n+1}^B),$$ just as you claimed.

  • $\begingroup$ Nice. I learnt something new. $\endgroup$
    – yoyostein
    Jan 26 '17 at 1:35

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.