Let f:M→Nf:M\to N be a . The cokernel of ff is the

coker⁡(f)=N/im⁡(f),\operatorname{coker}(f)=N/\operatorname{im}(f),

where im⁡(f)\operatorname{im}(f) is the of ff. It comes with a canonical surjection N→coker⁡(f)N\to \operatorname{coker}(f), and one always has an exact tail

M→ f N→coker⁡(f)→0.M \xrightarrow{\,f\,} N \to \operatorname{coker}(f) \to 0.

Cokernels are the natural “targets” that make maps surjective by force, dual to how kernels make maps injective by force.

Examples
  • For f:Z→Zf:\mathbb Z\to\mathbb Z given by f(n)=knf(n)=kn with k≠0k\ne 0, one has coker⁡(f)≅Z/kZ\operatorname{coker}(f)\cong \mathbb Z/k\mathbb Z.
  • If i:N↪Mi:N\hookrightarrow M is the inclusion of a submodule, then coker⁡(i)≅M/N\operatorname{coker}(i)\cong M/N.
  • (Edge case) If ff is surjective, then coker⁡(f)=0\operatorname{coker}(f)=0.