Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

The concept of "image" in Category Theory is, depending on source, defined in two possible ways: Either as a factorization of a morphism, or as the kernel of a cokernel.

More precisely, let some category with zero morphisms be given. If $f: A \to B$ is a morphism in that category, we can define the image of $f$ via factorization: $$ A \to I \underset{\operatorname{im}(f)}\hookrightarrow B = f$$ where $I$ is the smallest object with this property (this is the definition that Wikipedia and nLab give) or as the kernel of the cokernel: $$ \operatorname{im}(f) = \ker(\operatorname{coker}(f))$$

What I would like to understand is when and why these definitions are equivalent. Thus, my (two-part) question is:

  • Are these definitions equivalent in general, or only for Abelian categories?
  • Where can I find a simple proof of the equivalence?

I have seen a (very sketchy) proof in the Abelian case that I did not fully understand, and I could prove that the kernel-cokernel definition gives a factorization for any category with zero morphisms, but not that it is minimal.

share|cite|improve this question
They're not equivalent in general. It's true in an abelian category, but that's basically one of the axioms. – Zhen Lin Sep 27 '13 at 21:24
How do you define kernel and cokernel in a general category? – Thomas Andrews Sep 27 '13 at 21:25
@ThomasAndrews The non-additive version is "equaliser of the cokernel pair". – Zhen Lin Sep 27 '13 at 21:27
I use the definition that $f \circ \ker f = 0$ and if $f \circ k = 0$, there is a unique $g$ such that $g = \ker f \circ g$, similarily for cokernels. Note that I require my category to have zero morphisms. – Johannes Kloos Sep 27 '13 at 21:28
up vote 3 down vote accepted

The two concept aren't always equivalent. For example, in category of groups this is no longer true. In particular, the image of a group homomorphism is the kernel of its cokernel if and only if the image is a normal subgroup of the codomain.

You can prove easily that a kernel is always the kernel of its cokernel.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.