5
$\begingroup$

Let $G$ and $H$ are groups and $f:G\to H$ is group homomorphism. The kernel $f$ is defined by \begin{align*} \ker f =\{ g\in G \ : \ f(g)=1_{H}\}. \end{align*} And let $S$ and $T$ are Semigroups and $\varphi:S\to T$ is semigroup homomorphism. The kernel $\varphi$ is defined by \begin{align*} \ker \varphi=\{(s, s') \in S\times S \ : \ \varphi(s)=\varphi(s')\}. \end{align*} I know category of groups has a zero morphism, then the equalizer of zero morphism $0_{GH}:G\to H$ and $f:G\to H$ is kernel of $f$. But category of Semigroups doesn't has zero morphism.

I have two questions:

  1. What is the relation between these kernels.

  2. What is the definition of kernels in the category has no zero object or zero morphism.

Thank you

$\endgroup$
  • 2
    $\begingroup$ Well, if $\varphi$ is a group homomorphism, and writing $\ker_G(\varphi)$ for the group definition of kernel and $\ker_S(\varphi)$ for the semigroup definition, we have that $(s,s')\in\ker_S(\varphi)\Leftrightarrow s^{-1}s'\in\ker_G(\varphi)$. (In semigroups, $s^{-1}$ does not necessarily make sense.) Does that help with your question (1)? $\endgroup$ – user1729 Sep 24 '19 at 10:07
4
$\begingroup$

In general algebraic structures, the kernel of a homomorphism $f:X\to Y$ is defined just as for semigroups: $$\ker f:=\{(x, x') : f(x) =f(x')\}$$ which is always a congruence relation on $X$, i.e. an equivalence relation closed under the operations (meaning that it's a subalgebra of $X\times X$).

For groups [or rings or vector spaces, Boolean algebras, etc.], there is a one-to-one correspondence between congruence relations and normal subgroups [ideals, subspaces, Boolean ideals, etc], namely the equivalence class of the identity element [or, of $0$] already determines the whole relation.

In category theory, we can reflect this general notion of kernel by a kernel pair: a pair of arrows $k_1,k_2:K\to X$ (where $K$ plays the role of the congruence relation by the induced arrow $K\to X\times X$), which satisfy $f\circ k_1=f\circ k_2$, and whenever $f\circ u=f\circ v$ with $u,v:A\to X$, there's a unique $s:A\to K$ satisfying $u=k_1\circ s$ and $v=k_2\circ s$.

Working in a category of general algebras, $s$ is simply $A\ni\ a\mapsto (u(a),\, v(a))\ \in\ker f$.

$\endgroup$

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.