Let $G$ be a group and $f_g : G \to G, f_g(h) = ghg^{-1}$ be conjugation by $g$. Let $M = \{f_g|g\in G\}$ be the set of all conjugations.

Given $t \in M$, what can be said about "extracting" $g$? (That is finding the element $g$ such that $t(g)(h) = ghg^{-1}$.)

So far I have convinced myself of the following:

  • Let $f : G \to M, f(g) = f_g$. Then $f$ is a homomorphism and I am looking for $f^{-1}$ (the inverse of $f$).
  • I suspect the inverse does not exist in general but might exist for some groups. I wonder if something more could be said though.
  • In the case of $\tilde{f}_g(h)=gh$, I could "extract" $g$ with the neutral element $e$, since $\tilde{f}_g(e) = g$, similarly for $g^{-1}$ one has $f_g(g^{-1}) = g^{-1}$, that is if I have a way of "testing" all elements of $G$ I could identify $g$ by searching for a fixpoint (?!).
  • 3
    $\begingroup$ $M$ is the group ${\rm Inn}(G)$ of inner automorphisms of $G$, and the map $f$ has kernel the centre $Z(G)$ of $G$, so $G/Z(G) \cong {\rm Inn}(G)$. $\endgroup$ – Derek Holt Mar 9 at 9:45

You cannot find $g$ from $f_g$ in general because if $g$ is in the center of $G$, so $gh=hg$ for all $h\in G$, then you have that $f_g(h)=ghg^{-1}=hgg^{-1}=h$ so $f_g=\textrm{Id}_G$ and the identity of $G$ has not information about the element $g$ that induces $f_g$.

In general you cannot find $g$ also if $g$ is not in the center of $G$ because, for example, if you consider $r\in Z(G)$ then $f_g=f_{gr}$ so $g$ can be identified up to elements of $Z(G)$.

If $g,r\in G$ are such that $f_g=f_r$ then you can prove that $g^{-1}r\in Z(G)$ so you have that $M$ can be identify as the quotient $G/Z(G)$


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.