Is there a reasonable notion of an inner automorphism of a Hopf algebra $H$ which in the case of a group ring $H=\mathbb KG$ for a group $G$ reduces to a conjugation by $g\in G$?


I am not sure if the following is the kind of answer you are expecting, but take the (left) adjoint action $(ad_l h)\triangleright k=\sum h_1 kS(h_2)$ of a hopf algebra $H$ on itself.
(It is known that the left adjoint action of $H$ on itself is always inner (in the sense of the def. 6.1.1, p. 87, in Montgomery's book "Hopf algebras and their action on rings") and turns $H$ into a $H$-module algebra).
If you set $H=kG$, then, the left adjoint action for $g\in G$ reduces to action by conjugation: $(ad_l g)\triangleright k=gkg^{-1}$.

More generally, for any inner action $\ \triangleright : H\otimes A\rightarrow A$, it is easy to show that any grouplike element $g\in G(H)$, acts as an inner automorphism of $A$. In case $H=A=kG$ and $\triangleright$ is the left adjoint action then this inner automorphism is the action by conjugation (of the corresponding group element).

Concluding, i would say that the notion of an "inner action" (p.674-675) may be the reasonable notion you are looking for.

  • $\begingroup$ Many thanks! That is indeed on the spot. I was however hoping to generalize inner automorphisms beyond conjugation by grouplike elements and I got confused. By Montgomery's book, for $a,x\in H,$ $a\cdot x=\sum a_1xS(a_2)$ is a measuring and, hence, $a\cdot (xy)=\sum (a_1\cdot x)(a_2\cdot y)$. If $a$ is primitive then $a\cdot x=[a,x]=ax-xa$ (Eg. 6.1.6) implying $[a, xy]=\sum [a_1,x][a_2,y].$ But this is not true, since this sum is $[a,x] [1,y]+[1,x][a,y]=0.$ Where do I make a mistake? $\endgroup$ – Adam Nov 13 '17 at 0:07
  • $\begingroup$ Your claim is correct in that, if $a$ is primitive then $a\cdot x=[a,x]$. However, if $a$ is primitive then $a_1$ and $a_2$ cannot both be primitives. One of them is equal to $1$. Unlike the primitive case: $1\cdot x=x \neq [1,x]=0$. Consequently, your mistake lies at the following: $[a,xy]=a\cdot(xy)=\sum (a_1\cdot x)(a_2\cdot y)\neq \sum [a_1,x][a_2,y]$. $\endgroup$ – Konstantinos Kanakoglou Nov 13 '17 at 3:52
  • $\begingroup$ In fact: $[a,xy]=a\cdot(xy)=\sum (a_1\cdot x)(a_2\cdot y)=x[a,y]+[a,x]y$. $\endgroup$ – Konstantinos Kanakoglou Nov 13 '17 at 4:00
  • 1
    $\begingroup$ Oops, of course! :-) $\endgroup$ – Adam Nov 13 '17 at 4:36

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.