Part of Aluffi II.7.11 suggests proving this claim.

It is actually quite straightforward (especially after some earlier problems in the book): I've managed to prove it by noticing that for arbitrary $a, b, g$: $$g aba^{-1} b^{-1} g^{-1} = (gag^{-1})(gbg^{-1})(gag^{-1})^{-1}(gbg^{-1})^{-1}$$ which is in $[G, G]$, thus it's trivial to show that for arbitrary $n \in [G, G]$ the element $gng^{-1}$ is also in $[G, G]$ by interspersing it with some number of $g^{-1}g$'s.

My question, though, is a bit different (and that's what's different from all other questions about normality of commutators that MSE suggests as similar). Aluffi gives a hint that $g \cdot aba^{-1}b^{-1} \cdot g^{-1} = \gamma_g(aba^{-1}b^{-1})$, where $\gamma_g : a \mapsto gag^{-1}$ is an inner automorphism on $G$ defined by the element $g$. I don't think I've followed this hint and thus I haven't built a "more categorical" proof that Aluffi might have expected here. So, is there a way to prove the claim using this hint?

  • 3
    $\begingroup$ I think you have followed the hint. $\endgroup$ – arctic tern Jun 3 at 23:13
  • $\begingroup$ Aluffi's hint is, as explained in Berci's answer, just a different way of organizing your calculation. But this reorganization has real value, because it shows that the commutator subgroup is invariant under all automorphisms of $G$, not just the inner ones. (In other words, it shows that the commutator subgroup is a characteristic subgroup, which is a stronger property than normality.) $\endgroup$ – Andreas Blass Jun 4 at 2:16

You implicitly did follow the hint:

The point is that $\gamma_g$ is an inner automorphism, implying it's a group homomorphism, so that $$\gamma_g(xy) =\gamma_g(x)\gamma_g(y)$$ which was the crucial part in your proof, as well.


I think @Andreas Blass in his comment has the key to the "more categorical" approach.

The commutator can be viewed as a functor on the category of groups. The Wikipedia page is good on this.

It follows that the commutator subgroup is a fully characteristic subgroup,(ie, it is stable under any endomorphism).

Thus it is automatically a normal subgroup. This is because the normality condition can be phrased as invariance under inner automorphisms.

  • 1
    $\begingroup$ I believe the usual terminology is “fully invariant”, rather than “fully characteristic”; at least, that’s the one I normally encounter. (In fact, the commutator subgroup is even more than fully invariant: it is verbal, which is what yields the functoriality) $\endgroup$ – Arturo Magidin Jun 5 at 4:42

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.