I am trying to show that $[A,B^n] = nB^{n-1}[A,B]$ where A and B are two Hermitian operators that commute with their commutator. However, I am running into a little problem and would like a hint of how to proceed.

If A and B commute then $[A,B] = ABA^{-1}B^{-1} = e$ where e is the identity element of the group.

$$\therefore AB=BA$$

$$n=1; [A,B^1] = (1)B^0[A,B] = e$$ This statement is certainly true. however moving on to $n = 2$ I find...

$$[A,B^2] = AB^2A^{-1}B^{-2} = ABBA^{-1}B^{-1}B^{-1} = BBAA^{-1}B^{-1}B^{-1}$$

Where in the last step I have used the fact that A and B commute to rearange the terms. However, it is plain to see that this last term simply reduces to the identity as well and for the n = 2 case we have:

$$[A,B^2] = e \ne 2B[A,B] = 2Be = 2B$$

Clearly I have assumed something I shouldn't have. The fact that there is a multiplicative factor of n implies I should be adding things, but I thought if I kept it as general as possible, the answer should just fall out naturally. I don't want an answer please, only guidance.

  • 6
    $\begingroup$ I'm confused. Isn't $[A,B] = AB - BA$? $\endgroup$ – Prahar Sep 22 '13 at 19:27

It seems that the question (v1) is caused by the fact that there are two different notions of the commutator:

  1. One for group theory: $$\tag{1} [A,B] ~:=~ ABA^{-1}B^{-1}$$ (or sometimes $[A,B] := A^{-1}B^{-1}AB$, depending on convention), which is relatively seldom used in physics.

  2. One for rings/associative algebras: $$\tag{2} [A,B]:=AB-BA,$$ which is the definition usually used in physics. (This latter definition (2) generalizes to a supercommutator in superalgebras.)

The identity

$$\tag{*} [A,B^n] ~=~ nB^{n-1}[A,B]$$

holds in the latter sense (2), if $[[A,B],B]=0$. (It is not necessary to demand $[A,[A,B]]=0$.) More generally, for a sufficiently well-behaved function $f$, we have

$$\tag{**} [A,f(B)] ~=~ f^{\prime}(B)[A,B], $$

if $[[A,B],B]=0$.

The group commutator (1) is dimensionless, which (among other things) makes the identity (*) unnatural to demand for group commutators.

  • 2
    $\begingroup$ I got it, you have to use the identity [A,BC] = [A,B]C + [B,A]C coupled with what you said above. I am so foolish. Thanks! $\endgroup$ – user28823 Sep 22 '13 at 19:46
  • $\begingroup$ if $[A,B]$ doesn't commute with $B$, then does it go before or after $B$? $\endgroup$ – psitae Jun 11 '18 at 16:38
  • $\begingroup$ Eqs. ($\ast$) & ($\ast\ast$) do generically not hold in either ordering (or even linear combinations thereof) if $[A,B]$ doesn't commute with $B$. $\endgroup$ – Qmechanic Jun 11 '18 at 17:21

Maybe you could show this relation in an other way:


Then you take $[A,B^{n-1}]B$ and repeat the process:


So $B$ commute with $[A,B]$


Repeat for $n$ steps



Here's another way to prove this relationship, by induction:

  1. Verify that the statement holds for $n = 1$

$$ [A,B^1]=1 \cdot B^{1-1} [A,B] = [A,B] $$

  1. Show that, if the formula holds for $n=k ~ \big(I\big)$, then it also holds for $n=k+1$, using the identity $[X, YZ] = [X,Y]Z + Y[X,Z] ~\big(II\big)$ and the fact that $B$ commutes with $[A,B] ~\big(III\big)$

$$ \begin{align} [A,B^{(k+1)}] &= [A,B^k B] \\ & = [A,B^k]B + B^k[A, B] ~\big(II\big)\\ & = k B^{k-1}[A,B]B + B^k[A, B] ~\big(I\big) \\ & = k B^{k-1}B[A,B] + B^k[A, B] ~\big(III\big) \\ & = k B^k[A,B] + B^k[A, B] \\ & = (k + 1) B^k[A,B] \\ & = (k + 1) B^{(k+1)-1}[A,B] \\ \end{align} $$

  1. Since both the basis and the inductive step have been performed, by mathematical induction, the statement holds for all natural numbers $n$. $\Box$

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.