I observed a similarity. Is this a co-incidence?:

$$(I+\epsilon P)|x\rangle =|x+\epsilon\rangle$$



Here, $|x\rangle$ is an eigenfunction of position. $|n\rangle$ is an eigenfunction of the Hamiltonian $X^2+P^2$

The similarity that I observe is that $(I+\epsilon P)$ works as an "infinitesmial ladder operator" for $|x\rangle$

Is the motivation for ladder operators rooted in this similarity? If so, how? Can ladder operators be systematically derived by exploiting this similarity? I've only seen ladder operators introduced abruptly as a "mathematical trick".

EDIT There's one more clue. The first equation works only because of the value of the commutator $[X,P]$. The second equation also works because of the commutator $[a,H]$

Given all these clues, can we systematically motivate ladder operators?


1 Answer 1


The defining equation for a ladder (or lowering) operator $\hat{L}$ with respect to some observable $\hat{X}$ is* $$[\hat{X},\hat{L}] = -\Delta \hat{L},$$ where $\Delta$ is some difference in eigenvalues of $\hat{X}$. It follows that if $|\xi\rangle$ is some eigenvector of $\hat{X}$ with eigenvalue $\xi$, then $\hat{L}|\xi\rangle$ is another eigenvector with eigenvalue $\xi-\Delta$, since $$ \hat{X}\hat{L}|\xi\rangle = [\hat{X},\hat{L}]|\xi\rangle +\hat{L}\hat{X}|\xi\rangle = -\Delta \hat{L}|\xi\rangle + \hat{L}\xi |\xi\rangle = (\xi-\Delta)\hat{L} |\xi\rangle.$$

So far this is completely general, but there are numerous famous and familiar examples in quantum physics. For the simple harmonic oscillator, the relevant ladder operator is $\hat{L} = \hat{a} = \hat{x}+{\rm i}\hat{p}$ (in appropriate adimensional units), the observable is the Hamiltonian $\hat{X} = \hat{H} = \omega\hat{a}^\dagger\hat{a},$ and we have $[\hat{H},\hat{a}] = -\omega\hat{a}$ (taking $\hbar=1$). The OP raises another example: the unitary translation operator $\hat{T}(y) = {\rm e}^{{\rm i}y \hat{p}}$, whose action is usually written as $$\hat{T}^\dagger(y) \hat{x}\hat{T}(y) = \hat{x}+y,$$ but this is equivalent to a commutation relation $$ [\hat{x},\hat{T}(y)] = y \hat{T}(y),$$ which defines $\hat{T}(y)$ as a "ladder operator" that "lowers" the observable $\hat{x}$ by an amount $\Delta = -y$. (The OP describes the example of an infinitesimal transformation $\hat{T}(\varepsilon)$ up to order $O(\varepsilon)$ but the relation actually holds in general).

It should be obvious from the above argument that any other continuous unitary symmetry operator is also a ladder/lowering operator with respect to the transformed observable, which accords intuitively with our notion of translations, rotations etc. as a shift of spatial, angular coordinate etc.

*It is not always so easy to gain physical intuition from commutator relations but you can check out this nice question for some insights. Mathematically, the defining commutation relation of a ladder operator encapsulates the fact that $\hat{L}$ changes the (eigen)value of $\hat{X}$.

  • $\begingroup$ 1. Can you also edit in a physical intuition for the general ladder operator requirement that you wrote at the top? 2. $X$ and $P$ are canonical conjugates. Are $H$ and the Ladder operator also somehow canonical conjugates or something similar? $\endgroup$
    – Ryder Rude
    Feb 15, 2022 at 11:52
  • $\begingroup$ Isn't $[I+i\epsilon P, X]=X+i\epsilon XP - X -i\epsilon PX= \epsilon i [X,P] =-\epsilon h\neq \epsilon (I+i\epsilon P)$ $\endgroup$
    – Ryder Rude
    Feb 15, 2022 at 15:16
  • 2
    $\begingroup$ @RyderRude 1. I added an edit regarding physical intuition. 2. No, $\hat{X}$ and $\hat{L}$ are not canonically conjugate. 3. $[I-{\rm i}\epsilon \hat{p},\hat{x}] = -\epsilon I = -\epsilon (I -{\rm i}\epsilon \hat{p})$ up to order $\epsilon$, which is the same order that your equations in the OP are valid to. That is, $I -{\rm i}\epsilon \hat{p}$ is only the translation operator at lowest order in $\epsilon$. $\endgroup$ Feb 16, 2022 at 17:34
  • $\begingroup$ What about noncontinuous unitary symmetries? Is the ladder operator for a harmonic oscillator unitary? $\endgroup$
    – R. Rankin
    Jul 15, 2022 at 19:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.