5
$\begingroup$

In the context of quantum mechanics we postulate that every observable operator $A$ acting on the corresponding Hilbert space $\mathcal{H}$ is self-adjoint (Hermitian), i.e. $$\forall \Psi,\varphi\in\mathcal{H}:\langle\Psi,A\varphi\rangle=\langle A\Psi,\varphi\rangle,$$ which is equivalent to saying $A=A^{\dagger}$.

The map $(\cdot)\mapsto(\cdot)^{\dagger}$ is transposing and complex conjugating; i.e., in matrix notation: $(A^{\dagger})_{ij}=A_{ji}^{*}$ (the star denoting complex conjugation). Finally, we have the bra-ket notation $(|\Psi\rangle)^{\dagger}:=\langle\Psi|$.

My question is: How can I prove rigorously, with the above, that $$(A|\Psi\rangle)^{\dagger}=\langle\Psi|A?$$ Is it possible to do so without using the matrix notation (i.e. general statement including continuum states)?

$\endgroup$
2
  • 2
    $\begingroup$ Not every operator on the Hilbert space is Hermitian. For example, the time-evolution operator is unitary. $\endgroup$ Commented Sep 18, 2021 at 18:50
  • $\begingroup$ @RichardMyers While the time-evolution operator is non-Hermitian, it's worth noting that unitarity and Hermiticity are not mutually exclusive. $\endgroup$
    – Sandejo
    Commented Sep 19, 2021 at 19:40

3 Answers 3

6
$\begingroup$

I will approach your problem from a pure mathematical perspective:

  1. Let me first say that in your post you use the same notation $\dagger$ for two distinct mappings which have two different definitions and notations. One mapping acts on the space of linear densely defined operators in a complex separable Hilbert space and the other mapping is from the Hilbert space itself to its (topological) dual with respect to the topology induced by the norm (this mapping assignes a vector to a continuous functional). The first mapping is regularly denoted in physics by the dagger $\dagger$, while the second one by the tilde $\widetilde{,,,}$

  2. Let me redefine your statement to be proved by using proper mathematical notation and dismissing the mathematically complicated Dirac braket notation. You wish to prove that for a self-adjoint $A:D(A)\subseteq\mathscr{H} \rightarrow \mathscr{H}$ we have the following equality of functionals:

$$\widetilde{A\Psi} =A^{\times}\widetilde{\Psi}.$$

Denote by $ F_{\Psi}\in\widetilde{\mathscr{H}}$ a continuous functional on $\mathscr{H}$ assigned to an arbitrary vector $\Psi\in D(A)\subseteq\mathscr{H}$. We then have:

$$\widetilde{A\Psi} (\varphi) \equiv F_{A\Psi} (\varphi) = \langle A\Psi, \varphi\rangle = \langle \Psi, A^{\dagger}\varphi\rangle = \langle \Psi, A\varphi\rangle, ~ \forall \varphi\in D(A)\tag{1}$$

  1. Turning to the right hand side, $A^{\times}$ is called the dual operator assigned to a linear operator acting in the Hilbert space ($A^{\times}:\widetilde{\mathscr{H}}\rightarrow\widetilde{\mathscr{H}}$). We then use its definition:

$$ \left(A^{\times} \widetilde{\Psi}\right) (\varphi) =\left(A^{\times}F_{\Psi}\right)(\varphi)=: F_{\Psi} (A\varphi) = \langle\Psi, A\phi\rangle, ~ \forall \varphi\in D(A) \tag{2} $$

From $(1)$ and $(2)$ we obtain what we needed to prove.

$\endgroup$
1
  • 2
    $\begingroup$ These things are indeed much clearer without the Dirac notation. $\endgroup$
    – Gold
    Commented Sep 18, 2021 at 23:38
4
$\begingroup$

In general

\begin{equation} \langle \varphi| A |\psi \rangle^* = \langle \psi| A^\dagger |\varphi \rangle \end{equation}

If $A$ is hermitian $\Rightarrow A = A^\dagger$

\begin{align} &\Rightarrow \langle \varphi| A |\psi \rangle^* = \langle \psi| A |\varphi \rangle\\ &\iff (|\psi \rangle)^\dagger A^\dagger (\langle \varphi|)^\dagger = \langle \psi| A |\varphi \rangle\\ &\iff (A |\psi \rangle)^\dagger |\varphi \rangle = \langle \psi| A |\varphi \rangle \end{align}

Since this is true for any $|\varphi \rangle$

\begin{equation} \therefore \quad (A |\psi \rangle)^\dagger = \langle \psi| A \end{equation}

$\endgroup$
2
$\begingroup$

More generally, the adjoint can be defined for an operator $A:\mathcal H_1\to \mathcal H_2$ between different Hilbert spaces, and gives an operator $A^\dagger : \mathcal H_2 \to\mathcal H_1$. It is then easy to prove that : $$(AB)^\dagger = B^\dagger A^\dagger\tag{1}$$

To apply this definition to $|\psi\rangle$, we can see it as an operator $\mathbb C\to \mathcal H$ (which sends a complex number $\lambda$ to the ket $\lambda|\psi\rangle$). Its adjoint is an operator $|\psi\rangle^\dagger = X: \mathcal H \to \mathbb C$ such that : $$\forall \lambda \in \mathbb C,\forall |\varphi\rangle \in \mathcal H , (|\varphi\rangle,\lambda|\psi\rangle)_{\mathcal H} = (X|\varphi\rangle,\lambda)_{\mathbb C}$$ ie $\lambda \langle \varphi|\psi\rangle= (X|\varphi\rangle)^*\lambda$. This gives $X|\varphi\rangle= \langle\psi|\varphi\rangle$ and $|\psi\rangle^\dagger = \langle \psi|$.

Using $(1)$ with $A$ hermitian, we get : $$(A|\psi\rangle)^\dagger = |\psi\rangle^\dagger A ^\dagger = \langle \psi| A$$

$\endgroup$

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.