Rigorous Notation Translation Table

The same physics notation can represent a Hilbert vector, a continuous functional on test functions, a spectral integral, or a quadratic form. A sound translation preserves that object type and the conditions under which the expression is evaluated. This table is a lookup companion to the worked method in Notation Translations; it does not replace its convention checks.

Formal notation and the object it denotes

Expression	Rigorous interpretation	Condition to restore
$\lvert\psi\rangle$	A vector in the chosen Hilbert space, or its ray when global phase is irrelevant	Specify normalization and whether vectors or rays are compared
$\lvert x\rangle$	A generalized evaluation eigenfunctional in a chosen rigging	Evaluation is continuous on a suitable test space, not on all $L^2$; position kets
$\langle x\mid y\rangle=\delta(x-y)$	A distributional normalization tied to the coordinate measure	It is not a finite Hilbert-space inner product; delta normalization
$\int dx\,\lvert x\rangle\langle x\rvert=I$	A resolution interpreted weakly on appropriate vectors or through a PVM	No bounded rank-one projector exists at each continuous point; resolution of identity
$H=-\Delta+V$	A differential expression used to define an operator or a form	Specify domain, boundary conditions, and realization; Hamiltonian domains
$A^\dagger=A$	Equality of operators, including their domains	A formal integration-by-parts identity proves less; self-adjoint operators
$[A,B]\psi$	$AB\psi-BA\psi$ on $D(AB)\cap D(BA)$	Both products must exist; commutation on domains
$f(A)$	A Borel spectral function of a self-adjoint operator	$D(f(A))$ is fixed by spectral square-integrability; functional calculus
$\operatorname{Tr}(\rho A)$	A trace pairing for bounded $A$, or a qualified spectral/form expectation for unbounded $A$	State integrability and avoid unsupported cyclic rearrangements; expectation functionals
$\rho_A=\operatorname{Tr}_B\rho$	The unique trace-class operator dual to $A\mapsto A\otimes I$	Trace over a subsystem, not over a preferred ensemble; partial trace
$\dot\rho=\mathcal L\rho$	A derivative in the specified state-space topology	The initial state must lie in the generator domain for a classical solution; CP generators

The notation itself cannot decide these conditions. For example, “$\sum_n$” can denote a norm-convergent vector series, a strongly convergent sum of positive operators, or a trace-norm sum of channel outputs. The neighboring Operator Classes and Spectral Theory sheets identify which estimates justify which reading.

Translate equalities together with their topology

An equality involving all vectors in a Hilbert space is different from equality on a dense test domain. Likewise, a distributional identity is tested against test functions; it does not claim equality of ordinary functions point by point.

For an integral kernel $K$, the statement

$$(Tf)(x)=\int K(x,y)f(y)\,dy$$

must say in what sense the integral and output exist. A square-integrable kernel defines a Hilbert–Schmidt operator on the corresponding $L^2$ spaces. The identity on an infinite-dimensional $L^2$ space is not Hilbert–Schmidt, so its delta kernel cannot be an ordinary square-integrable kernel. This is a quick type check, not an alternative proof of distribution theory.

Similarly, $e^{-itH/\hbar}$ is bounded and unitary for self-adjoint $H$ even when $H$ is unbounded. Replacing the exponential by its power series on an arbitrary vector is a further assertion about domains and analytic vectors. Spectral calculus defines the exponential without requiring every power $H^n\psi$ to exist.

Worked translation: a mean without an operator-domain vector

The finite-mean, infinite-second-moment example constructs a normalized vector for the diagonal observable $Ne_n=ne_n$ with $\psi\in D(N^{1/2})$ but $\psi\notin D(N)$. Its weighted-series calculation belongs to that measure-theory treatment. Here the task is to translate the resulting expectation notation correctly.

The symbol $\langle N\rangle_\psi$ is meaningful as the form value $\|N^{1/2}\psi\|^2$. The literal vector pairing $\langle\psi,N\psi\rangle$ is not defined because $N\psi$ is not a Hilbert vector. Replacing the former by the latter without mentioning the form extension changes the mathematical assertion.

For $\rho=|\psi\rangle\langle\psi|$, the same distinction applies to a formal $\operatorname{Tr}(\rho N)$. The finite spectral mean does not authorize treating every ordering of this unbounded product as an ordinary trace-class product.

Exercises

A smooth kernel masquerading as an identity. Let $g\in L^2(\mathbb R)$ have norm one and set $K(x,y)=g(x)\overline{g(y)}$. Determine the associated operator and explain why its integral is not a resolution of the identity on all of $L^2$.

Solution

The output is $g(x)\langle g,f\rangle$, so $T=|g\rangle\langle g|$. It is the rank-one orthogonal projection onto $\operatorname{span}\{g\}$. It annihilates every vector orthogonal to $g$, whereas the identity fixes them. A smooth separable kernel does not acquire the delta kernel’s meaning by being inserted between integral signs.

A physical trace convention. A calculation uses root fidelity $F(\rho,\sigma)$, while another uses its square. Which notation check is needed before applying a trace-distance inequality?

Solution

Write the defining expression and the inequality explicitly before substituting numerical values. For root fidelity, $1-F\le D\le\sqrt{1-F^2}$; if a source calls $F^2$ the fidelity, the same bounds become $1-\sqrt{F_{\rm sq}}\le D\le\sqrt{1-F_{\rm sq}}$. See Fidelity and Trace Distance and the linked quantum-information treatment for the conventions and proof.

References

• Hall, Brian C. Quantum Theory for Mathematicians. Graduate Texts in Mathematics 267, Springer, 2013. doi:10.1007/978-1-4614-7116-5. Operators, spectral integrals, and conventions.
• Teschl, Gerald. Mathematical Methods in Quantum Mechanics: With Applications to Schrödinger Operators, 2nd ed. Graduate Studies in Mathematics 157, American Mathematical Society, 2014. Author’s edition. Domains, forms, and spectral expectations.