Guide for Physicists

For a physics reader, the most productive way to read a rigorous argument is to start from a prediction and ask what guarantees it. Domain conditions, compactness, and continuity are tools for answering that question. Read them as information about the model and its approximations, rather than as a separate vocabulary to memorize before doing quantum mechanics.

Start from the quantity to be predicted

Write down the target: an energy level, a probability, a scattering amplitude, or an error after finite time. Then identify the mathematical object that produces it. An energy eigenvalue requires a specified self-adjoint Hamiltonian; an outcome probability requires an effect or spectral projection; a transition amplitude requires states and a propagator.

This first step prevents an existence theorem from being mistaken for a calculation. Proving that a Hamiltonian is self-adjoint supplies a real spectral measure and unitary dynamics. It does not tell you its eigenvalues, whether it has bound states, or whether a perturbative estimate is accurate.

The object map and formalism crosswalk are useful when notation hides the distinction.

Read the hypotheses in physical groups

Group the assumptions into three kinds. Model data include the configuration space, potential, particle content, and boundary conditions. Analytic control includes self-adjointness, integrability, continuity, and decay. Regime restrictions include finite volume, a spectral gap, bounded energy, or a small expansion parameter.

For each assumption, ask what fails if it is dropped. Boundary conditions can change a spectrum. A missing gap can invalidate an isolated-eigenvalue approximation. A loss of trace-classness can make a proposed Gibbs density operator impossible to normalize. Such questions make an abstract hypothesis easier to remember because they attach it to a consequence.

Do not infer an omitted assumption from a familiar formula. In particular, a displayed kinetic-plus-potential expression is not yet a domain specification. Use Tracking Domains when a calculation differentiates, multiplies, or integrates by parts.

Worked reading example: a variational estimate

Suppose a Hamiltonian is self-adjoint and bounded below. A normalized trial state can provide an upper bound on the bottom of its spectrum. The useful version is formulated on its closed quadratic-form domain, which can be larger than the operator domain.

Let $q_H$ be that form. The variational statement is

$$\inf\sigma(H) =\inf_{\substack{\psi\in D(q_H)\\\|\psi\|=1}}q_H[\psi].$$

For a Schrödinger operator, the trial energy is often written as an integral involving $|\nabla\psi|^2$ and $V|\psi|^2$. Reading it as a form explains why a trial function may be admissible even when its second weak derivative is not square integrable. However, the trial must satisfy the form-domain boundary conditions and the potential integrability requirement.

The reading procedure now yields three separate conclusions. The theorem gives an upper bound from any admissible trial. It does not prove that the trial is an eigenvector. It does not prove the spectral bottom is an eigenvalue; additional spectral information is needed. Quadratic Forms and Spectra of Operators supply those distinctions.

Read the proof in two passes

On a first pass, locate the mechanism: a positivity estimate, a compactness argument, a spectral integral, or an iteration bound. Mark which theorem is imported and what it contributes. Work one concrete example far enough to test signs, units, normalization, and a limiting case.

On a second pass, inspect the delicate interchange: a derivative with an integral, a limit with an unbounded operator, or a trace with an infinite sum. Ask for the norm estimate or convergence theorem that justifies it. Reading Proof Sketches explains how to distinguish a genuinely imported lemma from an unproved leap.

This approach preserves the intuition of a physical calculation while testing the step most likely to fail. It is rarely productive to expand every elementary algebraic step while leaving the central domain or limiting assumption implicit.

Exercise: what does one trial energy prove?

A normalized admissible trial state has form energy $E_{\rm trial}$. A calculation finds no explicit eigenfunction. State the conclusion that follows and two conclusions that do not.

Solution

The conclusion is $\inf\sigma(H)\le E_{\rm trial}$. It does not establish equality, and it does not establish that the spectral bottom is attained by a normalizable vector. To certify a bound state, one can seek a trial value below a separately established essential-spectrum threshold, together with the relevant variational spectral theorem. That extra argument is substantive, not an automatic consequence of a low trial number.

References

• Hall, Brian C., Quantum Theory for Mathematicians, Graduate Texts in Mathematics 267, Springer, 2013, doi:10.1007/978-1-4614-7116-5 — physical examples tied to operator statements.
• 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 — quadratic forms, spectral bounds, and Schrödinger models.