From Old Quantum Theory to Hilbert Space

Old quantum theory used quantum numbers and action rules before physicists had the modern language of Hilbert spaces, operators, and state vectors. It was historically powerful because it found real patterns. It was also unstable because its rules were attached to classical orbits that did not survive the full theory.

Follow how restrictions on classical motion gave way to a general state-space framework. For the mathematical definitions, use Hilbert Spaces and Quantum States.

Old Quantum Numbers

Old quantum theory introduced integer labels before it had a general theory of states. The Bohr model used a principal quantum number $n$ to label allowed hydrogen energies. Sommerfeld’s extension introduced additional labels for radial and angular motion. Magnetic fields introduced orientation labels.

Those labels were not meaningless. They captured real regularities:

• hydrogen line frequencies depend strongly on a principal quantum number;
• angular momentum projections appear discretely in field experiments;
• spectra organize into patterns that are not continuous classical radiation;
• semiclassical quantization works well in some large-quantum-number limits.

The modern caution is that a useful label is not yet a state concept. An old quantum number often labeled a classical orbit or an imposed condition on an orbit. In modern quantum mechanics, quantum numbers usually label simultaneous eigenstates of chosen commuting operators.

Orbits and Action Variables

The old theory’s most systematic rule was action quantization. For periodic classical motion, one imposed conditions of the form

$$J_i = \oint p_i\,dq_i = n_i h.$$

The integral follows one complete cycle of the corresponding separable motion; $p_i$ is conjugate to $q_i$, and $n_i$ is an integer. The coordinate need not be an angle: a radial coordinate can oscillate between turning points. In favorable systems, especially separable integrable systems, these action variables gave allowed energies that matched parts of spectroscopy.

This rule is why old quantum theory was more than numerology. It identified Planck’s constant as a unit of action and connected quantum discreteness to classical phase-space geometry. The descendant is not the naive old rule itself, but controlled semiclassical methods such as WKB and EBK quantization. See Bohr–Sommerfeld Quantization.

The weakness was also built in. Applying the old rule required classical motion with suitable periodic or quasiperiodic cycles. Modern action variables are defined on independent cycles of an integrable invariant torus; arbitrary coordinate choices do not produce independent quantization rules. It did not say what to do for generic nonintegrable motion, multi-electron atoms, spin, transition probabilities, or measurement outcomes.

Breakdown of Old Rules

Old quantum theory broke down in several different ways:

• it could fit hydrogen better than helium or many-electron atoms;
• it predicted frequencies more naturally than intensities;
• it treated electron orbits as real while experiments increasingly pointed toward amplitudes and probabilities;
• it lacked intrinsic spin and the correct angular-momentum algebra;
• it did not provide a general composition rule for identical particles;
• it had no unified measurement probability rule.

The Limits of Old Quantum Theory page gives the historical survey. The bridge to Hilbert space is the structural lesson: the theory needed a new predictive object. Classical phase-space points and old orbits were too restrictive.

Modern quantum mechanics replaces the orbit-first strategy with a state-first strategy:

$$\text{preparation} \longrightarrow \text{state} \longrightarrow \text{probabilities for specified measurements}.$$

A pure state can be represented by a wavefunction or by coefficients in a complete basis, or equivalently by its rank-one density operator. Mixed states require more general density operators; they are not simply another basis representation of a single state vector.

Wavefunctions and Matrices

Wave mechanics and matrix mechanics looked different at first.

Wave mechanics represented states by functions such as $\psi(x)$ and found allowed energies from differential equations:

$$H\psi_n = E_n\psi_n.$$

Matrix mechanics organized transition quantities and noncommuting products. In modern notation, observables became operators and states could be represented by vectors in a basis. A state might be written as coefficients

$$\lvert\psi\rangle = \sum_n c_n\lvert n\rangle$$

in an energy basis, or as a wavefunction

$$\psi(x) = \langle x\vert\psi\rangle$$

in a position basis.

The equivalence of matrix and wave mechanics became intelligible once both were seen as representations of the same abstract structure. The matrix of an operator depends on the basis. The wavefunction is a coordinate representation of a state. Neither is the whole ontology of the theory by itself.

For the historical route, see Heisenberg’s Matrix Mechanics, Schrödinger’s Wave Mechanics, and Equivalence of Matrix and Wave Mechanics.

Hilbert-Space Unification

Hilbert space supplied the unifying grammar:

• states are rays or density operators;
• coherent superposition is a linear combination of state vectors, followed by normalization;
• probabilities come from inner products and positive operators;
• observables are represented by suitable operators;
• spectra replace ad hoc lists of allowed values;
• basis changes connect wave and matrix descriptions;
• tensor products describe composite systems.

The old orbit labels did not disappear completely. They were reinterpreted. In the nonrelativistic, spinless Coulomb approximation to hydrogen, labels such as $n$, $\ell$, and $m$ are eigenvalue labels for a compatible set of operators, not parameters of a literal Kepler ellipse:

$$H\lvert n\ell m\rangle = E_n\lvert n\ell m\rangle,$$ $$L^2\lvert n\ell m\rangle = \hbar^2\ell(\ell+1)\lvert n\ell m\rangle,$$ $$L_z\lvert n\ell m\rangle = \hbar m\lvert n\ell m\rangle.$$

This is a deeper change than replacing one formula with another. The modern labels are tied to operators, spectra, commutation relations, and measurement contexts.

What Survived

Several old-theory insights survived in changed form:

Old quantum idea	Modern descendant
Planck’s constant as an action scale	commutators, phases $e^{iS/\hbar}$, semiclassics
action quantization	WKB, EBK, Bohr–Sommerfeld limits
stationary states	Hamiltonian eigenstates and spectral theory
transition frequencies	energy differences and time-dependent perturbation theory
correspondence principle	classical limit, semiclassical approximations
quantum numbers	eigenvalue labels for commuting operators

The survival is not a simple vindication of classical orbits. It is a reinterpretation inside a broader formalism.

For a concrete comparison, solve the harmonic oscillator in Wave Mechanics and then study its state-space representation through Operators. The exercise below shows a correction that already separates modern semiclassics from the earliest action rule.

Exercises

1. In one sentence each, distinguish an old quantum number from a modern eigenvalue label.

Solution

An old quantum number often labels an imposed condition on a classical orbit or action variable. A modern eigenvalue label identifies a state within the spectrum of a specified operator, usually as part of a compatible set of commuting observables.

2. Why did success for hydrogen not make old quantum theory a complete mechanics?

Solution

Hydrogen is an unusually symmetric one-electron Coulomb problem. Old quantum rules could exploit its separability and periodic motion. A complete mechanics also had to handle many-electron atoms, transition probabilities, spin, noncommuting observables, measurement probabilities, and systems without simple classical orbits.

3. Explain why wavefunctions and matrices can represent the same state.

Solution

In Hilbert-space language, the state is an abstract vector or ray. Choosing the position basis gives a wavefunction $\psi(x)=\langle x\vert\psi\rangle$. Choosing a discrete basis gives components $c_n=\langle n\vert\psi\rangle$. Operators likewise have basis-dependent matrix or differential representations. The representation changes, but the state and operator structure are the same.

4. A classical harmonic oscillator has action $\oint p\,dq=2\pi E/\omega$. Compare the old rule $J=nh$ with the corrected semiclassical rule $J=(n+1/2)h$ for two smooth turning points.

Solution

The old rule gives $E=n\hbar\omega$. The turning-point phase correction gives $E=(n+1/2)\hbar\omega$, which happens to be exact for the harmonic oscillator. It restores the nonzero ground energy while approaching the same relative spacing at large $n$. For more general potentials, the corrected rule is an approximation, not an exact spectrum.

References

• O. Darrigol, From c-Numbers to q-Numbers: The Classical Analogy in the History of Quantum Theory, University of California Press, 1992.
• P. A. M. Dirac, The Principles of Quantum Mechanics, 4th ed., Oxford University Press, 1958.
• M. Jammer, The Conceptual Development of Quantum Mechanics, 2nd ed., American Institute of Physics, 1989.
• J. Mehra and H. Rechenberg, The Historical Development of Quantum Theory, Springer, 1982-2001.
• M. Reed and B. Simon, Methods of Modern Mathematical Physics I: Functional Analysis, 2nd ed., Academic Press, 1980.
• J. von Neumann, Mathematical Foundations of Quantum Mechanics, Princeton University Press, 1955.