🔬 physics

The History of Hilbert-Space Formulations of Classical Physics

This paper distinguishes between the Koopman-von Neumann formulation and the method of "classical" wave functions in classical physics, clarifying that the latter was developed independently by later researchers such as Mario Schönberg, Angelo Loinger, Giacomo Della Riccia, Norbert Wiener, and E. C. George Sudarshan, rather than by Koopman and von Neumann.

Original authors: Jacob A. Barandes

Published 2026-07-21

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Jacob A. Barandes

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe as a giant, cosmic game of pool. In the quantum version of this game—the one that governs atoms and light—the balls don't just roll; they act like ripples in a pond, described by mysterious "wave functions" that tell us where a ball might be. This is the realm of quantum mechanics, a place where math feels like magic. But what about the balls on a regular pool table? The ones we can see and touch? For centuries, we've described them using "classical physics," a set of rules that feels solid and predictable, dealing with probabilities that are just numbers, not waves.

For a long time, scientists thought these two worlds—quantum waves and classical balls—were speaking completely different languages. But in the 20th century, a few clever mathematicians asked a wild question: "What if we could describe the classical pool table using the same wave-like math as the quantum one?" This idea led to a fascinating, albeit messy, history of who actually invented the "classical wave function." It's a story about how a brilliant idea got mislabeled, like a famous painting being hung in the wrong museum for decades, and how a group of researchers tried to set the record straight.

This paper, written by Jacob Barandes, is a historical detective story that clears up a major mix-up in the physics community. The author investigates two different ways scientists tried to use "Hilbert spaces" (a fancy mathematical toolbox for handling waves and vectors) to describe classical physics. The first method, developed in the 1930s by Bernard Koopman and John von Neumann, treated classical observables (like position or speed) as vectors in a new kind of mathematical space. The second method, which came decades later, treated the probability of finding a particle as a "classical wave function," where the square of the wave's height gives you the probability, just like in quantum mechanics.

The paper finds that for a long time, people have been incorrectly calling this second method the "Koopman–von Neumann (KvN) formulation." It's as if everyone started calling a new type of car a "Ford" just because the original Ford engine was used in a different vehicle years ago. Barandes shows that Koopman and von Neumann never actually proposed using classical wave functions; they were working with something else entirely. The credit for the "classical wave function" idea actually belongs to a different cast of characters, starting with Mario Schönberg in the 1950s, followed by others like Angelo Loinger, Giacomo Della Riccia, Norbert Wiener, and E. C. George Sudarshan.

The author argues that while the two mathematical approaches eventually turn out to be equivalent (like two different roads leading to the same town), they are conceptually distinct. Koopman and von Neumann were describing how the rules of the game evolve, while the later researchers were describing the players (the probabilities) as waves. The paper concludes that we need to stop giving Koopman and von Neumann credit for the wave-function idea and start giving it to the people who actually invented it, ensuring that the history of science gets the names right.

Problem
The paper addresses a significant historical and conceptual conflation in the literature regarding Hilbert-space formulations of classical physics. Specifically, it identifies a widespread misattribution where the "Koopman–von Neumann (KvN) formulation" is incorrectly used to describe the method of "classical wave functions." In this misattributed view, Koopman and von Neumann are credited with introducing complex-valued wave functions $\psi$ on phase space, where the classical probability density $\rho$ is recovered via the modulus-square operation ( $\rho = |\psi|^2$ ), analogous to the Born rule in quantum mechanics. The author argues that this attribution is historically inaccurate and conceptually misleading, as Koopman and von Neumann's original work did not involve classical wave functions or the modulus-square relationship. Instead, their formulation treated classical observables as vectors in a Hilbert space, distinct from the later development of classical wave functions by other researchers.

Methodology
The paper employs a rigorous historical and technical analysis of primary sources, tracing the development of Hilbert-space techniques in classical mechanics from the 1930s through the 2000s. The methodology involves:

Textual Analysis: Examining the original 1931 paper by Bernard Koopman and the 1932 papers by John von Neumann to determine the precise mathematical definitions of their phase-space functions, inner products, and time-evolution operators.
Comparative Analysis: Contrasting the original KvN formalism with the "classical wave function" method developed by later researchers, specifically Mario Schönberg (1950s), Angelo Loinger (1962), Giacomo Della Riccia and Norbert Wiener (1966), and E. C. George Sudarshan (1970s).
Technical Verification: Verifying the time-evolution equations and the nature of the Hilbert spaces (e.g., whether functions represent observables or probability amplitudes) to demonstrate conceptual differences despite mathematical equivalence.
Literature Review: Tracking the propagation of the misattribution in modern literature, including the works of Danilo Mauro and Ennio Gozzi (1999–2004), to show how the error became entrenched in the field.

Key Contributions and Results
The paper provides a detailed chronology and technical clarification of two distinct approaches:

The Original Koopman–von Neumann Formulation (1931–1932):
- Koopman and von Neumann constructed a Hilbert space where the vectors represent classical observables (functions on phase space), not state vectors or wave functions.
- The inner product in their formulation includes a fixed, positive density function $\rho$ (representing a fluid density or measure) separate from the functions being integrated: $(\phi, \psi) = \int \rho \phi \bar{\psi} d\omega$ .
- The time evolution of these functions follows the Heisenberg picture, evolving as observables according to the Liouville operator $L$ (specifically, $d\phi/dt = \{ \phi, H \} = iL\phi$ ).
- Crucially, they did not propose that these functions were complex-valued wave functions whose modulus squared yielded a probability distribution.
The Method of Classical Wave Functions (Post-1950s):
- This approach, likely first introduced by Mario Schönberg in the early 1950s, introduces a complex-valued function $\Theta$ (or $\psi$ ) such that the classical probability density is $\rho = |\Theta|^2$ .
- This method treats $\psi$ as a "probability amplitude" analogous to the quantum wave function.
- Because the Liouville operator contains only first-order derivatives, if $\psi$ satisfies the Liouville equation, then $|\psi|^2$ automatically satisfies the same equation.
- Key contributors to this independent development include Angelo Loinger (who explicitly distinguished the two approaches in 1962), Giacomo Della Riccia and Norbert Wiener (1966), and E. C. George Sudarshan (1976), who used this method to embed classical systems into quantum systems with superselection sectors.
Mathematical Equivalence vs. Conceptual Distinction:
- The paper acknowledges that the Hilbert spaces arising from the observables-as-vectors method and the classical wave-function method are mathematically equivalent, a result grounded in the Gelfand–Naimark–Segal (GNS) construction.
- However, the author emphasizes that mathematical equivalence does not imply conceptual identity. Just as Heisenberg's matrix mechanics and Schrödinger's wave mechanics were distinct frameworks despite their eventual unification, the original KvN formulation and the classical wave-function method represent different conceptual starting points.
Identification of Misattribution:
- The paper documents how the term "Koopman–von Neumann wave functions" became a standard, albeit incorrect, label in recent literature (e.g., works by Mauro and Gozzi starting in 2001). These authors incorrectly claimed that Koopman and von Neumann postulated the evolution of complex wave functions $\psi$ via the Liouville equation and the relation $\rho = |\psi|^2$ .

Significance
The primary significance of the paper is the correction of the historical record and the assignment of appropriate credit. The author argues that while the "classical wave function" method is a valid and useful technique, attributing its invention to Koopman and von Neumann obscures the contributions of later researchers like Schönberg, Loinger, Della Riccia, Wiener, and Sudarshan.

The paper concludes that getting the history right is essential for intellectual honesty, ensuring that the researchers who developed the specific idea of using classical wave functions to capture probability distributions receive the recognition they deserve. The work serves as a corrective to the prevailing narrative in contemporary physics literature, clarifying that Koopman and von Neumann developed an operatorial approach for observables, not a wave-function approach for states.

More like this