Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2604.03214

Nelson's Stochastic Mechanics: Measurement, Nonlocality, and the Classical Limit

Partha Ghose

quant-phSubmitted Apr 3, 2026• Announced Apr 6, 2026• 0 citations• Revised Apr 8, 2026 (v2)
v1 · Apr 3, 2026v2 · Apr 8, 2026
arXivHTMLPDF

Abstract

Nelson's stochastic mechanics may be understood as a stochastic underpinning, or reconstruction, of nonrelativistic quantum mechanics, once the diffusion scale is fixed by $\hbar$ and the admissible states are restricted by the usual single-valuedness condition on the wavefunction. In this note I briefly indicate what this route achieves and why it remains conceptually attractive. Four advantages are emphasized. First, it supplies a clear configuration-space stochastic picture of the underlying processes. Second, the Born rule is built in from the outset, with $|ψ|^2$ arising as the probability density $ρ$ of the underlying diffusion process rather than as an independent postulate. Third, it offers a markedly different perspective on measurement and nonlocality: in particular, collapse need not be treated as an extra axiom, and the nonlocality associated with entangled states is softened relative to the deterministic Bohmian guidance picture. Fourth, by tying quantumness to a diffusion scale, it naturally suggests a continuum of physical descriptions ranging from the strictly classical to the strictly quantum-mechanical regime. I conclude by proposing a natural distance scale in stochastic mechanics and examining its implications for testing possible limits of Bell correlations.

Comments
6 pages, no figures; additional supporting paras added

Related papers

A Possible Explanation of Quantum Mechanics

Any limit to observer self-knowledge implies an uncertainty of external observation. It is proposed that for "homogeneous" observers made up of only a few types of mathematically equal particles, the external world suffers a distortion. The underlying reason is that the observer is effectively replaced by an "observers' ensemble" since he does not know his own state exactly. Between the elements of the ensemble the observer jumps back and forth with a finite rate of "self-permutation." Since the elements of the ensemble are pairwise identical under time reversal, "micro-time-reversals" exist. The consequence is that all measuring chains become "effectively reversible" for the observer. Hence all external objects are subject to "pseudodiffusion." The latter resembles the diffusion postulated as an axiom in stochastic mechanics. This axiom (Nelson's axiom) implies a version of quantum mechanics, as is well known. In consequence a version of quantum mechanics can be derived from classical-deterministic first principles. The question arises of whether a "bootstrap principle" (fixed point theorem) is needed in addition, fixing the value of the diffusion coefficient and hence "Planck's constant" in a universal fashion. However, a fixed value of Planck's constant exists already — in the world of the observer. Other properties of quantum mechanics (like Wheeler's "delayed choice" and Bell's "nonlocality") appear to be implicit, too. Thus, a deterministic local explanation of quantum mechanics is accessible in principle under one condition: that a distinction between two levels of reality (exo and endo), one valid externally and the other only for the internal observer, is introduced. In this case an external reality can, paradoxically, be chosen again, which is not unfamiliar since J. Willard Gibbs made serious use of it almost a century ago. A new principle of parsimony of the world would thereby become visible: The "fancy" things would be true only on the endo level. Quantum mechanics would become an observer-specific endo reality.

WORLD SCIENTIFIC eBooks•May 1, 1998•4 citations

Stochastic Theory of Quantum Mechanics

M. Godart

The stochastic theory of non-relativistic quantum mechanics presented here relies heavily upon the theory of stochastic processes, with its definitions, theorems and specific vocabulary as well. Its main hypothesis states indeed that the classical trajectories of the particles are identical to the sample functions of a diffusion Markov process, whose conditional probability density is proposed as a substitute for the wave function. The Schroedinger equation and the so-called Nelson equations are used to determine the diffusion tensor and the drift vectors characteristic of such a process. It is then possible to write down the forward and backward Kolmogorov equations that are used to determine the conditional probability density, as well as the Fokker-Planck equation that is used to determine the normal probability density. This method is applied to several simple cases and the results obtained are compared with those of the orthodox theory. Among the most important differences let us mention that any system evolving freely from any original state returns spontaneously to its ground state, that the definitions of the particles velocities and momenta are impossible because the sample functions of a diffusion Markov process nowhere possess a derivative with respect to time and that the so-called collapse of the wave function is a mere mirage explained by the updating choice between new and older initial conditions in the resolution of partial differential equations. We finally attempt to extend the proposed theory to the domain of the relativistic quantum mechanics. It is promising but is clearly unfinished because it has not been possible up to now to solve the Nelson and Kolmogorov equations, except in the very simple case of a free particle.

arXiv:1206.2917•Jun 13, 2012•4 citations

On the Stochastic Mechanics Foundation of Quantum Mechanics

Michael Beyer, Wolfgang Paul

Among the famous formulations of quantum mechanics, the stochastic picture developed since the middle of the last century remains one of the less known ones. It is possible to describe quantum mechanical systems with kinetic equations of motion in configuration space based on conservative diffusion processes. This leads to the representation of physical observables through stochastic processes instead of self-adjoint operators. The mathematical foundations of this approach were laid by Edward Nelson in 1966. It allows a different perspective on quantum phenomena without necessarily using the wave-function. This article recaps the development of stochastic mechanics with a focus on variational and extremal principles. Furthermore, based on recent developments of optimal control theory, the derivation of generalized canonical equations of motion for quantum systems within the stochastic picture are discussed. These so-called quantum Hamilton equations add another layer to the different formalisms from classical mechanics that find their counterpart in quantum mechanics.

Universe•May 27, 2021•17 citations

Dynamical probability, particle trajectories and completion of traditional quantum mechanics

Tulsi Dass

Maintaining the position that the wave function $ψ$ provides a complete description of state, the traditional formalism of quantum mechanics is augmented by introducing continuous trajectories for particles which are sample paths of a stochastic process determined (including the underlying probability space) by $ψ$. In the resulting formalism, problems relating to measurements and objective reality are solved as in Bohmian mechanics (without sharing its weak points). The pitfalls of Nelson's stochastic mechanics are also avoided.

arXiv:quant-ph/0505190•May 25, 2005•1 citation

A generalization of stochastic mechanics and its relation to quantum mechanics

Reinhard F. Werner

Nelson's assignment of a Markov process in configuration space to each Schr\"odinger Hamiltonian and each pure state is generalized to mixed states, arbitrary configuration spaces given by a maximal set of commuting observables, and general Hamiltonians given by integral kernels. Conditions for the quantum-mechanical description to be reconstructable from this stochastic description are discussed. It is shown that those stochastic expectation values which are stable with respect to the operation of coupling a dynamically independent but statistically correlated system to the given one are precisely those already expressible as quantum-mechanical expectation values.

Physical review. D. Particles, fields, gravitation, and cosmology/Physical review. D. Particles and fields•Jul 15, 1986•48 citations

A CONCEPTUAL INTRODUCTION TO NELSON’S MECHANICS

Guido Bacciagaluppi

Nelson's programme for a stochastic mechanics aims to derive the wave function and the Schroedinger equation from natural conditions on a diffusion process in configuration space. If successful, this pro- gramme might have some advantages over the better-known determin- istic pilot-wave theory of de Broglie and Bohm. The essential points of Nelson's strategy are reviewed, with particular emphasis on concep- tual issues relating to the role of time symmetry. The main problem in Nelson's approach is the lack of strict equivalence between the cou- pled Madelung equations and the Schroedinger equation. After a brief discussion, the paper concludes with a possible suggestion for trying to overcome this problem.

Endophysics, Time, Quantum and the Subjective•Oct 1, 2005•48 citations
Nelson's stochastic mechanics may be understood as a stochastic underpinning, or reconstruction, of nonrelativistic quantum mechanics, once the diffusion scale is fixed by ℏ\hbarℏ and the admissible states are restricted by the usual single-valuedness condition on the wavefunction. In this note I briefly indicate what this route achieves and why it remains conceptually attractive. Four advantages are emphasized. First, it supplies a clear configuration-space stochastic picture of the underlying processes. Second, the Born rule is built in from the outset, with ∣ψ∣2|ψ|^2∣ψ∣2 arising as the probability density ρρρ of the underlying diffusion process rather than as an independent postulate. Third, it offers a markedly different perspective on measurement and nonlocality: in particular, collapse need not be treated as an extra axiom, and the nonlocality associated with entangled states is softened relative to the deterministic Bohmian guidance picture. Fourth, by tying quantumness to a diffusion scale, it naturally suggests a continuum of physical descriptions ranging from the strictly classical to the strictly quantum-mechanical regime. I conclude by proposing a natural distance scale in stochastic mechanics and examining its implications for testing possible limits of Bell correlations.
Any limit to observer self-knowledge implies an uncertainty of external observation. It is proposed that for "homogeneous" observers made up of only a few types of mathematically equal particles, the external world suffers a distortion. The underlying reason is that the observer is effectively replaced by an "observers' ensemble" since he does not know his own state exactly. Between the elements of the ensemble the observer jumps back and forth with a finite rate of "self-permutation." Since the elements of the ensemble are pairwise identical under time reversal, "micro-time-reversals" exist. The consequence is that all measuring chains become "effectively reversible" for the observer. Hence all external objects are subject to "pseudodiffusion." The latter resembles the diffusion postulated as an axiom in stochastic mechanics. This axiom (Nelson's axiom) implies a version of quantum mechanics, as is well known. In consequence a version of quantum mechanics can be derived from classical-deterministic first principles. The question arises of whether a "bootstrap principle" (fixed point theorem) is needed in addition, fixing the value of the diffusion coefficient and hence "Planck's constant" in a universal fashion. However, a fixed value of Planck's constant exists already — in the world of the observer. Other properties of quantum mechanics (like Wheeler's "delayed choice" and Bell's "nonlocality") appear to be implicit, too. Thus, a deterministic local explanation of quantum mechanics is accessible in principle under one condition: that a distinction between two levels of reality (exo and endo), one valid externally and the other only for the internal observer, is introduced. In this case an external reality can, paradoxically, be chosen again, which is not unfamiliar since J. Willard Gibbs made serious use of it almost a century ago. A new principle of parsimony of the world would thereby become visible: The "fancy" things would be true only on the endo level. Quantum mechanics would become an observer-specific endo reality.
The stochastic theory of non-relativistic quantum mechanics presented here relies heavily upon the theory of stochastic processes, with its definitions, theorems and specific vocabulary as well. Its main hypothesis states indeed that the classical trajectories of the particles are identical to the sample functions of a diffusion Markov process, whose conditional probability density is proposed as a substitute for the wave function. The Schroedinger equation and the so-called Nelson equations are used to determine the diffusion tensor and the drift vectors characteristic of such a process. It is then possible to write down the forward and backward Kolmogorov equations that are used to determine the conditional probability density, as well as the Fokker-Planck equation that is used to determine the normal probability density. This method is applied to several simple cases and the results obtained are compared with those of the orthodox theory. Among the most important differences let us mention that any system evolving freely from any original state returns spontaneously to its ground state, that the definitions of the particles velocities and momenta are impossible because the sample functions of a diffusion Markov process nowhere possess a derivative with respect to time and that the so-called collapse of the wave function is a mere mirage explained by the updating choice between new and older initial conditions in the resolution of partial differential equations. We finally attempt to extend the proposed theory to the domain of the relativistic quantum mechanics. It is promising but is clearly unfinished because it has not been possible up to now to solve the Nelson and Kolmogorov equations, except in the very simple case of a free particle.
Among the famous formulations of quantum mechanics, the stochastic picture developed since the middle of the last century remains one of the less known ones. It is possible to describe quantum mechanical systems with kinetic equations of motion in configuration space based on conservative diffusion processes. This leads to the representation of physical observables through stochastic processes instead of self-adjoint operators. The mathematical foundations of this approach were laid by Edward Nelson in 1966. It allows a different perspective on quantum phenomena without necessarily using the wave-function. This article recaps the development of stochastic mechanics with a focus on variational and extremal principles. Furthermore, based on recent developments of optimal control theory, the derivation of generalized canonical equations of motion for quantum systems within the stochastic picture are discussed. These so-called quantum Hamilton equations add another layer to the different formalisms from classical mechanics that find their counterpart in quantum mechanics.
Maintaining the position that the wave function ψψψ provides a complete description of state, the traditional formalism of quantum mechanics is augmented by introducing continuous trajectories for particles which are sample paths of a stochastic process determined (including the underlying probability space) by ψψψ. In the resulting formalism, problems relating to measurements and objective reality are solved as in Bohmian mechanics (without sharing its weak points). The pitfalls of Nelson's stochastic mechanics are also avoided.
Nelson's assignment of a Markov process in configuration space to each Schr\"odinger Hamiltonian and each pure state is generalized to mixed states, arbitrary configuration spaces given by a maximal set of commuting observables, and general Hamiltonians given by integral kernels. Conditions for the quantum-mechanical description to be reconstructable from this stochastic description are discussed. It is shown that those stochastic expectation values which are stable with respect to the operation of coupling a dynamically independent but statistically correlated system to the given one are precisely those already expressible as quantum-mechanical expectation values.
Nelson's programme for a stochastic mechanics aims to derive the wave function and the Schroedinger equation from natural conditions on a diffusion process in configuration space. If successful, this pro- gramme might have some advantages over the better-known determin- istic pilot-wave theory of de Broglie and Bohm. The essential points of Nelson's strategy are reviewed, with particular emphasis on concep- tual issues relating to the role of time symmetry. The main problem in Nelson's approach is the lack of strict equivalence between the cou- pled Madelung equations and the Schroedinger equation. After a brief discussion, the paper concludes with a possible suggestion for trying to overcome this problem.