PhyslibAlpha.ClassicalMechanics.NortonDome.Determinism
The Norton dome and the determinism of Newtonian mechanics
i. Overview
A particle at rest on the apex of the Norton dome may stay there forever or slide off at any instant, in obedience to Newton's second law with a continuous force. This file states what that does and does not show, in the terms of `NortonDome.NewtonianSystem`.
The dome is such a system (`toNewtonianSystem`). It is not deterministic (`not_isDeterministic`) and it violates the first law (`not_satisfiesFirstLaw`). Since a locally Lipschitz force gives determinism, the dome must fail that hypothesis and everything implying it, and it does (`not_hasLocallyLipschitzForce`, `not_hasLipschitzForce`, `not_hasC2Potential`). Its force `m √r` is continuous but has infinite slope at the apex, and its potential is `C¹` but not `C²`.
Put together, this is `exists_continuous_force_not_isDeterministic`: a continuous force does not make a Newtonian system deterministic, while a locally Lipschitz force, or a `C²` potential, does. Newton's laws do not say which requirement, if any, belongs in the definition of a Newtonian system. The theorems say what each requirement buys and that the dome is what each excludes, and leave the choice open. Malament's proposal, a regularity condition on the constraint surface in physical space rather than on the potential, is not formulated here. With Peano's theorem, whose proof is pending in Mathlib, the dome has solutions from every initial datum (`hasLocalSolutions`), so its failure is one of uniqueness alone.
ii. Key results
- `NortonDome.not_isDeterministic`: the dome is not deterministic. - `NortonDome.not_satisfiesFirstLaw`: the dome violates the first law. - `NortonDome.exists_continuous_force_not_isDeterministic`: Norton's claim, a Newtonian system with a continuous force that is not deterministic. - `NortonDome.not_hasLocallyLipschitzForce`, `NortonDome.not_hasLipschitzForce` and `NortonDome.not_hasC2Potential`: the dome satisfies none of the regularity conditions. - `NortonDome.toNewtonianSystem` is the dome as a Newtonian system, with `toNewtonianSystem_force`, `toNewtonianSystem_equationOfMotion_iff` and `toNewtonianSystem_isSolution_iff` identifying its force, equation of motion and solutions with those of `NortonDome.Basic`. - `NortonDome.hasLocalSolutions` and `NortonDome.exists_hasLocalSolutions_not_isDeterministic`: the dome has local solutions from every initial datum, by Peano's theorem, so its failure of determinism is a failure of uniqueness alone (pending the upstream proof).
iii. Table of contents
- A. The dome as a Newtonian system - B. The properties the dome fails - B.1. Determinism and the first law - B.2. The regularity conditions - C. Norton's claim - D. Existence without uniqueness
iv. References
- Norton, J. D., *The dome: an unexpectedly simple failure of determinism*, Philosophy of Science 75 (2008), 786–798. - Malament, D. B., *Norton's slippery slope*, Philosophy of Science 75 (2008), 799–816.
A. The dome as a Newtonian system
The dome has a differentiable potential, so it is a Newtonian system whose force, equation of motion and solutions are by definition those of `NortonDome.Basic`.
B. The properties the dome fails
B.1. Determinism and the first law
Determinism fails by the non-uniqueness proved in `NortonDome.Solution`. The first law fails along the motion leaving the apex at the instant `1`: at the instant `0` it is at rest at the apex, where the force vanishes, and at the instant `2` it is off the apex.
B.2. The regularity conditions
The force is not Lipschitz on any closed ball about the apex, `not_lipschitzOnWith_force`, hence not locally Lipschitz; the two stronger conditions imply a locally Lipschitz force, so they fail too.
C. Norton's claim
D. Existence without uniqueness
The force is continuous and the configuration space is finite-dimensional, so by Peano's theorem, `NewtonianSystem.hasLocalSolutions_of_continuous_force`, every initial datum has a local solution: the failure of determinism is purely one of uniqueness. Pending the proof of Peano's theorem in Mathlib, these results are marked `@[sorryful]`.
12 declarations
The Norton dome as a Newtonian system
The Norton dome is modeled as a conservative Newtonian system in a one-dimensional Euclidean space (representing the arc length along the dome's surface). The system is defined by a mass and a potential energy function , both provided by the underlying parameters of the dome . This definition ensures that the mass is positive () and that the potential energy function is differentiable.
The Force of the Norton Dome as a Newtonian System Equals its Basic Force
For a Norton dome , the force function derived from its representation as a Newtonian system (defined as the negative gradient of its potential energy ) is equal to the specific force function defined in the dome's basic model.
The Equation of Motion of the Norton Dome as a Newtonian System is Equivalent to its Basic Equation of Motion
For a Norton dome , a trajectory (where the position is represented as a point in a one-dimensional Euclidean space) satisfies the equation of motion derived from the dome's formulation as a general Newtonian system if and only if it satisfies the equation of motion defined in the dome's basic model.
The Solutions of the Norton Dome as a Newtonian System are Equivalent to its Basic Solutions
For a Norton dome , a trajectory (where the position is represented as a point in a one-dimensional Euclidean space) is a solution of the dome viewed as a Newtonian system if and only if it is a solution of the dome as defined in its basic model.
The Norton dome is not deterministic
The Newtonian system associated with the Norton dome is not deterministic. That is, it is not the case that any two solutions to the equations of motion which share the same position and velocity at some initial time must be identical for all time .
The Norton dome violates Newton's first law
For a Norton dome modeled as a Newtonian system, Newton's first law is not satisfied. Specifically, a particle at rest at the apex of the dome, where the net force vanishes, does not necessarily remain at rest for all time.
The force of the Norton dome is not locally Lipschitz
For a Norton dome , the force function of its associated Newtonian system is not locally Lipschitz continuous. Specifically, the force is not Lipschitz on any neighborhood of the apex of the dome.
The force of the Norton dome is not globally Lipschitz
For a Norton dome , the force function of its associated Newtonian system is not globally Lipschitz continuous.
The potential of the Norton dome is not
For a Norton dome , the potential energy function of its associated Newtonian system is not twice continuously differentiable (i.e., ).
Existence of a non-deterministic Newtonian system with a continuous force
There exists a Newtonian system in one-dimensional Euclidean space such that its associated force function is continuous, but the system is not deterministic. In this context, a system is non-deterministic if there exist at least two distinct trajectories and that satisfy the equations of motion and share the same position and velocity at some initial time , yet diverge at other times.
Existence of Local Solutions for the Norton Dome
For a Norton dome , modeled as a Newtonian system in a one-dimensional Euclidean space , there exist local solutions for every initial position and velocity. Specifically, for any initial position and initial velocity , there exists a duration and a trajectory such that: 1. The initial conditions are satisfied: and . 2. The trajectory and its velocity are differentiable on the interval. 3. The Newtonian equation of motion is satisfied for all , where is the mass and is the force function associated with the dome.
Existence of a Non-Deterministic Newtonian System with Continuous Force and Local Solutions
There exists a Newtonian system with a one-dimensional configuration space such that the force function is continuous and the system admits local solutions for every initial position and initial velocity , yet the system is not deterministic (i.e., the solutions to the equation of motion are not unique for all initial conditions).
