Norton's Dome and Physical Determinism: Singular Geometry, Regularization, and Nonuniqueness
Abstract
Norton's dome (Norton 2008) has been presented as a demonstration that Newtonian mechanics fails to be deterministic: a particle at rest at the apex of a specially shaped surface can remain there indefinitely or spontaneously begin to move after an arbitrary delay, consistent with Newton's laws. This paper argues that the dome does not establish physical indeterminism. It takes seriously, and answers directly, the three defenses Norton mounts in his original paper, in the order he presents them: (1) that no principled "fourth law" can be added to Newtonian theory to exclude the dome; (2) that none of the standard senses of "unphysical" apply, because excluding the indeterministic solutions would require independent knowledge of a target system that the dome, as a pure idealization, cannot supply; and (3) that the dome is a legitimate abstractive idealization for which the demand of approximability is simply the wrong demand.We answer all three in turn, and show that two independent criteria, dynamical instability and physical unrealizability, converge on exactly the same boundary, reinforcing rather than merely repeating one another. First, a criterion of local uniqueness-robustness, in the spirit of, though weaker than, Andronov and Pontryagin's (1937) structural stability, supplies a non-ad-hoc fourth law: within the natural oneparameter family of power-law force laws generalizing the dome, the boundary at which unique evolution from rest is guaranteed coincides exactly with the boundary between finite and infinite apex curvature, so that dynamical instability and physical inadmissibility are not two separate complaints but a single fact examined from two directions. Second, continuum mechanics supplies the independent knowledge Norton's own underdescription criterion requires: unbounded curvature entails unbounded bending stress, placing the dome's apex outside the class of surfaces classical continuum theory itself licenses, and this same fact explains why the dimensional incoherence visible in Norton's founding equation cannot be repaired by any physically grounded constant, in a way that distinguishes the dome from ordinary unrealizable idealizations such as the perfectly rigid rod. Third, Norton's own precedents for tolerating discontinuity at a singular limit, the rigid beam and the infinite rotor, undercut rather than support him: in both cases he relies on independent theoretical knowledge to treat the singular behavior as an artifact, a resource equally available, and for the same reason, in the case of the dome. A rigorous regularization argument further shows that no sequence of smooth, physically realizable domes has dynamics converging to the dome's indeterministic solutions; the indeterminism is generated by the singular idealization rather than revealed by it. We close by locating this reply relative to Korolev (2007), Malament (2008), and Fletcher (2012). The dome remains an elegant illustration of non-uniqueness in ordinary differential equations with non-Lipschitz right-hand sides. It does not show that Newtonian mechanics is physically indeterministic.
Related
- Regularization effects on the observational appearance of asymmetric thin-shell wormholes from a new string-cloud geometry
- Physical properties of charged black holes from the nonlinear electrodynamics model based on electric potential regularization
- Supporting functionals and singular boundary geometry of two-qubit entanglement of formation
- Geometry-induced chaos and scaling near two singular limits of an eye-shaped billiard
- Schrödinger's unpublished hidden-variable model in the 1944-45 debate over the "Collapse of Determinism"
BibTeX
@misc{baravi2026nortons,
title = {{Norton's Dome and Physical Determinism: Singular Geometry, Regularization, and Nonuniqueness}},
author = {Patrick Baravi},
howpublished = {HAL},
year = {2026},
month = sep,
url = {https://hal.science/hal-05747953v1},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Norton's Dome and Physical Determinism: Singular Geometry, Regularization, and Nonuniqueness
AU - Patrick Baravi
PY - 2026
DA - 2026/09/12
PB - HAL
UR - https://hal.science/hal-05747953v1
AB - Norton's dome (Norton 2008) has been presented as a demonstration that Newtonian mechanics fails to be deterministic: a particle at rest at the apex of a specially shaped surface can remain there indefinitely or spontaneously begin to move after an arbitrary delay, consistent with Newton's laws. This paper argues that the dome does not establish physical indeterminism. It takes seriously, and answers directly, the three defenses Norton mounts in his original paper, in the order he presents them: (1) that no principled "fourth law" can be added to Newtonian theory to exclude the dome; (2) that none of the standard senses of "unphysical" apply, because excluding the indeterministic solutions would require independent knowledge of a target system that the dome, as a pure idealization, cannot supply; and (3) that the dome is a legitimate abstractive idealization for which the demand of approximability is simply the wrong demand.We answer all three in turn, and show that two independent criteria, dynamical instability and physical unrealizability, converge on exactly the same boundary, reinforcing rather than merely repeating one another. First, a criterion of local uniqueness-robustness, in the spirit of, though weaker than, Andronov and Pontryagin's (1937) structural stability, supplies a non-ad-hoc fourth law: within the natural oneparameter family of power-law force laws generalizing the dome, the boundary at which unique evolution from rest is guaranteed coincides exactly with the boundary between finite and infinite apex curvature, so that dynamical instability and physical inadmissibility are not two separate complaints but a single fact examined from two directions. Second, continuum mechanics supplies the independent knowledge Norton's own underdescription criterion requires: unbounded curvature entails unbounded bending stress, placing the dome's apex outside the class of surfaces classical continuum theory itself licenses, and this same fact explains why the dimensional incoherence visible in Norton's founding equation cannot be repaired by any physically grounded constant, in a way that distinguishes the dome from ordinary unrealizable idealizations such as the perfectly rigid rod. Third, Norton's own precedents for tolerating discontinuity at a singular limit, the rigid beam and the infinite rotor, undercut rather than support him: in both cases he relies on independent theoretical knowledge to treat the singular behavior as an artifact, a resource equally available, and for the same reason, in the case of the dome. A rigorous regularization argument further shows that no sequence of smooth, physically realizable domes has dynamics converging to the dome's indeterministic solutions; the indeterminism is generated by the singular idealization rather than revealed by it. We close by locating this reply relative to Korolev (2007), Malament (2008), and Fletcher (2012). The dome remains an elegant illustration of non-uniqueness in ordinary differential equations with non-Lipschitz right-hand sides. It does not show that Newtonian mechanics is physically indeterministic.
ER -