Does Newtonian dynamics need Euclidean space?

This paper presents an elementary deduction of Newtonian force from Kepler's laws by generalizing Keplerian motion to non-Euclidean spaces with homogeneous functions, thereby revealing new convexity properties of orbits and describing their generalized hodographs.

The Gist

Imagine you are a cosmic cartographer trying to map the dance of planets around a star. For centuries, we've used a very specific map: Euclidean space. Think of this as a perfectly flat, infinite sheet of graph paper where distances are measured with a standard ruler, and gravity pulls everything straight toward the center, like a magnet.

Alain Albouy's paper asks a fascinating question: Does the universe have to be drawn on this flat graph paper? Or could the laws of planetary motion work on a different kind of "fabric" entirely?

Here is the story of the paper, broken down into simple concepts and analogies.

1. The Standard Recipe (Newton and Kepler)

First, let's look at the recipe we already know.

• The Setup: A planet orbits a Sun.
• The Rules (Kepler's Laws):
  1. The path is an oval (ellipse).
  2. The planet sweeps out equal areas in equal times (it speeds up when close to the Sun, slows down when far).
  3. There is a strict math relationship between how long a trip takes and how big the orbit is.
• The Result: Newton figured out that if you follow these rules on a flat sheet of paper (Euclidean space), the force pulling the planet must be gravity, which gets weaker the farther you get ($1/r^2$).

Albouy starts by showing that you can reverse this process. If you assume Kepler's rules are true, you can mathematically prove that the force must be Newton's gravity. It's like deducing that a cake must have been baked with flour because it tastes like flour.

2. The Big Twist: Changing the "Ruler"

Now, Albouy asks: What if we change the definition of "distance"?

In our normal world, the distance from the center to a point $(x, y)$ is $r = \sqrt{x^2 + y^2}$. This creates a perfect circle.
But what if we used a different "ruler"? Imagine a world where the distance is defined by a weird shape, like a square or a diamond.

• The Analogy: Imagine you are walking in a city with a grid of streets (like Manhattan). The distance isn't a straight line through buildings; it's the number of blocks you walk. If you walk 3 blocks East and 4 blocks North, the "distance" is 7, not 5.
• The Math: Albouy replaces the standard distance formula ($r$) with a flexible, custom shape function called $\rho$ (rho). This function can be stretched, squashed, or shaped like a four-leaf clover.

3. The "Kepler-Jacobi" Universe

Albouy discovers that even if you use this weird, non-standard ruler ($\rho$), Kepler's laws still work!

• The planets still move in closed loops.
• They still sweep out equal areas.
• The math connecting the orbit size and the time it takes still holds, but the formula changes slightly to fit the new shape.

He calls this the Kepler-Jacobi problem. It's like saying, "If I change the shape of the track, the runners still follow the same basic rules of racing, even if the track looks like a square instead of a circle."

4. What Changes? (The "Hodograph" and Energy)

While the rules of the race stay similar, the nature of the race changes in two surprising ways:

A. The Velocity Map (The Hodograph)
In the normal universe, if you plot the planet's speed and direction on a separate map, it draws a perfect circle. This is a beautiful geometric fact discovered by Hamilton.

• In the new universe: That circle disappears. Depending on the shape of your "ruler" ($\rho$), the speed map might look like a squashed oval, a star, or a weird blob. The "circle" is replaced by whatever shape your ruler dictates.

B. The Loss of "Energy"
In our normal universe, we have a concept called Energy (Kinetic + Potential). It's a conserved quantity, like a bank account that never changes balance.

• In the new universe: Albouy shows that for these weird shapes, Energy doesn't exist anymore. You can't define a "potential energy" that stays constant. The system is too distorted.
• The Metaphor: Imagine a roller coaster. In the real world, the total energy (height + speed) is constant. In this new world, the track is so weirdly shaped that the concept of "height" breaks down. You can't say the coaster has "potential energy" because the ground itself is bending in a way that defies our usual physics.

5. So, Is This Real Physics?

The paper ends with a reality check.

• Math vs. Reality: This is a brilliant piece of mathematics. It shows that the logic of planetary motion is more flexible than we thought. It works on "convex" shapes (shapes that bulge outward, like a ball or a cube, but not a donut).
• The Catch: Real gravity (and electricity) seems to be perfectly symmetrical (isotropic). It pulls equally in all directions. A "weird ruler" universe implies gravity pulls harder in some directions than others.
• Conclusion: While we probably don't live in a Kepler-Jacobi universe (our gravity looks too round), this math helps us understand the deep structure of how forces and orbits relate. It's like studying a fictional alien physics to better understand our own.

Summary

Albouy's paper is a journey into a parallel mathematical universe.

1. Standard Universe: Flat space, circular orbits, standard gravity, energy is conserved.
2. Kepler-Jacobi Universe: Distorted space, "squashed" orbits, weird gravity, energy is lost.

The paper proves that the beautiful dance of the planets is robust; it can survive even if we change the very fabric of space, provided we adjust the rules of the dance accordingly. It's a reminder that while our universe is specific, the mathematical logic behind it is surprisingly flexible.

Technical Summary

1. Problem Statement

The paper investigates the foundational geometric requirements of Newtonian dynamics, specifically the inverse-square law of gravitation (Kepler's problem). The central question is whether the formulation of Newton's equations of motion and the resulting Keplerian laws are intrinsically tied to Euclidean space (specifically the Euclidean norm $r = \sqrt{x^2+y^2+z^2}$) or if they can be generalized to other geometric structures while preserving the qualitative nature of the solutions (e.g., closed orbits, area laws).

The author seeks to:

• Re-derive Newton's law of attraction directly from Kepler's laws without relying on standard "tricks" (like the $u=1/r$ substitution).
• Generalize these laws to spaces defined by a homogeneous function $\rho$ rather than the Euclidean distance $r$.
• Determine which properties of the Kepler problem (planarity, convexity, hodographs, energy conservation) survive in this generalized "Kepler-Jacobi" setting.

2. Methodology

The author employs a pedagogical and geometric approach, reversing the historical derivation:

• Derivation from Kepler to Newton: Instead of starting with $F=ma$, the author starts with Kepler's three laws.

  • Law (K1): Uses the focus-directrix equation of a conic section ($r = \alpha x + \beta y + p$) rather than the two-foci definition. This formulation handles parabolas and hyperbolas more naturally.
  • Law (K2): Uses the conservation of areal velocity (angular momentum $C$) to determine the velocity vector components.
  • Law (K3): Uses the relationship between the period, semi-major axis, and mass to fix the proportionality constant in the acceleration.
  • Differentiation: By differentiating the focus-directrix equation twice with respect to time and utilizing the homogeneity of the distance function $r$, the author derives the Newtonian acceleration vector $\ddot{\mathbf{r}} = -m \mathbf{r}/r^3$ directly.

• Generalization (Kepler-Jacobi): The author replaces the Euclidean distance $r$ with a general function $\rho(x, y)$ that is positively homogeneous of degree 1 ($\rho(\lambda \mathbf{q}) = \lambda \rho(\mathbf{q})$).

  • The orbit equation becomes $\rho = \alpha x + \beta y + p$.
  • The acceleration is derived using the Hessian matrix of $\rho$.
  • The author analyzes the convexity properties of $\rho$ to determine the stability and shape of the orbits.

3. Key Contributions

A. A Trick-Free Derivation of Newton's Law

The paper presents a direct derivation of the inverse-square law from Kepler's laws using the focus-directrix formulation. This method avoids the standard substitution $u=1/r$ and the change of independent variable to the polar angle $\theta$, offering a more transparent geometric link between the laws of motion and the geometry of conics.

B. The Kepler-Jacobi Generalization

The paper introduces a generalized dynamical system where the "distance" is defined by a homogeneous function $\rho$.

• Generalized Equation of Motion: The acceleration is given by:
  $$\ddot{\mathbf{r}} = -m \rho_{\{2\}} \mathbf{r}$$
  where $\rho_{\{2\}}$ is a scalar factor derived from the Hessian of $\rho$ (specifically, the factor relating the Hessian matrix to the projection matrix).
• Generalized Kepler's Third Law: The standard law $n^2 a^3 = m$ is replaced by a condition on the areal velocity $C$ and the semi-parameter $p$:
  $$\frac{C^2}{p} = m$$
  This formulation is shown to be more robust, applying to parabolic and hyperbolic orbits where the standard semi-major axis $a$ is undefined or infinite.

C. Geometric and Convexity Analysis

The author establishes that the qualitative behavior of the orbits depends on the strict convexity of the function $\rho$:

• Homogeneous Strict Convexity: If $\rho$ satisfies specific convexity inequalities (related to the Hahn-Banach theorem), the orbits exhibit properties similar to the Euclidean case:
  • Non-radial orbits cover the boundary of a strictly convex domain containing the origin.
  • Orbits do not self-intersect.
  • There exist open sets of periodic orbits.
• Intersection Property: Two distinct orbits intersect at most twice.

D. Modification of the Hodograph

In the classical Kepler problem, Hamilton showed that the velocity vector (hodograph) traces a circle.

• In the Kepler-Jacobi generalization, the hodograph is no longer a circle.
• Instead, the vector $(-\partial \rho/\partial y, \partial \rho/\partial x)$ traces a unique curve determined solely by the geometry of $\rho$. The actual hodograph is a translation and scaling of this curve.

E. Loss of Energy Conservation

A critical finding is that energy conservation is not guaranteed in the generalized system.

• If $\rho$ is defined by a quadratic form (equivalent to a Euclidean metric with a linear transformation), an energy integral exists.
• If $\rho$ is a general homogeneous function (e.g., $\rho = \sqrt[4]{x^4+y^4}$), no scalar "energy" first integral can be constructed. The system is defined purely by the geometric constraints and the generalized force law.

4. Results

• Existence of Generalized Dynamics: Newtonian dynamics can be formulated in non-Euclidean vector spaces defined by homogeneous norms, provided the force law is adjusted to depend on the Hessian of the defining function $\rho$.
• Robustness of Kepler's Laws: The laws of areas (K2) and the specific relation between angular momentum and the semi-parameter (K3') hold in the generalized space. The law of ellipses (K1) is replaced by the law of $\rho$-conics.
• Convexity as the Key: The "Euclidean-ness" of the orbits (convexity, lack of self-intersection) is preserved if and only if the defining function $\rho$ is homogeneously strictly convex.
• Physical Limitations: The author notes that while mathematically consistent, these generalized systems lack a clear physical interpretation for gravity or electromagnetism, as real forces are typically isotropic at infinity. However, they may have relevance in anisotropic optical or light-particle contexts.

5. Significance

• Pedagogical Value: The paper offers a fresh, rigorous, and "trick-free" proof of the equivalence between Kepler's laws and Newton's inverse-square law, making the connection more accessible to students of geometry and analysis.
• Mathematical Physics: It bridges classical mechanics with convex geometry and functional analysis (specifically the Hahn-Banach theorem and Minkowski geometry). It demonstrates that the "elliptical" nature of orbits is a consequence of the convexity of the underlying metric, not necessarily the Euclidean metric itself.
• Historical Context: The work connects modern analysis with historical insights from Jacobi, Darboux, and Gauss, highlighting that Gauss had already emphasized the generality of the focus-directrix formulation and the $C^2/p$ law.
• Theoretical Limits: It clarifies the boundary conditions under which conservation laws (like energy) hold, showing that energy is a specific feature of quadratic metrics (Euclidean/Riemannian) and not a universal feature of central force problems in homogeneous spaces.

In conclusion, the paper argues that while Newtonian dynamics can be formulated in non-Euclidean spaces, the specific "Euclidean" features (circular hodographs, energy conservation, standard $1/r^2$ force) are special cases of a broader class of "Kepler-Jacobi" systems governed by convex homogeneous functions.