Newton From Moving Parts: What This Mechanism Explains, and the Ten Open Items
July 2026
Newton wrote down how gravity behaves in 1687, and it has worked ever since. He declined to say why — “hypotheses non fingo,” I feign no hypotheses. Einstein later gave a complete account of gravity as curved spacetime: geometry, not machinery. Neither man offered a picture of gravity as a physical process with moving parts.
This is an attempt at that picture. Paper 17 works it out formally. This article says plainly what it produces, and lists what it doesn’t — so the open items can be worked one at a time instead of argued about all at once.
The idea
Space is full of tiny particles moving in every direction, striking everything from all sides. On its own that pushes nothing anywhere: the pushes cancel.
Put two objects near each other and they stop cancelling. Each object blocks a little of the rain that would otherwise have struck the other. Each now sits in a slight shadow, pushed a bit less from the direction of its neighbour than from everywhere else. Less push from one side is indistinguishable from a pull toward it.
That’s the whole mechanism. Gravity as a shortfall of pushing, not a pull.
What it produces: Newton, exactly
Worked through carefully, the shadow between two bodies produces a force that is:
- Inverse-square — twice the distance, one quarter the force.
- Proportional to each body’s mass — double either mass, double the force.
- Blind to how that mass is arranged — a small dense body and a large diffuse one of the same mass pull identically.
Those three properties are Newton’s law of gravitation. There is nothing left over.
Two things are worth emphasising, because they’re stronger than they sound.
It’s exact, not approximate. This isn’t a formula that works far away and drifts up close. The result holds precisely, at every separation where the two bodies aren’t overlapping.
The inverse-square isn’t assumed — it appears on its own. The calculation starts by adding up how much material lies along each line of sight. Converting that bookkeeping into a sum over volume introduces a factor of one-over-distance-squared automatically, purely from the geometry of the change. Nobody puts it in. It arrives.
Why agreeing with Newton is the goal, not a weakness
It would be easy to read “it reproduces Newton” as “so it tells us nothing new.” That gets the purpose backwards.
The kinetic theory of gases doesn’t earn its keep by disagreeing with the gas laws. It earns it by producing them — showing that pressure and temperature are what you get when countless molecules bounce around. The law was already known. Explaining where it comes from was the achievement.
Same here. This isn’t a competitor to Newton trying to out-predict him. It’s an attempt to explain him. Reproducing the known law exactly is the pass condition. A mechanism that couldn’t reproduce Newton would be finished on the spot.
Where Einstein fits
General Relativity and Newton agree almost everywhere in everyday conditions — around stars, planets, and everything in between. Throughout that range, this mechanism agrees with both.
It’s worth putting a number on “agree.” Across the whole everyday range — from weights on a laboratory bench, out through the Earth, the Moon, and the planets, to the edge of the solar system — this mechanism, Newton, and Einstein all give the same answer to better than 99.9999975%.
That worst case is Mercury, the fastest-moving and most deeply-sunk planet, where the three differ by about one part in 39 million. Everywhere else they agree more closely still — at the scale of a laboratory experiment, to roughly one part in 10²⁶.
Against Newton specifically the agreement isn’t merely close, it’s exact. There’s no distance, mass, or density in that range where the two give different answers at all.
What defines that range needs stating carefully, because the obvious way to say it is wrong. It isn’t simply “any distance from a centimetre to a trillion metres, and any mass up to the Sun’s.” The real condition is about how strong the gravity is at the place you’re standing — and you can stay well inside those distance and mass figures while badly breaking it.
The Sun’s own surface is the easy example. It’s obviously within “up to the Sun’s mass,” but gravity there is 83 times stronger than anywhere the figure above applies, and agreement drops to 99.9998%. A white dwarf — a burnt-out star with less mass than the Sun packed into something Earth-sized — breaks it by a factor of 5,000, and agreement falls to 99.987%.
So the honest statement is: ordinary conditions — planets, moons, orbits, laboratory benches — where gravity is weak and things move slowly compared to light. Everyday solar-system gravity, in other words. Not the surface of a star, and not anything collapsed and dense.
They part company in a small set of effects Newton doesn’t capture and Einstein does: the slow rotation of Mercury’s orbit, starlight bending twice as much as Newton predicts, clocks running slower deeper in a gravitational field, and a few others.
This mechanism doesn’t produce those — but it doesn’t contradict them either. It gives no answer at all, because the calculation assumes nothing is moving and everything happens instantly. Those two assumptions remove exactly the effects Einstein’s corrections describe. That’s silence, not disagreement. Nothing here conflicts with any measurement ever taken.
Getting into that territory means relaxing those assumptions, which is real work, not a footnote. It’s item 10 below.
Two things about that 99.9999975% are worth being straight about.
Tiny doesn’t mean invisible. At any given instant, Newton and Einstein differ at Mercury by about one part in 39 million — but that difference always leans the same way, so it piles up. Over the 415 orbits Mercury completes in a century it adds to roughly 43 arcseconds of drift in the orbit’s orientation, which astronomers measured in the 1800s and couldn’t explain. Einstein’s theory gets it right; Newton’s doesn’t. Since this mechanism reproduces Newton exactly, it inherits that miss exactly too.
Step outside the range and agreement doesn’t fade — it breaks. Light travels at the speed of light, which violates the slow-motion condition outright. There, the naive Newtonian answer for starlight bending past the Sun is 0.88 arcseconds and the correct answer is 1.75 — off by a factor of two, a 100% disagreement, not one part in millions. That’s why “everything moving slowly” is a hard boundary rather than fine print.
The ten open items
Everything above holds under a specific set of assumptions. Here is everything those assumptions leave out, in plain terms, with an honest status on each.
1. Double shadowing. When two bodies both block the rain, the second sits in flux the first already thinned. Only the first-order effect was calculated. Open — not yet computed, and the most clearly defined next calculation.
2. Odd shapes. The exact result assumes spheres. Newton’s own shell theorem carries the same restriction, so this isn’t a special weakness here. Open.
3. The strength dial. The mechanism produces gravity’s shape but not its strength. Nothing yet connects the absorption constant to the measured value of G. Until that’s fixed, “the rain must be weak” can’t be turned into an actual number for the Earth or a laboratory weight. Open, and it gates several others.
4. Drag. Moving through the rain should feel like a headwind, slowly decaying orbits. Two fixes were tried against real data — letting particles bounce instead of being absorbed, and letting the rain get dragged along near a planet. Both failed; the second is ruled out by starlight measurements dating to 1728. Closed, negative.
5. Heating. Rain strong enough to make gravity should also cook everything it strikes. The maths does work in one specific range, staying under Earth’s measured heat output. But checked against how white dwarf and red giant stars actually cool, it requires the rain particles to weigh between 70 grams and 70 kilograms — objects you could hold. Contradicted wherever it has been tested. Open only at a shorter range nobody has computed yet.
6. Lag and aberration. If the rain travels at any finite speed, the pull points slightly behind where the other body actually is, and orbits destabilise. This is Laplace’s original objection. Untouched.
7. Where the rain comes from. Something must supply this flux and replenish it indefinitely, everywhere, forever. Untouched.
8. No candidate particle, no relativistic version. No known particle has the required properties, and there’s no formulation compatible with relativity’s symmetry requirements. Untouched.
9. Every other scale. Beyond the everyday range — galaxy rotation, black hole collapse, atomic and nuclear scales — the wider theory fails, in some cases by enormous margins. Documented in detail in Paper 14. Known and recorded.
10. Einstein’s corrections. Mercury’s orbit, the doubled light bending, gravitational time dilation. Excluded by the assumptions rather than attempted and missed. Open.
One related tension worth naming, since it sits at the boundary: earlier papers in this series escape items 4 and 6 by proposing the rain moves at effectively infinite speed. But gravity’s own influence has been measured to travel at the speed of light, to extraordinary precision, from a neutron star collision observed in 2017. If the pull is carried by the rain, those two claims pull against each other. This isn’t part of the result above — the formal work assumes nothing about speed — but it’s a real problem waiting in item 6.
Where that leaves things
The mechanism produces Newton’s law exactly, from moving parts, with the inverse-square emerging rather than assumed. Within everyday conditions it agrees with both Newton and Einstein, and contradicts no measurement.
Of the ten items, one is closed negatively (drag), one is contradicted wherever it has been tested (heating), one is documented as failing at other scales, and seven remain genuinely open.
Item 1 is the place to start. It’s fully specified, needs no new assumption, and settles decisively either way: extend the calculation to second order and find out whether mass-proportionality survives double shadowing — and if so, at what point it would breach the equivalence-principle limit. Combined with item 5’s flux requirement, that turns a 150-year-old qualitative argument into a closed inequality: either a real surviving window, or a proof there isn’t one.
That’s the honest position. The foundation is fixed and checked; the list above is the work.
The formal treatment, including the full derivation, the numbered assumptions, and independent numerical verification, is in Paper 17.