Newton's Bucket and Absolute Motion

The Leibniz–Clarke correspondence left the absolute–relational debate with one unresolved residue: absolute acceleration and rotation. Position and uniform velocity through absolute space are undetectable, and the relationist is right to deny them; but rotation seems to make a real, observable difference — one that appears to need something more than the relations among bodies to explain. Newton dramatised this with two thought experiments in the Principia's Scholium: the rotating bucket and the two globes. Together they form the most durable argument for the reality of something absolute in space, and they set the problem that Mach, Einstein, and modern philosophy of physics have all had to answer.

This page presents the experiments and the interpretive options; the relational reply is developed on the Mach's principle page, and the modern spacetime resolution on the absolute vs. relational and general relativity pages.

The rotating bucket

Hang a bucket of water from a twisted cord, and let it go. Newton describes four stages:

1. Bucket and water both at rest. The water's surface is flat. There is no relative motion between water and bucket.
2. Bucket rotating, water not yet. The cord unwinds and the bucket spins, but friction has not yet dragged the water along. Now there is maximal relative motion between water and bucket — yet the water's surface is still flat.
3. Water caught up, rotating with the bucket. Friction has brought the water up to the bucket's speed. Now there is no relative motion between water and bucket — yet the water's surface is concave, climbing the walls.
4. Bucket stopped, water still spinning. Again there is large relative motion between water and bucket, but the surface stays concave until the water slows.

The moral is stark. The concavity of the surface — the inertial effect that reveals the water is "truly" rotating — correlates not at all with the water's motion relative to the bucket. At stage 2 there is much relative motion and no effect; at stage 3 there is no relative motion and maximal effect. So the water's true rotation cannot be its rotation relative to its immediate surroundings. Relative to what, then, is it rotating? Newton's answer: relative to absolute space.

The two globes

Newton reinforced the point with a scenario stripped of all landmarks. Imagine two globes joined by a cord, alone in an otherwise empty universe, and suppose the cord is under tension. In a completely empty universe there are no other bodies, so there is no relative motion of anything — yet the tension in the cord would tell us the pair is rotating, and even let us measure the rate. Since there are no bodies for the rotation to be "relative to," the rotation must be absolute. Here the relationist seems to have nothing at all to appeal to: the effect persists when every relatum has been removed.

The structure of the argument

Newton's inference can be laid out as:

1. True rotation produces real inertial effects (concave surface, cord tension) that are detectable without reference to any external body.
2. These effects do not track motion relative to surrounding bodies (the bucket) — indeed they persist when no bodies are present (the globes).
3. Therefore true rotation is not relative motion with respect to bodies.
4. Therefore there is a non-bodily standard of rotation — absolute space.

The relationist must block the move from (3) to (4): grant that rotation is not relative to nearby bodies, but deny that the only alternative is a substantival absolute space.

Interpretive options

Three broad responses have shaped the debate:

• Newtonian substantivalism accepts the argument: absolute space (or absolute rotation with respect to it) is real, and the inertial effects are its fingerprints. The cost is the shift arguments — a full absolute space also implies undetectable absolute position and velocity, which we have independent reason to reject.
• The Galilean-spacetime compromise grants that the effects are absolute but denies that a standard of rest is needed. What the bucket reveals is absolute acceleration, and that requires only the affine (inertial) structure of neo-Newtonian spacetime — a distinction between straight and curved worldlines — not Newton's superfluous privileged rest frame. This concedes the relationist's point about position and velocity while keeping exactly what the bucket demands. It is the mainstream modern reading. (See absolute vs. relational.)
• Machian relationism attacks premise (2): it denies that the globes in a truly empty universe would exhibit tension at all. Inertia, on this view, is not a relation to space but a relation to the total matter of the universe — the "fixed stars." The water's surface curves because it rotates relative to the great mass of distant matter; empty the universe and the effect disappears. This is Mach's principle, and its (partial, contested) vindication in general relativity is one of the deepest threads in the subject.

Why it still matters

The bucket is not a historical curiosity. It poses a question every theory of spacetime must answer — what distinguishes inertial from accelerated motion? — and the answers track deep commitments. Newtonian mechanics posits absolute space; special relativity replaces it with the Minkowski metric and its inertial structure; general relativity makes that structure dynamical, so that the "fixed stars" and the local inertial frames are related by the field equations — a partial realisation of Mach's dream, though whether general relativity is fully Machian remains disputed. In every case the bucket is the test the theory must pass.

Where this sits

Newton's bucket is the empirical heart of the ontological question: where the shift arguments favour the relationist, the bucket favours the substantivalist, and the two together isolate acceleration as the true battleground of substantivalism and relationism. The relationist rescue is Mach's principle; the substantivalist retreat is Galilean spacetime; and both are transformed once the metric becomes a dynamical field. The next page pursues the relational response and its fortunes in modern physics.