A plate can help move itself by sinking. Where cold oceanic lithosphere descends into the mantle, its weight can pull on the plate still at the surface. This is slab pull, a major part of the answer to why Earth’s tectonic plates keep moving.
It is not the whole answer. Gravity also acts on elevated oceanic ridges, and slowly moving mantle exerts forces on plates. The useful picture is a system of interacting forces, rather than one engine pushing every continent in the same way.
The moving object includes more than crust
A tectonic plate is a piece of lithosphere: Earth’s crust and the rigid uppermost mantle beneath it. A continent can ride within a plate that also includes ocean floor. It does not have to plow independently through the surrounding seabed.
Below the rigid plates, hotter mantle can deform over geological time. That does not require a worldwide ocean of magma. The British Geological Survey describes a solid mantle that transmits both compressional and shear waves, while also flowing over very long timescales.
There is no contradiction in those two behaviors. Whether something appears rigid depends partly on the time over which a force acts. Rock that transmits the rapid disturbance of an earthquake can still change shape under sustained stress over millions of years.
Our companion on how we know what’s inside Earth explains the seismic evidence. Here, the question is what supplies the force that keeps the outer pieces moving.
Follow the cold edge
New oceanic lithosphere near a spreading ridge is hot and relatively buoyant. Farther from the ridge, it cools and becomes denser. At a subduction zone, an oceanic plate bends downward and enters the mantle beneath another plate.
The sinking section is the slab. When it is denser than its surroundings, gravity pulls it downward. Because that slab remains connected to the plate behind it, the sinking section can exert a pull on the rest.
The video’s slipping-blanket comparison captures this connection. A heavy section hanging over the side of a bed can pull on the fabric still on top. The descending section supplies a force through an attached object; someone does not have to push the far end continuously.
Earth adds complications that the blanket leaves out. Rock must bend, surrounding mantle must move, and resistance acts along the system. The comparison explains a direction of force, not the complete mechanics of a subduction zone.
The BGS account identifies slab pull as a major driver for most plate movement and notes the association between more extensive subducting edges and faster plates. That does not establish a universal percentage for how much motion slab pull explains. A plate’s geometry and surroundings matter.
Ridge push is another gravitational effect
Mid-ocean ridges stand higher than the older, colder ocean floor away from them. Gravity acting on this elevated lithosphere contributes a force directed away from the ridge. This is called ridge push.
The name can suggest magma acting like a piston. The gravitational explanation is more useful: hot, elevated lithosphere has a different arrangement of weight from cooler lithosphere farther away. BGS describes that elevation difference as the basis of the force.
Slab pull and ridge push therefore work at different parts of a plate. One concerns a dense section descending at a subduction zone; the other concerns elevated lithosphere near a ridge. Neither phrase means that every plate must move at the same rate or have the same balance of forces.
The mantle moves too
It would also be misleading to turn the mantle into a stationary surface across which plates slide. Hot mantle circulates slowly, with warmer, more buoyant material rising and colder, denser material sinking. This is convection.
The USGS chapter Some unanswered questions connects plate motion to mantle flow while emphasizing the importance of sinking slabs. Its history of the competing explanations is useful, but some of its illustrations and descriptions are simplified. In particular, its discussion should not be read as evidence for a globally liquid mantle.
A sinking slab moves through mantle and affects its surroundings. Mantle flow can in turn exert stresses on plates. Treating the surface and interior as interacting parts helps explain why a tidy drawing of circular currents carrying passive plates cannot describe every setting.
The remaining question is how those interactions balance in a particular place. Knowing that convection occurs does not identify the size and direction of every force on every plate.
Why the motion can continue
Earth’s interior contains heat left over from its formation, and radioactive decay supplies additional heat. The USGS account identifies uranium, thorium and potassium among the heat-producing elements. Heat escapes toward the much colder surface.
That temperature contrast allows density differences and circulation to persist. Cooling produces denser oceanic lithosphere, helping sustain the descending side of the system. Plates are not simply coasting from one ancient shove.
Their slow speed can hide the scale of the process. As an illustrative calculation, a steady three centimeters per year would accumulate to 30 kilometers in a million years. This is a unit conversion, not a forecast for a particular plate: actual rates and directions can change.
Nor does a slow average speed mean continuous smooth sliding at every boundary. BGS explains how faults can remain stuck while surrounding rock deforms and stores elastic energy. Sudden slip releases some of that energy as seismic waves. Long-term plate movement and a brief earthquake describe different timescales within the same system.
Keeping subduction going is different from starting it
An established sinking slab offers a clear mechanism for continued motion. Explaining how a new subduction zone begins is harder. A dense plate does not automatically bend downward wherever it becomes cold enough.
A 2020 Nature Communications study assembled evidence for more than a dozen initiation events during the last hundred million years. It found horizontally forced initiation dominant in that collection, with most events near existing subduction zones. That result concerns the studied events; it is not a demonstration that every subduction zone began the same way, or an answer to how Earth’s earliest subduction started.
The distinction prevents a common shortcut: explaining the pull of an already sinking slab does not fully explain how the first downward bend formed. Plate motion is well established. The history and local balance of the forces still give geologists specific problems to investigate.
