Something that puzzles me if gravity is just bending of space time near a mass then what is gravitational force?

If say two massive bodies were perfectly at rest relative to each other they would bend space time around themselves but they shouldn't ever move towards each other since they are not moving in the bent space.

In short why should one object fall into bent space of other object leading it to follow the bent space to collide with the mass, why did it move let alone its path and direction?

Is Space time bending in 4th dimension real or just a visualization of the actual process?

  • 1
    $\begingroup$ See Why would spacetime curvature cause gravity? and all its linked questions. $\endgroup$ – ACuriousMind Mar 30 '16 at 19:26
  • $\begingroup$ i really don't understand your statement:'If say two massive bodies were perfectly at rest relative to each other they would bend space time around themselves but they shouldn't ever move towards each other since they are not moving in the bent space.' Some other force prevent these bodies to move-this doesn't mean that spacetime is not 'curved' due to their masses. $\endgroup$ – user98038 Mar 30 '16 at 22:06
  • 3
    $\begingroup$ Possible duplicate of: How does “curved space” explain gravitational attraction? $\endgroup$ – John Rennie Mar 31 '16 at 9:09

A. Spacetime bending is described by the metric tensor.

B. By the equivalence principle, a particle's motion at "free fall", i.e. moving as dictated by gravity, is indistinguishable from an inertial motion

C. The inertial motion of particles in a curved space-time is descibed by geodesics which, unless space-time is flat, they are not straight lines but a generization of them.

D. The goedesic equation is includes the functions (Christoffel symbols) defined through the metric tensor and its derivatives.

Put the above ingredients together and then you should deduce what is the gravitational force.


Since particles that spin don't follow geodesics and particles with mass or energy or momentum or stress don't follow geodesics, the linked questions and answers about geodesics don't answer your question.

You want to know how real objects move, not how test particles move.

And this is handled by the Cauchy problem in General Relativity. Basically, you specify a 3d space that has some 3d spatial curvature, and you also specify some additional structure that indicates, not the full spacetime curvature, but just a little but about how it is changing right now.

It's like specifying both the initial position and the initial velocity, it's hopeless to get the motion without first specifying both. And just like you can use $\vec F=m\vec a$ to hope to get motion from the initial position, initial velocity, and the force, so you can use Einstein's Equation to try to get the full spacetime.

The Einstein Equation has a left hand side that has the Einstein Tensor. And that Einstein tensor has second derivatives of the metric tensor. And the Einstein Equation has a right hand side that is the Stress-Energy tensor.

If you know the Stress-Energy tensor and you know the metric and the first derivatives of the metric then you can hope to solve for the second derivatives of the metric and then find the full evolution over time.

It's not as simple as in Newton's theory. The first complication is similar to electromagnetism. In electromagnetism you can use the Maxwell Equation to find how he initial electromagnetic field evolves into the electromagnetic field in all of spacetime, but the initial electromagnetic field can not be arbitrary. The divergence of the initial magnetic part of the electromagnetic field has to be zero. So you can't write down just any fields and then evolve them.

Similarly, you can't write down just any initial manifold and any initial Stress-Energy tensor and any initial spatial metric and any arbitrary information about first derivatives.

And secondly, even when you have good initial conditions that meet the constraints. You have to extract the second derivatives (which isn't easy becasue the equations are not linear) and then use them to make a complete solution. Which is also hard.

And furthermore, not every manifold that satisfies the Einstein Equation has a Cauchy problem setup. There are even solutions to Einstein's Equation that have spatial slices that are perfectly flat, perfectly empty, and perfectly infinite and look just like a spatial slice of empty flat Minkowksi space. And yet some of those solutions are different than each other.

So even the simplest possible case of a flat empty infinite universe with nothing in right now, could evolve differently, and there simply is no law of physics telling it that it has to have a particular future development or a particular past development.

Physics is a science. And the theories tell you to restrict your models to only certain models. Electromagnetism tells us to restrict our initial magnetic field to have zero divergence and it tells us to restrict our electromagnetic fields to have their eceokve and their initial values to related to each other in a particular way. So it says to only use certain fields for your models.

And you use the models to make your predictions. So part of the theory is that some particular one of the restricted initial magnetic fields will be able to match your observations, and that one of its evolutions will match your later observations.

So in General Relativity the equation and the theory really tell you to select manifolds and metrics and Stress-Energy tensors that satisfy the Einstein Equation and the evolution equations for the specific things that make up the Stress-Energy tensor (which then cause the evolution of the Stress-Energy tensor).

So the theory says to restrict yourself to just those models. And the theory is good if there are models that match observations.

Think of science as tell you what the world does not do. That's what it has to do, becasue we want science to be falsifiable.

So science hypothesizes that the universe doesn't evolve in a way to look like a manifold that does not satisfy the Einstein Equation. So if it ever did, we'd know the theory is wrong and we could modify it or replace it. It doesn't really do anything more than that.


Not the answer you're looking for? Browse other questions tagged or ask your own question.