Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

GR and Gravitional Field

  1. Sep 5, 2006 #1
    Some say that the reason Einstein developed the GR is because he wanted to avoid the spooky "action at a distance," since two massive bodies exerting forces on each other at a large distance with no "mediator" seems rather odd. Therefore Einstein invented the concept of spacetime and gravity could be defined as the curvature of spacetime.

    Now my question is that why didn't he just invented the concept of gravitaional field? That would resolve the spooky "action at a distance" right?
  2. jcsd
  3. Sep 5, 2006 #2
    a big part of his reasoning behind the curves in space-time come from his equivalence principle. Gravity and Acceleration feel the same. An accelerating object has a curved space-time line. Therefore it is only natural to assume that gravity must exist as curves in spacetime.
  4. Sep 5, 2006 #3
    Well I am not at all familiar with those "some" you are referring to. Perhaps you could provide some sources, otherwise this seems nothing more than rumor spreading.

    Anyway, even in GR there is no explanation for the "pulling" action of gravity, since the pulling is not part of the curvature, only the volume reducing and tidal effects (Ricci and Weyl tensors) are.
  5. Sep 5, 2006 #4


    User Avatar
    Staff Emeritus
    Science Advisor

    How do we measure an electric field? We take a test charge, and measure the force on it, relative to an uncharged particle.

    How do we generalize this to a "gravitational field"? We take a test mass and we want to measure the force on it due to gravity. But by the principle of equivalence, we can't distinguish forces due to gravity from forces due to acceleration. And there is no "gravitationally neutral" particle that we can use as a reference. So we are stuck, right from the get-go, in trying to define the gravitational field in the manner we are used to. The problem is that gravity acts on everything, so there isn't any reference about what force is due to "gravity" and what force is due to "inertia". And if the equivalence principle is to work properly, this is as it should be.

    So traditional ideas of gravity as a field, just like the classical electromagnetic field, don't pan out, just because gravity is universal and affects everything. Einstein's solution to this dilemna was to say that the gravitational force was due to geometry.

    I believe that Einstein, in his later years, came to regard the metric tensor as a sort of "gravitational field". It's not a field in the usual sense, but it is a tensor, and from it everything you want to know about gravity can be calculated if you know the metric. so it makes some abstract sense to call it a "field". Personally I think this is confusing, so I prefer to call the metric tensor the metric tensor, and let people think of "gravitational fields" as the traditional Newtonian sort of field.

    One thing that one can't do is attribute energy to any specific location of the "field", regardless of whehter one choses to call the metric tensor a field or just the metric tensor. This happens because of Noether's theorem. It turns out that GR is a bit general for its own good - because it is diffeomorphism invariant, it in general lacks the sort of single parameter time translation symmetry that usually results in a conserved energy. Because the time translation symmetry is an infinte group rather than a single parameter group, energy conservation does not automatically arise in general relativity.
  6. Sep 5, 2006 #5
    The topic of energy density in the "gravitational field" is of special interest to me. Indeed, it seems not possible to define an energy density similar to the energy density of the electromagnetic field. But would it maybe not be possible to talk (or define) about a gradient in energy density in the gravitational field ?
  7. Sep 6, 2006 #6


    User Avatar
    Science Advisor

    I thought the "spooky action at a distance" with which Einstein was uncomfortabe was quantum entanglement, not gravity. Have I got my history mixed up?
  8. Sep 6, 2006 #7


    User Avatar
    Staff Emeritus
    Science Advisor

    If you could determine the gradient of the energy density of the field, you could determine the energy density up to a scalar, so no, you can't find the gradient of the density either.

    The problem is not just an additive scalar. See for instance the following article about Noether's theorem:


    Last edited by a moderator: Apr 22, 2017
  9. Sep 7, 2006 #8
    Thank you Pervect for this detailed answer. I'll need some time to study it in more detail.
    What puzzles me though is that Einsteins field equations are nonlinear which seems to imply that there exists something like "gravitational energy". And if this can not be localized, one would naively expect that the strength or direction of the field would also not be defined correctly. How can one reconcile this ?
  10. Sep 7, 2006 #9


    User Avatar
    Gold Member

    In his 1924 essay "On the Ether", Einstein rejected Machian concepts in which gravitation and inertial effects are the result of "action at a distance". He insisted that gravitation and inertia arise from matter's interaction with the space in which it is embedded - an entirely local effect with no consideration of force acting over a distance. This "etheric" concept did not sit well with his contemporaries, though there a people pursuing it today. For instance, Thanu Padmanabhan, who is quite interested in the intersection of gravitation and quantum theory models vacuum as if it is an elastic solid with which matter interacts. Links to some of his writings on this subject are on his homepage.

    Last edited by a moderator: Apr 22, 2017
  11. Sep 7, 2006 #10


    User Avatar
    Staff Emeritus
    Science Advisor

    Defining gravity as a field does have similar problems. What works in GR is geometry, not fields.

    If you have an observer at a particular point, there exists an observer at that point who sees no "field". This is an observer following a geodesic, i.e. in free fall, at that point.

    This implies that "fields" can't transform as a tensor. If all components of a tensor are zero at a pint for one observer, they must be zero at that point for all observers.

    When we translate the math of GR to talk about "fields", this translation requires us to have some particular symmetry about the problem. Usually this is a static metric, one that isn't changing with time.

    If we could single out some particular background frame on physical reasons, we could use this to define the symmetry needed to define a field unambiguously in all cases. In GR this isn't possible - all frames are just as good, and there isn't any way to talk about the "field" at a point. This goes back to my earlier remarks, about how we lack a "gravitationally uncharged" reference parrticle. If we had such a particle, it would create the sort of symmetry one needs to define a field.

    So this is a short version of why, in GR, we have geometry rather than fields.

    I'm not sure if I should add this, but I will anyway. There is another approach to the issue. This is to assume that space-time is flat, and that gravity is a field. (Just ignore the fact that I just got through saying one couldn't do that :-)).

    The result one gets from this approach is that these "fields" cannot actually be observed. So we see that there is a point, we can "create" a field with the right assumptions, but we can't define one without making some extra assumptions. The problem is that we have to spell out all our assumptions to work with this notion of "field".

    In other words, this "field" description is coordinate dependent.

    This is a bit like assuming that there is some sort of special frame of reference in special relativity that defines "absolute motion", and then finding that there is no way to observe it.

    For a tutorial but rather technical paper on this, see http://xxx.lanl.gov/abs/astro-ph/0006423.
    Last edited: Sep 7, 2006
  12. Sep 7, 2006 #11
    Since I found out about this equivalence principle, I have been trying to understand it. I looked up some books and several online articles/texts on it, but I was still confused...until I found this text:

    I think this is a great online text for any beginners of GR who want to know what Einstein's equivalence principle is all about. A very through and clear text. :wink:

    edit1: Actually, I have one more article that might be very useful. I think that it would be best if the beginner reads the above article first and then read s the following article, which shows how the equivalence principle leads to Gravitaional time dilation.

    edit2: OK, after reading the above two articles, I think that the following article would be the best way to go for a beginner. It shows that why does it make sense for gravity to bend light according to the equivalence principle. Its really amazing! :cool:
    Last edited by a moderator: Apr 22, 2017
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook