On the Transition from Newtonian Gravity to General Relativity ()
1. Introduction
Those who wish to master the special and general theories of relativity commonly have to pass certain milestones in questioning and re-thinking some of their most dearly held reality assumption. The first of these, without a doubt, is the realization that the apparent constancy and observer-independence of the speed of light in Maxwell’s electromagnetic wave equation causes space and time in special relativity to lose their familiar, Newtonian absoluteness. Once this vital fact has been grasped, the students of relativity can move on to explore in more detail the inner workings of special-relativistic dynamics, but they are liable to again feel confounded when these dynamics are subsequently integrated into a general-relativistic theory of gravity that ultimately aims to achieve nothing less than the rigorous mathematical elucidation of the large-scale geometric structure of the entire physical universe.
In facing this latter conceptual challenge, it is important, first and foremost, to thoroughly appreciate the fundamental significance of the equivalence of inertial and gravitational masses. For it is precisely this equivalence that renders the spacetime environment in a constant gravitational field indistinguishable from the spacetime environment that is experienced by a gravity-free observer who undergoes a steadily accelerated motion. And it is this equivalence as well and in consequence, that makes it possible to re-conceive a seeming acceleration under the influence of a gravitational force as a force-free uniform motion relative to an accelerated observer and that, thereby, also makes it possible to re-conceive the universe as a whole as a patchwork of infinitesimal constant-gravity regions that are each equal in structure to the spacetime world within a constant-acceleration frame of reference.
However, when it comes to the problem of how such a patchwork ought to be organized so as to yield a theory of gravity that is compatible, in the classical limit, with Newton’s absolute-spacetime conception of a gravitational force that acts at a distance instantaneously, the matter quickly gets confusing. For not only must those who endeavor to tackle this problem be familiar with the mathematical formalisms of differential geometry, but they also must learn to bridge in their minds the seemingly deep divide between these formalisms on the one hand and the more elementary mathematical tools employed by Newton on the other. And the purpose of the present paper, therefore, is to suggest a derivation of the field equations of general relativity that greatly narrows that mathematical divide or even closes it completely.
That said, we must hasten to add that our purpose is not to overrule or discredit common approaches to the bridging of the gap between the theories of gravity of Newton and Einstein but merely to offer an alternative point of view. The method of stratification (as explained in Chapter 12 of [1] ) and the consideration of weak-field limits (as discussed in Section 8.1 of [2] ), for example, are perfectly valid and well-known means of establishing the inherent compatibility of Newton’s and Einstein’s theories, but they do not demonstrate their actual formal identity—as we hope to do in Sections 3 and 4 below. Moreover, for further discussions of the relation between Newtonian gravitation and general relativity the reader is referred to [3] [4] [5] and [6].
2. Prerequisites and Result Summary
Since general relativity cannot be divorced from differential geometry, it behooves us to recall to begin with some pertinent mathematical facts and constructions: denoting by TM the tangent space bundle of a
-manifold M, it can be shown (see for instance [7], pp. 70-73) that there exists an open set
and a differentiable map
with the following properties:
1) For all
and all
the set
is an open interval containing zero, and
,
is a geodesic that satisfies the initial conditions
and
. Furthermore,
is maximal and unique in the sense that any geodesic that satisfies the same initial conditions is a restriction of
to a subinterval of
.
2) For all
there exists an open neighborhood
of
in
such that the restriction
is a diffeomorphism from
onto an open neighborhood
of p in M. (Note: w.l.o.g. we may assume that for all
and all
the set
is an open interval.)
Assuming further that M is equipped with a Lorentz inner product
(with the sign convention (+−−−)) and that p is a given point in M, we denote by
the set of timelike unit vectors in
, i.e.,
.
Moreover, by
we denote the set of all points
(with
as defined above) for which there exists a
and a
such that
, and by
we denote the geodesic velocity field that the exponential map induces on
, i.e., for
we set
(1)
Since
is a geodesic, it follows that
.
To complete our setup, we pick an arbitrary Lorentz frame
at p and create a frame field
on
by parallel shifting the vectors
along the geodesics that originate at p (and are described by
). This construction readily implies that
for all
and that
(2)
for all
and all
. Setting further
a dual basis field on
is
, and the general-relativistic analogue of the Jacobian derivative matrix of
is
,
where
is the row index and
the column index.
Using brackets, as usual, to indicate the taking of the directional derivative (that is,
), we will show in Sections 3 and 4, that the matrix
satisfies the equation
(3)
in the classical Newtonian case, in which particle motion is governed by the Laplace equation
, as well as in the free-particle special-relativistic case. Furthermore, since the validity of (3) in the classical Newtonian case turns out to be equivalent to the validity of the classical vacuum field equation
, it is perfectly reasonable to require, in light of these results, that Equation (3) be satisfied as well in general-relativistic vacuum spacetime. And it is precisely this latter requirement that turns out to be equivalent to the validity of the vacuum field equation
or equivalently,
Moreover, in Section 5 we will show that a similar unifying description—analogous to (3)—can be given for the classical and general-relativistic matter field equations as well. That is to say, in using velocity-field divergences, we will be able to reveal that the classical Newtonian and general-relativistic theories of gravity are formally essentially identical.
Introducing the natural, generalizing notation
,
we summarize our findings in the following table so as to illustrate thereby how seemingly disparate classical and relativistic field equations become perfectly unified when represented by velocity-field divergence equations:
As a note of caution, we wish to add that all the results derived in the present paper appear to be so elementary in character that it is difficult to imagine that they have not been previously established. However, since the present author is not aware of any pertinent reference, the results in question are here being offered—with considerable hesitation—as provisional novelties.
3. The Classical Vacuum Field Equation
To begin with, we consider the free-particle Newtonian case in which the possible particle trajectories are straight lines in four-dimensional absolute spacetime. Thus, we define the analogue of
in (1) (with the base point p at the origin
) via the equation
and observe that
This yields
and Equation (3) has therefore been shown to be valid.
To proceed, we assume that a Newtonian test particle of mass m moves in a gradient force field
that satisfies the field equation
Furthermore, setting
, we denote by
the spacetime curve, traced out by the test particle, that satisfies the boundary conditions
as well as the time-component condition
. Given this definition, it is natural and plausible to assume that there exists a constant
such that the boundary and time-component conditions above determine
uniquely for all
(or perhaps
for some large’ set
that contains
). That is to say, we will in essence assume that
is a well-defined vector field on a suitable subset of
that contains
. Given this assumption, it follows that
for all
, and that, by implication,
Hence
and, by implication,
(4)
Consequently, the classical vacuum field equation
is indeed satisfied if and only if
4. The Relativistic Vacuum Field Equation
In order to prove that Equation (3) is satisfied as well in the free-particle special relativistic case, we set
and
Given these definitions, we find (by way of a trivial computation) that
and since
it follows (again by way of a trivial computation) that
as desired.
In the light of this result and in the light as well of the preceding result concerning classical free particles, it is perfectly reasonable to expect that general-relativistic vacuum spacetime should be structured in such a way that
(5)
In order to show that this equation is indeed equivalent to the vacuum field equation
, we will now proceed to prove the following theorem:
Theorem 4.1. For all spacetime vector fields
and
it is the case that
Note:the term on the right is indeed a tensor because it is easily seen to be
-bilinear.
Proof. Given a spacetime event p, and given the basis fields
, defined in Section 2, we may employ Einstein’s summation convention in conjunction with (4) to infer that
as desired. □
For later reference and in order to further explore this result, we introduce the following definition:
Definition 4.2. Given a spacetime event
and given a tangent vector
, we say that a vector field
is a geodesic extension of
if
and if the domain of
contains an open neighborhood U of q such that
for all
.
Lemma 4.3. Every spacetime tangent vector
admits a geodesic extension.
Proof. This fact is essentially well known and hardly requires a proof, but one way to establish it is to use a local coordinate system at q to translate the geodesic equation
into a first-order system of differential equations and to find a solution of this system that satisfies the condition
. Alternatively, a proof using the exponential map is feasible as well. □
Corollary 4.4. If
is a geodesic extension of a vector
, then
Proof. This is a trivial consequence of Theorem 4.1 and the fact that
on an open neighborhood of q. □
Since the geodesic velocity field
, defined in (1), satisfies the equation
, we may apply Theorem 4.1 to infer that
if and only if (5) is satisfied. Using polarization in conjunction with the familiar symmetry of the Ricci-tensor (as well as the fact that the time-like unit vectors
span
as the base point p in the definition of
is properly varied), it follows that indeed Equation (5) is satisfied if and only if
.
To conclude our discussion in this section, it is worth noting that both terms in Equation (5) are frame-independent. That is to say, if
is some other arbitrary frame field, then
and
The first of these two equations is well known, and the second can be derived as follows:
5. The Matter Field Equations
According to (4), the classical matter field equation
is equivalent to
(6)
Thus, we only need to show that an analogous representation is valid as well for the matter field equation of general relativity. To this end we will consider to begin with the very simple special case where the curvature of spacetime is induced by the gravitational interactions of a swarm of free particles whose rest-frame mass density is
and whose unit-length geodesic velocity field is
(i.e.,
and
). In other words, we will assume that the stress-energy tensor is
or equivalently, that
To proceed, we require that (6) be valid as well in the general-relativistic case in which
is replaced by
. That is to say, we demand that
Inspired by this natural requirement, we establish the following general theorem:
Theorem 5.1. Let q be a fixed spacetime event in the domain of
. Then the matter field equation
(7)
is satisfied for all
if and only if
(8)
for any geodesic vector field
(i.e.,
),defined on an open neighborhood of q,that satisfies one of the following conditions:
a)
,
b)
for some spacelike unit vector
that is Lorentz-perpendicular to
, that is,
and
,
c)
for some vector
as described in (b).
Proof. Throughout the proof below we will assume that
is a Lorentz frame based at q such that
. (Note: this latter assumption will be needed only in the second part of the proof, not in the first.)
“
” If (7) is valid, then the geodesic equation
in conjunction with Theorem 4.1 implies that
Since (7) also implies that
it follows that
and therefore,
.
Given this equation, it is easy to verify that
whenever one of the conditions (a), (b), or (c) above is satisfied.
“
” Since all the tensors in Equation (7) are symmetric, it is sufficient—by polarization—to show that
for all
. To do so, we will show to begin with that
(9)
for all
. If
, then we pick a geodesic extension
of
(see Lemma 4.3) and apply Corollary 4.4 in conjunction with (8) and either (a) or (b) to infer that
(10)
as desired. If
and
, then
Consequently, if
is a geodesic extension of
, then Corollary 4.4 in conjunction with (8), (10), and (b) implies that
and therefore,
as desired. Finally, if
and if
is a geodesic extension of
then Corollary 4.4 in conjunction with (8), (10), and (c) implies that
and again we find that
as desired. Having thus established Equation (9), it follows that
and
Hence
as desired. □
Concerning the conditions (a), (b), and (c) in Theorem 5.1, we wish to remark that the validity of (6) in the Newtonian case naturally suggests that Equation (8) ought to be valid if (a) and (b) are satisfied because
is the rest-frame density and
and
are timelike and spacelike rest-frame vectors, respectively. Moreover, regarding the validity of (8) in the remaining case, where
satisfies (c), it is helpful to notice that Newtonian gravity can be regarded as a classical limit that emerges from general relativity as the speed of light diverges to infinity. For in adopting this point of view, the Newtonian spatial rest frame at q merges with the relativistic lightcone at q, and the lightlike vector
, multiplied with the Euclidean scaling factor
, may therefore be considered to be a spatial rest-frame vector in the Newtonian limit. However, regardless of whether we consider this latter interpretation to be convincing or not, Theorem 5.1 remains perfectly valid as a mathematical fact. So ultimately (c) is simply a condition that needs to be added in order to guarantee that (7) and (8) are mathematically equivalent and that, by implication, Newtonian gravity and general relativity may justifiably be viewed to be formally identical.
Furthermore, the somewhat unsatisfactory restriction to a swarm of particles moving on geodesics can easily be lifted by considering an entire family of swarms in dependence on the geodesic unit vector fields
—that is, by considering a perfect fluid in which particles of common rest mass m move along geodesics so that at each point q the mass density of particles moving with velocity
is equal to
(in the particles’ rest frames) for all spherical coordinate triples
. (Note: the suggested dependence of
on v alone—rather than on v,
, and
—is due to the fact that in a perfect fluid the distribution of particle velocities may be assumed to be isotropic relative to the fluid’s rest frame
.) Given this setup, it follows that the mass density corresponding to
relative to the fluid’s rest frame
is
(11)
where the first factor
is due to relativistic length contraction (in the direction of motion) and the second accounts for the relativistic increase in mass. Furthermore, the stress energy tensor
corresponding to
is given by the equation
and the matrix representing
at q with respect to the frame
is
That said, we now are justified in asserting that Equation (8) plausibly suggests—by way of its proven equivalence to Equation (7)—that the spacetime environment of an ideal gas or perfect fluid is described by the equation
where the matrix representing
(as defined by the equation
) is
.
Since the definition of
readily implies (by way of elementary integration) that
whenever
, it follows that
is a diagonal matrix with diagonal elements
and
Setting
, we may infer that the matrix representing T at q with reference to the frame
is
and that, by implication,
Thus we have arrived at the well-known representation of the stress-energy tensor of a perfect fluid (with the metric sign convention (+−−−)) because, according to (11),
is the total mass density as measured in
, and
is easily seen to be and also well-known to be the fluid’s pressure.
Finally, to round up our discussion of the relativistic matter field equation, we wish to point out that the equation
always allows us to compute all the components of the stress-energy tensor T by properly choosing
in the term on the right-hand side of this equation, but recovering the components of T from a single equation of the form
is not always possible. There are special cases in which it is possible because the components of T are all equal to a constant factor multiplied by the energy density
, but in general, of course, the components of T may contain a wide variety of quantities other than
. Two prominent special cases in which
completely characterizes T are encountered when the curvature-generating gravitational energy is produced by an electric field
or by a plane electromagnetic wave. For in the former case, the electromagnetic field tensor
is represented by the matrix
(relative to the basis
), and the electromagnetic stress-energy tensor
(12)
is easily seen to be represented by the matrix
(13)
and in the latter case, of a plane wave, the electromagnetic field tensor is represented by the matrix
(relative to the basis
because for a plane electromagnetic wave it is the case that
and
) and the matrix representing the corresponding stress-energy tensor turns out to be
(14)
(because
). So in either case, the components of the matrix representing T are either 0 or
, and theorems analogous to Theorem 5.1 can therefore be formulated.
In order to see this more clearly, we may want to take another look at Theorem 5.1: the central step in proving this theorem was to demonstrate that the equation
in conjunction with the assumed validity of the equation
for all geodesic vector fields
, as specified in (a), (b), and (c), completely characterizes the Ricci tensor, that is, it implies that for all tangent vectors
it is the case that
In essence, therefore, the statement of Theorem 5.1 is a somewhat stronger version of the assertion that the equation
must be satisfied for all
if the equation
(15)
is valid for all vectors
for which
(16)
(Note: Equation (16) is easily seen to be satisfied for all vectors specified in (a), (b), and (c), and therefore, the assertion above is slightly weaker than the statement of Theorem 5.1 because the vectors specified in (a), (b), and (c) form a strict subset of the vectors that satisfy (16)).
That said, we will now proceed to consider the electromagnetic stress-energy tensors given in (13) and (14): for the former of these we readily find that
and for the latter, we find that
Consequently, since (12) implies that the trace of T is equal to zero and that therefore R is equal to zero as well, the assertion above, concerning the sufficiency of Equations (15) and (16) for the generation of the Ricci tensor, leads us to assert, by analogy, that
for all
if the equation
is valid for all vectors
for which either
(in the case where T is given as in (13)) or
(in the case where T is given as in (13)). Not surprisingly, both of these respective assertions can be readily established by using essentially the same methods as in the proof of Theorem 5.1. But since the pertinent details would add little value to the results derived in this paper, we will not attempt to include them.
6. Conclusion
In reinterpreting the classical and relativistic gravitational field equations as velocity-field divergence equations, we were able to show that classical and relativistic descriptions of gravitation may be considered to be formally strictly analogous. This observation in itself does not elucidate many of the deeper structural questions that limit-like transitions from general relativity to Newtonian gravity are commonly thought to raise, but it does appear to bring to light a surprisingly simple and straightforward mathematical kinship—and hence it does appear to be worth mentioning.
Acknowledgements
I would like to thank my former colleague, the late Gaston Griggs, for many stimulating discussions concerning the nature of time and the structure of modern physical theories.