General Relativity The theory of special relativity, developed by Einstein in 1905, gave rise to a truly major revolution in our notions of space and time. The effects of this revolution on the known laws of physi were soon felt, as theorists began to reformulate them within a framework compatible with special relativity. The attempt to reformulate the theory of gravitation produced a second major revolution in our notions of space and time: Einstein's theory of general relativity, developed in 1915. 3.1 Special Relativity and Physical Theories As I previously remarked, the notion of absolute simultaneity is so deeply ingrained in the way most people think about spacetime that it takes a great deal of effort to be consciously aware of when and how one is using this assumption. Hence, it is not surprising that the known laws of physics prior to special relativity implicitly assumed absolute simultaneity in their formulation. For example, Newton’s law for the gravi- tational force between two bodies states that Here F is the force, G is the gravitational constant, , and are the masses of the two bodies, and r is the distance be- 2930 Chapter 3 tween them. More precisely, F is supposed to be the force between the bodies at a given time and r is supposed to be the distance between the bodies at this same time. Within the framework of the prerelativity notions of space and time, this is a perfectly meaningful statement. But in special relativity, the phrase “at the same time” has no absolute meaning. New- ton’s theory of gravity does not make sense because it is phrased in terms of spacetime structure which does not exist! Clearly Newton’s theory of gravity, as well as most theories of physics prior to special relativity, can at best be only approximations—valid when all relevant velocities are small compared with c—to the true exact laws of physics. The aim now is to find these true laws of physics. What are the properties of the new laws of physics which we are seeking? First, they must be formulated within the framework of the spacetime structure of special relativity. If notions such as relative simultaneity are used in the formu- lation, it must be demonstrated that the theory is actually independent of which inertial observer is chosen to define simultaneity. (Otherwise, the principle of relativity would be violated.) This is usually referred to as “relativistic invari- ance” or “covariance” of the theory. Second, the predictions of the theory must be consistent with special relativity. In particular, the theory must not predict that one can send a body or signal between two events which are spacelike sep- arated. The theory of electricity and magnetism given by Maxwell over a century ago satisfies these requirements. Indeed, his- torically the analysis of Maxwell's theory and its predictions concerning the propagation of electromagnetic waves (that is, light) led to the discovery of special relativity. What is the situation for the theory of gravitation? The close analogy between Newton’s theory of gravitation and the theory of electrostatics is well known. Coulomb’s law for the force between two charges differs from Newton’sGenerel Relativity aa force law only by a change in sign. Hence, one might hope that one could formulate a special relativistic theory of grav- itation closely analogous to Maxwell’s theory of electromag- netism. In the limit where the motion of the masses is slow compared with the speed of light, this theory should reduce to Newton’s theory, just as in the slow-motion limit Maxwell’s theory reduces to Coulomb's electrostatics. However, this attempt fails. It turns out that the sign difference between gravity and electromagnetism also enters the formula for the energy of radiation. Gravitational radiation in a theory of gravity analogous to Maxwell’s theory would carry away neg- ative energy. This is physically unacceptable, since it means that a system of masses that radiates would gain energy, which would cause it to radiate more and gain more energy ad infinitum. One might not become terribly discouraged by this and try to formulate a theory of gravitation in the framework of spe- cial relativity which does not suffer from the above difficulty. In fact, it is possible to do so, though other difficulties (or contradictions with observations) then arise. However, there are a number of reasons for believing that perhaps one should not simply try to fit gravity into the framework of special relativity. These reasons motivated Einstein to abandon some of the assumptions of special relativity and formulate a new theory of space, time, and gravity: general relativity. 3.2 Space, Time, and Gravity As I remarked above, in Newtonian theory gravity is treated as a field which is closely analogous in many of its properties to an electric field. However, there are two basic ideas which suggest that perhaps one should think of gravitation in a different manner from other fields. First, gravity is universal; all bodies are affected by gravity. (By contrast, electric fields, for example, do not affect neutral bodies.) Furthermore, all objects fall the same way in a grav-32 Chapter 3 itational field. This fact, known as the equivalence principle, is expressed in Newtonian theory by the statement that the gravitational force on a body is proportional to its mass, so that the acceleration of the body is independent of its mass. The equivalence principle was first demonstrated by Galileo’s famous (though possibly apocryphal) experiment of dropping two bodies of different weight from the Leaning Tower of Pisa. Much more recently it has been verified to very high precision by the experiments of Robert Dicke and others. What is the significance of this universality of gravity? It suggests the possibility of ascribing the properties of the grav- itational field to spacetime itself! A second rather vague set of ideas deals with inertia and goes under the name of Mach’s principle, though many of the notions predate Mach’s nineteenth-century writings. The ba- sic idea is as follows: Consider a rotating body. One can measure its rotation in two independent ways: (1) in an ab- solute, local manner by measuring the stresses in the body required to keep the outer parts of it from flying off, or (2) in relation to distant matter (the “fixed stars”). Why is it that these two notions of rotation agree? Mach’s idea was that perhaps the distant matter in the universe determines local inertial effects such as rotation. Thus, if someone were some- how to accelerate the distant matter in the universe, our local determinations of nonaccelerating and nonrotating should be affected. If there were no matter in the rest of the universe, there should be no such thing as inertia or rotation. Einstein accepted the basic idea of Mach’s principle. Hlow- ever, Mach’s ideas do not have any expression in special relativity where the structure of spacetime is not allected by the matter which is present. Thus, Einstein was motivated to seek a new theory in which the effects of gravitation would be expressed in terms of the structure of spacetime (equivalence principle) and the structure of spacetime would be affected by the matter present in spacetime (Mach’s principle). HisGeneral Relativity 33 success in this regard was probably the greatest single achieve- ment that has yet occurred in theoretical physics. Interestingly, although some of Mach’s ideas are reflected in Einstein’s theory of general relativity (see 8.4), it is certainly not true that general relativity fully incorporates all of them. In general relativity, it can be said that the motion of distant matter affects some local inertial properties but it does not fully determine them. In particular, rotation can always be locally determined even if there is no distant matter. 3.3 General Relativity I now shall describe the theory of general relativity, Einstein’s theory of space, time, and gravitation. Recall, first, two basic results of our discussion of special relativity: (1) The intrinsic properties of spacetime are fully described by the spacetime metric (see 2.5). (2) Inertial observers are precisely those ob- servers who move on timelike geodesics of the spacetime metric (see 2.7). Recall also two of the key assumptions which went into the formulation of special relativity (see 2.1): (ii) The spatial relationships between (relatively) simultaneous events are described by Euclidean geometry. (‘v ) Inertial ob- servers all move with uniform velocity with respect to each other; furthermore, any observer who moves with uniform velocity with respect to an inertial observer is himself an inertial observer. Here is Einstein’s brilliant idea: Identify freely falling bod- ies in a gravitational field with the inertial observers of special relativity. This clearly necessitates dropping assumption (iv ) since freely falling bodies in a nonuniform gravitational field do not maintain a uniform velocity with respect to each other; drop assumption (#i) as well. However, retain the two key conclusions, 1 and 2, stated above. In other words, we continue to assume that all the observer- independent information concerning the spacetime relation- ship between events is contained in the metric of spacetime.34 Chapter 3 However, we no longer assume that the spacetime metric is given by the simple expression dst = — di + 3 de? + dy? + de?) as it is in special relativity (see 2.4). Such a spacetime metric corresponds to a flat geometry in that initially parallel geode- sics remain parallel (see 2.7). In general relativity we wish to retain the identification of the world lines of inertial (that is, freely falling) observers with geodesics of the spacetime met- ric. However, in a (nonuniform) gravitational field, two freely falling observers who are initially at rest do not remain at rest with respect to each other. Thus, in general relativity the presence of a gravitational field is reflected in the fact that initially parallel geodesics of the spacetime metric do not remain parallel. In other words, spacetime is curved. Note the dual role of the spacetime metric in general relativ- ity: (a) Just as in special relativity, it carries all the information concerning the spacetime relationship of events; (b) It fully de- scribes the gravitational field by specifying, via its timelike geo- desics, the motion of all freely falling observers. Thus, the de- scription of gravitation is inextricably tied into the properties of space and time in that both are described by a single quantity, the spacetime metric. The presence of a gravitational field cor- responds to curvature of the spacetime geometry. 3.4 Einstein’s Equation The theoretical framework of general relativity was outlined in the previous section. The gravitational field is to be de- scribed in terms of the curved geometry of spacetime. To complete the specification of the theory, it remains to specify what spacetime geometry (that is, gravitational field) is asso- ciated with a given configuration of matter. Einstein did this by postulating an equation which, in essence, says, “curvature of spacetime” = “energy density of matter.”General Relativity 35 Thus, Einstein provided us with a truly remarkable and beautiful theory of gravitation where, in accordance with the ideas of 3.2, the effects of gravitation are fully expressed in terms of the structure of spacetime and, in accordance with some of Mach’s ideas, the structure of spacetime is related to the distribution of matter via Einstein’s equation. This description of Einstein’s equation is somewhat of an oversimplification, and some words of caution should be given. First, the left side of the equation is not the entire curvature of spacetime but only a part of it, Thus, outside the matter distribution (where the right side of Einstein's equa tion will be zero) spacetime will in general still be curved; that is, a gravitational field will be present. Furthermore, gravita- tional radiation—ripples in the curvature which propagate through spacetime—can exist (see 9.3). Second, the right side contains contributions from other properties of matter be- sides energy density. In particular, pressures and stresses con- tribute to the curvature of spacetime in general relativity. If no matter is present, the right side of Einstein’s equation vanishes and a perfectly valid solution is the flat spacetime geometry of special relativity. (However, there also are many other solutions describing possible exterior gravitational fields of bodies, gravitational waves, and so on.) In this sense, general relativity includes special relativity as a special case. In our description of general relativity it was postulated that freely falling bodies follow geodesics of the spacetime metric. It turns out, however, that—as discovered more than ten years after the formulation of general relativity— Einstein's equation itself actually determines the motion of matter in spacetime. The “geodesic hypothesis” actually fol- lows as a consequence of Einstein's equation and does not have to be postulated separately. Finally, it should be mentioned that in practice it has proven very difficult to obtain exact solutions of Einstein's equation. Fortunately, many solutions of great physical inter-36 Chapter 3 est (for example, solutions describing black holes) are known and the properties of these spacetimes have been investigated in detail. But the study of what possible structures of space- time are permitted in general relativity continues to be ham- pered by our general inability to solve Einstein's equation. 3.5 Consequences of Einstein’s Theory Einstein’s theory of general relativity is unquestionably one of the most beautiful theories ever devised, but that, of course, does not mean that nature conforms to it. In this section we shall briefly describe four key predictions of the theory which are subject to experimental and observational check. All these predictions have been verified. One would not yet be justified in saying that the validity of general relativity has been proven beyond any doubt, since the theory has not yet been tested for very strong gravitational fields and the accuracy of the exper- iments and observations which have been done is not great enough to rule out small modifications to Einstein’s equation. However, the agreement between theory and observation in these four cases described below greatly increases one’s con- fidence that general relativity is a true theory of nature. The Motion of the Planets Tf the energy density of matter is not too large, it is possible to obtain approximate solutions of Einstein’s equation for the spacetime geometry. One can then calculate the geodesics of the spacetime metric to obtain the motion of freely falling bodies. In the limit where the relative velocities of the bodies of matter are much smaller than ¢, one finds that the predic- tions of general relativity reduce to the predictions of New- ton’s theory of gravitation. Thus, in particular, in this approx- imation general relativity predicts that the planets should move around the sun in elliptical orbits which obey Kepler's laws.Generel Relativity 37 However, when corrections to the above approximations are put back in, one finds that the predictions of general relativity deviate from those of Newtonian theory. In partic- ular, the elliptical orbits of planets don’t quite close but pre- cess instead. This precession is much too small to observe for all planets except Mercury, where general relativity predicts an orbital precession of 43 seconds of arc per century. (A second of arc is ¥%,00 of a degree.) This precession of Mer- cury’s orbit has been observed. Indeed, it had been observed long before the development of general relativity and had been an unexplained mystery until then. Light “Bending” As discussed above, in general relativity, just as in special relativity, a freely falling (that is, nonaccelerating) material body moves on a timelike geodesic of the spacetime metric. Similarly, in general relativity it remains true that the path of a light ray in spacetime is a null (lightlike) geodesic (see 2.7). Because spacetime is not flat in the vicinity of the sun, a light ray passing near the sun will appear to a distant observer to be deflected (see fig. 15). One can think of this physically as the light being bent by the gravitational attraction of the sun, although, in fact, light is actually traveling on the “straightest possible path” in the curved spacetime geometry. All other light ray stor (nut! geodesic) image . curved spacetime stor curvature negligible > observer Figure 15. A space diagram (not a spacetime diagram) illustrating the “light- bending” effect. Because of the spacetime curvature in the vicinity of the sun, the direction of motion of the light is changed. The deflection angle is greaty exaggerated in this diagram. The actual deflection angle of alight ray passing by the surface of the sun is only 1.75 seconds of arc.38 Chepter 3 forms of electromagnetic radiation, such as x-rays and radio waves, will undergo the same deflection as light. General relativity predicts a deflection angle of 1.75 sec- onds of arc for light rays which just pass by the surface of the sun. It is difficult to make an accurate observation of this effect. One must wait for a total eclipse of the sun in order to observe the deflected light, and a number of observational problems arise. However, the light-bending effect has been observed and found to agree with the predictions of general relativity to within the uncertainty of the observation (about 10% of the predicted value). The light-bending effect also has been investigated for radio waves coming from a quasar as it passes behind the sun. This approach has the advantage of not requiring that the sun be eclipsed by the moon. Recent observations made using the technique of very long baseline interferometry have verified the light-bending effect pre- dicted by general relativity to .1% accuracy. The Gravitational Redshift Suppose an observer O, sends out two signals with a time interval At, between them. Let observer O, receive these signals. In curved spacetime—or even in flat spacetime if there is relative motion between O, and O,—there is no reason why the time interval Ar, between O,’s reception of the signals must equal A¢,. Thus, in particular, if O, emits light at frequency 2, O, will, in general, observe it to have frequency v, ¥ v;. If O, and O, are “at rest” and v, ¥ v4, this effect is known as the gravitational redshift. One can show that light emitted in a region of strong gravitational attraction will be seen by a distant observer to have a lower frequency. Actually, it is not difficult to see that the gravitational redshift must occur if energy is conserved. Quantum theory tells us that the energy, E, of a photon (that is, light) is proportional to its frequency, E = hv, where h is Planck’sGenere! Relativity 39 constant. If the frequency of light remained unchanged as it left a region of strong gravitational attraction, its energy would not be decreased. One could then convert the energy of the photons in the light beam to rest mass energy and lower the mass back into the strong-field region using the gravitational attraction to do work. If one then converts the rest mass energy back into photons, one will return to the same configuration as one began with, but energy will have been gained in the process of lowering the mass. Thus, the quantum formula E = hv, together with conservation of energy, requires the existence of the gravitational redshift phenomenon predicted by general relativity. The gravita tional redshift effect has been confirmed to a very high precision (01%) by the tracking of hydrogen masers launched by rockets. Time Delay In general relativity, as in special relativity, any observer who measures the velocity of a light ray as it passes by him always will obtain the result c. However, if a light ray emitted from a distant object propagates through a region of (relatively) high spacetime curvature, the curvature, of course, can affect the total elapsed time of propagation of the ray. General relativity predicts that light rays emitted from a planet or spacecraft as it passes behind the sun will suffer a small, additional “gravitational time delay” as compared with what would be predicted using Newtonian theory, due to the spacetime curvature near the sun. The best method available for observing the gravitational time delay effect is to bounce radar signals off a reflector which was placed on the surface of Mars by the Viking lander and to track accurately the arrival times of the reflected sig- nals as Mars passes (nearly) behind the sun. These measure- ments have confirmed this prediction of general relativity to a very high precision.