arXiv:gr-qc/9607009v1 4 Jul 1996 The Lense–Thirring Effect and Mach’s Principle Hermann Bondi†and Joseph Samuel Raman Research Institute, Bangalore 560 080, India Abstract We respond to a recent paper by Rindler on the “Anti–Machian” nature of the Lense–Thirring effect.We remark that his conclusion depends crucially on the particular formulation of Mach’s principle used. †Permanent address: Churchill College, Cambridge, CB3 0DS, England 1
1Introduction In a recent paper, Rindler [1] has analysed the Lense–Thirring effect [2, 3, 4] and concluded that the result is anti-Machian.Rindler uses a particular interpretation of Mach’s principle.We wish to stress here that Rindler’s interpretation is only one amongst many. Indeed, the literature on this topic is so diffuse that we think it desirable to set out a list of interpretations that come to mind. Our list is far from exhaustive, but it is long enough to make numbering different versions necessary. We begin with Mach0, which is the basis of the whole idea:The universe, as represented by the average motion of distant galaxies [5] does not appear to rotate relative to local inertial frames. We illustrate this point by a modern version of Newton’s famous bucket experiment: the Sagnac effect.This effect provides an operational method for an observer to decide, by local measurements, if she is rotating. Consider an astronaut in an enclosed spaceship with angular velocityω. The astronaut takes a closed circular fibre optic tube at rest with respect to the spaceship and sends two rays of monochromatic laser light in opposite directions around the tube. These rays are made to interfere [6] after each ray has gone round once. If the spaceship is rotating, the corotating ray will take longer to come around than the counter–rotating one, leading to an arrival time difference, which can be observed as a fringe shift.The time difference is given by: ∆t=−4Aω/c2, whereωis the angular velocity of the spaceship andAis the area enclosed by the tube. Using the Sagnac effect, one can by experiments internal to the spaceship, so arrange the angular velocity of the spaceship that the Sagnac shift (defined as ∆t/2) vanishes. A frame at rest with respect to such a spaceship is called a locally non rotating frame. Sit in this frame, look 2
up at the sky and note that the distant galaxies are still. Mach’s principle (Mach0) is the experimental observation that the inertial frame defined by local physics (zero Sagnac shift) coincides with the frame in which the distant objects are at rest. Mach0 is an experimental observation and not a principle.One could interpret Mach’s writings as a suggestion to construct a theory in which Mach0 appears as a natural consequence.But Mach’s writings have been variously interpreted. Our purpose here is to list a number of interpretations of Mach’s principle and view them in the light of currently accepted theories in an effort to refine and clarify the idea. We do have at our disposal two well established theories of space, time, gravity, matter and motion –Newton’s and Einstein’s– both experimentally succesfull in their respective domains of validity. Newton’s holds that space and time are absolute. Einstein’s holds that space time geometry is affected by matter. There is no question (as Rindler observes) that these experimen- tally successful theories are here to stay regardless of whether they satisfy any of the rather philosophical criteria embodied in Mach’s principle. 2Versions of Mach’s Principle Recent discussions of Mach’s Principle, including this one, have greatly ben- efitted from the 1993 Conference organised by J. Barbour and H. Pfister and the excellent book [7] resulting from it. A glance at the book (note especially J. Barbour’s list on page 530) will show that there have been numerous in- terpretations of Mach’s writings. For an authoritative account of the history of Machian ideas, the reader is referred to [7]. We now list a few versions of Mach’s principle which appear in the lit- 3
erature. Each statement of Mach’s principle, will be accompanied by a dec- laration of the theoretical framework in which it is intended to apply. Two levels of compatibility will be considered: Does the particular statement of Mach’s Principle make sense in the theory, and secondly, is it satisfied by it? We use the letters N and E to refer to Newtonian and Einsteinian space time. Even within Einstein’s theory there is a further dichotomy– is one dis- cussing cosmology (the whole universe) or an isolated system embedded in an asymptotically flat space time? This distinction is made by the notation EA for asyptotically flat spacetimes and EC for relativistic Cosmologies. Our purpose in compiling this list is to draw attention to the diversity of ideas that pass under the guise of “Mach’s principle”. (Page numbers refer to [7] unless otherwise indicated.) •Mach1:Newton’s gravitational constantGis a dynamical field.(Makes sense in N, EA, EC.) Mach1 is not true in N or E. This version applied to Einstein’s theory has led to Brans–Dicke Theory[8, 9]. •Mach2:An isolated body in otherwise empty space has no inertia (pp 11,39,181, 185).(Makes sense in N, EA, EC.) Neither Newtonian nor Einsteinian gravity satisfy this version. In both theories the motion of an isolated body is determined and not arbitrary. •Mach3:local inertial frames are affected by the cosmic motion and distribution of matter (p92).(Makes sense in N, EA, EC [10] .) This version is closest to the bucket experiment.In this form, Newton’s theory is in clear conflict with Mach3.Einstein’s theory is not (see section 4 below). •Mach4:The universe is spatially Closed (p 79).(Makes sense only in 4
EC.) We do not know if Mach4 is true. •Mach5:the total energy, angular and linear momentum of the universe are zero (p237).(Makes sense in N, EA, EC.) It is not true in N and EA. In EC it is claimed [11] that the total angular momentum of a closed universe must vanish. •Mach6:Inertial mass is affected by the global distribution of matter (pp 91,249).Makes sense in (N, EA, EC). Is not true in any of them. Hoyle and Narlikar [12] proposed a theory in which implements Mach6. •Mach7:If you take away all matter, there is no more space [13].Makes sense in (N,EA,EC). Not true in any of them. •Mach8: Ω = 4πρGT2is a definite number of order unity (p475).(Here, ρis the mean density matter in the universe andTis the Hubble time. Makes sense in EC only.) Ω does seem to be of order unity in our present universe, but note that of all EC models, only the Einstein–DeSitter makes this number a constant, if Ω is notexactlyone. Making a theory in which this approximate equality appears natural is a worthwhile and ongoing effort (eg inflationary cosmologies). •Mach9:The theory contains no absolute elements ([14].(Makes sense in N, EA and EC) This version is clearly explained by J¨urgen Ehlers in [7] p 458. The elements (fields, for example) appearing in the theory can be divided into dynamical (those that are varied in an Action prin- ciple) and absolute (those that are not). The Action principle leads to equations for the dynamical fields to satisfy. The absolute elements are predetermined and unaffected by the dynamics. 5
Newton’s theory does not satisfy Mach9 (space and time are abso- lute) and neither does EA (asymptotic flatness introduces an absolute element–the flat metric at infinity). EC does satisfy Mach9 [15]. From the point of view of invariance groups (J.L Anderson, A. Trautmann, quoted on p 468 [7]) Mach9 is the requirement that the invariance group of the theory is theentirediffeomorphism group of spacetime. Viewed in this light Mach9 is just the principle of general covariance. •Mach10:Overall rigid rotations and translations of a system are unob- servable.(This version makes senseonlyin N; In Einsteinian spacetime one has no idea what a rigid rotation is anymore than one knows what a rigid body is.) This isnotsatisfied in Newtonian theory. If one insists on the principle and constructs a theory which satisfies it, one is led [16] to a class of models (called “relational” by Barbour and Bertotti [16]). There is considerable literature on these models [7, 17].We spend a few words on these models and their connection with Newonian theory. Relational Models:Letxi a,i= 1,2,3,a= 1...Nbe the positions of N particles in Newtonian spacetime andpiatheir conjugate momenta. The HamiltonianH(x, p) determines the time evolution of (xi a, pia) via Hamilton’s equations. The transformation xi a(t)→Ri j(t)xj a(t) pia(t)→Rj i(t)pja(t),(1) whereRi j(t) is an arbitrary time dependent rotation matrix maintains the distance relations between theNparticles. If a model is relational [16], such a transformation is unobservable, like a “gauge transforma- tion” in electrodynamics. From Dirac’s theory of constrained systems 6
[18, 19], it follows that the transformations (1) must be generated by first class constraints. The generator of overall rotations of the system is the total angular momentum: Ji= Σaǫijkxajpak. Thus the system is subject to the constraints φi(x, p) :=Ji−Ci≈0, whereCiare constants. The requirement that the constraints be first class in the sense of Dirac [18] forces the constantsCito vanish. The extended Hamiltonian in the sense of Dirac is HE(x, p) =H(x, p) +ωiJi, whereωiare arbitrary functions. While we have only dealt with overall rotations in (1), one can similarly deal with arbitrary translations and arbitrary time reparametrizations. Relational models can be thus de- rived from Newtonian Hamiltonian mechanics by imposing constraints on the phase space so that the total angular momentum, momentum and Energy vanish. These relational models are clearly distinct from Newtonian theory. For instance, Newtonian theory admits solutions with nonzero angular momentum (like the solar system in an otherwise empty universe) while relational models do not permit such solutions. 3Rindler’s Criticism We now briefly summarise Rindler’s argument. Consider the earth in an otherwise empty universe. LetObe a reference frame rigidly attached 7
to the earth. Suppose that a gyroscope G is taken around the earth in the equatorial plane along a circle of radiusrwith a constant clockwise angular velocity Ω. To keep track of orientations, we suppose the earth and the gyroscope marked with cross hairs (as in Fig.1 of Rindler). We arrange that the orientation of G relative to the earth’s is constant during the motion. (Rindler uses the Schwarzschild metric outside the earth to computeαthe precession rate of the gyroscope.We choose the radiusrto setαto zero. It simplifies the argument.) Now view the situation from the point of view of an observerO′, who rotates rigidly relative toOwith constant clockwise angular velocity Ω.O′sees the earth rotating anticlockwise with angular velocity Ω, the centre of the gyroscope at rest. Notice however, that the gyroscope (which was not rotating with respect toO) now rotates anticlockwise with angular velocity Ω relative toO′. Thus the gyroscope rotatesin the same senseas the earth. It follows from Mach10 that a rotating body in otherwise empty space makes the local compass of inertia take upallof the body’s angular velocity. Applied to the earth, which is not in empty space but in the universe, one would expect that the effect of the earth on the gyroscope should be considerably diluted by the effect of the rest of the universe. Thus one would expect that the local compass of inertia would take up a smallpositivefraction of the earth’s angular velocity. The sign of this effect is everywhere positive unlike the sign of the Lense–Thirring effect. This is the basis for Rindler’s conclusion that the Lense–Thirring effect is Anti–Machian. 8
4The Lense-Thirring effect as Machian We now show that one can arrive at the opposite conclusion from Rindler’s by using a different version of Mach’s Principle. We use the often employed exact analogy between rotation in General Relativity and magnetic fields [20] to deduce that the slight influence of a spinning body on the rotation of the near-by compass of inertia goes with that of the body near the poles and in the opposite sense in the equatorial plane. The Lense–Thirring effect:Consider a stationary spacetimei.eone with a timelike Killing vectorξ:∇aξb+∇bξa= 0. One can adapt the time coordinate toξso thatξ=∂/∂tand the metric assumes the form: ds2=g00(dt+Aidxi)2−γijdxidxj, whereAi=g0i/g00. The coordinate transformations that preserve this form includet→t+α(xi), which physically represents the resetting of clocks. Under such transformationsAitransforms asAi→Ai+∇iαlike the vector potential in electrodynamics.Consequently its curlFij:=∂iAj−∂jAiis invariant and represents rotation of the spacetime (more geometrically, the failure ofξto be hypersurface orthogonal). It is easily seen that a stationary Sagnac tube will measure a Sagnac shift of∮Aidxi.A locally nonrotating Sagnac tube (one that measures zero Sagnac shift) will appear to rotate as viewed from infinity.The angular velocity of rotation has the spatial distribution of a dipole magnetic field andreverses sign between the equator and the poles. As we show below this isexactlywhat one expects from Mach’s principle (Mach3). If one applies Mach’s Principle in the form Mach3 to understanding ro- tation in General Relativity, one sees that the prediction of Mach3 agrees 9
100%
Scan to connect with one of our mobile apps
Coinbase Wallet app
Connect with your self-custody wallet
Coinbase app
Connect with your Coinbase account
Open Coinbase Wallet app
Tap Scan
Or try the Coinbase Wallet browser extension
Connect with dapps with just one click on your desktop browser
Add an additional layer of security by using a supported Ledger hardware wallet