5.3 Mach’s Principle
5.3.1 Overview
reasonable to expect the following, at the center of this idealized universe :
• Very soon after the big-bang expansion begins (when the cosmological horizon is tiny compared to the radius of the universe’s outer edge, so an observer at the center can see only a tiny fraction of the distance to the edge), the universe’s matter will almost completely control the inertial axes at its center; i.e., the angular velocityωcof the central inertial axes (relative to radial infinity) will be almost identical to the angular velocityΩcof the central matter’s rotation: (Ωc−ωc)/Ωc1.
• As the universe expands, the inertial grip of its rotating matter will weaken and the inertial influence of the external empty space (radial “infinity”) will correspondingly grow, causing the fractional slippage of the inertial axes from the universe’s grasp, (Ωc−ωc)/Ωc, to increase, approaching unity as the universe becomes very large and its cosmological horizon becomes enormously larger.
Indeed, this is what our calculation in Sec. 5.3.3 reveals, though there is a complicated subtlety: in a dynamically expanding spacetime such as this one, we must find a preferred way to “map” angular positions at radial infinity (in the asymptotically flat, vacuum region outside the universe) onto angular positions ¯ϕat the center, so that the angular velocityΩc=dϕ/dτ¯ of the central matter with respect to radial infinity (with τ=proper time as measured by the matter) will be well defined, and so will beωc.4We achieve this preferred mapping via a preferred family of spacelike geodesics introduced by Schmid [18], which connect the center of the universe to radial infinity (see Appendix 5.A for details). With this choice of mapping, our results are gauge-invariant and take a very simple and elegant form [Eq. (5.41)]. Figure 5.1 shows the corresponding fractional slippage of the inertial axes, (Ωc−ωc)/Ωc.
Notice that the slippage at the universe’s center begins immediately after the big bang and long before the center has had any causal contact with the universe’s edge—i.e., before the cosmological horizon has been reached. How is this possible? At such early times there is no way, through electromagnetic or any other light-speed observations, for a central observer to discover that the universe has an edge. The observer sees only a homogeneous universe with small, homogeneous vorticity. But frame dragging appearsto tell the observer that the universe is finite.
This appearance is an illusion. Since the central observer cannot see out into the external universe, there is no way for that observer to learn what the standard of rotational inertia is out there and correspondingly, no way to learn how the angular coordinate ¯ϕat the center should be chosen so as to tie it to inertia at infinity. As a result, there is no way for the observer to know whatΩ =(dϕ/dτ)¯ matteris (nor what is the frame-dragging angular velocityω). The only thing the central observer can measure is the matter’s vorticity, i.e. (twice) the angular velocity of the matter relative to local gyroscopes. [In our spatially homogeneous model, this turns out, not surprisingly, to be 2(Ω−ω).]
Even though the central observer cannot measureΩor ωindividually at early times, we as physicists probing this model mathematically can calculate them; and they show a remarkably smooth slippage as a
4The differenceΩc−ωcis uniquely defined independent of any such mapping. It is half the vorticity of the matter, i.e. it is the spatially constant angular velocity of the matter relative to local inertial axes.
0.2 0.4 0.6 0.8 1 H W
c-Ω
cL W
c0.5 1 1.5 2 2.5
R H Τ L M
cosmological horizon white - hole horizon
Figure 5.1: The fractional slippage of the inertial axes at the center of a model universe with a spherical edge, (Ωc−ωc)/Ωcat timeτ, plotted against the radiusR(τ) of the expanding universe’s outer edge (which depends on time τ) in unit of its mass M [see Eq. (5.41)]. Hereτ is proper time measured in the rest frame of the universe’s slowly rotating matter,Ωcis the angular velocity (dφ/dτ)cof the matter at the center relative to inertial frames at infinity,ωcis the corresponding angular velocity of inertial axes at the center, and (it turns out) their differenceΩc−ωcis half the (locally measurable) vorticity of the matter, i.e. it is the (spatially homogeneous) locally mea- sured angular velocity of the matter relative to gyroscopes. This figure assumes the mapping of angles, between the center of the universe and infinity, that is embodied in the matching of the 3- metric (5.34)–(5.35) and the extrinsic curvature (5.37)–(5.38). The radius marked “cosmological horizon” corresponds to the moment when an observer at the center of the universe can first see out to the universe’s edge and discover thereisan edge [see Eq. (5.42) and surrounding discussion].
The radius marked “white-hole horizon” corresponds to the moment when the universe’s surface passes through its Schwarzschild radius,R=2M.
function of time: no lock-step rotation [(Ωc−ωc)/Ωc=0] before causal contact with the universe’s edge is achieved; no sudden change when causal contact is reached. This seems rather non-Machian.
This non-Machian behavior is intimately tied to a truly non-causal aspect of frame dragging. In any spherically symmetric situation, such as this one, the frame-dragging angular velocityω(viewed as a vector when its direction is taken into account) has a dipolar angular dependence; and in general relativity, dipolar fields are not radiative, i.e. they do not obey wave equations. This remarkable fact is tied to the spin-two nature of the graviton: only fields with quadrupolar and higher-order angular forms (i.e. with angular order greater than or equal to the graviton spins=2) are governed by wave equations.
The frame-dragging angular velocity ω, like the non-propagating, spherical coulomb field of a charge distribution in electromagnetic theory, is laid down at some initial time and evolves forward thereafter in a nonradiative manner. For a detailed general relativistic analysis and discussion of how the matter’s angular momentum distribution governs the evolution ofω, see Refs. [16, 18] and references therein.
The most serious way in which our model problem differs from the physical universe is in its neglect of
“dark energy”, which constitutes today about 65% of the universe’s energy density. In Sec. 5.3.4, we rectify this neglect by inserting dark energy into our universe-with-edge model. We assume (as is somewhat likely) that the dark energy takes the mathematical form of a cosmological constant in Einstein’s equations.
By contrast with the matter, we cannot cut offthis dark energy at an outer edge for the universe. It extends out of the universe into the vacuum exterior, undiminished, i.e. with a continuing-constant cosmological term in Einstein’s equations. This radically alters the geometry of spacetime outside the universe: the geometry is no longer asymptotically flat and so no longer has the asymptotic inertial properties and influence familiar from special relativity. Instead, the geometry is asymptotically that of the DeSitter solution to Einstein’s equations, and it has an asymptotic “DeSitter horizon”. Despite this change of spacetime geometry, we find that the fractional slippage of the inertial axes, (Ωc−ωc)/Ωc, remains qualitatively the same as in the absence of dark energy [Eq. (5.57) and Fig. 5.2 of Sec. 5.3.4 below].