Foundations o f Physics, Vol. 25, No.

7, 1995

Cosmological Relativity: A Special Relativity for Cosmology

M. Carmeli ]
Received January 3, 1995 Under the assumption that Hubble's constant H o is constant in cosmic time, there is a l l analogy between the equation o f propagation o f light and that o f expansion o f the universe. Us#1g this analogy, and assuming that the laws o f physics are the same at all cosmic thnes, a new special relativity, a cosmological relativity, is developed. As a result, a transformation is obtabled that relates physical quantities at different cosmic times. In a one-dimensional motion, the new transformation is given by xt~ x--To (1 -- T2/T~) I/2-' I)~=
(1 --

v- xT/T o T2/T~) j/'-

where x and v are the coordinate and velocity, T is the cosmic time measured backward with respect to out" present time T = O, and To is Hubble's time. Some consequences o f this transformation are given, and its applicabili O, limitation is pohlted out.

1. I N T R O D U C T I O N In prerelativistic physics it was assumed that space is not related to time; a "stationary" frame of reference was presumed to exist with respect to which all physical phenomena can be described. A. Einstein, (]) in his famous paper of 1905, showed that this picture was wrong: Space has no preference for a particular frame or any other one that moves with a constant velocity, and in this way one can accommodate the fact that light propagates with a constant velocity in all moving
Department of Physics, Ben-Gurion University of the Negev, Beer Sheva, 84105 Israel. 1029 0015-9018/95/0700.1029507.50/0 9 1995 Plenum Publishing Corporation

1030

Carmeli

systems. The mixture of space and time became a necessity in order to preserve the constancy of the propagation of light in all inertial frames. The mathematical expression of this fact is given by the familiar Lorentz transformation, which was rederived by Einstein, who also gave to it the correct physical interpretation. We wish to point out that nowadays we have a similar situation in cosmology to that which existed in prerelativistic times with respect to space and (not velocity but) cosmic time, in conjunction with the constancy of expansion of the universe (and not propagation of light). If we make the convention according to which cosmic time, denoted by T, is measured backward, then our present time ( T = 0) is a preferred time to which all cosmological physical phenomena are referred. This is exactly analogous to the prerelativity assumption that physical phenomena referred to only one "stationary" ( v = 0) system. Actually space has no such a preference: When we consider an astronomical object and say that it is, let us say at T = T,,/2, where To = 1/Ho is Hubble's time, that faraway object has the same right to say that he is at cosmic time zero ( T = 0 ) and we are at T = To~2 with respect to him, exactly as in relativistic physics but with the roles of cosmic time and velocity exchanged. We will assume that such a reciprocity relationship between cosmological objects is a universal property of space and cosmic time, just as Einstein did with respect to space and velocity in special relativity.

2. COSMIC FRAMES In addition, we will make two assumptions which will be elevated to postulates. These are: (1) The pr#wiple of the constancy of the expansion of the universe (expressed by Hubble's law) at all cosmic times (analogous to the principle of the constancy of propagation of light in all moving frames); and (2) The principle of cosmological relativity (analogous to the principle of special relativity) according to which the laws of physics are the same at all cosmic times (at moving frames in special relativity). In this way the universe has cosmic frames of reference located at fixed cosmic times and differ from each other by relative constant cosmic times, similar to the situation in special relativity, but now cosmic times replace velocities. Observers in each cosmic frame are equipped with a ruler to measure distance (like in special relativity) and with a small radar device (similar to that used by highway patrol) for velocity measurements (instead of clocks in special relativity).

Cosmological Relativity

1031

Notice the analogy between the relation [To] =distance/velocity in the present theory and [c] =distance/time in special relativity, which suggests the choice of distance and velocity as our fundamental variables as compared with distance and time in special relativity.
Remark. The constant To is used by us just as the constant c is used in special relativity even though it is well known that both the speed of light and the rate of expansion of the universe change their values due to gravity. This is possible since local measurements of both the velocity of light and the rate of expansion of the universe always yield constant c and Ho, respectively. With the above postulates, and by comparison with special relativity, it is obvious that space and velocity cannot be independent if Hubble's law is to be preserved at all cosmic times. In fact, this will enable us to derive a transformation that relates space points and velocities (and other quantities) measured in different cosmic frames of reference that differ in relative cosmic times, just like the Lorentz transformation which relates space points and time (and other quantities) measured in different inertial frames that differ in relative velocities. Space coordinates and velocities become unified in cosmology just as space and time are unified in local (noncosmological) physics.

3. D E V E L O P I N G THE THEORY With the above preliminaries we are now in a position to develop our theory. To begin with, we repeat very briefly what preceded special relativity. The Galileo transformation between two inertial systems K and K', where K' moves relative to K with a constant velocity v along the x-axis, is given by x'=x-vt, t'=t (1)

when the other coordinates are unchanged. Here x and x' represent the coordinates of a particle in the systems K and K', respectively. The trouble with the transformation (I) is its incompatibility with the equation of propagation of light which satisfies
c2t '2 -- X '2 -~ c2t 2 -- X 2

(2)

Hence the transformation (1) should be abolished in favor of a new one that relates not only x' to x leaving t unchanged but relates x' and t' to x and t. And this immediately leads to the familiar Lorentz transformation.

1032

Carmeli

In cosmology one is not interested in comparing quantities at two reference frames moving with a constant velocity with respect to each other. Rather, one is interested in comparing quantities in two different cosmic times. For example, one often asks what was the density of matter or the temperature of the universe at an earlier time T as compared to the values of these quantities at our present time now ( T = 0). The backward time T is the relative time with respect to our present time. The concept of the relative time is not restricted only to the backward time T with respect to the present time ( T = 0 ) . Every two observers with times T 1 and T2 with respect to us are related to each other by a relative time T. Thus T plays the role of the velocity v in special relativity and we will see in the sequel that T has an upper limit which is the Hubble time T o = H o I just as the maximum velocity permitted in special relativity is c. The variables (coordinates) in this theory are naturally the Hubble variables, i.e., the velocity v and the distance x. To derive the transformation between these variables in the system K and K', where K' has a relative time T with respect to K, we proceed as follows. We first do this classically, and for simplicity it is assumed that the motion is one-dimensional. Denoting the coordinates and velocities in the systems K and K' by x, v and x', v', respectively, we have
x' = x - - T v, v'= v

(la)

where v was assumed to be constant. The x's and v's in Eq. (la) represent the coordinates and velocities not for just one particle but for as many as one wishes, with T the same for all of them. The transformation (la) does not satisfy the equation of expansion of the universe which, according to the principle of the constancy of expansion of the universe and the principle of cosmological relativity demanding the laws of physics (and in particular Hubble's law) to be valid at all cosmic times, satisfies

T2oV,2- - x

,2 = T o2Y 2

x2

(2a)

The situation here is similar to that we had at the beginning of the century where the Galileo transformation (1) could not accommodate both of the principle of special relativity and the principle of the constancy of the speed of light, whence leading to the Lorentz transformation. A new transformation here also has to be found which relates not only x' to x leaving v unchanged but relates x' a n d v' to x a n d v.

Cosmological Relativity
4. DERIVATION OF THE TRANSFORMATION

1033

For simplicity we again assume that the motion is along the x-axis. Hence Hubble's Law in the systems K and K' is given by
x = Tov, x ' = Tov'

(3)

where x, v and x', v' are measured in K and K'. Assuming now that x, v and x', v' transform linearly, we have
x' = ax by

(4)
(5)

x = a x ' + by'

where a and b are some variables which are independent of the coordinates. At x ' = 0 and x = 0, Eqs. (4) and (5) yield, respectively, ....
a b x /)

(6)

and . . . . .
a b x' I)'

(7)

Using now Eqs. (3), (4), and (5), we obtain

Tov=X=ax' +by' =aTov' +by' =(aTo+b)v'

(8a)

and similarly
Tov'=(aTo-b)v

(8b) from Eq. (6),

Eliminating v and v' from Eqs. (8), and using b = a T we get

T ] = a 2 ( T ] - T 2)
or

(9)

a = l / ( 1 -- T -~ T o~ i/, / ) and therefore

b = T/(1 -- T:/T2o) '/2

(10)

(ll)

825/25/7-5

1034

Carmeli

Inserting these results in Eqs. (4) and (5), we obtain

X t __

(1

x - Tv T2/T~) 1/2
-~ 9 Ip

(12a) (12b)

v'and

v--xT/T2~
(1 -- T-/To) -

x' + Tv'
x - (1 - T2/T~) u2 (13a)

v' + x'T/T2o
v - (1 - T2/T~)1/2

(13b)

for the inverse transformation. Equations (12) give the transformed values of x and v as measured in the system K' with a relative time T with respect to K. The roles of the time and the velocity are exchanged as compared to special relativity. This fits our needs in cosmology where one measures distances and velocities at different times in the past. TIT o replaces v/c of special relativity. If one considers the whole three-dimensional motion, one then has three spatial coordinates and three components of the velocity, and the transformation will be six-dimensional. It will have 15 parameters, three for the coordinate rotations, three for the velocity rotations, and nine "boosts" between the x's and the v's, as compared to the 6-parameter Lorentz group. Similarly to the Lorentz group, 0(3, 1), the new group, 0(3, 3), is not compact. Its covering group is not the familiar SL(2, C) group, the covering group of the Lorentz group, anymore but rather the group SL(4, R). As is seen, not only the physics of the present theory is different from Einstein's relativity, but also the mathematics as well.

5. A N O T H E R

DERIVATION

The transformation (12) could also have been derived as in deriving the Lorentz transformation in standard texts by writing
Xr2 ~2 --~o vt2 = x -9- - T o ~ -"~ y

(14)

whose solution is
x' = x c o s h ~ - vTo sinh ~/,

v' = v cosh ~b- (x/To) sinh ~b

(15)

Cosmological Relativity

1035

At x ' = 0 we obtain tanh ~b = and therefore

x/Tov = TIT o

(16)

sinh ~b - (1 cosh ~ / - ( 1

T/To
T2/TZo) vz

(17a) (17b)

1 ~2"~'~" 1/2 1 / 1 o)

which lead to the transformation (12). The geometrical description of the galaxy cone in the present theory and its comparable light cone in special relativity are given in Figs. 1 and 2.

ToV EXP/~NDING
TqC
i

UNIVERSE
i i ,<.

T > To T>T o

:
T < TO
i

CONTRACTING

-ToC ',
I

UNIVERSE

Fig. 1. Galaxy cone,

~ x 2 - T O o ~' =

0, in cosmology.

1036

Carmeli

v,c

Fig. 2. Lighl cone, x 2 - c~-t2=O, in special relativity.

6. C O N S E Q U E N C E S O F T H E T R A N S F O R M A T I O N S (12) AND (13) In the following we draw some consequences of the transformations (12) and (13).
( 1 ) Classical Limit. Assuming that T is much smaller than T o, one can neglect T 2 with respect to T ] , and the transformation (12) gives x'=x-Tv, v'=v

(18)

which is exactly the transformation obtained from classical mechanics. (2) Length Contraction. Suppose there is a rod located in the K system parallel to the x-axis. Let its length, measured in this system, be Ax=x2-x~, where x~ and x~ are the coordinates of the two ends of the rod. To determine the length of this rod as measured in the K' system, we

Cosmological Relativity

1037

must find the coordinates of the two ends of the rod x', and x / in this system at the same velocity v'. From Eq. (13a) we have
xl=(1
x'~ + To' T 2/ T o 2 1/2, ) -

x ~ - - (1 -

x',_ + To' - - T 2/ T , , ) 2 i/2

The length of the rod in the K' system is A x ' = x ' 2 - x ' ~ ; thus zJx-Ax' "

(1

--

T-~T:)

"

I/'~ -

The proper length of a rod is its length in a system in which it is located. Let us denote it by L o = A x and the length of the rod in any other system K' by L. Then
L = Lo( 1 -- T2/T~,)1/2 (19)

Thus a rod has its greatest length in the system in which its relative time with respect to the system is zero; its length in a system in which it is located at a relative time T with respect to that system is decreased by the factor (1 - T 2 / T ] ) ~/2. This result of the present theory is exactly similar to the familiar Lorentz contraction with the factor (1-v2/c2) ~/2 in special relativity. (3) V e l o c i t y C o n t r a c t i o n . Suppose a velocity-measuring instrument is located at x' = 0 in the K' system. Then from Eq. (13b) we have
I) t

v - ( 1 - T2/T~)1/2

(20)

Denoting now v by v o and v' by v, we obtain

v = Vo( 1 - T 2 / T ~ ) 1 / 2

(21)

The above result is like the time dilation in special relativity and was expected since time in special relativity goes over to velocity in the present theory. The velocity measured by an observer with a relative time T with respect to us is smaller by the factor ( 1 - T2/T2o) 1/2 than what is observed by us at T = 0 . Remark on Dark Matter. As is well known, much of the support for the existence of the hypothetical dark matter is due to the observed very high velocities of gas molecules or galaxies. For example, galaxies in the far-off Coma cluster are observed whirling around one another faster than

1038

Carmeli

the laws of physics would allow. So is the mysteriously rapid rotation of spiral galaxies. Equation (21) clearly shows that the velocity observed by us is not the velocity measured by a local observer at a relative time T with respect to us. He measures a smaller velocity, and the more back in time the more the velocity decreases. Does this mean that the hypothetical dark matter can be abolished just as the "luminiferous ether" was proved to be superfluous by special relativity? We will see in a coming paper that this is not so.

(4) Addition o f Times. Dividing the first of Eqs. (13) by the second,
we find
x
V V'"]-(

x' + Tv' T/ T o) X t 2

(22)

or, dividing the numerator and the denominator of the right-hand side of this equation by v', we obtain

7"

T+T
1+

(23)

Where 7 = x/v a n d / ' = x'/v'. Equation (23) determines the transformation of time and describes the law of composition of times in cosmological relativity. In the limiting case of T much smaller than T o, Eq. (23) goes over into the formula 7'= T' + T of classical mechanics. We see that the simple law of adding and subtracting cosmic times is no longer valid or, more precisely, is only approximately valid for short times with respect to us, but not for those near the Hubble time. Two consecutive events that occur at T~ = (9/lO)To and T = (180/181)To both with respect to us (at T = 0), for example, then with respect to the first event the second one does not occur at T - T~ ~ To/lO but rather at T - T1 9 T2 = 1 -- TT,/T2o = 1--0To which is much longer that T - T ~ and happens to be exactly equal to T 1. We also notice that the past cosmic time cannot be greater than To. This is similar to what we have in special relativity where the velocity cannot exceed c. It will be noted that one may add as many consecutive time intervals as one wishes without ever reaching the limiting time To. (5)
The Line Element.

This is given by
T2o d v
2 -- dx 2 =

ds 2

(24)

Cosmological Relativity
Hence

1039

(dq2_(dx?(aq =(To-Tb\&,/ - - I rO\as) \ & ) \ds) (

dv Po

(25)

Multiplying now this equation by p2 the matter density of the universe at o, the present time, we obtain for the matter density at a past time T
p = Topo ~ = ( 1 - T 2 / T 2) t/2

(26)

Similarly, one can calculate the temperature at a past time T in terms of the temperature at the present time (2.7 ~ K), but we will leave this, and other possible applications, to the reader. Remark. Since the volume of the universe is inversely proportional to its density, it follows that the ratio of the volumes at two backward times T1 and T2 with respect to us is given by (T2 < T, ).

-~-

T ~_/T~o'~ T1/r2oJ

= ( ( T o - T2)( ro + \(To T,)(To+ T,))

For times T 1 and T 2 very close to To we can assume that T o + T 2 To + Tl ~ 2T,,. Hence
V2 (t2"~ 1/2

Vl= 7,/
where tl = T o - T l we then have and t 2 = T o - T 2 . For t2--t, ~ I0 -32 sec and t2 @ l sec,

V2--. (1 + l O - 3 2 / t 1 ) m ~ (10-32/t,) '/2= lO-'6/tV 2 V1 F o r tl ~ 10 -132 sec we obtain V2 ~ 10S~ This result conforms with inflationary universe theory without assuming any model (such as the universe is propelled by a sort of antigravity). (6)

Minimal Acceleration in Nature.

F r o m Eq. (6) we have

T = x/v = dx/dv = via

where a is the acceleration. Hence

Tm,x = H o 1 = (v/a) max = c/a min

Consequently, in nature there is a minimal acceleration
amin = C l i o ~ 10 - s cm/sec 2

1040

Carmeli

7. C O N C L U D I N G REMARKS It is worth mentioning that the theory presented here is not a dynamical theory that describes the universe in any way since it does not take into account gravitation. Our theory, like Einstein's special relativity, is rather a skeleton theory which might be used as a guideline to dynamical theories that include gravitation and probably other things. Further study should certainly be carried out to find the equations of matter (dark matter, maybe) which are invariant under the transformation (12). This might give a completely new direction to cosmology in our attempt to understand the universe, since it involves at the outset the Hubble constant. Finally, it would be of great interest to find the representations of the cosmological group 0(3, 3) (or its covering group SL(4, R)). Since the cosmological group is not compact, it should have finite as well as O~nite-dimensional representations, which will need an extensive mathematical study.

ACKNOWLEDGMENTS I am indebted to Gidon Erez for useful discussions and encouragement, and to Yuval Ne'eman for pointing out to me that SL(4, R) is the covering group to the cosmological group.
REFERENCE

1. A. Einstein, //nn. Phys. (Leipzig) 17, 891 (1905); English translation in: A. Einstein et al., The Prhwiple of Relativity (Dover, New York, 1923).