Vous êtes sur la page 1sur 21

Cosmic acceleration as the solution to the cosmological constant

problem
Philip D. Mannheim
Department of Physics, University of Connecticut, Storrs, CT 06269
mannheim@uconnvm.uconn.edu

arXiv:astro-ph/9910093 v2 4 Apr 2000

Abstract
In this paper we provide both a diagnosis and resolution of the cosmological
constant problem, one in which a large (as opposed to a small) cosmological
constant can be made compatible with observation. We trace the origin of
the cosmological constant problem to the assumption that Newtons constant
G sets the scale for cosmology. And then we show that once this assumption
is relaxed (so that the local G as measured in a local Cavendish experiment
is no longer to be associated with global cosmology), the very same cosmic
acceleration which has served to make the cosmological constant problem so
very severe instead then serves to provide us with its potential resolution. In
addition, we present an alternate cosmology, one based on conformal gravity
(a theory which explicitly possesses no fundamental G), and show that once
given only that there is to be cosmic acceleration in the conformal theory
(i.e. once given only that in the theory the sign of is to specifically be
the negative one suggested by spontaneous symmetry breaking), then that
alone, no matter how big might actually be in magnitude, is sufficient to
not only make the actually measurable contribution (t0) of to current
era cosmology naturally be of order one today, but to even do so in a way
which is fully compatible with the recent high z supernovae cosmology data.
Cosmology can thus live with either a fundamental G or with the large (and
even potentially negative) associated with elementary particle physics phase
transitions but not with both.
I. DIAGNOSIS OF THE COSMOLOGICAL CONSTANT PROBLEM

The recent discovery [1,2] of a cosmic acceleration has made the already extremely disturbing cosmological constant problem even more vexing than before. Specifically, a phenomenological fitting to the new high z supernovae Hubble plot data using the standard
Einstein-Friedmann cosmological evolution equation
R 2 (t) + kc2 = R 2 (t)(M (t) + (t))

astro-ph/9910093/v2,

April 4, 2000

(1)

where M (t) = 8GM (t)/3c2 H 2 (t) is due to ordinary matter (viz. matter for which M (t) =
A/Rn (t) where A > 0 and 3 n 4) and where (t) = 8G/3cH 2 (t) is due to a
cosmological constant , has revealed that not only must the current era (t0 ) actually
be non-zero today, it is even explicitly required to be of order one. Typically, the allowed
parameter space compatible with the available data is found to be centered on the line
(t0) = M (t0 ) + 1/2 or so with (the presumed positive) M (t0) being found to be limited
to the range (0, 1) and (t0) to the range (1/2, 3/2) or so, with the current (n=3) era
deceleration parameter q(t0) = (n/2 1)M (t0) (t0 ) thus having to approximately lie
within the (1/2, 1) interval. Thus, not only do we find that the universe is currently
accelerating, but additionally we see that with there being no allowed (t0) = 0 solution
at all (unless M (t0 ) could somehow be allowed to go negative), the longstanding problem
of trying to find some way by which (t0 ) could be quenched by many orders of magnitude
from both its quantum gravity and particle physics expectations (perhaps by making it
vanish altogether) has now been replaced by the need to find a specific such mechanism
which in practice (rather than just in principle) would explicitly put (t0 ) into this very
narrow (1/2, 3/2) box. Not only is it not currently known how it might be possible to
actually do this, up to the present time no mechanism has been identified which might even
fix the sign of the standard model (t0 ) let alone its magnitude.
As such, the new supernovae data pose problems for both quantum and classical gravitational physics, but since the problems (familiar as they are) are somewhat different in the
two cases it is useful to discuss them independently. As regards first the quantum gravity
and elementary physics cosmological constant problem, we note that while it actually preceded the new high z data (see e.g. [3,4] for a review of the prior situation) the new data
have actually compounded the problem. Specifically, with quantum gravity being associated
with the Planck temperature kTP L = (
hc5 /G)1/2 and with elementary particle physics being
associated with a hadronic physics vacuum breaking scale,1 the expectation of fundamental
physics is that the ratio (t)/M (t) should be absolutely enormous. With the insertion
of this ratio into the Friedmann cosmological evolution equation of Eq. (1) then leading to
violent conflict with observation (even prior to the new high z data in fact), we see that as
it stands, the standard theory is explicitly in disagreement with data, and stress the point
now to emphasize that this is not merely a fine tuning problem (such as the one we will
encounter below when we discuss the classical gravitational physics cosmological constant
problem), but rather it constitutes the actual observational failure of an explicit prediction
of the standard theory (with a possible 120 orders of magnitude discrepancy potentially
being the largest disagreement ever recorded by any known theory). Now prior to the new
high z data it was widely hoped that might somehow be quenched, with some symmetry
principle perhaps bringing it down to zero identically, or with some dynamical mechanism
perhaps making it negligibly small. While neither of these possibilities has yet been achieved
in any convincing, universally accepted way, the situation was suddenly made all the more

1 We

can represent such a vacuum breaking scale by a typical effective temperature TV , a temperature which is then overwhelmingly larger than the effective temperature (of order a factor of ten
or so times the current temperature T (t0 ) of the universe) that would be associated with a black
body with critical density C (t0 ) = 3c2H 2(t0 )/8G.

complicated by the new high z data since this yet to be established mechanism would now
not only have to quench both the quantum gravity and the particle physics contributions
to (i.e. not just either but both), but also it would have to actually leave the quenched
with a very small but nonetheless non-vanishing component so that the current value of
the (t0)/M (t0) ratio might then be of order one instead.
Beyond the fact that such a quenching has yet to be achieved, it is important to note
that cosmology actually puts additional explicit demands on any such mechanism above and
beyond simply requiring the quenched to be such that the current value of (t0)/M (t0 )
comes out correctly. Specifically, suppose that some explicit quenching does take place in
the very early Planck temperature dominated quantum gravity universe (some candidate
mechanisms such as wormhole effects [5] which might be able to do this are discussed in
[3,4]), so that at the end of that era (or at the end of a subsequent but still early inflationary
universe era [6]) takes some particular value (this value might even be due to a fundamental
or anthropically induced cosmological constant which simply appears in the fundamental
gravitational action as an a priori fundamental constant). As the ensuing universe then
expands and cools it will potentially go through a whole sequence of elementary particle
physics phase transitions, in each one of which the vacuum energy would be lowered (this
being the definition of a phase transition). The residual from the early universe quenching
would then have to be such that it would just almost (but not quite completely) cancel the
net drop in vacuum energy due to all the particle physics phase transitions (transitions that
would occur only after the early universe quenching had already taken place - unless each
such phase transition is to be accompanied by its own quenching that is), so that just today,
i.e. conveniently just for our own particular epoch, (t0) would then be of order one.
Difficult as this is to even conceive of let alone demonstrate in an explicit dynamical model,
we note, however, that even if such a delicate balancing were to actually take place, the
resulting universe would then be one in which there could not have been any such delicate
cancellation prior (or near) to the very last phase transition. Thus in epochs prior or near to
the very last phase transition there could well have been substantial cosmological constant
contributions, to thus potentially give the universe a history and cosmology very different
from the standard one.
As an explicit example of how potentially severe a problem this might be, we note that
a standard electroweak or grand-unified symmetry breaking phase transition, for instance,
can generically be described by an effective Ginzburg-Landau theory with potential
V (, T ) = g4/2 2 (T )2,

(2)

where is the relevant order parameter and where 2 (T ) is typically given [7] by a form
such as 2 (T ) = g(TV2 T 2) in a convenient normalization. When the temperature T is less
than the transition temperature TV , the potential V (, T ) possesses a non-trivial minimum
away from the origin in which it takes the temperature dependent value
Vmin (T < TV ) = g(TV2 T 2)2 /2,

(3)

a value which is expressly negative (with respect to the zero value which Vmin (T > TV )
takes above the critical temperature), with the zero temperature Vmin (T = 0) taking the
convenient negative value gTV4 /2 in our normalization. To this potential we must now add
on an extra residual potential from the early universe of the form
3

Vres = g(TV2 T 2(t0 ))2/2 + T 4(t0),

(4)

to thus give us a total potential


Vtot (T < TV ) = Vmin (T < TV ) + Vres = gTV2 (T 2 T 2(t0 )) g(T 4 T 4 (t0))/2 + T 4(t0 ) (5)
and a current era Vtot (T (t0)) which would then nicely be of order the energy density in order
matter (generically given as T 4(t0 )) today.2 However, as we immediately see, in the nucleosynthesis era where TV  T  T (t0), the total vacuum energy density Vtot (T ) would then
be given by gTV2 T 2 in leading order and (with g being related to the standard Higgs and vec2
tor boson masses as g e2 MH2 /MW
) would thus be substantially larger than the black body
energy density at the same temperature T . Since the temperature dependence of a standard
k = 0 radiation era cosmology with non-zero is given by T 2/T 2 = (8G/3c2 )(T 4 + c),
we see that the very validity of standard big bang nucleosynthesis is now contingent on an
explicit demonstration that vacuum terms such as gTV2 T 2 are not in fact of relevance in the
nucleosynthesis era.3 It thus not sufficient to simply find a mechanism which quenches the
cosmological constant once (in some particular chosen epoch). Rather one needs renewed,
temperature dependent, quenching each and every time there is phase transition or new contribution to the vacuum energy.4 This then is quantum gravitational cosmological constant
problem.
Further, even independent of any of the above quantum considerations, the new high
z data pose a problem for Eq. (1) even when considered purely from the viewpoint of
phenomenological classical physics alone. Specifically, even if we ignore the above phase
transition issue, so that the temperature T (t) then uninterruptedly evolves adiabatically as
1/R(t), the current closeness to one of the ratio (t)/M (t) entails that in the early universe
this same ratio would have had to have been fantastically small, with a standard Friedmann
universe then only being able to evolve into its current state if this ratio had been extremely
fine tuned in the early universe (this is the fine tuning problem to which we referred above,
being one of having to adjust initial conditions to incredible accuracy, rather than one of
having to remove an explicit conflict with data). Moreover, this particular fine tuning would
have to be above and beyond that imposed by the flat (k (t) = kc2/R 2 (t) ' 0) inflationary
universe model since inflation only constrains the sum of M (t) and (t) to be one and
does not fix their ratio.5 This then is the classical gravitational cosmological constant

2 Why

exactly the residual early universe Vres should explicitly depend on the current temperature
T (t0 ) so that just our particular epoch of observers would see only the quenched energy density
T 4(t0 ) remains of course to be explained.
3

While it had generally been recognized [8] that Vtot(T > TV ) Vres would not compete with the
black body energy density at temperatures above the phase transition temperature TV , it appears
not to have been appreciated that there could still be a problem below it.
4 While

Vres can cancel the leading term in Vmin (T < TV ), because it is temperature independent
it leaves the next to leading term untouched.
5 While

there is still some question as to the extent of the region in the (M (t0 ), (t0 )) parameter

problem, and as we see, at the present time neither inflationary nor quantum cosmology
can readily accommodate the new high z data at all (i.e. neither a very early quantum
cosmology phase nor an early universe inflationary phase have yet been shown capable of
producing a subsequent Robertson-Walker phase whose onset value of /M would then
be the particular fantastically small one now required by data), with the early universe
now needing to be fine tuned all over again. Now while a solution to both the classical
and the quantum mechanical cosmological constant problems might yet be found within
standard gravity, the above described situation is so disquieting (with the cosmological
constant problem having resisted solution for such a very long time now) as to suggest that
in fact there might actually be something basically wrong with the whole standard picture,
a viewpoint to which we shall return below. Moreover, since the elementary particle physics
cosmological constant problem is a clash between two different branches of physics, gravity
and particle physics, we should not immediately assume that it is the particle physics side
which needs addressing. Rather, the indications of particle physics could well be correct,
with its contribution to actually being as big as it would appear to be, with the problem
then having to lie on the gravitational side instead. And indeed, in the following we shall
explicitly explore the implications for cosmology of actually being a very big rather than
a very small quantity. Thus, with attempts to quench not having been successful thus far,
we instead turn the issue around and explore below whether it is possible for cosmology to
accommodate a large instead.
In order to try to isolate the primary cause of all the above problems (so as to then know
exactly where it is that the standard theory is running into problems, with a view to then
identifying what explicitly needs to be done - either in standard gravity itself or beyond),
we shall now present as general a diagnosis of the problem as would appear possible. Thus
= 0) would have
we note that in any cosmology with a big bang, the early universe R(t
to be divergent (or at least be extremely large), with Eq. (1) then requiring the quantity
(M (t = 0) + (t = 0)) to be equal to one no matter what the value of the spatial
curvature k. Thus, given the radically different temporal behaviors of M (t) and (t),
in standard gravity we see directly that no cosmology, flat or non-flat, could ever evolve
into one in which (t0) ' M (t0) ' O(1) today without extreme fine tuning. Thus it
is actually the very fact of the big bang itself which is bringing standard cosmology to so
= 0) does start off divergent, it must diminish as
severe a current impasse. Further, if R(t
the universe begins to evolve, with the early universe thus decelerating. Since the current
universe now appears to be accelerating, the Friedmann universe fine tuning problem can

space which is allowed by the supernovae data, the one thing in the data that appears to be definitive
is that the point (1, 0) is overwhelmingly excluded. Since this is the only point which standard
classical cosmology is capable of reaching today without a fine tuning of Eq. (1), no change in the
supernovae data would appear likely to lead to a standard cosmology without some form or other
of fine tuning problem. Moreover, even a change in the data that might actually eliminate any
need for an explicit non-vanishing (t0 ) would still leave us with an M (t0) which would need to
be fine tuned to some value less than one, as well as with an k (t0) which could then no longer be
negligible. And even in that case we would of course still need to explain why would then have
been quenched to zero.

be viewed as the need to adjust parameters in such a way that the cosmology can exhibit
diametrically opposite deceleration and acceleration behaviors in differing epochs. Since the
big bang singularity itself derives from the fact that standard gravity is always attractive
(since G controls standard gravity on all distance scales including those much larger than
the solar system one on which standard gravity was first established) while acceleration is
more naturally associated with repulsion (cf. the high z data solution in which M (t0) is
taken to be negative), it is thus suggestive that we might be able to more readily balance
the early and current universes if there were no initial singularity at all, and if cosmological
gravity in fact got to be repulsive in all epochs, with the universe then expanding from some
= 0) = 0 instead. To achieve such
initial (but still very hot) state characterized by R(t
a singularity free cosmology, and to hope to decouple locally attractive gravity from such
cosmological repulsion would thus appear to require the removal of G from the fundamental
= 0) were indeed to vanish, the initial values of
gravitational action.6 Moreover, if R(t
(t = 0) and M (t = 0) would then both be infinite, and thus potentially never require
fine tuning at all. Thus we identify the very presence of G in the fundamental gravitational
action as a primary cause of the cosmological constant problem, an issue whose implications
we shall address in detail below.
To continue with the above analysis, we note further that if the apparently currently
accelerating universe continues to accelerate indefinitely, then, no matter what may or may

not have occurred in the early universe, in the very late universe R(t)
will actually become
arbitrarily large, with Eq. (1) then requiring the quantity (M (t) + (t)) to tend to one
at very late times, again independent of the value of k. However, because of their differing
time behaviors, we see that in the very late universe it would precisely be (t) which would
then have to tend to one no matter what its early universe value. Thus at very late times
the cosmological constant problem would actually get solved, and would in fact get solved
by cosmology itself (i.e. no matter how big might actually be, in universes whose fate
it is to accelerate at late times, there will eventually come a time in which the measurable
consequence of (t) will be that it will make a contribution to the expansion of the universe
which will be of order one). Thus even while the discovery of cosmic acceleration makes
the cosmological constant problem more acute, nonetheless, its very existence also suggests
a possible resolution of the issue. We shall thus explore this option, first (Sec. II) as a
generic cosmological effect, and then (in Sec. III) use it to obtain an explicit solution to the
cosmological constant problem in an explicitly solvable cosmological model.
II. GENERIC SOLUTION TO THE COSMOLOGICAL CONSTANT PROBLEM

In order to see how we might be able to capitalize on the behavior of universes which
accelerate at late times, it is very instructive [9] to analyze de Sitter geometry in a purely
kinematic way which requires no commitment to any particular dynamical equation of mo-

6 If

G is not to be a fundamental parameter, the G as measured in a Cavendish experiment would


then only be an effective, low energy parameter, with the very high temperature universe then
being controlled by some altogether different (and possibly even repulsive) scale instead.

tion. Specifically, suppose we know only that a given geometry is de Sitter, i.e. that its
Riemann tensor is given by
R = (g g g g ).

(6)

For such a geometry contraction then yields the kinematic relation


R g R /2 = 3g ,

(7)

R 2 (t) + kc2 = c2 R2 (t)

(8)

a relation which reduces to

when expressed in Robertson-Walker coordinates. On defining (t) = c2 R2 (t)/R 2 (t) we


obtain q(t) = (t) = 1 + kc2 /R 2 (t), with q(t) and (t) then being found [9] to be given
by
(t, < 0, k < 0) = q(t, < 0, k < 0) = tan2(()1/2ct),
(t, = 0, k < 0) = q(t, = 0, k < 0) = 0,
(t, > 0, k < 0) = q(t, > 0, k < 0) = tanh2(1/2 ct),
(t, > 0, k = 0) = q(t, > 0, k = 0) = 1,
(t, > 0, k > 0) = q(t, > 0, k > 0) = coth2 (1/2ct)

(9)

in the various allowed cases. As we thus see, when the parameter is positive, each associated solution corresponds (because of the absence of any explicit M matter term) to a
permanently accelerating universe, and that in each such universe (t, > 0) will eventually reach one no matter how big the parameter might be, and independent in fact of
whether or not G even appears in the cosmological evolution equations at all.7 Moreover,
while (t, > 0, k > 0) will only come down to one at very late times, quite remarkably, the
negative spatial curvature (t, > 0, k < 0) is bounded between zero and plus one at all
times, no matter how large might be. Thus unlike the unbounded < 0 case, we see that
when is greater or equal to zero, (t) is either bounded at all times or approaches a bound
at late times. Late time accelerating 0 de Sitter cosmologies will thus, without any
fine tuning at all, always quench the contribution of a cosmological constant to cosmology
no matter how large may be, and even no matter what form the underlying gravitational
theory might take. We thus we distinguish between quenching and quenching (t), while
noting that only the latter is actually required by known cosmological observations. Thus,
the parameter need not itself be small, and in fact the larger it is, the faster (t, > 0)
will then approach its asymptotic value.
In passing it is instructive to note that this particular quenching of (t) is also familiar in
the standard inflationary universe model where a positive cosmological constant dominated

I.e. no matter how large might be, the Hubble parameter always adjusts itself to be accordingly
large so that (t = , > 0) is then equal to one - for instance in the familiar flat k = 0 case

R(t)/R(t)
= 1/2c at all times.

cosmology rapidly quenches k (t) = 1 (t) with (t) rapidly becoming one. Now
while this quenching of k (t) is the major achievement of inflation, inflationary cosmology
was always faced with the (still not fully resolved) task of then disposing of this now large
(equal to unity in fact) (t) some time during the inflation exit transition to a subsequent
Robertson-Walker era, an era which would (ideally) then acquire an M (t) with the desired
value of one. In addition to this still not fully understood task, we now see that given the
new high z data, this exit transition must now not quite quench (t) completely, but must
instead make it fantastically small, so that it could then build itself back up to order one
today. Inflationary cosmology thus requires an (t) which first rises to one, then fine tunes
itself almost to zero, and then rises back to one once again. Thus even while an epoch of
inflationary acceleration nicely provides for an (t) equal to one, it is naturally only able
to do so at what might now possibly be the wrong time (unless inflation can naturally do it
twice that is), since independent of whether or not one should have had an (t) of order
one in the early universe, it is precisely an (t0) of order one today which is what appears
to be needed now.
Thus we propose that no matter what may or may not have happened in the early
universe (standard inflation or whatever), it is late universe acceleration which can naturally
resolve the cosmological constant problem, with a late universe de Sitter phase (either in
addition to or instead of an early universe one) then being able to naturally solve the
cosmological constant problem. Moreover, in such a situation need not itself be small,
and thus even if its associated temperature TV really is as big as suggested by elementary
particle physics, even so, (t) will still eventually approach one. Thus given these remarks,
it would appear that the key task then is to find a cosmology in which the current era
is already sufficiently late. As regards finding any such cosmology we are immediately
confronted with the fact that M (t0 ) is not zero today. Thus, in order to able to implement
our above ideas, the cosmological constant problem has to be converted from being one of
needing to quench into being one of needing to find a way in which M (t0) can be naturally
quenched instead. In order to actually try to achieve this alternate quenching, we note first
that the ratio M (t)/ (t) is actually independent of Newtons constant G, with its current
value being the gravity independent quantity M /c ' T 4 (t0)/TV4 . Moreover, not only is this
ratio independent of G, the only place where G does appear in cosmology is in the Friedmann
evolution equation where it fixes the overall normalization of both (t) and M (t), i.e. it
determines exactly just how big a contribution to cosmological evolution is to be made by
any given source of energy density. Thus we see that if we want to dominate cosmology
today with an effective contribution to cosmological evolution which is to be of order one,
we must in turn suppress the contribution of M to current cosmology, something we can
(drastic as it may initially seem) do if we can remove G from the cosmological evolution
equation and replace it by an effective coupling constant which is altogether smaller. Thus
we need not modify or M themselves. Rather, we need only change their effect on
cosmology. Thus just as in our discussion of the implications of the existence of a big bang,
we are again led to consider removing G from cosmology, and are again led to a G which is
to only be an effective low energy parameter, with this then enabling itself to actually be
as big as elementary particle physics seems to imply, while still providing us with a current

era (t0) of order one.8


With these remarks we have now concluded our diagnosis of the cosmological constant
problem and provided what we believe to be useful pointers for future attacks on the problem,
pointers that could prove helpful in trying to find a solution even within standard gravity
itself (a G which evolves with temperature is certainly conceivable within standard gravity though perhaps not one which might also change sign as it evolves). However, since we are
not currently aware of how to actually naturally implement the above ideas within standard
gravity, in the following, we shall instead turn to an alternate gravitational theory and show
that the above (t) bounding mechanism will precisely be found to occur in the conformal
gravity theory, a theory which has recently been advanced as an alternative to standard
gravity and its familiar dark matter paradigm, a theory in which G does not in fact set the
scale for cosmology. Beyond being of interest in and of itself (with it being quite remarkable
that there even exists any theory at all in which (t) and M (t) can both be naturally
quenched), our study can also be viewed simply as an existence theorem which shows that
it is actually possible to construct a model which explicitly realizes all the ideas we have
presented above.
III. EXPLICIT SOLUTION TO THE COSMOLOGICAL CONSTANT PROBLEM

In attempting to go beyond standard second order gravity, even within the confines of
covariant, pure metric based theories of gravity, we immediately realize that the choice is
vast, since we can in principle consider covariant theories based on derivative functions of the
metric of arbitrarily high order. However, given our previous remarks, within this infinite
family of higher order derivative gravitational theories, one of them is immediately singled
out, namely conformal gravity,9 a theory which can immediately lead to a cosmology free of
intrinsic scales at sufficiently high enough temperatures precisely because its gravitational
action possesses no fundamental scale at all. As such, conformal gravity emerges as a
potential gravitational analog of the Weinberg-Salam-Glashow electroweak theory, with it
immediately being suggested [11,12] that in it Newtons constant G might be generated as a

8 An

additional virtue of not having G be a fundamental parameter which is to control gravity


on all distance scales, is that then the Planck length no longer need control quantum gravity.
Consequently, as long as the mechanism which is to remove G from cosmology does so without
inducing any new quantum gravity scale itself, the quantum gravity contribution to the cosmological
constant would then be under control, leaving us with only the need to have to deal with the
elementary particle physics contribution TV .
9

Conformal gravity is a locally conformal invariant theory of gravity (i.e. one which is invariant
under any and all local conformal stretchings g 2 (x)g of the geometry), which has as its
R
uniquely allowed gravitational action the Weyl action IW = g d4x(g)1/2C C where
C is the conformal Weyl tensor [10] and where g is a purely dimensionless gravitational
coupling constant, and which consequently has gravitational field equations which are fourth order
derivative equations of motion.

low energy effective parameter in much the same manner as Fermis constant GF is generated
in the electroweak theory, with G as measured in a low energy Cavendish experiment then
indeed nicely being decoupled from the hot early universe. However, it turns out that the low
energy limit of conformal gravity need not emerge in precisely this fashion since [13] it is not
in fact necessary to spontaneously break the conformal gravity action down to the EinsteinHilbert action (i.e. down to the standard theory equations of motion). Rather [13] it is
only necessary to obtain the solutions to those equations in the kinematic region (viz. solar
system distance scales) where those standard solutions have been tested. Thus, as had been
noted by Eddington [14] already in the very early days of relativity, the standard gravity
vacuum Schwarzschild solution is just as equally a vacuum solution to higher derivative
gravity theories as well, since the vanishing of the Ricci tensor entails the vanishing of its
derivatives as well. And indeed, with variation of the conformal gravity action leading to
the equation of motion [15]
(g)1/2IW /g = 2g W = T /2

(10)

where W is given by
W = g (R );; /2 + R; ; R; ; R;; 2R R + g R R /2
2g (R); ; /3 + 2(R );; /3 + 2R R /3 g (R)2 /6,

(11)

and where T is the associated energy-momentum tensor, we confirm immediately that the
Schwarzschild solution is indeed a vacuum solution to conformal gravity despite the total
absence of the Einstein-Hilbert action in the purely gravitational piece IW of the conformal
action. Standard gravity is thus seen to be only sufficient to give the standard Schwarzschild
metric phenomenology but not at all necessary, with it thus being possible to bypass the
Einstein-Hilbert action altogether as far as low energy phenomena are concerned. 10 As we
shall show in detail below, it is precisely this aspect of the theory which will lead us to
a demarcation between the high and low energy regions which is very different from that
present in the electroweak case.
While conformal gravity itself is indeed an old idea, almost as old as General Relativity
itself in fact, it is only recently that its potential role in cosmology and astrophysics appears
to have been emphasized, with it having been found capable of addressing so many of the
problems (such as dark matter) which currently afflict standard gravity (see [9] and references
therein). As noted above, it has as a motivation the desire to give gravity a dimensionless
coupling constant and a local invariance structure, to thereby make it analogous to the three

10

Since higher derivative theories have different continuations to larger distances [13], we see that
the standard Schwarzschild solar system distance scale wisdom is compatible with many differing
extrapolations to larger distances, with standard gravity giving only one particular possible such
extrapolation. And indeed, it has been argued [16,13] that this is actually the origin of the dark
matter problem, with standard gravity simply giving an unsatisfactory extrapolation to galactic
distance scales and beyond. Interestingly, the conformal gravity extrapolation [17] has been found
to provide for a satisfactory explanation of galactic rotation curve systematics without the need to
introduce any galactic dark matter.

10

other fundamental interactions. And indeed as stressed in [18], the local conformal symmetry
invoked to do this then not only excludes the existence of any fundamental mass scales such
as a fundamental cosmological constant, even after mass scales are induced by spontaneous
breakdown of the conformal symmetry, the (still) traceless energy-momentum tensor then
constrains any induced cosmological constant term to be of the same order of magnitude
as all the other terms in T , neither smaller nor larger. Thus, unlike standard gravity,
precisely because of its additional symmetry, conformal gravity has a great deal of control
over the cosmological constant (essentially, with all mass scales - of gravity and particle
physics both - being jointly generated by spontaneous breakdown of the scale symmetry,
conformal gravity knows exactly where the zero of energy is), and it is our purpose now
to show that it is this very control which then provides for both a natural solution to the
cosmological constant problem and for a complete accounting of the new high z data.
The cosmology associated with conformal gravity was first presented in [19] where it
was shown, well in advance of the recent high z data, to be a cosmology with an effective
negative G, with the associated repulsion yielding a cosmology which was then found to
possess no flatness problem and thus be free of the copious amounts of cosmological dark
matter required in the standard theory. Subsequently [9], the cosmology was shown to
possess no horizon problem or universe age problem, and to also be capable of producing
cosmic repulsion through negative spatial curvature. To discuss conformal cosmology it is
convenient to consider the conformal matter action
IM =
h

(x)( + (x)) gS ]

d4 x(g)1/2[S S /2 + S 4 S 2 R /12 + i
(12)

for generic massless scalar and fermionic fields, a matter action which contains up to only
second order derivative functions of the matter fields.11 In this matter action we have
introduced a dimensionful scalar field S(x) which is to serve to spontaneously break the
conformal symmetry, and have, for simplicity, taken it to be a fundamental, conformally
coupled, purely macroscopic field. Nonetheless, this scalar field could just as easily be
generated microscopically as the dynamical expectation value of some elementary particle
multilinear condensate, with S(x) then being an effective Ginzburg-Landau field. Moreover,
in such a case, not only would there be the standard T = 0 elementary particle physics
contribution to S(x) due to a condensation of the filled negative energy modes of the relevant
elementary particles, but at non-zero temperatures there could also be a condensation of the

11

The same conformal invariance which forces the pure gravity Weyl action IW to be fourth order
also obliges the matter action IM to be a familiar second order quantity. In fact since actions such
R
as IM = d4x(g)1/2(F F )2 are both coordinate invariant and locally gauge invariant, they
are not in fact excluded by these particular invariances, with it being only scale invariance which
actually ensures their (otherwise simply assumed) absence in the standard SU (3) SU (2) U (1)
model of quarks and leptons. Hence it is actually scale invariance which obliges the standard
model to actually be second order, and thus renormalizable, in the first place. And moreover, if
the spontaneous breakdown of the standard model is generated by a gauge mediated interaction
such as hypercolor rather than by a fundamental tachyonic scalar field, the entire standard particle
physics IM would then actually possess no fundamental scale at all.

11

N occupied positive energy modes of these same particles as well. This latter condensation
would then lead to an effective, temperature dependent order parameter which would be
of order N in magnitude (the magnetization of a ferromagnet is, for example, proportional
to the number of positive energy spins present in a crystal), as well as to an effective
Vmin (T < TV ) which would, according to Eq. (3), explicitly be negative.12 In such a case the
effective would be negative, and would necessarily have to be so in fact in conformal gravity,
since in a theory with an underlying conformal symmetry, the (dimensionful) vacuum energy
density has to be zero identically in the scale invariant S(x) = 0 high temperature regime
above all scale generating phase transitions since an exact conformal symmetry ensures the
absence of any fundamental cosmological constant.13 Consequently, we shall incorporate
this aspect of conformal gravity in the following by considering cosmologies based on a
fundamental scalar field whose quartic self coupling coefficient is preferentially taken to
actually be negative, and find that it is precisely such negative which leads to cosmic
acceleration in the conformal theory.
For the above matter action, when the scalar field acquires a non-zero vacuum expectation
value (an expectation value which can always be rotated into a spacetime constant S 0 by
an appropriate local conformal transformation), the entire energy-momentum tensor of the

theory is found (for a perfect matter fluid Tkin


of the fermions) to take the form

T = Tkin
h
S02 (R g R /2)/6 g h
S04 ;

(13)

12 Since

the free energy density of a ferromagnet is a non-extensive function of the number of spins
which is dependent only on the particle number density and the mean magnetization per spin, the
identification of S(x) with a cosmological analog of the total magnetization of a ferromagnet and
the simulation of the analog of the large quantity Vmin (T < TV ) by the equally large c
hS 4(x)
together then entail that a positive frequency mode cosmological condensation would generate an
effective coupling constant which itself would be of order 1/N 4.
13 In

passing we note that this particular aspect of the theory is even manifest in the event of
dynamical symmetry breaking in scale invariant theories such as QED with no fundamental electron
mass, theories which also enjoy the scale invariance that is being considered here for gravity. In
such theories, it was found [2025] that even though the scale symmetry is destroyed by the bad
ultraviolet behavior of perturbation theory, nonetheless it gets restored non-perturbatively by a
Gell-Mann Low eigenvalue for the coupling constant, at which the Greens functions of the theory
then scale with anomalous dimensions. Additionally, it was noted [26] that in such theories the
same anomalous dimensions which serve to soften the short distance behavior of the massless
theory then cause the massless theory to become more badly behaved in the infrared, with such
infrared behavior then being found to lead to a dynamical double well effective potential of the
form (m) = m2 [ln(m2/M 2 ) 1], and to drive the theory to the non-trivial minimum at m = M
where the potential takes the expressly negative value min (m = M ) = M 2 , with the fermion
then acquiring a non-zero dynamical mass M . Since the unbroken (m = 0) is zero, we see that
non-perturbative dynamical symmetry breaking in scale invariant theories leads to an explicitly
negative vacuum energy density (and even to an explicit Ginzburg-Landau structure [27]), just as
desired.

12

with the complete solution to the scalar field, fermionic field, and gravitational field equations
of motion in a background Robertson-Walker geometry (viz. a geometry in which the Weyl
tensor and W both vanish) then reducing [9] to just one relevant equation, namely
T = 0,

(14)

a remarkably simple condition which immediately fixes the zero of energy. On rewriting this
condition in the form

h
S02(R g R /2)/6 = Tkin
g h
S04 ,

(15)

we thus see that the evolution equation of conformal cosmology looks identical to that of
standard gravity save only that the quantity
hS02 /12 has replaced the familiar c3 /16G.
With this homogeneous and isotropic, global scalar field S0 filling all space and acting
cosmologically, we see that this change in sign compared with standard gravity leads to
a cosmology in which gravity is globally repulsive rather than attractive. Because of this
change in sign, conformal cosmology thus has no initial singularity (i.e. it expands from a
finite minimum radius), and is thus precisely released from the standard big bang model
constraints described earlier. Similarly, because of this change in sign the contribution of
M (t) to the expansion of the universe is now effectively repulsive, to nicely mesh with the
phenomenological high z data fits in which M (t) was allowed to go negative. Apart from
a change in sign, we see that through S0 there is also a change in the strength of gravity
compared to the standard theory. It is this feature which will prove central to the solution
to the cosmological constant problem which we present below.
Despite the fact that conformal gravity has now been found to be globally repulsive,
nonetheless, it is important to note that in the conformal theory local solar system gravity
can still be attractive; with it having been specifically found [13] that for a static, spherically
symmetric source such as a star, the conformal gravity field equation of Eq. (10) reduces to a
fourth order (i.e. not a second order) Poisson equation 4g00 = 3(T 00 T rr )/4g g00 f (r),
R
R
with solution g00(r) = 1 2 /r + r where = drf (r)r4 /12 and = drf (r)r2 /2.
With the coupling constant g in the Weyl action IW simply making no contribution in
highly symmetric cosmologically relevant geometries where C and W vanish, and
with the sign of being directly given by the sign of this thus cosmologically irrelevant
g , we see that locally attractive and globally repulsive gravity are now decoupled and
thus able to coexist. Local gravity is thus fixed by local sources alone, sources which are
only gravitational inhomogeneities in the otherwise homogeneous global cosmological background, i.e. sources which are characterized by small, local variations in the background
scalar field S(x), variations which themselves are completely decoupled from the homogeneous, constant, cosmological background field S0 itself. It is thus the distinction between
homogeneity and inhomogeneity which provides the demarcation between local and global
gravity, to thus now enable us to consider repulsive cosmologies which are not incompatible
with the attractive gravity observed on solar system distance scales.
Given the equation of motion T = 0, the conformal cosmology evolution equation is
then found to take the form (on setting = h
S04 )
M (t) +
(t))
R 2 (t) + kc2 = 3R 2 (t)(M (t) + (t))/4S02 L2P L R 2 (t)(
13

(16)

M (t) and
(t)), with the deceleration parameter now being
(Eq. (16) serves to define
M (t)
(t). As we see, precisely because the underlying
given as q(t) = (n/2 1)
conformal invariance forces the conformal T to be of the standard second order form, Eq.
(16) is found to be remarkably similar in form to Eq. (1), with conformal cosmology thus
only containing familiar ingredients. As an alternate cosmology then, conformal gravity
thus gets about as close to standard gravity as it is possible for an alternative to get while
nonetheless still being different. Moreover, even though that had not been its intent, because
of this similarity, we see that phenomenological fits in which M (t) and (t) are allowed
to vary freely in Eq. (1) are thus also in fact phenomenological fits to Eq. (16), with the
various (t) simply being replaced by their barred counterparts. In order to see whether
(t0 ) =
M (t0)+1/2 window, it is necessary
conformal gravity can thus fit into the relevant
to analyze the solutions to Eq. (16). Such solutions are readily obtained [9], and can be
classified according to the signs of and k. In the simpler to treat high temperature era
where M (t) = A/R4 = T 4 the complete family of solutions is given as
R2 (t, < 0, k < 0) = k(1 )/2 + ksin2(()1/2ct)/,
R2 (t, = 0, k < 0) = 2A/k
hcS02 kc2t2 ,
R2 (t, > 0, k < 0) = k( 1)/2 ksinh2(1/2ct)/,
R2 (t, > 0, k = 0) = (A/
hcS04 )1/2cosh(21/2 ct),
R2 (t, > 0, k > 0) = k(1 + )/2 + ksinh2(1/2ct)/,

(17)

where we have introduced the parameters = 2S02 and = (116A/k 2 h


c)1/2. Similarly
the associated deceleration parameters take the form
q(t, < 0, k < 0) = tan2(()1/2ct) 2(1 )cos(2()1/2ct)/sin2(2()1/2 ct),
q(t, = 0, k < 0) = 2A/k 2h
c3S02 t2,
q(t, > 0, k < 0) = tanh2(1/2 ct) + 2(1 )cosh(21/2ct)/sinh2(21/2 ct),
q(t, > 0, k = 0) = 1 2/sinh2 (21/2 ct),
q(t, > 0, k > 0) = coth2 (1/2ct) 2(1 )cosh(21/2 ct)/sinh2(21/2 ct).

(18)

Now while Eq. (17) yields a variety of temporal behaviors for R(t), it is of great interest
= 0) being zero (rather than infinite)
to note that every single one of them begins with R(t
just as desired above, and that each one of the solutions in which is negative (viz. > 0)
is associated with a universe which permanently expands (only the > 0 solution can
recollapse, with conformal cosmology thus correlating the long time behavior of R(t) with
the sign of rather than with the sign of k). We thus need to determine the degree to which
the permanently expanding universes have by now already become permanently accelerating.
To this end we note first from Eq. (18) that with being greater than one when is
negative, both the > 0, k < 0 and the > 0, k = 0 cosmologies are in fact permanently
accelerating ones no matter what the values of their parameters. To explore the degree to
which they have by now already become asymptotic, as well as to determine the acceleration
properties of the > 0, k > 0 cosmology, we note that since each of the solutions given in
Eq. (17) has a non-zero minimum radius, each associated > 0 cosmology has some very
large but finite maximum temperature Tmax given by
14

2
Tmax
( > 0, k < 0)/T 2 (t, > 0, k < 0) = 1 + 2sinh2(1/2ct)/( 1),
2
Tmax
( > 0, k = 0)/T 2 (t, > 0, k = 0) = cosh(21/2 ct),
2
Tmax
( > 0, k > 0)/T 2 (t, > 0, k > 0) = 1 + 2sinh2 (1/2ct)/( + 1),

(19)

with all the permanently expanding ones thus necessarily being way below their maximum
temperatures once given enough time. To obtain further insight into these solutions it is
convenient to introduce an effective temperature according to c
hS04 = TV4 . In terms of
this TV we then find that in all the < 0 cosmologies the energy density terms take the
form
(t) = (1 T 2/T 2 )1 (1 + T 2T 2 /T 4 )1 ,

max
max
V
4
4

M (t) = (T /TV ) (t),

(20)

4
4
for the k < 0 case, and where ( 1)/( + 1) = Tmax
/TV4
where ( 1)/( + 1) = TV4 /Tmax
for the k > 0 case. With being greater than one, we find that for the k > 0 case TV
is greater than Tmax, for k = 0 TV is equal to Tmax, and for k < 0 TV is less than Tmax,
with the energy in curvature (viz. the energy in the gravitational field itself) thus making
a direct contribution to the maximum temperature of the universe. Hence, simply because
the temperature Tmax is overwhelmingly larger than the current temperature T (t0) (i.e.
simply because the universe has been expanding and cooling for such a long time now), we
see that, without any fine tuning at all, in both the k > 0 and k = 0 cases (i.e. cases
(t0 ) is already at its asymptotic limit of one
where TV Tmax  T (t0)), the quantity

today, that M (t0) is completely suppressed, and that the deceleration parameter is given
by q(t0) = 1.
For the > 0, k < 0 case (the only > 0 case where TV is less than Tmax) a very different
2
outcome is possible however. Specifically, since in this case the quantity (1 + T 2Tmax
/TV4 )1
is always bounded between zero and one no matter what the relative magnitudes of TV ,
Tmax and T (t), we see that as long as Tmax is very much greater than T (t0), rather than
(t0 )
having had to have already reached its asymptotic limit of one by now, the quantity
is instead only required to be bounded by it. With it thus being expressly bounded from
(t0 ) thus has to lie somewhere between zero and one today no
above, the current value of
matter how big or small TV might be; with the simple additional requirement that TV also
M (t0) will yet again be completely
be very much greater than T (t0) then entailing that

suppressed in the current era. Moreover, (t0) will take a typical value of one half should
2
(t0) not
the value of the quantity T 2(t0)Tmax
/TV4 currently be close to one. Values of
merely less than one but even appreciably so are thus readily achievable in the k < 0 case
for a continuous range of temperature parameters which obey Tmax  TV  T (t0) without
the need for any fine-tuning at all.14 Noting from Eq. (19) that the temporal evolution of
the > 0, k < 0 case is given by

14 Since

an S0 6= 0 negative curvature conformal cosmology has a maximum temperature even in


2
the absence of any cosmological constant contribution at all [9], viz. Tmax
= k
hS02c/2(A)1/2,
the magnitudes of Tmax and TV are thus fixable independently in the TV 6= 0 case, with the > 0,
k < 0 case TV thus not merely being smaller than Tmax, but also being capable of being naturally
very much smaller than Tmax even while still being altogether greater than T (t0 ).

15

2
4
sinh2(1/2ct) = (Tmax
/T 2 1)/(Tmax
/TV4 + 1),

(21)

(t),
R 2 (t) + kc2 = R 2 (t)

(22)

(t0 ) is then given by


we see that in the Tmax  TV  T (t0) case, the current value of
2
1/2
the nicely bounded form tanh ( ct0), i.e. given precisely by the form found in the model
independent analysis of de Sitter space that was presented above. Additionally, in this
case k (t0) is then given by sech2 (1/2ct0), with negative spatial curvature then explicitly
contributing to current era cosmology (and even doing so repulsively according to [9]). 15 In
(t0 ) less than one are thus naturally achievable
the k < 0 case then we find that values of
(t0) not being able to be larger than one (no matter what the
in our model, and with
value of k) given only that is positive and that the universe is as old as it is, we see that
a conformal cosmology universe solves the cosmological constant problem simply by living
for a very long time.
Thus we see that in all three of the > 0 cases the simple requirement that Tmax  T (t0),
M (t0) is completely negligible at current temperatures (it can
TV  T (t0) ensures that
thus only be relevant in the early universe), with the current era Eq. (16) then reducing to

to thus not only yield as a current era conformal cosmology what in the standard theory
could only possibly occur as a very late one, but to also yield one which enjoys all the nice
purely kinematic properties of a de Sitter geometry which we identified above. Since studies
of galaxy counts indicate that the purely visible matter contribution to M (t0) is of order
one (actually of order 102 or so in theories in which dark matter is not considered), it
M (t0) will in fact be achieved if the
follows from Eq. (16) that current era suppression of
conformal cosmology scale parameter S0 is altogether larger than the inverse Planck length
L1
P L , a condition which is naturally imposable (i.e. for a continuous, non-fine-tuned, range
of parameters of the theory) and which is explicitly precisely compatible with a large rather
than a small TV , i.e. with a large rather than a small S0 .16 Comparison with Eq. (1)

15 For

completeness, we note that in the case where > 0 and k < 0, the Hubble param2
eter obeys H(t) = 1/2c(1 T 2 (t)/Tmax
)/tanh(1/2ct), with its current value thus obeying
2
2
q(t0 )H (t0 ) = c , and with the current age of the universe then being given by H(t0 )t0 =
arctanh[(q(t0 ))1/2]/(q(t0))1/2. Thus in general we see that t0 is greater than 1/H(t0) (t0 =
1/H(t0) in the k < 0 case in which = 0), with it taking the value t0 = 1.25/H(t0) when
q(t0 ) = 1/2. Thus as already noted in [9] conformal cosmologies have no universe age problem.
M (t0 ) is suppressed by large S0 it no longer matters whether
passing we note that once
M (t) is itself dominated by n = 3 matter or n = 4 radiation, since neither of them makes any
substantial contribution to the full current era conformal gravity energy-momentum tensor (with
M (t) is only of relevance in the early universe). As regards this
this suppression, in our model
suppression we recall, as noted earlier, that if the scalar field is microscopic, this same suppression
M (t0) can be generated simply by virtue of the number, N , of occupied positive energy states
of
in the universe being very large, with the negative energy modes not themselves then needing to
generate a temperature way in excess of TP L. Moreover, in such a case the parameter = 2S02
would be of order 1/N 2, with the quantity tanh2 (1/2ct0 ) not then having to be asymptotic in the
current epoch.
16 In

16

shows that current era < 0 conformal cosmology looks exactly like a low mass standard
model cosmology, except that instead of M (t0) being negligibly small (something difficult
M (t0 ) = 3M (t0)/4S 2 L2 which is negligibly
to understand in the standard theory) it is
0 PL
small instead (M (t0) itself need not actually be negligible in conformal gravity - rather, it
is only the contribution of M (t) to the evolution of the current universe which needs be
small). Hence, we see that the very essence of our work is that the same mechanism which
(t0) to be of order one today, viz. a large rather than a small = h
causes
S04 , serves at
M (t0) to decouple from current era
the same time, and without any fine tuning, to cause
17
cosmology. Thus to conclude we see that when is negative, that fact alone is sufficient
M (t0) = 0 and to 0
(t0) 1, with
(t0) coming closer
to automatically lead us to
to one half the more negative the spatial curvature of the universe gets to be.
Now while we have seen that all the three negative conformal cosmologies lead us to a
(t0 ) of order one, in order to be able to choose between them it is necessary to
current era
try to determine k. To this end we appeal [17] to an at first highly unlikely source, namely
galactic rotation curve data. Recalling that in conformal gravity the metric outside of a static
spherically symmetric source such as a star is given by g00 (r) = 1 2 /r + r, we see
that in the conformal theory the departure from Newton is found to be given by a potential
that actually grows (linearly) with distance. Hence, unlike the situation in standard gravity,
in conformal gravity it is not possible to ever neglect the matter exterior to any region of
interest, with the rest of the universe (viz. the Hubble flow) then also contributing to galactic
motions (i.e. a test particle in a galaxy not only samples the local galactic gravitational field,
it also samples that of the global Hubble flow as well). And indeed, it was found [17] that
the effect on galaxies of the global Hubble flow was to generate an additional linear potential
with a universal coefficient given by 0 /2 = (k)1/2, i.e. one which is generated explicitly
by the negative scalar curvature of the universe18 (heuristically, the repulsion associated
with negative scalar curvature pushes galactic matter deeper into any given galaxy, an effect
which an observer inside that galaxy interprets as attraction), with conformal gravity then
being found able to give an acceptable accounting of galactic rotation curve systematics (in
the data fitting 0 is numerically found to be given by 3.06 1030 cm1, i.e. to be explicitly
given by a cosmologically significant length scale) without recourse to dark matter at all.
We thus identify an explicit imprint of cosmology on galactic rotation curves, recognize
that it is its neglect which may have led to the need for dark matter, and for our purposes

M (t0) = 0 are within the region allowed by the new high z data, and while
Cosmologies with
such a situation would be hard to understand within standard gravity where M (t0 ) = 0 would
correspond to an empty universe, we see that conformal cosmology can contain ordinary matter
(M > 0) and yet still have it decouple from its current evolution.
17

18 Essentially,

under the general coordinate transformation r = /(1 0/4)2, t = d /R( ),


a static, Schwarzschild coordinate observer in the rest frame of a given galaxy recognizes the
(conformally transformed) comoving Robertson-Walker metric ds2 = (, )[c2d 2 R2 ( )(d2 +
2d)/(1 2 02/16)2] (where (, ) = (1 + 0/4)2/R2( )(1 0/4)2) as being conformally
equivalent to the metric ds2 = (1 + 0r)c2dt2 dr2/(1 + 0r) r2 d.

17

here confirm that k is indeed negative.19 Conformal cosmology thus leads us directly to
M (t0) = 0,
(t0) = tanh2(1/2ct0), and would thus appear to lead us naturally (providing

only that < 0) right in to the region favored by the new high z data.
As regards the role played by negative in conformal cosmology, we recall that the
effect of elementary particle physics phase transitions is to lead to a vacuum energy density
Vmin (T < TV ) which is typically expected to be negative rather than positive since each
one of the many particle physics phase transitions acts to lower the vacuum energy density
some more as the universe cools. Once given such a negative (something we simulate by
taking the quartic scalar self-coupling coefficient to be negative), its effect on the evolution
of the universe then depends on the sign of the effective G. Thus, for repulsive conformal
(t) and thus to cosmic acceleration, whereas
cosmology, negative translates into positive
for the attractive standard cosmology it translates into negative (t), to then not lead to
any cosmic acceleration at all. Thus added to the challenges faced by the standard theory is
the need to explain not only why should be small, but also to explain why it should also
not in fact be negative, with the very fact of cosmic acceleration providing some support for
the central theme of our work, namely that cosmologically, the effective G is in fact negative.
As regards such an effective negative cosmological G, we note that are essentially two
primary arguments which have in the past supported the contrary, G positive, cosmological
position, namely the current value of M (t0) and big bang nucleosynthesis. However, of
these two, the M (t0 ) argument now has to be discounted. Specifically, with earlier (i.e.
pre high z) data having led to a current value of M (t0 ) = 8GM (t0)/3c2 H 2 (t0) which was
tantalizingly close to one (provided one included dark matter that is), it strongly suggested
that cosmology was indeed normalized to the gravitational constant G, with cosmological
theory otherwise having to explain this closeness as an accident. And, indeed, the great
appeal of inflation was that it provided a rationale for having M (t0) be close to one today
by having M (t) be identically equal to one in each and every epoch. However, with the new
high z data, we now know that M (t0) is unambiguously less than one, and more, that it
will get ever smaller as the universe continues to accelerate. Thus, for observers sufficiently
far enough into the future M (t) will be nowhere near one, with its current closeness to
one being only an artifact of the particular epoch in which current observers happen to be
making observations (and of course without dark matter, by itself known explicitly detected
luminous matter only yields for M (t0) a value which is actually a few orders of magnitude
or so below one today).
With regard to nucleosynthesis, we note that, in principle, it only requires that the
universe had once been hot enough to have been able to trigger nuclear reactions, with it
not being at all necessary that even earlier there had been an altogether hotter big bang
phase. And indeed, it has been found [2931] that since the universe has been expanding
and cooling for such a very long time now, conformal cosmology is also capable of having
once been hot enough to have undergone nucleosynthesis; with the latest calculations [31]

19 Given

the presence of the imprint of such a cosmological scale on galaxies, it thus becomes
necessary (see also [28] for related discussion) for dark matter models to equally produce such a
scale, something which may not be all that easy in standard flat k = 0 models where no curvature
scale is available.

18

yielding the requisite amount of helium as well as the metallicity which is explicitly seen in
population II stars,20 with the inability [2931] of conformal cosmology to yield a sufficient
amount of deuterium being its only outstanding nucleosynthesis problem. Now, as regards
the production of deuterium, we note that while it is generally thought difficult to produce
post-primordially, this is not quite the case, as it is actually fairly easy to both produce and
then retain deuterium by spallation or fragmentation of light nuclei [3234], particularly if
the spallation is pre rather than post galactic.21 In fact the problem is then not one of an
underproduction of deuterium, but rather of an overproduction of the other light elements.
However, as noted by Epstein [34], if the spallation is to also take place in the early universe
with its onset occurring after the nucleosynthesis itself, then (i) in such a situation only
hydrogen and helium interactions would be of any significance, with only Z 3 nuclei then
being producible, and (ii) that in such a case the high energies involved would serve to favor
deuterium production over the lithium production which is favored at ordinary energies.
In addition to this, the authors of [31], on having found the helium abundance in their
nucleosynthesis calculations to be a rather sensitive function of the baryon to entropy ratio,
have suggested that lithium production could also be suppressed if the spallation were to take
place inhomogeneously with helium deficient clouds then spallating with helium rich ones.
In such a case, deuterium would then be produced not during nucleosynthesis itself but some
time afterwards just as inhomogeneities first begin to form in the universe.22 Since a theory
for the growth of inhomogeneities in conformal cosmology has not yet been developed, it is
not possible to currently provide a detailed analysis of this issue or assess its implications
for conformal gravity. And while one should not understate the seriousness of the deuterium
problem in conformal cosmology (indeed its viability as a cosmological model is contingent
upon a successful resolution of this very issue), nonetheless, the relative ease with which
conformal gravity deals with cosmological constant problem, the most severe problem the
standard theory faces, would appear to entitle the conformal theory to further consideration.
And even if the conformal gravity alternative were to fall by the wayside, nonetheless our
analysis of the role that G plays in the standard model cosmological constant problem would
still remain valid.

20 Even

though the cosmology expands far more slowly than the standard cosmology, nonetheless,
this gets compensated for in the conformal case due to the fact that weak interactions are then found
[31] to remain in thermal equilibrium down to much lower temperatures than in the standard case,
with its metallicity predictions then being found [31] to actually outperform those of the standard
model.
21

Even though spallation models of deuterium production were never actually ruled out, the
models were quickly set aside once it became apparent that deuterium could be produced by
standard big bang nucleosynthesis.
22

It could thus be of interest to measure the lithium to deuterium abundance ratio of high z quasar
absorbers, with the obtaining of a value for this ratio different from that expected in standard big
bang nucleosynthesis then possibly indicating the occurrence of inhomogeneous but still fairly early
universe spallation.

19

To conclude this paper, we note once again that spontaneous breakdown effects such as
those associated with a Goldstone boson pion or with massive intermediate vector bosons
seem to be very much in evidence in current era particle physics experiments, and are
thus not quenched at all apparently. Hence all the evidence of particle physics is that its
contribution to should in fact be large rather than small today. However, since in such
(t0) to still be small today, we see that the standard
a case it is nonetheless possible for
gravity fine tuning problem associated with having M (t0) ' (t0) today can be viewed as
being not so much one of trying to understand why it is (t0) which is of order one after
15 or so billion years, but rather of trying to explain why the matter density contribution to
cosmology should be of order one after that much time rather than a factor T 4/TV4 smaller.
Since this latter problem is readily resolved if G does not in fact control cosmology, but
if cosmology is instead controlled by some altogether smaller length squared scale such as
1/S02 , we see that the origin of the entire cosmological constant problem can be directly
traced to the assumption that gravity is controlled by Newtons constant G on each and
every distance scale; with the very existence of the cosmological constant problem possibly
being an indicator that the extrapolation of standard gravity from its solar system origins
all the way to cosmology might be a lot less reliable than is commonly believed.
The author wishes to thank Dr. D. Lohiya for useful discussions. This work has been
supported in part by the Department of Energy under grant No. DE-FG02-92ER40716.00.
Added Notes
(1.) In Sec. (I) we noted that the canceling of the cosmological constant term in the
current epoch does not in and of itself guarantee its irrelevance in all earlier ones. However,
as regards the particular Ginzburg-Landau mean field model discussed in Sec. (I), the
author is indebted to Dr. M. Sher for pointing out to him that, at least in the absence
of any chemical potential (such as that generatable by the condensation contemplated in
this paper of a large number of occupied positive energy modes), the one loop correction
to the classical tree approximation effective potential can actually serve to suppress the
temperature dependence of the cosmological constant contribution at temperatures below
the critical temperature TV (see L. Dolan and R. Jackiw, Phys. Rev. D 9, 3320 (1974); M.
Sher, Phys. Rept. 179, 273 (1989)); with it thus being necessary to check for any possible
temperature dependence to the vacuum energy in general on a case by case basis.
(2.) It is of interest to note that within the general framework of our proposal of trying
to suppress the effective cosmological G, a new mechanism may have potentially just become
available within standard gravity itself. Specifically, it was suggested recently (L. Randall
and R. Sundrum, Phys. Rev. Lett. 83, 3370 (1999); ibid. 83, 4690 (1999)) that standard
four dimensional gravity might only be a brane-localized limit of gravity in some higher large
extra dimension with the effective four-dimensional G then being determined dynamically
by the structure of the embedding in the higher dimensional space. It would thus be of
interest to explore whether such embeddings could lead to a different effective G in the
four-dimensional high and low energy limits.
(3.) Since this paper was finished we have been able to show from a study of conformal
cosmology at temperatures above all phase transitions that the sign of the spatial curvature
k of the universe is then uniquely fixed to be negative, just as desired in this paper. This
result as well as further related discussion of our work may found in P. D. Mannheim, Found.
Phys. 30, (2000), in press (gr/qc 0001011).
20

REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]

[29]
[30]
[31]

[32]
[33]
[34]

A. G. Riess et. al., Astronom. J. 116, 1009 (1998).


S. Perlmutter et. al., Astrophys. J. 517, 565 (1999).
S. Weinberg, Rev. Mod. Phys. 61, 1 (1989).
Y. J. Ng, Int. J. Mod. Phys. D 1, 145 (1992).
S. Coleman, Nucl. Phys. B 310, 643 (1988).
A. H. Guth, Phys. Rev. D 23, 347 (1981).
S. Weinberg, Phys. Rev. D 9, 3357 (1974).
S. A. Bludman and M. A. Ruderman, Phys. Rev. Lett. 38, 255 (1977).
P. D. Mannheim, Phys. Rev. D 58, 103511 (1998).
H. Weyl, Math. Zeit., 2, 384 (1918).
S. L. Adler, Rev. Mod. Phys. 54, 729 (1982).
A. Zee, Ann. Phys. 151, 431 (1983).
P. D. Mannheim and D. Kazanas, Gen. Relativ. Gravit. 26, 337 (1994).
A. S. Eddington, The Mathematical Theory of Relativity, Eighth Edition (First Edition
1922), Cambridge University Press, Cambridge, U. K. (1960).
B. S. DeWitt, Dynamical Theory of Groups and Fields, Gordon and Breach, New York
(1965).
P. D. Mannheim and D. Kazanas, Astrophys. J. 342, 635 (1989).
P. D. Mannheim, Astrophys. J. 479, 659 (1997).
P. D. Mannheim, Gen. Relativ. Gravit. 22, 289 (1990).
P. D. Mannheim, Astrophys. J. 391, 429 (1992).
K. Johnson, M. Baker and R. Willey, Phys. Rev. 136, B1111 (1964).
K. Johnson, R. Willey and M. Baker, Phys. Rev. 163, 1699 (1967).
M. Baker and K. Johnson, Phys. Rev. 183, 1292 (1969).
M. Baker and K. Johnson, Phys. Rev. D 3, 2516 (1971).
M. Baker and K. Johnson, Phys. Rev. D 3, 2541 (1971).
K. Johnson and M. Baker, Phys. Rev. D 8, 1110 (1973).
P. D. Mannheim, Phys. Rev. D 12, 1772 (1975).
P. D. Mannheim, Nucl. Phys. B 143, 285 (1978).
S. S. McGaugh, in Proceedings of Galaxy Dynamics, A Rutgers Symposium, August
1998. Astronomical Society of the Pacific Conference Series, Vol. 182, edited by D.
Merritt, J. A. Sellwood, and M. Valluri, A. S. P., San Francisco (1999).
L. Knox and A. Kosowsky, Primordial Nucleosynthesis in Conformal Weyl Gravity,
Fermilab-Pub-93/322-A, astro-ph/9311006 (1993).
D. Elizondo and G. Yepes, Astrophys. J. 428, 17 (1994).
D. Lohiya, A. Batra, S. Mahajan, and A. Mukherjee, Nucleosynthesis in a Simmering
Universe, nucl-th/9902022 (1999); M. Sethi, A. Batra, and D. Lohiya, Phys. Rev. D
60, 108301 (1999).
F. Hoyle and W. A. Fowler, Nature 241, 384 (1973).
R. I. Epstein, J. M. Lattimer, and D. N. Schramm, Nature 263, 198 (1976).
R. I. Epstein, Astrophys. J. 212, 595 (1977).

21