From global time to local physics

Syed Moeez Hassan,1, ∗ Viqar Husain,2, † and Babar Qureshi3, ‡

1
Department of Physics, Syed Babar Ali School of Science and Engineering,
Lahore University of Management Sciences, Lahore 54792, Pakistan
2
Department of Mathematics and Statistics,
University of New Brunswick, Fredericton, NB, Canada E3B 5A3
3
Department of Mathematical Sciences, IBA, Karachi, Pakistan

(Dated: July 12, 2024)

arXiv:2407.08283v1 [gr-qc] 11 Jul 2024

Global time is a gauge or relational choice of time variable in canonical gravity. Local
time is the time used in a flat patch of spacetime. We compare the dynamics of a scalar field
with respect to choices of global time and Minkowski patch time in an expanding cosmology.
Our main results are that evolutions starting from the same initial conditions are similar
on the time scales of terrestrial experiments, and that global time leads to a mechanism for
evolving coupling constants.

I. INTRODUCTION

Clocks are oscillatory physical systems. For all terrestrial experiments their readings coincide with
either Newtonian absolute time or the special relativistic time of a local inertial frame. For cos-
mological theories and observations described by a spacetime metric theory, the notion of time is
neither of these. Instead, it is a “time” constructed out of dynamical variables, such as the Hubble
time TH = a(t)/ȧ(t) coming from the scale factor a(t) of homogeneous isotropic models. This is
fundamentally different from the notion of time used in laboratory experiments, where the clock
is “external” to the system under observation, not interacting with the system, and monotonically
increasing on the scale of laboratory experiments.
∗
syed [email protected]
†
[email protected]
‡
[email protected]
These differences between local clocks external to the system used in labs and global internal clocks
such as Hubble time, pose interesting foundational questions about the origin of time, in particular
the compatibility of global versus local notions of time. Every laboratory system is a flat Minkowski
patch of an expanding Universe in which the time scale of experiments is a tiny fraction of the age
of the Universe as measured by Hubble time.

From the perspective of general relativity coupled to matter described as a phase space system,
there is no natural notion of time that might describe physical evolution. It is well-known that
this “problem of time” [1–4] arises ultimately from the fact that the Hamiltonian of the system is
a constraint condition on the phase space – it represents a fine balance between the matter and
geometry variables. The problem of time is “resolved” through a choice of time and other coordinate
gauges, or by introducing a mechanism of relational time where one phase variable is used as the
clock to measure evolution of other phase space variables. We refer to such a choice as “global
time”.

In contrast, the time variable that appears in classical and quantum physics is either Newtonian
absolute time or the time in a special relativistic inertial frame. It happens to coincide with the time
we measure despite there being no physical description of the clock itself in these theories. This
comes from a fixed background metric structure and coincides with the natural time in a Minkowski
patch in general relativity. We refer to this as “local time”. Related questions concerning what is
a physical clock and what is a cosmological one have been discussed in the philosophy of science
literature, e.g. [5–7].

In this paper we show that, starting from general relativity coupled to matter, there is a way
to recover the dynamics with respect to Minkowski Killing-time only for certain choices of global
clocks. We also show that a global choice of clock also provides a natural mechanism for “evolving”
coupling constants; this is unlike other ideas and mechanisms for evolving constants [8–12]. All our
results are obtained for general relativity coupled to dust and a scalar field; while our results are
classical, this is a well studied model at the quantum level [13, 14].

An interesting related question is whether a global choice of time, either fixed through matter or
geometric variables, can be consistent with Lorentz invariance. This question has been posed for
general relativity coupled to a scalar field where the latter is chosen as the clock [15]; this choice

2
gives a corresponding physical Hamiltonian that leads in a certain expansion to a theory that is
“Poincare invariant, even if the starting point is not,” as the author notes. We also address this
question, and show that the global time defined by dust respects Poincare invariance, without any
approximation.

In the next section (II) we review the canonical formulation of general relativity minimally coupled
to a pressureless dust and scalar field. In Sec. III we consider this model in the volume and dust
time gauges in the cosmological setting and compare the evolution of the scalar field with these
two clocks with the same evolution in Minkowski spacetime. In Sec. IV, we discuss whether such
a procedure is possible beyond cosmological models in the full theory. We summarize and discuss
our conclusions in the last section (V).

II. SCALAR FIELD AND DUST IN GENERAL RELATIVITY

We start with general relativity written in the Arnowitt-Deser-Misner (ADM) form minimally
coupled to a pressureless dust field T , and a scalar field ϕ with a potential V (ϕ) (we work in
c = ℏ = 8πG = 1 units),
Z
S = d3 xdt [π ab q̇ab + Pϕ ϕ̇ + PT Ṫ − N (HG + Hϕ + HD ) − N a (CaG + Caϕ + CaD )], (1)

where,
!
1 1 √
HG = √ π ab πab − π 2 + q(Λ − R),
q 2
Pϕ2 1 √ ab √
Hϕ = √ + qq ∂a ϕ∂b ϕ + qV (ϕ),
2 q 2
p
HD = PT 1 + q ab ∂a T ∂b T , (2)

CaG = Db πab ,

Caϕ = Pϕ ∂a ϕ,

CaD = −PT ∂a T, (3)

(qab , π ab ), (ϕ, Pϕ ) and (T, PT ) are the gravitational, scalar field and dust phase space variables
respectively, q is the determinant of the spatial metric qab , π is the trace of π ab , R is the 3-Ricci
scalar curvature, Λ is the cosmological constant, N is the lapse and N a is the shift.

3
We now reduce to a spatially flat (k = 0) Friedmann-Lemaitre-Robertson-Walker (FLRW) Universe
with the ansatz

Pa (t) ab
qab = a2 (t)eab , π ab = e , (4)
6a(t)

where a(t) is the scale factor, Pa is the momentum canonically conjugate to the scale factor, and
eab = diag(1, 1, 1) is the flat metric. The scalar and dust fields must therefore also be homogeneous:

ϕ = ϕ(t) , Pϕ = Pϕ (t), (5)

T = T (t), PT = PT (t). (6)

With this homogeneity ansatz, the diffeomorphism constraint vanishes identically and we are left
with
Z !
S = V0 dt Pa ȧ + Pϕ ϕ̇ + PT Ṫ − N H , (7)

after the spatial integration, with the Hamiltonian constraint

Pa2 Pϕ2
H=− + a3 Λ + 3 + a3 V (ϕ) + PT ≈ 0, (8)
24a 2a
Z
and V0 = d3 x.

The FLRW line element is invariant under the scaling

x → sx, a → a/s. (9)

This induces the following scale transformations of variables in the canonical action

(a, Pa , ϕ, Pϕ , T, PT ) → a/s, Pa /s2 , ϕ, Pϕ /s3 , T, PT /s3

(10)

The scalar potential V (ϕ) and the cosmological constant Λ are invariant under this scale transfor-
mation. This naturally suggests the following definitions of scale-invariant variables:

1/3 2/3
a = V0 a, Pa = V0 Pa , P ϕ = V 0 Pϕ , P T = V 0 PT (11)

(Effectively, this sets V0 = 1 in the action, with everything taking the same form as before but now
in terms of scale-invariant variables).

4
 III. LOCAL PHYSICS FROM GLOBAL TIME

A global time gauge fixing proceeds by choosing some combination of the phase space variables
as a clock, i.e. f (p, q) = t. Then the Hamiltonian constraint is solved (strongly) for the variable
canonically conjugate to this choice of time. This leads to a non-vanishing physical Hamiltonian
from which dynamics can be derived. We illustrate this procedure in the next subsection for two
choices of time.

A. Volume time

The first global clock we consider is made up of the gravitational configuration degree of freedom
t = f (a). For now, we leave this function arbitrary. This variable is canonically conjugate to
Pa /f ′ (a) (where the prime indicates a derivative with respect to a), since their canonical Poisson
n o
bracket is f (a), Pa /f ′ (a) = 1. The physical Hamiltonian is the negative of the variable canonically
conjugate to time, and is obtained after solving the Hamiltonian constraint and eliminating any
gravitational degrees of freedom through a = f −1 (t),
Pa
Hp = − . (12)
f ′ (a) t=f (a),H=0

The lapse function N is fixed by the requirement that the time gauge fixing must be dynamically
preserved,

1 = ṫ = {t, N H} = N {t, H} (13)

and the gravitational symplectic term becomes

ṫ
Pa ȧ = Pa = −Hp . (14)
f ′ (a)
These steps lead to the (time) gauge fixed action,
Z  
GF
S = dt Pϕ ϕ̇ + PT Ṫ − Hp . (15)

Solving for Hp from the Hamiltonian constraint gives

v " #
√ 2
u 2 √
a u P ϕ P T a2 √
Hp = 24 ′ tΛ + + V (ϕ) + 3 ≡ 24 ρΛ + ρM , (16)
f (a) 2a6 a f ′ (a)
t=f (a) t=f (a)

5
where in the last line, we have identified the usual matter and cosmological constant energy densities.
This completes the procedure of fixing a global time gauge and obtaining the corresponding physical
Hamiltonian. We note two key features of this Hamiltonian: it is a square root, and it is explicitly
time dependent (through substituting a = f −1 (t)).

We now come to the central assumption in this subsection that will allow us to make a connection
with local physics. In its current form (with the square root), the Hamiltonian does not look like
any known Hamiltonian of a theory describing local physics. To obtain a standard form – a sum of
kinetic and potential terms – we expand the square root assuming that

ρM < ρΛ . (17)

This gives,
! " #
√ Pϕ2
r
a2 6 a2 PT
Hp = 2 6Λ ′ + + V (ϕ) + 3 . (18)
f (a) Λ f ′ (a) 2a6 a

The first term in the parenthesis above, is purely a function of time, and will lead to a boundary
term in the action, which can be ignored; the rest takes the form,

Pϕ2
   
1  2  1
Hp = + V (ϕ) y (t) + PT , (19)
2 x2 (t) z(t)

where
r
2 Λ 4 ′
x (t) = a f (a)
6 !
r
6 1
y(t) = x(t)
Λ af ′ (a)
r
′ Λ
z(t) = af (a) . (20)
6

For concreteness, we fix the potential to be V (ϕ) = 12 m2 ϕ2 + λϕ4 .

For the purpose of comparing the dynamics of the scalar field with the standard one on flat space-
time, we seek a time-dependent canonical transformation. This is accomplished by defining

Pϕ PT
Peϕ = , ϕe = ϕx(t); PeT = , Te = T z(t), (21)
x(t) z(t)

6
which is generated by the function F2 (ϕ, Peϕ , t) = ϕPeϕ x(t) for the scalar field (and similarly for dust:
F2 (T, PeT , t) = T PeT z(t)). Under this transformation, the action becomes,
Z
˙ ˙
S= dt [Peϕ ϕe + PeT Te − H
ep] (22)

with,

Peϕ21 2 e2 e e4 e  ẋ   ż 
H
ep = + me ϕ + λϕ + PT + ϕePeϕ + TePeT (23)
2 2 x z

where,
"r # "r #3  
f ′′ (a) 1 f ′′ (a)
     
6 1 6 1 ẋ 2 ż
m
e =m , λ=λ
e , = ȧ + , = ȧ + ′
Λ af ′ (a) Λ a2 f ′ (a) x a 2f ′ (a) z a f (a)
(24)

We now specialize to a particular choice of time, the volume time, t = f (a) = a3 , which measures
time as the size (volume) of the universe. This gives

Peϕ21 2 e2 e e4 e ϕePeϕ TePeT

H
ep = + me (t)ϕ + λ(t)ϕ + PT + + , (25)
2 2 t t

where
r r !3
m 2 e λ 2
m(t)
e = , λ(t) = 4 . (26)
t 3Λ t 3Λ

This Hamiltonian has some interesting features: (i) the coupling constants (mass m, and λ) acquire
a time dependence; (ii) if the universe is expanding, t increases monotonically, which means that
these coupling constants are dynamically driven to zero; (iii) apart from the last two terms (which
also have a 1/t dependence), this Hamiltonian is just the usual Hamiltonian of a scalar field and
a dust field (albeit with time dependent coupling constants) that one would write down in flat
spacetime.

It is this last point that we wish to emphasize further here: starting from a global choice of time
(volume time), we have been able to get to a theory describing physics with respect to Minkowski
patch time. To see how closely this reduced theory matches local physics, we look at the dynamics
through the equations of motion (an over-dot indicates a derivative with respect to the global time

7
t),

˙ ϕe
ϕe = Peϕ + (27)
t
˙ e ϕe3 − Pϕ
e
Peϕ = −m e 2 (t)ϕe − 4λ(t) (28)
t
˙ Te
Te = 1 + (29)
t
˙ P
e T
PeT = − . (30)
t

e˙ p = ϕ Pe˙ϕ + T Pe˙T . We want to compare this dynamics to the

e e
The Hamiltonian also evolves in time: H
t t
dynamics of a theory of a scalar field and a dust field written down in a local region of spacetime
(laboratory) with respect to the (local) Minkowski patch time. It is clear from the above equations
that (apart from the time-dependence of the coupling constants m and λ) the last term in each of
them (with the 1/t dependence) is responsible for any deviations from what might be seen locally.
However, all of these terms, and even the coupling constants, contain a 1/t dependence or higher.
This means that as the universe expands, t (which is the volume of the universe) increases, and all
of these terms go to zero, making any deviations from local physics also go to zero.

To elucidate the role of this last term, let us look at the solutions to these equations. There is no
coupling between the scalar and the dust fields, therefore, we can solve them separately. The dust
equations have solutions,

Te = t(log(t) + c1 ), (31)
c2
PeT = , (32)
t

where c1 and c2 are some constants of integration. To connect this with local physics, the global
volume time t can be written as t = t0 + δt where t0 is the time now, and δt is the measure of time
in some small neighbourhood of t0 . The key point is that in this global volume time gauge, t0 is
very large (the universe is very large) and δt is very small (the size of the universe does not change
by much over the timescales of local experiments). For instance, if we plug in the current numerical
values, we find (in natural units), t0 ∼ 10184 . If we consider local experiments on timescales of
∼ 1000 years, we have δt ∼ 1054 , which gives δt/t0 ∼ 10−130 . This allows us to expand the above

8
solutions near t = t0 to obtain,
   2 !
δt δt
Te ≈ [(c1 + log(t0 ))t0 ] + [(c1 + log(t0 ) + 1)t0 ] +O (33)
t0 t0
′
≈ c1 + c̃1 δt, (34)
      2 !
c2 c2 δt δt
PeT ≈ − +O (35)
t0 t0 t0 t0
′
≈ c2 (36)

where in the last equation for PeT , we have ignored the linear term because it contains a factor of
1/t20 in the denominator. Such behaviour is not present in the linear term in Te, hence we must keep
that term. These results reproduce exactly the behaviour we expect from a free dust field in flat
spacetime.

The analysis for the scalar field proceeds in a similar fashion. If there is no potential (m = λ = 0),
then we can find exact analytical solutions and expand them near t = t0 to recover flat space
behavior. In the presence of a potential, we note that both the mass and the quartic coupling terms
become very small when the universe is large due to their time dependence (Eqn. 26). To look at
the exact behaviour, we numerically solved the scalar field equations of motion (Eqns. 27, 28), and
compared these solutions with the usual dynamics of a (local) scalar field in Minkowski spacetime
(ϕloc , Ploc ) with mass mloc and quartic coupling λloc . These new parameters (mloc and λloc ) were
chosen to match the values the time dependent coupling constants m(t) and λ(t) (in Eqn. 26) would
take now (at t = t0 ). Our results are shown in Figures (1 - 3).

Figure 1 shows the time evolution of ∆ϕ = ϕe − ϕloc (where ϕe is the global scalar field) on timescales
of local experiments. Similarly, Figure 2 shows the time evolution of ∆Pϕ = Peϕ − Ploc on these
timescales. As is clear from these figures, there is negligible difference between the dynamics of the
global scalar field and the local one during this period of time.

Figure 3 shows the differences between the two fields on a phase portrait with a wide range of
initial conditions. The phase portrait is plotted with the differences (∆ϕ and ∆Pϕ ) to show any
deviations. Again, there is almost no difference on these scales, and the results are robust with
respect to choices of different initial conditions.

Since the Hamiltonian obtained from this procedure (Eqn. 25) is explicitly time dependent, whereas
the Hamiltonian describing the dynamics of a local scalar field is not, we also computed the time

9
 FIG. 1. The time evolution of ∆ϕ = ϕe − ϕloc . Different colours represent different initial conditions.

FIG. 2. The time evolution of ∆Pϕ = Peϕ − Ploc . Different colours represent different initial conditions.

evolution of this Hamiltonian on relevant timescales. The result is shown in Figure 4 (different
colors represent different initial conditions) which shows that although there is time evolution in
the Hamiltonian, it is negligible on these timescales, ensuring that for all practical purposes, it does
behave as the usual Hamiltonian of a scalar field in flat spacetime.

Finally, we also test the validity of our initial assumption for the square root expansion (Eqn. 17),
and compute the evolution of ρM /ρΛ on relevant timescales for various initial conditions. Figure

10
 FIG. 3. Phase portrait of (∆ϕ, ∆Pϕ ). Different colours represent different initial conditions.

FIG. 4. Time evolution of the Hamiltonian. The evolution with different initial conditions is very close to
one another for any differences to be visible on this scale.

5 shows the results which indicate that it is indeed a good assumption for a large range of initial
conditions. Hence, we have shown that using the volume of the universe as global time, we can
recover local dynamics of a scalar field. In the next subsection, we show how this can also be
achieved with a special matter clock: dust as time.

11
FIG. 5. Time evolution of ρM /ρΛ . It should remain below 1 to justify our assumption of expanding the
square root. Different colours represent different initial conditions, but some are too close to one another to
be visible on this scale.

B. Dust time

We now choose a matter field (dust) as our clock: t = T [16]. The momentum conjugate to this is
PT , and therefore, the Physical Hamiltonian is,
Pa2 Pϕ2
Hp = − + a Λ + 3 + a3 V (ϕ),
3
(37)
24a 2a
with the (time) gauge-fixed action,
Z
S GF = dt[Pa ȧ + Pϕ ϕ̇ − Hp ]. (38)

Notice that this physical Hamiltonian is not a square-root as the Hamiltonian constraint (Eqn. 8)
is linear in the dust momentum. With this choice the scalar field (ϕ) and gravity (a) are dynamical.

To connect with local physics (where gravity is not dynamical), we consider a fixed background
limit of this theory (this is equivalent to doing quantum field theory on a fixed background but in
the dust time gauge). Since we are fixing the background, we can evaluate the first (symplectic)
term in the action (Pa ȧ) on this fixed background, and include it in our physical Hamiltonian. This
gives
Z
GF
Sfixed bg = dt[Pϕ ϕ̇ − H̄p ] (39)

12
with,

P2 Pa2 Pϕ2
 
H̄p = −Pa ȧ + Hp = a + Hp = + a Λ + 3 + a3 V (ϕ)
3
(40)
12a 24a 2a

where it is understood that a and Pa are to be evaluated on the fixed background, and the only
dynamical degree of freedom left is ϕ.

On this fixed background, the term in the brackets is a fixed function of time (since a(t), Pa (t) are
fixed functions of t); the resulting physical Hamiltonian is

Pϕ2
H̄p = + a3 V (ϕ) (41)
2a3

which is the standard Hamiltonian of a scalar field on an FLRW background, which can be used
to describe local (non-gravitational) physics. Now, since a(t) is arbitrary, it is possible to choose
a(t) =constant; this obviously gives the flat spacetime scalar field Hamiltonian. As a brief summary,
what this shows is that the global dust time gauge and the choice a(t) = constant gives the standard
scalar field Hamiltonian.

At this stage one may ask what exactly has been achieved? If we had started from general relativity
coupled to only a scalar field and set a(t)=constant and pa (t) = 0 in the canonical equations, solving
the remaining equations would yield ϕ = 0 and Pϕ = 0, i.e. the trivial solution. Thus, we see that in
the dust + scalar + gravity system, choosing dust time and a background gives scalar field theory
on that background. This holds for any background, not just FLRW.

What is unique about dust time is that the physical Hamiltonian is not a square root; it provides a
direct way to connect global time to local time; and produces a matter theory on a fixed background
if the gravitational variables are taken to be fixed functions.

C. Other clocks

Let us look at three other choices of time that can be made in this theory, scalar field time t =
√
ϕ, Hubble time t = 1/H, and York time t = 2π/3 q = Pa /3a2 (where π is the trace of the
ADM momentum and q is the determinant of the spatial metric). The corresponding physical

13
Hamiltonians are
r
a2 Pa2
Hpϕ = − 2a6 Λ − 2a6 V (t) − 2a3 PT , (42)
12 q
P2
PT2 + 2 t2ϕ [6 − t2 (Λ + V (ϕ))]
PT ∓
HpH = 2 , (43)
[6 − t2 (Λ + V (ϕ))]
q
PT ± PT2 + 2Pϕ2 38 t2 − (Λ + V (ϕ))
 
HpY = . (44)
2 83 t2 − (Λ + V (ϕ))
 

In each of these cases, it is not at all clear how to establish a correspondence with the standard
scalar field Hamiltonian. Thus we see that it is only with certain rather special choices of a global
time gauge, volume and dust, with further simplifying assumptions, that we are able to establish a
correspondence with the local Minkowski patch scalar field Hamiltonian.

IV. BEYOND COSMOLOGY

We now ask whether we can get to a theory describing local physics starting from full general
relativity coupled to matter fields without reducing to cosmology. In general, we have seen that
with any arbitrary choice of global clock, it is not possible to do this, even in the restricted setting
of cosmology. However, it seems possible with the dust time gauge, where the physical Hamiltonian
is not a square root. We start with the full theory (Eqns. 1, 2) and set the gauge T = t.

With this choice, the dust contribution to the diffeomorphism constraint vanishes, the dust Hamil-
tonian reduces to HD = PT , and the physical Hamiltonian (density) becomes
!
1 ab 1 2 √ Pϕ2 1 √ ab √
Hp = HG + Hϕ = √ π πab − π + q(Λ − R) + √ + qq ∂a ϕ∂b ϕ + qV (ϕ), (45)
q 2 2 q 2

with the action

Z
S GF = d3 xdt [π ab q̇ab + Pϕ ϕ̇ − Hp − N a (CaG + Caϕ )], (46)

where

CaG = Db πab , Caϕ = Pϕ ∂a ϕ. (47)

We note that Minkowski spacetime (qab = δab , π ab = 0 = N a , N = 1, Λ = 0), along with zero scalar
field (ϕ = 0 = Pϕ ) is a solution to the equations of motion of this theory. To make a connection

14
with local physics, we consider a linearization of this theory about the fixed Minkowski background.
The following discussion closely follows [17], where the authors consider a linearization of gravity
about Minkowski spacetime in the dust time gauge (the main difference here is the addition of a
scalar field). We cite their main results, show what differences occur with the addition of the scalar
field, and refer to [17] for technical details.

The linearization proceeds by writing,

qab = δab + hab (x, t), π ab = 0 + pab (x, t), N a = 0 + ξ a (x, t),

ϕ = 0 + φ(x, t), Pϕ = 0 + Pφ (x, t), (48)

where hab , pab , ξ a , φ, and Pφ are small perturbations, and we expand the Hamiltonian to second order
in these perturbations to obtain equations of motion at the linear order. Under this linearization,
the scalar field perturbations (φ, Pφ ) decouple from the gravitational perturbations, and the scalar
field Hamiltonian becomes,
Pφ2 1 ab
Hφ = + δ ∂a φ∂b φ + V (φ), (49)
2 2

which is the usual Hamiltonian for a scalar field on a flat background. The contribution of the
scalar field to the diffeomorphism constraint (Pφ ∂a φ) also vanishes at the linear order, and hence
only the gravitational contribution remains.

The gravitational sector of this theory gives rise to two graviton modes, and an ultralocal scalar
mode [17]. Hence we find that, after fixing dust as global time in the full theory, Minkowski
spacetime is a solution, and that perturbations about this fixed background Minkowski solution
give rise to two graviton modes, an ultralocal scalar mode, and the usual scalar field dynamics that
we would expect from local physics.

V. SUMMARY AND DISCUSSION

In the foregoing we posed and discussed the evolution of a scalar field with respect to “global time”
as defined by a canonical gauge fixing in gravity-matter system, and with respect to the unique
time variable provided by the Killing vector field of Minkowski spacetime. While the resulting
Hamiltonians are distinct at the structural level, we find these different times lead to very similar

15
evolution on the time scales of terrestrial experiments. Of particular note is the dust time choice,
where Poincare invariance of the scalar field remains unscathed despite the apparent intuition that
any global choice of time should break it.

We used the volume and dust time in cosmology for this comparison of scalar field dynamics. There
is of course no known general method available for extracting a unique time variable—if there were
one, we would have a solution to the problem of time, as for instance in the parametrized particle
toy model where the Hamiltonian constraint is linear in one momentum. Dust time is apparently
the closest one can come to this feature for gravity-matter systems. And dust time retains global
Poincare invariance despite making an explicit time choice even in the full theory (i.e. without
reduction to cosmology), a remarkable feature already noted in [17] where linearized gravity is
developed in the dust time gauge.

We also noted that “evolving coupling constants” arise naturally in the volume time gauge. The
basic “mechanism” is straightforward: matter potentials in the Hamiltonian have the universal
√
form qV (ϕ) = a3 V (ϕ) = tV (ϕ) in homogeneous cosmology in the volume time gauge. Hence all
coupling constants in V (ϕ) acquire time dependence. We described the details of this mechanism,
where a canonical transformation is necessary to bring the free part of the scalar Hamiltonian to its
familiar form. Our main result following such an analysis is that all coupling constants are driven to
zero with evolution in volume time. Thus, while one might be tempted to argue that cosmological
evolution makes all matter interactions “asymptotically free,” we will not do so here.

Lastly we note another feature of global time that is relevant for the so-called cosmological constant
(Λ)/vacuum energy problem. Its conventional formulation is in the framework of effective field
theory on Minkowski spacetime, where a cosmological constant is generated in perturbation theory;
(see e.g. [18, 19] for foundational discussion). From the perspective of a global time, the corre-
sponding physical matter-gravity Hamiltonian contains Λ as a coupling constant. In the volume
time gauge the physical Hamiltonian is not linear in Λ, whereas in the dust time gauge it is linear
[20, 21]. The vacuum energy on the other hand would be, by the usual definition, the ground state
energy E0 of such a Hamiltonian; this quantity that would be a function of all coupling constants gi ,
including Λ, i.e. E0 (Λ, g1 , g2 , · · · gn ). Thus, from the perspective of global time, there is no reason
to associate vacuum energy (so defined) with the cosmological constant. This observation provides

16
a point of contrast on the consequences of global time in a gravity-matter system on the one hand,
and Minkowski Killing time for matter fields on flat spacetime on the other.

Acknowledgments The work of VH was supported in part by a Discovery Grant from the Natural
Sciences and Engineering Research Council of Canada.

[1] C. J. Isham, Canonical quantum gravity and the problem of time, in 19th International Colloquium on
Group Theoretical Methods in Physics (GROUP 19) Salamanca, Spain, June 29-July 5, 1992 (1992)
arXiv:gr-qc/9210011 [gr-qc].
[2] K. V. Kuchar, Time and interpretations of quantum gravity, Int. J. Mod. Phys. D 20, 3 (2011).
[3] E. Anderson, The Problem of Time in Quantum Gravity, (2010), arXiv:1009.2157 [gr-qc].
[4] E. Anderson, The Problem of Time, Vol. 190 (Springer, 2017).
[5] P. W. Evans, G. J. Milburn, and S. Shrapnel, How clocks define physical time (2023).
[6] S. E. Rugh and H. Zinkernagel, On the physical basis of cosmic time, Stud. Hist. Phil. Sci. B 40, 1
(2009), arXiv:0805.1947 [gr-qc].
[7] S. E. Rugh and H. Zinkernagel, Limits of time in cosmology, in The Philosophy of Cosmology, edited
by S. Saunders, J. Silk, J. D. Barrow, and K. Chamcham (2016) pp. 377–395, arXiv:1603.05449 [gr-qc].
[8] P. A. M. Dirac, The Cosmological constants, Nature 139, 323 (1937).
[9] J. D. Bekenstein, Fine Structure Constant: Is It Really a Constant?, Phys. Rev. D25, 1527 (1982).
[10] H. B. Sandvik, J. D. Barrow, and J. Magueijo, A simple cosmology with a varying fine structure constant,
Phys. Rev. Lett. 88, 031302 (2002), arXiv:astro-ph/0107512 [astro-ph].
[11] J. D. Barrow, Varying constants, Phil. Trans. Roy. Soc. Lond. A363, 2139 (2005), arXiv:astro-
ph/0511440 [astro-ph].
[12] J.-P. Uzan, Varying constants, gravitation and cosmology, Living Reviews in Relativity 14, 2 (2011).
[13] V. Husain and S. Singh, Matter-Geometry entanglement in quantum cosmology, Class. Quant. Grav.
37, 15LT01 (2020), arXiv:1907.03776 [gr-qc].
[14] S. Gielen and L. Menéndez-Pidal, Unitarity, clock dependence and quantum recollapse in quantum
cosmology, Class. Quant. Grav. 39, 075011 (2022), arXiv:2109.02660 [gr-qc].
[15] L. Smolin, Time, measurement and information loss in quantum cosmology, in Directions in General
Relativity: An International Symposium in Honor of the 60th Birthdays of Dieter Brill and Charles
Misner (1993) arXiv:gr-qc/9301016.
[16] V. Husain and T. Pawlowski, Time and a physical Hamiltonian for quantum gravity, Phys. Rev. Lett.
108, 141301 (2012), arXiv:1108.1145 [gr-qc].

17
[17] M. Ali, V. Husain, S. Rahmati, and J. Ziprick, Linearized gravity with matter time, Class. Quant. Grav.
33, 105012 (2016), arXiv:1512.07854 [gr-qc].
[18] S. E. Rugh and H. Zinkernagel, The quantum vacuum and the cosmological constant problem (2000),
arXiv:hep-th/0012253 [hep-th].
[19] A. Koberinski, B. Falck, and C. Smeenk, Contemporary philosophical perspectives on the cosmological
constant (2023), arXiv:2212.04335 [physics.hist-ph].
[20] V. Husain and B. Qureshi, Ground state of the universe and the cosmological constant. A nonpertur-
bative analysis, Phys. Rev. Lett. 116, 061302 (2016), arXiv:1508.07664 [gr-qc].
[21] S. M. Hassan, A Non-perturbative Hamiltonian Approach to the Cosmological Constant Problem, Found.
Phys. 49, 391 (2019).

18