Stochastic Gravity: Theory and Applications

B. L. Hu
Maryland Center for Fundamental Physics,
Department of Physics, University of Maryland,
College Park, Maryland 20742-4111, U.S.A.
e-mail:[email protected]
arXiv:0802.0658v1 [gr-qc] 5 Feb 2008

E. Verdaguer
Departament de Fı́sica Fonamental
and Institut de Ciències del Cosmos,
Universitat de Barcelona,
Av. Diagonal 647, 08028 Barcelona, Spain
e-mail:[email protected]

(February 5, 2008 update of the 2003 review in Living Reviews in Relativity)

Abstract
Stochastic semiclassical gravity of the 90’s is a theory naturally evolved from semiclassical
gravity of the 80’s and quantum field theory in curved spacetimes of the 70’s. Whereas semiclas-
sical gravity is based on the semiclassical Einstein equation with sources given by the expectation
value of the stress-energy tensor of quantum fields, stochastic semiclassical gravity is based on
the Einstein-Langevin equation, which has in addition sources due to the noise kernel. The noise
kernel is the vacuum expectation value of the (operator-valued) stress-energy bi-tensor which
describes the fluctuations of quantum matter fields in curved spacetimes. A new criterion for the
validity of semiclassical gravity may also be formulated from the viewpoint of this theory. In the
first part, we describe the fundamentals of this new theory via two approaches: the axiomatic
and the functional. The axiomatic approach is useful to see the structure of the theory from the
framework of semiclassical gravity, showing the link from the mean value of the stress-energy
tensor to their correlation functions. The functional approach uses the Feynman-Vernon influ-
ence functional and the Schwinger-Keldysh closed-time-path effective action methods which are
convenient for computations. It also brings out the open systems concepts and the statistical
and stochastic contents of the theory such as dissipation, fluctuations, noise and decoherence.
We then focus on the properties of the stress energy bi-tensor. We obtain a general expression
for the noise kernel of a quantum field defined at two distinct points in an arbitrary curved
spacetime as products of covariant derivatives of the quantum field’s Green function, and show
from this that the trace anomaly of the noise kernel is zero for massless conformal scalar fields.
In the second part, we describe three applications of stochastic gravity theory. First, we con-
sider metric perturbations in a Minkowski spacetime. We offer an analytical solution of the
Einstein-Langevin equation and compute the two-point correlation functions for the linearized
Einstein tensor and for the metric perturbations. We also prove that Minkowski spacetime
is a stable solution of semiclassical gravity. Second, we discuss structure formation from the

1
 stochastic gravity viewpoint, which can go beyond the standard treatment by incorporating the
full quantum effect of the inflaton fluctuations. Third, we discuss the backreaction of Hawking
radiation in the gravitational background of a quasi-static black hole (enclosed in a box). We
derive a fluctuation-dissipation relation between the fluctuations in the radiation and the dissi-
pative dynamics of metric fluctuations. Finally we describe the behavior of metric fluctuations
near the event horizon of an evaporating black hole using the Einstein-Langevin equation and
point out directions for further development.

1 Overview
Stochastic semiclassical gravity is a theory developed in the Nineties using semiclassical gravity
(quantum fields in classical spacetimes, the dynamics of both matter and spacetime are solved self-
consistently) as the starting point and aiming at a theory of quantum gravity as the goal. While
semiclassical gravity is based on the semiclassical Einstein equation with the source given by the
expectation value of the stress-energy tensor of quantum fields, stochastic semiclassical gravity, or
stochastic gravity for short, includes also its fluctuations in a new stochastic semiclassical Einstein-
Langevin equation (we will often use the shortened term stochastic gravity as there is no confusion
as to the nature and source of stochasticity in gravity being induced from the quantum fields and
not a priori from the classical spacetime). If the centerpiece in semiclassical gravity theory is the
vacuum expectation value of the stress-energy tensor of a quantum field, and the central issues
being how well the vacuum is defined and how the divergences can be controlled by regularization
and renormalization, the centerpiece in stochastic semiclassical gravity theory is the stress-energy
bi-tensor and its expectation value known as the noise kernel. The mathematical properties of this
quantity and its physical content in relation to the behavior of fluctuations of quantum fields in
curved spacetimes are the central issues of this new theory. How they induce metric fluctuations
and seed the structures of the universe, how they affect the black hole horizons and the backreaction
of Hawking radiance in black hole dynamics, including implications on trans-Planckian physics, are
new horizons to explore. On the theoretical issues, stochastic gravity is the necessary foundation to
investigate the validity of semiclassical gravity and the viability of inflationary cosmology based on
the appearance and sustenance of a vacuum energy-dominated phase. It is also a useful beachhead
supported by well-established low energy (sub-Planckian) physics to explore the connection with
high energy (Planckian) physics in the realm of quantum gravity.
In this review we summarize major work and results on this theory since 1998. It is in the nature
of a progress report rather than a review. In fact we will have room only to discuss a handful of
topics of basic importance. A review of ideas leading to stochastic gravity and further developments
originating from it can be found in Refs. [1, 2]; a set of lectures which include a discussion of the
issue of the validity of semiclassical gravity in Ref. [3]; a pedagogical introduction of stochastic
gravity theory with a more detailed treatment of backreaction problems in cosmology and black
holes in quasi-equilibrium in Ref. [4]. A comprehensive formal description of the fundamentals is
given in Refs. [5, 6] while that of the noise kernel in arbitrary spacetimes in Refs. [6, 7, 8]. Here
we will try to mention related work so the reader can at least trace out the parallel and sequential
developments. The references at the end of each topic below are representative work where one can
seek out further treatments.
Stochastic gravity theory is built on three pillars: general relativity, quantum fields and nonequi-
librium statistical mechanics. The first two uphold semiclassical gravity, the last two span statistical

2
field theory. Strictly speaking one can understand a great deal without appealing to statistical me-
chanics, and we will try to do so here. But concepts such as quantum open systems [9, 10, 11]
and techniques such as the influence functional [12, 13] (which is related to the closed-time-path
effective action [14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25]) were a great help in our understanding
of the physical meaning of issues involved toward the construction of this new theory, foremost
because quantum fluctuations and correlation have ascended the stage and in some new paradigm
become the focus. Quantum statistical field theory and the statistical mechanics of quantum field
theory [26, 27, 28, 29] also aided us in searching for the connection with quantum gravity through
the retrieval of correlations and coherence. We show the scope of stochastic gravity as follows:

1 Ingredients:

(a) From General Relativity [30, 31] to Semiclassical Gravity.

(b) Quantum Field Theory in Curved Spacetimes [32, 33, 34, 35]:
i. Stress-energy tensor: Regularization and renormalization.
ii. Self-consistent solution: Backreaction problems in early universe and black holes
[36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48], and analog gravity [49, 50, 51, 52, 53].
iii. Effective action: Closed time path, initial value formulation [14, 15, 16, 17, 18, 21,
23, 24, 19, 25, 20].
iv. Equation of motion: Real and causal [22].
(c) Nonequilibrium Statistical Mechanics (see [54] and references therein) :
i. Open quantum systems [9, 10, 11].
ii. Influence Functional: Stochastic equations [12, 13].
iii. Noise and Decoherence: Quantum to classical transition [55, 56, 57, 58, 59, 60, 61,
62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84,
85, 86, 87, 88].
(d) Decoherence in Quantum Cosmology and Emergence of Classical Spacetimes [89, 90, 91,
92, 93, 94, 95].

2 Theory:

(a) Dissipation from Particle Creation [21, 22, 23, 24, 25, 96];
Backreaction as Fluctuation-Dissipation Relation (FDR) [97, 98, 99, 100].
(b) Noise from Fluctuations of Quantum Fields [1, 101, 102].
(c) Einstein-Langevin Equations [102, 103, 99, 104, 105, 106, 107, 5, 6, 108].
(d) Metric Fluctuations in Minkowski spacetime [109].

3 Issues:

(a) Validity of Semiclassical Gravity [110, 111, 24, 112, 113, 114, 115, 116, 117, 118, 119].
(b) Viability of Vacuum Dominance and Inflationary Cosmology.
(c) Stress-Energy Bitensor and Noise Kernel: Regularization Reassessed [7, 8].

4 Applications: Early Universe and Black Holes:

3
 (a) Wave Propagation in Stochastic Geometry [120].
(b) Black Hole Horizon Fluctuations: Spontaneous/Active versus Induced/Passive [121, 122,
123, 124, 125, 126, 127, 128, 8].
(c) Noise induced inflation [129].
(d) Structure Formation [130, 131, 132, 133, 134];
trace anomaly-driven inflation [135, 136, 137].
(e) Black Hole Backreaction and Fluctuations [97, 138, 139, 98, 140, 100, 141, 142, 143, 144].

5 Related Topics:

(a) Metric Fluctuations and Trans-Planckian Problem [125, 126, 127, 128, 145].
(b) Spacetime Foam, Loop and Spin Foam [146, 147, 148, 149, 150, 151, 152, 153, 154, 155].
(c) Universal ‘Metric Conductance’ Fluctuations [156].

6 Ideas:

(a) General Relativity as Geometro-Hydrodynamics [157, 158, 159, 160, 161, 162, 163];
Emergent Gravity [164, 165, 166, 167].
(b) Semiclassical Gravity as Mesoscopic Physics [168, 169].
(c) From Stochastic to Quantum Gravity:
i. Via Correlation hierarchy of interacting quantum fields [2, 27, 29, 170].
ii. Possible relation to string theory and matrix theory.
iii. Other major approaches to quantum gravity [171].

For lack of space we list only the latest work in the respective topics above describing ongoing
research. The reader should consult the references therein for earlier work and the background
material. We do not seek a complete coverage here, but will discuss only those selected topics in
theory, issues and applications. We use the (+, +, +) sign conventions of Refs. [30, 31], and units
in which c = h̄ = 1.

2 From Semiclassical to Stochastic Gravity

There are three main steps that lead to the recent development of stochastic gravity. The first step
begins with quantum field theory in curved spacetime [172, 32, 33, 34, 35], which describes the
behavior of quantum matter fields propagating in a specified (not dynamically determined by the
quantum matter field as source) background gravitational field. In this theory the gravitational field
is given by the classical spacetime metric determined from classical sources by the classical Einstein
equations, and the quantum fields propagate as test fields in such a spacetime. An important process
described by quantum field theory in curved spacetime is indeed particle creation from the vacuum,
and effects of vacuum fluctuations and polarizations, in the early universe [173, 174, 175, 176, 177,
178, 179, 180, 172, 181, 182], and Hawking radiation in black holes [183, 184, 185, 186, 187].
The second step in the description of the interaction of gravity with quantum fields is back-
reaction, i.e., the effect of the quantum fields on the spacetime geometry. The source here is the

4
expectation value of the stress-energy operator for the matter fields in some quantum state in the
spacetime, a classical observable. However, since this object is quadratic in the field operators,
which are only well defined as distributions on the spacetime, it involves ill defined quantities. It
contains ultraviolet divergences the removal of which requires a renormalization procedure [172, 188,
189]. The final expectation value of the stress-energy operator using a reasonable regularization
technique is essentially unique, modulo some terms which depend on the spacetime curvature and
which are independent of the quantum state. This uniqueness was proved by Wald [190, 191] who
investigated the criteria that a physically meaningful expectation value of the stress-energy tensor
ought to satisfy.
The theory obtained from a self-consistent solution of the geometry of the spacetime and the
quantum field is known as semiclassical gravity. Incorporating the backreaction of the quantum
matter field on the spacetime is thus the central task in semiclassical gravity. One assumes a general
class of spacetime where the quantum fields live in and act on, and seek a solution which satisfies
simultaneously the Einstein equation for the spacetime and the field equations for the quantum
fields. The Einstein equation which has the expectation value of the stress-energy operator of the
quantum matter field as the source is known as the semiclassical Einstein equation. Semiclassical
gravity was first investigated in cosmological backreaction problems [36, 37, 38, 39, 40, 41, 42, 43,
44, 45, 46], an example is the damping of anisotropy in Bianchi universes by the backreaction of
vacuum particle creation. Using the effect of quantum field processes such as particle creation to
explain why the universe is so isotropic at the present was investigated in the context of chaotic
cosmology [192, 193, 194] in the late seventies prior to the inflationary cosmology proposal of the
eighties [195, 196, 197, 198], which assumes the vacuum expectation value of an inflaton field as
the source, another, perhaps more well-known, example of semiclassical gravity.

2.1 The importance of quantum fluctuations

For a free quantum field semiclassical gravity is fairly well understood. The theory is in some
sense unique, since the only reasonable c-number stress-energy tensor that one may construct
[190, 191] with the stress-energy operator is a renormalized expectation value. However the scope
and limitations of the theory are not so well understood. It is expected that the semiclassical
theory would break down at the Planck scale. One can conceivably assume that it would also break
down when the fluctuations of the stress-energy operator are large [114, 115]. Calculations of the
fluctuations of the energy density for Minkowski, Casimir and hot flat spaces as well as Einstein and
de Sitter universes are available [115, 199, 116, 117, 7, 8, 6, 109, 200, 134, 201, 202]. It is less clear,
however, how to quantify what a large fluctuation is, and different criteria have been proposed
[115, 203, 204, 116, 117, 205, 113, 206, 207]. The issue of the validity of the semiclassical gravity
viewed in the light of quantum fluctuations was discussed in our Erice lectures [3]. More recently
in Refs. [118, 118] a new criterion has been proposed for the validity of semiclassical gravity. It is
based based on the quantum fluctuations of the semiclassical metric and incorporates in a unified
and self-consistent way previous criteria that have been used [110, 113, 114, 115]. One can see the
essence of the validity problem by the following example inspired by Ford [114].
Let us assume a quantum state formed by an isolated system which consists of a superposition
with equal amplitude of one configuration of mass M with the center of mass at X1 and another
configuration of the same mass with the center of mass at X2 . The semiclassical theory as described
by the semiclassical Einstein equation predicts that the center of mass of the gravitational field of

5
the system is centered at (X1 + X2 )/2. However, one would expect that if we send a succession
of test particles to probe the gravitational field of the above system half of the time they would
react to a gravitational field of mass M centered at X1 and half of the time to the field centered at
X2 . The two predictions are clearly different, note that the fluctuation in the position of the center
of masses is of the order of (X1 − X2 )2 . Although this example raises the issue of how to place
the importance of fluctuations to the mean, a word of caution should be added to the effect that
it should not be taken too literally. In fact, if the previous masses are macroscopic the quantum
system decoheres very quickly [58, 62] and instead of being described by a pure quantum state
it is described by a density matrix which diagonalizes in a certain pointer basis. For observables
associated to such a pointer basis the density matrix description is equivalent to that provided by
a statistical ensemble. The results will differ, in any case, from the semiclassical prediction.
In other words, one would expect that a stochastic source that describes the quantum fluc-
tuations should enter into the semiclassical equations. A significant step in this direction was
made in Ref. [1] where it was proposed to view the back-reaction problem in the framework
of an open quantum system: the quantum fields seen as the “environment” and the gravita-
tional field as the “system”. Following this proposal a systematic study of the connection be-
tween semiclassical gravity and open quantum systems resulted in the development of a new con-
ceptual and technical framework where (semiclassical) Einstein-Langevin equations were derived
[102, 103, 99, 104, 105, 106, 107]. The key technical factor to most of these results was the use of
the influence functional method of Feynman and Vernon [12] when only the coarse-grained effect of
the environment on the system is of interest. Note that the word semiclassical put in parentheses
refers to the fact that the noise source in the Einstein-Langevin equation arises from the quantum
field, while the background spacetime is classical; generally we will not carry this word since there
is no confusion that the source which contributes to the stochastic features of this theory comes
from quantum fields.
In the language of the consistent histories formulation of quantum mechanics [64, 65, 66, 67,
68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88] for the existence
of a semiclassical regime for the dynamics of the system one needs two requirements: The first is
decoherence, which guarantees that probabilities can be consistently assigned to histories describing
the evolution of the system, and the second is that these probabilities should peak near histories
which correspond to solutions of classical equations of motion. The effect of the environment is
crucial, on the one hand, to provide decoherence and, on the other hand, to produce both dissipation
and noise to the system through back-reaction, thus inducing a semiclassical stochastic dynamics
on the system. As shown by different authors [208, 55, 56, 57, 58, 59, 60, 61, 62, 63], indeed
over a long history predating the current revival of decoherence, stochastic semiclassical equations
are obtained in an open quantum system after a coarse graining of the environmental degrees of
freedom and a further coarse graining in the system variables. It is expected but has not yet been
shown that this mechanism could also work for decoherence and classicalization of the metric field.
Thus far, the analogy could be made formally [108] or under certain assumptions, such as adopting
the Born-Oppenheimer approximation in quantum cosmology [209, 210].
An alternative axiomatic approach to the Einstein-Langevin equation without invoking the
open system paradigm was later suggested based on the formulation of a self-consistent dynamical
equation for a perturbative extension of semiclassical gravity able to account for the lowest order
stress-energy fluctuations of matter fields [5]. It was shown that the same equation could be derived,
in this general case, from the influence functional of Feynman and Vernon [6]. The field equation

6
is deduced via an effective action which is computed assuming that the gravitational field is a c-
number. The important new element in the derivation of the Einstein-Langevin equation, and of the
stochastic gravity theory, is the physical observable that measures the stress-energy fluctuations,
namely, the expectation value of the symmetrized bi-tensor constructed with the stress-energy
tensor operator: the noise kernel. It is interesting to note that the Einstein-Langevin equation can
also be understood as a useful intermediary tool to compute symmetrized two-point correlations
of the quantum metric perturbations on the semiclassical background, independent of a suitable
classicalization mechanism [211].

3 The Einstein-Langevin equation: Axiomatic approach

In this section we introduce stochastic semiclassical gravity, or stochastic gravity for short, in an
axiomatic way. It is introduced as an extension of semiclassical gravity motivated by the search of
self-consistent equations which describe the back-reaction of the quantum stress-energy fluctuations
on the gravitational field [5].

3.1 Semiclassical gravity

Semiclassical gravity describes the interaction of a classical gravitational field with quantum matter
fields. This theory can be formally derived as the leading 1/N approximation of quantum gravity
interacting with N independent and identical free quantum fields [212, 213, 214, 215] which interact
with gravity only. By keeping the value of N G finite, where G is Newton’s gravitational constant,
one arrives at a theory in which formally the gravitational field can be treated as a c-number field
(i.e. quantized at tree level) while matter fields are fully quantized. The semiclassical theory may
be summarized as follows.
Let (M, gab ) be a globally hyperbolic four-dimensional spacetime manifold M with metric gab
and consider a real scalar quantum field φ of mass m propagating on that manifold; we just assume
a scalar field for simplicity. The classical action Sm for this matter field is given by the functional
1 √ h
Z   i
Sm [g, φ] = − d4 x −g g ab ∇a φ∇b φ + m2 + ξR φ2 , (3.1)
2
where ∇a is the covariant derivative associated to the metric gab , ξ is a coupling parameter between
the field and the scalar curvature of the underlying spacetime R, and g = detgab .
The field may be quantized in the manifold using the standard canonical quantization formalism
[32, 33, 34]. The field operator in the Heisenberg representation φ̂ is an operator-valued distribution
solution of the Klein-Gordon equation, the field equation derived from Eq. (3.1),
(✷ − m2 − ξR)φ̂ = 0. (3.2)
We may write the field operator as φ̂[g; x) to indicate that it is a functional of the metric gab and a
function of the spacetime point x. This notation will be used also for other operators and tensors.
The classical stress-energy tensor is obtained by functional derivation of this action in the usual
√
way T ab (x) = (2/ −g)δSm /δgab , leading to
1  
T ab [g, φ] = ∇a φ∇b φ − gab ∇c φ∇c φ + m2 φ2
 2 
+ξ g ✷ − ∇a ∇b + Gab φ2 ,
ab
(3.3)

7
where ✷ = ∇a ∇a and Gab is the Einstein tensor. With the notation T ab [g, φ] we explicitly indicate
that the stress-energy tensor is a functional of the metric gab and the field φ.
The next step is to define a stress-energy tensor operator T̂ ab [g; x). Naively one would replace
the classical field φ[g; x) in the above functional by the quantum operator φ̂[g; x), but this procedure
involves taking the product of two distributions at the same spacetime point. This is ill-defined
and we need a regularization procedure. There are several regularization methods which one may
use, one is the point-splitting or point-separation regularization method [188, 189] in which one
introduces a point y in a neighborhood of the point x and then uses as the regulator the vector
tangent at the point x of the geodesic joining x and y; this method is discussed for instance in
Refs. [117, 7, 8] and in section 5. Another well known method is dimensional regularization in
which one works in arbitrary n dimensions, where n is not necessarily an integer, and then uses
as the regulator the parameter ǫ = n − 4; this method is implicitly used in this section. The
regularized stress-energy operator using the Weyl ordering prescription, i.e. symmetrical ordering,
can be written as
1
T̂ ab [g] = {∇a φ̂[g] , ∇b φ̂[g]} + D ab [g] φ̂2 [g], (3.4)
2
where D ab [g] is the differential operator:
 
D ab ≡ (ξ − 1/4) g ab ✷ + ξ Rab − ∇a ∇b . (3.5)

Note that if dimensional regularization is used, the field operator φ̂[g; x) propagates in a n-
dimensional spacetime. Once the regularization prescription has been introduced a regularized
R [g; x) may be defined as
and renormalized stress-energy operator T̂ab
R
T̂ab C
[g; x) = T̂ab [g; x) + Fab [g; x)Iˆ, (3.6)

which differs from the regularized T̂ab [g; x) by the identity operator times some tensor counterterms
C [g; x), which depend on the regulator and are local functionals of the metric, see Ref. [6] for
Fab
details. The field states can be chosen in such a way that for any pair of physically acceptable
states, i.e., Hadamard states in the sense of Ref. [34], |ψi, and |ϕi the matrix element hψ|Tab R |ϕi,

defined as the limit when the regulator takes the physical value, is finite and satisfies Wald’s axioms
[33, 190]. These counterterms can be extracted from the singular part of a Schwinger-DeWitt series
[33, 188, 189, 216]. The choice of these counterterms is not unique but this ambiguity can be
absorbed into the renormalized coupling constants which appear in the equations of motion for the
gravitational field.
The semiclassical Einstein equation for the metric gab can then be written as
R
Gab [g] + Λgab − 2(αAab + βBab )[g] = 8πGhT̂ab [g]i, (3.7)
R [g]i is the expectation value of the operator T̂ R [g, x) after the regulator takes the phys-
where hT̂ab ab
ical value in some physically acceptable state of the field on (M, gab ). Note that both the stress
tensor and the quantum state are functionals of the metric, hence the notation. The parameters G,
Λ, α and β are the renormalized coupling constants, respectively, the gravitational constant, the
cosmological constant and two dimensionless coupling constants which are zero in the classical Ein-
stein equation. These constants must be understood as the result of “dressing” the bare constants
which appear in the classical action before renormalization. The values of these constants must be

8
determined by experiment. The left hand side of Eq. (3.7) may be derived from the gravitational
action
1 4 √ 1
Z  
abcd 2
Sg [g] = d x −g R − Λ + αCabcd C + βR , (3.8)
8πG 2
where Cabcd is the Weyl tensor. The tensors Aab and Bab come from the functional derivatives with
respect to the metric of the terms quadratic in the curvature in Eq. (3.8), they are explicitly given
by
1 δ √
Z
Aab = √ d4 −gCcdef C cdef
−g δgab
1 ab 2
= g Ccdef C cdef − 2Racde Rbcde + 4Rac Rcb − RRab
2 3
2 1
−2✷Rab + ∇a ∇b R + gab ✷R, (3.9)
3 3
1 δ √
Z
B ab = √ d4 −gR2
−g δgab
1 ab 2
= g R − 2RRab + 2∇a ∇b R − 2gab ✷R, (3.10)
2
where Rabcd and Rab are the Riemann and Ricci tensors, respectively. These two tensors are, like
the Einstein and metric tensors, symmetric and divergenceless: ∇a Aab = 0 = ∇a Bab .
A solution of semiclassical gravity consists of a spacetime (M, gab ), a quantum field operator
φ̂[g] which satisfies the evolution equation (3.2), and a physically acceptable state |ψ[g]i for this
field, such that Eq. (3.7) is satisfied when the expectation value of the renormalized stress-energy
operator is evaluated in this state.
For a free quantum field this theory is robust in the sense that it is self-consistent and fairly well
understood. As long as the gravitational field is assumed to be described by a classical metric, the
above semiclassical Einstein equations seems to be the only plausible dynamical equation for this
metric: the metric couples to matter fields via the stress-energy tensor and for a given quantum
state the only physically observable c-number stress-energy tensor that one can construct is the
above renormalized expectation value. However, lacking a full quantum gravity theory the scope
and limits of the theory are not so well understood. It is assumed that the semiclassical theory
should break down at Planck scales, which is when simple order of magnitude estimates suggest that
the quantum effects of gravity should not be ignored because the energy of a quantum fluctuation
in a Planck size region, as determined by the Heisenberg uncertainty principle, is comparable to
the gravitational energy of that fluctuation.
The theory is expected to break down when the fluctuations of the stress-energy operator are
large [114]. A criterion based on the ratio of the fluctuations to the mean was proposed by Kuo
and Ford [115] (see also work via zeta-function methods [199, 202]). This proposal was questioned
by Phillips and Hu [116, 117, 7] because it does not contain a scale at which the theory is probed
or how accurately the theory can be resolved. They suggested the use of a smearing scale or point-
separation distance, for integrating over the bi-tensor quantities, equivalent to a stipulation of the
resolution level of measurements; see also the response by Ford [203, 204]. A different criterion
is recently suggested by Anderson et al. [205, 113] based on linear response theory. A partial
summary of this issue can be found in our Erice Lectures [3].
More recently, in collaboration with A. Roura [118, 119], we have proposed a criterion for the
validity of semiclassical gravity which is based on the stability of the solutions of the semiclassical

9
Einstein equations with respect to quantum metric fluctuations. The two-point correlations for
the metric perturbations can be described in the framework of stochastic gravity, which is closely
related to the quantum theory of gravity interacting with N matter fields, to leading order in a
1/N expansion. We will describe these developments in the following sections.

3.2 Stochastic gravity

The purpose of stochastic gravity is to extend the semiclassical theory to account for these fluctua-
tions in a self-consistent way. A physical observable that describes these fluctuations to lowest order
is the noise kernel bi-tensor, which is defined through the two-point correlation of the stress-energy
operator as
1
Nabcd [g; x, y) = h{t̂ab [g; x), t̂cd [g; y)}i, (3.11)
2
where the curly brackets mean anticommutator, and where

t̂ab [g; x) ≡ T̂ab [g; x) − hT̂ab [g; x)i. (3.12)

This bi-tensor can also be written Nab,c′ d′ [g; x, y), or Nab,c′ d′ (x, y) as we do in section 5, to emphasize
that it is a tensor with respect to the first two indices at the point x and a tensor with respect to
the last two indices at the point y, but we shall not follow this notation here. The noise kernel
is defined in terms of the unrenormalized stress-tensor operator T̂ab [g; x) on a given background
metric gab , thus a regulator is implicitly assumed on the right-hand side of Eq. (3.11). However,
for a linear quantum field the above kernel – the expectation function of a bi-tensor – is free of
ultraviolet divergences because the regularized Tab [g; x) differs from the renormalized Tab R [g; x) by

the identity operator times some tensor counterterms, see Eq. (3.6), so that in the subtraction
(3.12) the counterterms cancel. Consequently the ultraviolet behavior of hT̂ab (x)T̂cd (y)i is the
same as that of hT̂ab (x)ihT̂cd (y)i, and T̂ab can be replaced by the renormalized operator T̂ab R in

Eq. (3.11); an alternative proof of this result is given in Ref. [7, 8]. The noise kernel should
be thought of as a distribution function, the limit of coincidence points has meaning only in the
sense of distributions. The bi-tensor Nabcd [g; x, y), or Nabcd (x, y) for short, is real and positive
semi-definite, as a consequence of T̂abR being self-adjoint. A simple proof is given in Ref. [4].

Once the fluctuations of the stress-energy operator have been characterized we can perturba-
tively extend the semiclassical theory to account for such fluctuations. Thus we will assume that
the background spacetime metric gab is a solution of the semiclassical Einstein Eqs. (3.7) and we
will write the new metric for the extended theory as gab + hab , where we will assume that hab is a
perturbation to the background solution. The renormalized stress-energy operator and the state of
the quantum field may now be denoted by T̂ab R [g + h] and |ψ[g + h]i, respectively, and hT̂ R [g + h]i
ab
will be the corresponding expectation value.
Let us now introduce a Gaussian stochastic tensor field ξab [g; x) defined by the following corre-
lators:
hξab [g; x)is = 0, hξab [g; x)ξcd [g; y)is = Nabcd [g; x, y), (3.13)
where h. . .is means statistical average. The symmetry and positive semi-definite property of the
noise kernel guarantees that the stochastic field tensor ξab [g, x), or ξab (x) for short, just introduced
is well defined. Note that this stochastic tensor captures only partially the quantum nature of the
fluctuations of the stress-energy operator since it assumes that cumulants of higher order are zero.

10
 An important property of this stochastic tensor is that it is covariantly conserved in the back-
ground spacetime ∇a ξab [g; x) = 0. In fact, as a consequence of the conservation of T̂ab R [g] one can see

that ∇x Nabcd (x, y) = 0. Taking the divergence in Eq. (3.13) one can then show that h∇a ξab is = 0
a

and h∇ax ξab (x)ξcd (y)is = 0 so that ∇a ξab is deterministic and represents with certainty the zero
vector field in M.
For a conformal field, i.e., a field whose classical action is conformally invariant, ξab is traceless:
ab
g ξab [g; x) = 0; so that, for a conformal matter field the stochastic source gives no correction to the
trace anomaly. In fact, from the trace anomaly result which states that g ab T̂ab R [g] is, in this case, a

local c-number functional of gab times the identity operator, we have that gab (x)Nabcd [g; x, y) = 0.
It then follows from Eq. (3.13) that hgab ξab is = 0 and hgab (x)ξab (x)ξcd (y)is = 0; an alternative
proof based on the point-separation method is given in Ref. [7, 8], see also section 5.
All these properties make it quite natural to incorporate into the Einstein equations the stress-
energy fluctuations by using the stochastic tensor ξab [g; x) as the source of the metric perturbations.
Thus we will write the following equation.
 
R
Gab [g+h]+Λ(gab +hab ) − 2(αAab + βBab )[g+h] = 8πG hT̂ab [g+h]i+ξab [g] . (3.14)

This equation is in the form of a (semiclassical) Einstein-Langevin equation, it is a dynamical

equation for the metric perturbation hab to linear order. It describes the back-reaction of the
metric to the quantum fluctuations of the stress-energy tensor of matter fields, and gives a first
order extension to semiclassical gravity as described by the semiclassical Einstein equation (3.7).
Note that we refer to the Einstein-Langevin equation as a first order extension to semiclassical
Einstein equation of semiclassical gravity and the lowest level representation of stochastic gravity.
However, stochastic gravity has a much broader meaning, it refers to the range of theories based on
second and higher order correlation functions. Noise can be defined in effectively open systems (e.g.
correlation noise [29] in the Schwinger-Dyson equation hierarchy) to some degree but one should
not expect the Langevin form to prevail. In this sense we say stochastic gravity is the intermediate
theory between semiclassical gravity (a mean field theory based on the expectation values of the
energy momentum tensor of quantum fields) and quantum gravity (the full hierarchy of correlation
functions retaining complete quantum coherence [2, 170].
The renormalization of the operator T̂ab [g + h] is carried out exactly as in the previous case,
now in the perturbed metric gab + hab . Note that the stochastic source ξab [g; x) is not dynamical,
it is independent of hab since it describes the fluctuations of the stress tensor on the semiclassical
background gab .
An important property of the Einstein-Langevin equation is that it is gauge invariant under the
change of hab by h′ab = hab + ∇a ζb + ∇b ζa , where ζ a is a stochastic vector field on the background
manifold M. Note that a tensor such as Rab [g +h], transforms as Rab [g +h′ ] = Rab [g +h]+Lζ Rab [g]
to linear order in the perturbations, where Lζ is the Lie derivative with respect to ζ a . Now, let
us write the source tensors in Eqs. (3.14) and (3.7) to the left-hand sides of these equations. If we
substitute h by h′ in this new version of Eq. (3.14), we get the same expression, with h instead of h′ ,
plus the Lie derivative of the combination of tensors which appear on the left-hand side of the new
Eq. (3.7). This last combination vanishes when Eq. (3.7) is satisfied, i.e., when the background
metric gab is a solution of semiclassical gravity.
From the statistical average of equation (3.14) we have that gab + hhab is must be a solution of
the semiclassical Einstein equation linearized around the background gab ; this solution has been

11
proposed as a test for the validity of the semiclassical approximation [205, 113] a point that will be
further discussed in section 3.3.
The stochastic equation (3.14) predicts that the gravitational field has stochastic fluctuations
over the background gab . This equation is linear in hab , thus its solutions can be written as follows,
Z q
hab (x) = h0ab (x) + 8πG d x 4 ′
−g(x′ )Gret ′ cd ′
abcd (x, x )ξ (x ), (3.15)

where h0ab (x) is the solution of the homogeneous equation containing information on the initial
conditions and Gret ′
abcd (x, x ) is the retarded propagator of equation (3.14) with vanishing initial
conditions. Form this we obtain the two-point correlation functions for the metric perturbations:

hhab (x)hcd (y)is = hh0ab (x)h0cd (y)is +

Z q
2
(8πG) d4 x′ d4 y ′ g(x′ )g(y ′ )Gret ′
abef (x, x )N
ef gh ′ ′
(x , y )Gret ′
cdgh (y, y )

≡ hhab (x)hcd (y)iint + hhab (x)hcd (y)iind . (3.16)

There are two different contributions to the two-point correlations, which we have distinguished
in the second equality. The first one is connected to the fluctuations of the initial state of the
metric perturbations, and we will refer to them as intrinsic fluctuations. The second contribution
is proportional to the noise kernel and is thus connected with the fluctuations of the quantum
fields, we will refer to them as induced fluctuations. To find these two-point stochastic correlation
functions one needs to know the noise kernel Nabcd (x, y). Explicit expressions of this kernel in terms
of the two-point Wightman functions is given in [6], expressions based on point-splitting methods
have also been given in [217, 7]. Note that the noise kernel should be thought of as a distribution
function, the limit of coincidence points has meaning only in the sense of distributions.
The two-point stochastic correlation functions for the metric perturbations of Eq. (3.16) satisfy
a very important property. In fact, it can be shown that they correspond exactly to the symmetrized
two-point correlation functions for the quantum metric perturbations in the large N expansion, i.e.
the quantum theory describing the interaction of the gravitational field with N arbitrary free fields
and expanded in powers of 1/N . To leading order for the graviton propagator one finds that

h{ĥab (x), ĥcd (y)}i = 2 hhab (x)hcd (y)is , (3.17)

where ĥab (x) mean the quantum operator corresponding to the metric perturbations and the statis-
tical average in Eq. (3.16) for the homogeneous solutions is now taken with respect to the Wigner
distribution that describes the initial quantum state of the metric perturbations. The Lorentz
gauge condition ∇a (hab − (1/2)ηab hcc ) = 0 as well as some initial condition to completely fix the
gauge of the initial state should be implicitly understood, moreover since there are now N scalar
fields the stochastic source has been rescaled so that the two-point correlation defined by Eq. (3.13)
should be 1/N times the noise kernel of a single field. This result was implicitly obtained in the
Minkowski background in Ref. [109] where the two-point correlation in the stochastic context was
computed for the linearized metric perturbations. This stochastic correlation exactly agrees with
the symmetrized part of the graviton propagator computed by Tomboulis [215] in the quantum con-
text of gravity interacting with N Fermion fields, where the graviton propagator is of order 1/N .
This result can be extended to an arbitrary background in the context of the large N expansion, a
sketch of the proof with explicit details in the Minkowski background can be found in Ref. [118].

12
This connection between the stochastic correlations and the quantum correlations was noted and
studied in detail in the context of simpler open quantum systems [218]. Stochastic gravity goes
beyond semiclassical gravity in the following sense. The semiclassical theory, which is based on the
expectation value of the stress energy tensor, carries information on the field two-point correlations
only, since hT̂ab i is quadratic in the field operator φ̂. The stochastic theory on the other hand, is
based on the noise kernel (3.11) which is quartic in the field operator. However, it does not carry
information on the graviton-graviton interaction, which in the context of the large N expansion
gives diagrams of order 1/N 2 . This will be illustrated in section 3.3.1. Furthermore the retarded
propagator gives also information on the commutator
 
h[ĥab (x), ĥcd (y)]i = 16πiG Gret ret
abcd (y, x) − Gabcd (x, y) , (3.18)

so that combining the commutator with the anticommutator the quantum two-point correlation
functions are determined. Moreover, assuming a Gaussian initial state with vanishing expectation
value for the metric perturbations any n-point quantum correlation function is determined by the
two-point quantum correlations and thus by the stochastic approach. Consequently, one may regard
the Einstein-Langevin equation as a useful intermediary tool to compute the correlation functions
for the quantum metric perturbations.
We should, however, emphasize also that Langevin like equations are obtained to describe the
quantum to classical transition in open quantum systems, when quantum decoherence takes place
by coarse graining of the environment as well as by suitable coarse graining of the system variables
[208, 219, 73, 74, 75, 220]. In those cases the stochastic correlation functions describe actual classical
correlations of the system variables. Examples can be found in the case of a moving charged particle
in an electromagnetic field in quantum electrodynamics [221] and in several quantum Brownian
models [218, 222, 223].

3.3 Validity of semiclassical gravity

As we have emphasized earlier the scope and limits of semiclassical gravity are not well understood
because we still lack a fully well understood quantum theory of gravity. From the semiclassical
Einstein equations it seems also clear that the semiclassical theory should break down when the
quantum fluctuations of the stress tensor are large. Ford [114] was among the first to have empha-
sized the importance of these quantum fluctuations. It is less clear, however, how to quantify the
size of these fluctuations. Kuo and Ford [115] used the variance of the fluctuations of the stress
tensor operator compared to the mean value as a measure of the validity of semiclassical gravity.
Hu and Phillips pointed out [116, 117] that such a criterion should be refined by considering the
back-reaction of those fluctuations on the metric. Ford and collaborators also noticed that the
metric fluctuations associated to the matter fluctuations can be meaningfully classified as active
[121, 206, 207] and passive [114, 115, 224, 204], which correspond to our intrinsic and induced
fluctuations, respectively, and have studied their properties in different contexts [225, 226, 227].
However, these fluctuations are not treated in a unified way and their precise relation to the
quantum correlation function for the metric perturbations is not discussed. Furthermore, the full
averaged back-reaction of the matter fields is not included self-consistently, and the contribution
from the vacuum fluctuations in Minkowski space is discarded.
A different approach to the validity of semiclassical gravity was pioneered by Horowitz [110, 111]
who studied the stability of a semiclassical solution with respect to linear metric perturbations. In

13
the case of a free quantum matter field in its Minkowski vacuum state, flat spacetime is a solution
of semiclassical gravity. The equations describing those metric perturbations involve higher order
derivatives, and Horowitz found unstable runaway solutions that grow exponentially with char-
acteristic timescales comparable to the Planck time; see also the analysis by Jordan [24]. Later,
Simon [112, 228], argued that those unstable solutions lie beyond the expected domain of validity of
the theory and emphasized that only those solutions which resulted from truncating perturbative
expansions in terms of the square of the Planck length are physically acceptable [112, 228]. Further
discussion was provided by Flanagan and Wald [229], who advocated the use of an order reduction
prescription first introduced by Parker and Simon [230]. More recently Anderson, Molina-Parı́s
and Mottola have taken up the issue of the validity of semiclassical gravity [205, 113] again. Their
starting point is the fact that the semiclassical Einstein equation will fail to provide a valid de-
scription of the dynamics of the mean spacetime geometry whenever the higher order radiative
corrections to the effective action, involving loops of gravitons or internal graviton propagators, be-
come important. Next, they argue qualitatively that such higher order radiative corrections cannot
be neglected if the metric fluctuations grow without bound. Finally, they propose a criterion to
characterize the growth of the metric fluctuations, and hence the validity of semiclassical gravity,
based on the stability of the solutions of the linearized semiclassical equation. Following these ap-
proaches the Minkowski metric is shown to be a stable solution of semiclassical gravity with respect
to small metric perturbations.
As emphasized in Refs. [205, 113] the above criteria may be understood as criteria based on
semiclassical gravity itself. It is certainly true that stability is a necessary condition for the validity
of a semiclassical solution, but one may also look for criteria within extensions of semiclassical
gravity. In the absence of a quantum theory of gravity such criteria may be found in some more
modest extensions. Thus, Ford [114] considered graviton production in linearized quantum grav-
ity and compared the results with the production of gravitational waves in semiclassical gravity.
Ashtekar [231] and Beetle [232] found large quantum gravity effects in three-dimensional quantum
gravity models. In a more recent paper [118] (see also Ref. [119]), we advocate for a criteria within
the stochastic gravity approach, and since stochastic gravity extends semiclassical gravity by in-
corporating the quantum stress tensor fluctuations of the matter fields, this criteria is structurally
the most complete to date.
It turns out that this validity criteria is equivalent to the validity criteria that one might
advocate within the large N expansion, that is the quantum theory describing the interaction of
the gravitational field with N identical free matter fields. In the leading order, namely the limit
in which N goes to infinity and the gravitational constant is appropriately rescaled, the theory
reproduces semiclassical gravity. Thus, a natural extension of semiclassical gravity is provided
by the next to leading order. It turns out that the symmetrized two-point quantum correlation
functions of the metric perturbations in the large N expansion are equivalent to the two-point
stochastic metric correlation functions predicted by stochastic gravity. Our validity criterion can
then be formulated as follows: a solution of semiclassical gravity is valid when it is stable with
respect to quantum metric perturbations. This criterion involves the consideration of quantum
correlation functions of the metric perturbations, since the quantum field describing the metric
perturbations ĥab (x) is characterized not only by its expectation value but also by its n-point
correlation functions.
It is important to emphasize that the above validity criterion incorporates in a unified and self-
consistent way the two main ingredients of the criteria exposed above. Namely, the criteria based on

14
the quantum stress tensor fluctuations of the matter fields, and the criteria based on the stability of
semiclassical solutions against classical metric perturbations. The former is incorporated through
the induced metric fluctuations, and the later through the intrinsic fluctuations introduced in Eq.
(3.16). Whereas information on the stability of the intrisic metric fluctuations can be obtained from
an analysis of the solutions of the perturbed semiclassical Einstein equation, the homogeneous part
of Eq. (3.14), the effect of the induced metric fluctuations is accounted only in stochastic gravity,
the full inhomogeneous Eq. (3.14). We will illustrate this criteria in section 6.5 by studying the
stability of Minkowski spacetime as a solution of semiclassical gravity.

3.3.1 The large N expansion

To illustrate the relation between the semiclassical, stochastic and quantum theories, a simplified
model of scalar gravity interacting with N scalar fields is considered here.
The large N expansion has been successfully used in quantum chromodynamics to compute
some non-perturbative results. This expansion re-sums and rearranges Feynman perturbative series
including self-energies. For gravity interacting with N matter fields it shows that graviton loops
are of higher order than matter loops. To illustrate the large N expansion let us, first, consider the
following toy model of gravity, which we will simplify as a scalar field h, interacting with a single
scalar field φ described by the action
1
Z
S = d4 x (∂a h∂ a h + h ∂a h∂ a h + . . .)
κZ Z
 
− d4 x ∂a φ∂ a φ + m2 φ2 + d4 x (h ∂a φ∂ a φ + . . .) , (3.19)

where κ = 8πG, and we have assumed that the interaction is linear in the (dimensionless) scalar
gravitational field h and quadratic in the matter field φ to simulate in a simplified way the coupling
of the metric with the stress tensor of the matter fields. We have also included a self coupling
graviton term of O(h3 ) which also appears in perturbative gravity beyond the linear approximation.
We may now compute the dressed graviton propagator perturbatively as the following series
of Feynman diagrams. The first diagram is just the free graviton propagator which is of O(κ), as
one can see from the kinetic term for the graviton in equation (3.19). The next diagram is one
loop of matter with two external legs which are the graviton propagators. This diagram has two
vertices with one graviton propagator and two matter field propagators. Since the vertices and
the matter propagators contribute with 1 and each graviton propagator contributes with a κ this
diagram is of order O(κ2 ). The next diagram contains two loops of matter and three gravitons,
and consequently it is of order O(κ3 ). There will also be terms with one graviton loop and two
graviton propagators as external legs, with three graviton propagators at the two vertices due to
the O(h3 ) term in the action (3.19). Since there are four graviton propagators which carry a κ4
but two vertices which have κ−2 this diagram is of order O(κ2 ), like the term with one matter loop.
Thus, in this perturbative expansion a graviton loop and a matter loop both contribute at the same
order to the dressed graviton propagator.
Let us now consider the large N expansion. We assume that the gravitational field is coupled
with a large number of identical fields φj , j = 1, . . . , N which couple only with h. Next we rescale
the gravitational coupling in such a way that κ̄ = κN is finite even when N goes to infinity. The

15
action of this system is:
N
Z
S = d4 x (∂a h∂ a h + h ∂a h∂ a h + . . .)
κ̄
N Z   N Z
4 a 2 2
d4 x (h ∂a φj ∂ a φj + . . .) .
X X
− d x ∂a φj ∂ φj + m φ + (3.20)
j j

Now an expansion in powers of 1/N of the dressed graviton propagator is given by the following
series of Feynman diagrams. The first diagram is the free graviton propagator which is now of order
O(κ̄/N ) the following diagrams are N identical Feynman diagrams with one loop of matter and
two graviton propagators as external legs, each diagram due to the two graviton propagators is of
order O(κ̄2 /N 2 ) but since there are N of them the sum can be represented by a single diagram with
a loop of matter of weight N , and therefore this diagram is of order O(κ̄2 /N ). This means that it
is of the same order as the first diagram in an expansion in 1/N . Then there are the diagrams with
two loops of matter and three graviton propagators, as before we can assign a weight of N to each
loop and taking into account the three graviton propagators this diagram is of order O(κ̄3 /N ), and
so on. This means that to order 1/N the dressed graviton propagator contains all the perturbative
series in powers of κ̄ of the matter loops.
Next, there is a diagram with one graviton loop and two graviton legs. Let us count the order
of this diagram: it contains four graviton propagators and two vertices, the propagators contribute
as (κ̄/N )4 and the vertices as (N/κ̄)2 , thus this diagram is of O(κ̄2 /N 2 ). Therefore graviton loop
contributes to higher order in the 1/N expansion than matter loops. Similarly there are N diagrams
with one loop of matter with an internal graviton propagator and two external graviton legs. Thus
we have three graviton propagators and since there are N of them, their sum is of order O(κ̄3 /N 2 ).
To summarize, we have that when N → ∞ there are no graviton propagators and gravity is classical
yet the matter fields are quantized, this is semiclassical gravity as was first described in Ref. [214].
Then we go to the next to leading order, 1/N , now the graviton propagator includes all matter
loop contributions, but no contributions from graviton loops or internal graviton propagators in
matter loops. This is what stochastic gravity reproduces.
That stochastic gravity is connected to the large N expansion can be seen from the stochastic
correlations of linear metric perturbations on the Minkowski background computed in Ref. [109].
These correlations are in exact agreement with the imaginary part of the graviton propagator
found by Tomboulis in the large N expansion for the quantum theory of gravity interacting with
N Fermion fields [215]. This has been proved in detail in Ref. [118], see also [119], where the case
of a general background is also briefly discussed.

4 The Einstein-Langevin equation: Functional approach

The Einstein-Langevin equation (3.14) may also be derived by a method based on functional tech-
niques [6]. Here we will summarize these techniques starting with semiclassical gravity.
In semiclassical gravity functional methods were used to study the back-reaction of quantum
fields in cosmological models [38, 41, 42]. The primary advantage of the effective action approach
is, in addition to the well-known fact that it is easy to introduce perturbation schemes like loop
expansion and nPI formalisms, that it yields a fully self-consistent solution. (For a general discussion
in the semiclassical context of these two approaches, equation of motion versus effective action,

16
contrast e.g., [36, 37, 39, 40] with the above references and [43, 44, 45, 46]. See also comments in Sec.
5.6 on the black hole backreaction problem comparing the approach by York et al. [233, 234, 235]
versus that of Sinha, Raval and Hu [142].
The well known in-out effective action method treated in textbooks, however, led to equations of
motion which were not real because they were tailored to compute transition elements of quantum
operators rather than expectation values. The correct technique to use for the backreaction problem
is the Schwinger-Keldysh [14, 15, 16, 17, 18, 19, 20] closed-time-path (CTP) or ‘in-in’ effective
action. These techniques were adapted to the gravitational context [21, 22, 23, 24, 25, 96] and
applied to different problems in cosmology. One could deduce the semiclassical Einstein equation
from the CTP effective action for the gravitational field (at tree level) with quantum matter fields.
Furthermore, in this case the CTP functional formalism turns out to be related [18, 102, 104,
236, 237, 100, 141, 238, 239, 6, 108] to the influence functional formalism of Feynman and Vernon
[12] since the full quantum system may be understood as consisting of a distinguished subsystem
(the “system” of interest) interacting with the remaining degrees of freedom (the environment). In-
tegrating out the environment variables in a CTP path integral yields the influence functional, from
which one can define an effective action for the dynamics of the system [102, 99, 240, 237]. This
approach to semiclassical gravity is motivated by the observation [1] that in some open quantum
systems classicalization and decoherence [55, 56, 57, 58, 59, 60, 61, 62, 63] on the system may be
brought about by interaction with an environment, the environment being in this case the matter
fields and some “high-momentum” gravitational modes [89, 90, 91, 92, 93, 94, 95, 220]. Unfortu-
nately, since the form of a complete quantum theory of gravity interacting with matter is unknown,
we do not know what these “high-momentum” gravitational modes are. Such a fundamental quan-
tum theory might not even be a field theory, in which case the metric and scalar fields would not
be fundamental objects [2]. Thus, in this case, we cannot attempt to evaluate the influence action
of Feynman and Vernon starting from the fundamental quantum theory and performing the path
integrations in the environment variables. Instead, we introduce the influence action for an effec-
tive quantum field theory of gravity and matter [241, 242, 243, 244, 245, 209, 210], in which such
“high-momentum” gravitational modes are assumed to have already been “integrated out.”

4.1 Influence action for semiclassical gravity

Let us formulate semiclassical gravity in this functional framework. Adopting the usual procedure
of effective field theories [246, 247, 241, 242, 243, 244, 248], one has to take the effective action for
the metric and the scalar field of the most general local form compatible with general covariance:
S[g, φ] ≡ Sg [g] + Sm [g, φ] + · · · , where Sg [g] and Sm [g, φ] are given by Eqs. (3.8) and (3.1),
respectively, and the dots stand for terms of order higher than two in the curvature and in the
number of derivatives of the scalar field. Here, we shall neglect the higher order terms as well
as self-interaction terms for the scalar field. The second order terms are necessary to renormalize
one-loop ultraviolet divergences of the scalar field stress-energy tensor, as we have already seen.
Since M is a globally hyperbolic manifold, we can foliate it by a family of t = constant Cauchy
hypersurfaces Σt , and we will indicate the initial and final times by ti and tf , respectively.
The influence functional corresponding to the action (3.1) describing a scalar field in a spacetime
(coupled to a metric field) may be introduced as a functional of two copies of the metric, denoted
+ −
by gab and gab , which coincide at some final time t = tf . Let us assume that, in the quantum
effective theory, the state of the full system (the scalar and the metric fields) in the Schrödinger

17
picture at the initial time t = ti can be described by a density operator which can be written as
the tensor product of two operators on the Hilbert spaces of the metric and of the scalar field. Let
ρi (ti ) ≡ ρi [φ+ (ti ), φ− (ti )] be the matrix element of the density operator ρ̂S (ti ) describing the initial
state of the scalar field. The Feynman-Vernon influence functional is defined as the following path
integral over the two copies of the scalar field:
Z
Dφ+ Dφ− ρi (ti )δ[φ+ (tf )−φ− (tf )] ei(Sm [g ).
+ ,φ − ,φ
+ ]−Sm [g −]
FIF [g ± ] ≡ (4.1)

Alternatively, the above double path integral can be rewritten as a closed time path (CTP) integral,
namely, as a single path integral in a complex time contour with two different time branches, one
going forward in time from ti to tf , and the other going backward in time from tf to ti (in practice
one usually takes ti → −∞). From this influence functional, the influence action SIF [g+ , g− ], or
SIF [g± ] for short, defined by
±
FIF [g± ] ≡ eiSIF [g ] , (4.2)
carries all the information about the environment (the matter fields) relevant to the system (the
gravitational field). Then we can define the CTP effective action for the gravitational field, Seff [g± ],
as
Seff [g± ] ≡ Sg [g+ ] − Sg [g − ] + SIF [g± ]. (4.3)
This is the effective action for the classical gravitational field in the CTP formalism. However,
since the gravitational field is treated only at the tree level, this is also the effective classical action
from which the classical equations of motion can be derived.
Expression (4.1) contains ultraviolet divergences and must be regularized. We shall assume
that dimensional regularization can be applied, that is, it makes sense to dimensionally continue all
the quantities that appear in Eq. (4.1). For this we need to work with the n-dimensional actions
corresponding to Sm in (4.1) and Sg in (3.8). For example, the parameters G, Λ α and β of Eq. (3.8)
are the bare parameters GB , ΛB , αB and βB , and in Sg [g], instead of the square of the Weyl tensor
in Eq. (3.8), one must use (2/3)(Rabcd Rabcd − Rab Rab ) which by the Gauss-Bonnet theorem leads
to the same equations of motion as the action (3.8) when n = 4. The form of Sg in n dimensions
is suggested by the Schwinger-DeWitt analysis of the ultraviolet divergences in the matter stress-
energy tensor using dimensional regularization. One can then write the Feynman-Vernon effective
action Seff [g± ] in Eq. (4.3) in a form suitable for dimensional regularization. Since both Sm and Sg
contain second order derivatives of the metric, one should also add some boundary terms [31, 99].
The effect of these terms is to cancel out the boundary terms which appear when taking variations
+ −
of Seff [g± ] keeping the value of gab and gab fixed at Σti and Σtf . Alternatively, in order to obtain
the equations of motion for the metric in the semiclassical regime, we can work with the action
terms without boundary terms and neglect all boundary terms when taking variations with respect
±
to gab . From now on, all the functional derivatives with respect to the metric will be understood
in this sense.
The semiclassical Einstein equation (3.7) can now be derived. Using the definition of the stress-
√
energy tensor T ab (x) = (2/ −g)δSm /δgab and the definition of the influence functional, Eqs. (4.1)
and (4.2), we see that 
ab 2 δSIF [g ± ] 
hT̂ [g; x)i = p +  , (4.4)
−g(x) δgab (x) g± =g

18
where the expectation value is taken in the n-dimensional spacetime generalization of the state
+
described by ρ̂S (ti ). Therefore, differentiating Seff [g ± ] in Eq. (4.3) with respect to gab , and then
+ −
setting gab = gab = gab , we get the semiclassical Einstein equation in n dimensions. This equation is
then renormalized by absorbing the divergences in the regularized hT̂ ab [g]i into the bare parameters.
Taking the limit n → 4 we obtain the physical semiclassical Einstein equation (3.7).

4.2 Influence action for stochastic gravity

In the spirit of the previous derivation of the Einstein-Langevin equation, we now seek a dynamical
equation for a linear perturbation hab to the semiclassical metric gab , solution of Eq. (3.7). Strictly
speaking if we use dimensional regularization we must consider the n-dimensional version of that
equation. From the results just described, if such an equation were simply a linearized semiclassical
Einstein equation, it could be obtained from an expansion of the effective action Seff [g + h± ]. In
particular, since, from Eq. (4.4), we have that

2 δSIF [g+h± ] 
hT̂ ab [g + h; x)i = p , (4.5)
− det(g+h)(x) δh+

ab (x)
 ±
h =h

the expansion of hT̂ ab [g+h]i to linear order in hab can be obtained from an expansion of the influence
action SIF [g + h± ] up to second order in h± ab .
To perform the expansion of the influence action, we have to compute the first and second order
functional derivatives of SIF [g + h± ] and then set h+ −
ab = hab = hab . If we do so using the path integral
representation (4.1), we can interpret these derivatives as expectation values of operators. The
relevant second order derivatives are

4 δ2 SIF [g + h± ] 
=−HSabcd [g; x, y)−K abcd [g; x, y)
−g(x) −g(y) δh+ +
p p 
ab(x)δh cd(y)  ±
h =h
+iN abcd [g; x, y),

4 δ2 SIF [g± ] 
=−HAabcd [g; x, y)−iN abcd [g; x, y), (4.6)
−g(x) −g(y) δh+(x)δh− (y) 
p p 
ab cd h± =h

where
1 D ab E
N abcd [g; x, y) ≡ {t̂ [g; x), t̂cd [g; y)} ,
2
D  E
H abcd [g; x, y) ≡ Im T∗ T̂ ab [g; x)Tˆcd [g; y) ,
S

i D ab E
HAabcd [g; x, y) ≡ − [T̂ [g; x), T̂ cd [g; y)] ,
2
*  +
abcd −4 δ2 Sm [g + h, φ] 
K [g; x, y) ≡ p p ,
δhab (x)δhcd (y) φ=φ̂

−g(x) −g(y)

with t̂ab defined in Eq. (3.12), [ , ] denotes the commutator and { , } the anti-commutator. Here we
use a Weyl ordering prescription for the operators. The symbol T∗ denotes the following ordered
operations: First, time order the field operators φ̂ and then apply the derivative operators which
appear in each term of the product T ab (x)T cd (y), where T ab is the functional (3.3). This T∗ “time

19
ordering” arises because we have path integrals containing products of derivatives of the field, which
can be expressed as derivatives of the path integrals which do not contain such derivatives. Notice,
from their definitions, that all the kernels which appear in expressions (4.6) are real and also HAabcd
is free of ultraviolet divergences in the limit n → 4.
∗ [g + , g− ], we can write the
From (4.4) and (4.6), since SIF [g, g] = 0 and SIF [g− , g+ ] = −SIF
±
expansion for the influence action SIF [g + h ] around a background metric gab in terms of the
previous kernels. Taking into account that these kernels satisfy the symmetry relations

HSabcd (x, y) = HScdab (y, x), HAabcd (x, y) = −HAcdab (y, x), K abcd (x, y) = K cdab (y, x), (4.7)

and introducing the new kernel

H abcd (x, y) ≡ HSabcd (x, y) + HAabcd (x, y), (4.8)

the expansion of SIF can be finally written as

1
Z q
SIF [g+h± ] = d4 x −g(x) hT̂ ab [g; x)i [hab (x)]
2
1
Z q q
− d4 x d4 y −g(x) −g(y)
8  
× [hab (x)] H abcd [g; x, y)+K abcd [g; x, y) {hcd (y)}
i
Z q q
+ d4 x d4 y −g(x) −g(y)
8
× [hab (x)] N abcd [g; x, y) [hcd (y)] + 0(h3 ), (4.9)

where we have used the notation

[hab ] ≡ h+ −
ab −hab , {hab } ≡ h+ −
ab +hab . (4.10)

From Eqs. (4.9) and (4.5) it is clear that the imaginary part of the influence action does not
contribute to the perturbed semiclassical Einstein equation (the expectation value of the stress-
energy tensor is real), however, as it depends on the noise kernel, it contains information on the
fluctuations of the operator T̂ ab [g].
We are now in a position to carry out the derivation of the semiclassical Einstein-Langevin
equation. The procedure is well known [102, 99, 104, 249, 250, 251, 252]: it consists of deriving a
new “stochastic” effective action from the observation that the effect of the imaginary part of the
influence action (4.9) on the corresponding influence functional is equivalent to the averaged effect
of the stochastic source ξ ab coupled linearly to the perturbations h±
ab . This observation follows from
the identity first invoked by Feynman and Vernon for such purpose:
1
 Z q q 
4 4 abcd
exp − d xd y −g(x) −g(y) [hab (x)] N (x, y) [hcd (y)]
8
i
Z  Z q 
4 ab
= Dξ P[ξ] exp d x −g(x) ξ (x) [hab (x)] , (4.11)
2

where P[ξ] is the probability distribution functional of a Gaussian stochastic tensor ξ ab characterized
by the correlators (3.13) with N abcd given by Eq. (3.11), and where the path integration measure

20
is assumed to be a scalar under diffeomorphisms of (M, gab ). The above identity follows from the
identification of the right-hand side of (4.11) with the characteristic functional for the stochastic
field ξ ab . The probability distribution functional for ξ ab is explicitly given by
1 4 4 q
 Z q 
P[ξ] = det(2πN )−1/2exp − d xd y −g(x) −g(y)ξ ab(x)Nabcd
−1
(x, y)ξ cd(y) . (4.12)
2
We may now introduce the stochastic effective action as
s
Seff [g + h± , ξ] ≡ Sg [g + h+ ] − Sg [g + h− ] + SIF
s
[g + h± , ξ], (4.13)

where the “stochastic” influence action is defined as

1
Z q
s
SIF [g + h± , ξ] ≡ Re SIF [g+h± ] + d4 x −g(x) ξ ab (x) [hab (x)] + O(h3 ). (4.14)
2

Note that, in fact, the influence functional can now be written as a statistical average over ξ ab :
FIF [g + h± ] = hexp (iSIF
s [g + h± , ξ])i . The stochastic equation of motion for h
s ab reads

s [g + h± , ξ] 
δSeff
= 0, (4.15)

δh+

ab (x)  ±
h =h

which is the Einstein-Langevin equation (3.14); notice that only the real part of SIF contributes
to the expectation value (4.5). To be precise we get first the regularized n-dimensional equations
with the bare parameters, and where instead of the tensor Aab we get (2/3)D ab where

1 δ √ 
Z 
D ab ≡ √ dn x −g Rcdef Rcdef − Rcd Rcd
−g δgab
1 ab  
= g Rcdef Rcdef − Rcd Rcd + ✷R − 2Racde Rb cde
2
−2Racbd Rcd + 4Rac Rc b − 3✷Rab + ∇a ∇b R. (4.16)

Of course, when n = 4 these tensors are related, Aab = (2/3)D ab . After that we renormalize and
take the limit n → 4 to obtain the Einstein-Langevin equations in the physical spacetime.

4.3 Explicit form of the Einstein-Langevin equation

We can write the Einstein-Langevin equation in a more explicit form by working out the expansion
of hT̂ ab [g+h]i up to linear order in the perturbation hab . From Eq. (4.5), we see that this expansion
can be easily obtained from (4.9). The result is

hT̂nab [g+h; x)i = hT̂nab [g, x)i + hT̂n(1) ab [g, h; x)i

1 n q
Z
− d y −g(y)Hnabcd [g; x, y)hcd (y) + 0(h2 ). (4.17)
2
Here we use a subscript n on a given tensor to indicate that we are explicitly working in n-
dimensions, as we use dimensional regularization, and we also use the superindex (1) to generally
indicate that the tensor is the first order correction, linear in hab , in a perturbative expansion
around the background gab .

21
 Using the Klein-Gordon equation (3.2), and expressions (3.3) for the stress-energy tensor and
the corresponding operator, we can write
1 ab
 
(1) ab
T̂n [g, h] = g hcd − δca hbd − δcb had T̂ncd [g] + F ab [g, h] φ̂2n [g], (4.18)
2
where F ab [g; h] is the differential operator
1 1
  
ab
F ≡ ξ− hab − gab hcc ✷
4 2
ξh c a b
+ ∇ ∇ hc + ∇c ∇b hac −✷hab −∇a ∇b hcc −gab ∇c ∇d hcd + gab ✷hcc
2  i
+ ∇a hbc +∇b hac −∇c hab−2gab ∇d hcd +gab ∇c hdd ∇c−g ab hcd ∇c ∇d . (4.19)

It is understood that indices are raised with the background inverse metric gab and that all the
covariant derivatives are associated to the metric gab .
Substituting (4.17) into the n-dimensional version of the Einstein-Langevin Eq. (3.14), taking
into account that gab satisfies the semiclassical Einstein equation (3.7), and substituting expression
(4.18) we can write the Einstein-Langevin equation in dimensional regularization as
" #
1 1 1

G(1) ab − gab Gcd hcd + Gac hbc + Gbc hac + ΛB hab − gab hcc
8πGB 2 2
4αB 1
 
− D (1)ab − gab D cd hcd + D ac hbc + D bc hac
3 2
1
 
−2βB B (1)ab − gab B cd hcd + B ac hbc + B bc hac
2
1
Z q
− µ−(n−4) Fxab hφ̂2n [g; x)i + dn y −g(y) µ−(n−4) Hnabcd [g; x, y)hcd (y)
2
−(n−4) ab
=µ ξn , (4.20)

where the tensors Gab , D ab and B ab are computed from the semiclassical metric gab , and where
we have omitted the functional dependence on gab and hab in G(1)ab , D (1)ab , B (1)ab and F ab to
simplify the notation. The parameter µ is a mass scale which relates the dimensions of the physical
field φ with the dimensions of the corresponding field in n-dimensions, φn = µ(n−4)/2 φ. Notice
that, in Eq. (4.20), all the ultraviolet divergences in the limit n → 4, which must be removed
by renormalization of the coupling constants, are in hφ̂2n (x)i and the symmetric part HSabcd n (x, y)
of the kernel Hnabcd (x, y), whereas the kernels Nnabcd (x, y) and HAabcd
n (x, y) are free of ultraviolet
divergences. If we introduce the bi-tensor Fnabcd [g; x, y) defined by
D E
Fnabcd [g; x, y) ≡ t̂ab ρσ
n [g; x) t̂n [g; y) (4.21)

where t̂ab is defined by Eq. (3.12), then the kernels N and HA can be written as

Nnabcd [g; x, y) = Re Fnabcd [g; x, y), HAabcd abcd

n [g; x, y) = Im Fn [g; x, y), (4.22)

where we have used that 2ht̂ab (x) t̂cd (y)i = h{t̂ab (x), t̂cd (y)}i+h[t̂ab (x), t̂cd (y)]i, and the fact that the
first term on the right hand side of this identity is real, whereas the second one is pure imaginary.

22
Once we perform the renormalization procedure in Eq. (4.20), setting n = 4 will yield the physical
Einstein-Langevin equation. Due to the presence of the kernel Hnabcd (x, y), this equation will be
usually non-local in the metric perturbation. In section 6 we will carry out an explicit evaluation
of the physical Einstein-Langevin equation which will illustrate the procedure.

4.3.1 The kernels for the vacuum state

When the expectation values in the Einstein-Langevin equation are taken in a vacuum state |0i,
such as, for instance, an “in” vacuum, we can be more explicit, since we can write the expectation
values in terms of the Wightman and Feynman functions, defined as
 
G+
n [g; x, y) ≡ h0|φ̂n [g; x)φ̂n [g; y)|0i, iGF n [g; x, y) ≡ h0|T φ̂n [g; x)φ̂n [g; y) |0i. (4.23)

These expressions for the kernels in the Einstein-Langevin equation will be very useful for explicit
calculations. To simplify the notation, we omit the functional dependence on the semiclassical
metric gab , which will be understood in all the expressions below.
From Eqs. (4.22), we see that the kernels Nnabcd (x, y) and HAabcd
n
(x, y) are the real and imaginary
parts, respectively, of the bi-tensor Fnabcd (x, y). From the expression
n (3.4) we see that o the stress-
ab
energy operator T̂n can be written as a sum of terms of the form Ax φ̂n (x), Bx φ̂n (x) , where Ax
and Bx are some differential operators. It then follows that we can express the bi-tensor Fnabcd (x, y)
in terms of the Wightman function as
Fnabcd (x, y) = ∇ax ∇cy G+ b d + a d + b c +
n (x, y)∇x ∇y Gn (x, y) + ∇x ∇y Gn (x, y)∇x ∇y Gn (x, y)
+ 2Dxab (∇cy G+ d +
n (x, y)∇y Gn (x, y))
+2Dycd (∇ax G+ b + ab cd +2
n (x, y)∇x Gn (x, y)) + 2Dx Dy (Gn (x, y)), (4.24)

where Dxab is the differential operator (3.5). From this expression and the relations (4.22), we get
expressions for the kernels Nn and HAn in terms of the Wightman function G+ n (x, y).
Similarly the kernel HSabcd
n
(x, y), can be written in terms of the Feynman function as
h
HSabcd a c b d
n (x, y) = −Im ∇x ∇y GF n(x, y)∇x ∇y GF n(x, y)

+∇ax ∇dy GF n(x, y)∇bx ∇cy GF n(x, y)

− gab (x)∇ex ∇cy GF n(x, y)∇ex ∇dy GF n(x, y)
−gcd (y)∇ax ∇ey GF n(x, y)∇bx ∇ey GF n(x, y)
1
+ gab (x)gcd (y)∇ex∇fy GF n(x, y)∇ex ∇fy GF n(x, y)
2
+Kxab (2∇cy GF n(x, y)∇dy GF n(x, y)
− gcd (y)∇ey GF n(x, y)∇ey GF n(x, y))
+Kycd (2∇ax GF n(x, y)∇bx GF n(x, y)
i
−gab (x)∇ex GF n(x, y)∇ex GF n(x, y))+2Kxab Kycd(GF2n(x, y)) , (4.25)

where Kxab is the differential operator

  1 2 ab
Kxab ≡ ξ gab (x)✷x − ∇ax ∇bx + Gab (x) − m g (x). (4.26)
2

23
Note that, in the vacuum state |0i, the term hφ̂2n (x)i in equation (4.20) can also be written as
hφ̂2n (x)i = iGF n(x, x) = G+n (x, x).
Finally, the causality of the Einstein-Langevin equation (4.20) can be explicitly seen as follows.
The non-local terms in that equation are due to the kernel H(x, y) which is defined in Eq. (4.8) as
the sum of HS (x, y) and HA (x, y). Now, when the points x and y are spacelike separated, φ̂n (x) and
φ̂n (y) commute and, thus, G+ n (x, y) = iGF n(x, y) = (1/2)h0| {φ̂n (x), φ̂n (y)} |0i, which is real. Hence,
from the above expressions, we have that HAabcd abcd
n (x, y) = HSn (x, y) = 0, and thus Hn
abcd (x, y) = 0.

This fact is expected since, from the causality of the expectation value of the stress-energy operator
[190], we know that the non-local dependence on the metric perturbation in the Einstein-Langevin
equation, see Eq. (3.14), must be causal. See Ref. [4] for an alternative proof of the causal nature
of the Einstein-Langevin equation.

5 Noise Kernel and Point-Separation

In this section we explore further the properties of the noise kernel and the stress energy bi-tensor.
Similar to what was done for the stress energy tensor it is desirable to relate the noise kernel defined
at separated points to the Green function of a quantum field. We pointed out earlier [2] that field
quantities defined at two separated points may possess important information which could be the
starting point for probes into possible extended structures of spacetime. Of more practical concern
is how one can define a finite quantity at one point or in some small region around it from the noise
kernel defined at two separated points. When we refer to, say, the fluctuations of energy density
in ordinary (point-wise) quantum field theory, we are in actuality asking such a question. This is
essential for addressing fundamental issues like
• the validity of semiclassical gravity [115] – e.g., whether the fluctuations to mean ratio is a
correct criterion [116, 117, 203, 204, 205, 113]

• whether the fluctuations in the vacuum energy density which drives some models of inflation-
ary cosmology violates the positive energy condition;

• physical effects of black hole horizon fluctuations and Hawking radiation backreaction – to
begin with, is the fluctuations finite or infinite?

• general relativity as a low energy effective theory in the geometro-hydrodynamic limit towards
a kinetic theory approach to quantum gravity [157, 2, 170].
Thus, for comparison with ordinary phenomena at low energy we need to find a reasonable
prescription for obtaining a finite quantity of the noise kernel in the limit of ordinary (point-defined)
quantum field theory. It is well-known that several regularization methods can work equally well
for the removal of ultraviolet divergences in the stress energy tensor of quantum fields in curved
spacetime. Their mutual relations are known, and discrepancies explained. This formal structure of
regularization schemes for quantum fields in curved spacetime should remain intact when applied to
the regularization of the noise kernel in general curved spacetimes; it is the meaning and relevance
of regularization of the noise kernel which is more of a concern (see comments below). Specific
considerations will of course enter for each method. But for the methods employed so far, such
as zeta-function, point separation, dimensional, smeared-field, applied to simple cases (Casimir,
Einstein, thermal fields) there is no new inconsistency or discrepancy.

24
 Regularization schemes used in obtaining a finite expression for the stress energy tensor have
been applied to the noise kernel. This includes the simple normal ordering [115, 253] and smeared
field operator [117] methods applied to the Minkowski and Casimir spaces, zeta-function [254, 255,
256] for spacetimes with an Euclidean section, applied to the Casimir effect [202] and the Einstein
Universe [199], or the covariant point-separation methods applied to the Minkowski [117], hot flat
space and the Schwarzschild spacetime [8]. There are differences and deliberations on whether it is
meaningful to seek a point-wise expression for the noise kernel, and if so what is the correct way to
proceed – e.g., regularization by a subtraction scheme or by integrating over a test-field. Intuitively
the smear field method [117] may better preserve the integrity of the noise kernel as it provides a
sampling of the two point function rather than using a subtraction scheme which alters its innate
properties by forcing a nonlocal quantity into a local one. More investigation is needed to clarify
these points, which bear on important issues like the validity of semiclassical gravity. We shall set
a more modest goal here, to derive a general expression for the noise kernel for quantum fields in
an arbitrary curved spacetime in terms of Green functions and leave the discussion of point-wise
limit to a later date. For this purpose the covariant point-separation method which highlights the
bi-tensor features, when used not as a regularization scheme, is perhaps closest to the spirit of
stochastic gravity.
The task of finding a general expression of the noise-kernel for quantum fields in curved space-
times was carried out by Phillips and Hu in two papers using the “modified” point separation
scheme [187, 257, 191]. Their first paper [7] begins with a discussion of the procedures for dealing
with the quantum stress tensor bi-operator at two separated points, and ends with a general expres-
sion of the noise kernel defined at separated points expressed as products of covariant derivatives
up to the fourth order of the quantum field’s Green function. (The stress tensor involves up to two
covariant derivatives.) This result holds for x 6= y without recourse to renormalization of the Green
function, showing that Nabc′ d′ (x, y) is always finite for x 6= y (and off the light cone for massless
theories). In particular, for a massless conformally coupled free scalar field on a four dimensional
manifold they computed the trace of the noise kernel at both points and found this double trace
vanishes identically. This implies that there is no stochastic correction to the trace anomaly for
massless conformal fields, in agreement with results arrived at in Refs. [102, 104, 6] (see also section
3). In their second paper [8] a Gaussian approximation for the Green function (which is what limits
the accuracy of the results) is used to derive finite expressions for two specific classes of spacetimes,
ultrastatic spacetimes, such as the hot flat space, and the conformally- ultrastatic spacetimes, such
as the Schwarzschild spacetime. Again, the validity of these results may depend on how we view
the relevance and meaning of regularization. We will only report the result of their first paper here.

5.1 Point Separation

The point separation scheme introduced in the 60’s by DeWitt [258] was brought to more popular
use in the 70’s in the context of quantum field theory in curved spacetimes [172, 188, 189] as a
means for obtaining a finite quantum stress tensor. Since the stress-energy tensor is built from
the product of a pair of field operators evaluated at a single point, it is not well-defined. In this
scheme, one introduces an artificial separation of the single point x to a pair of closely separated
points x and x′ . The problematic terms involving field products such as φ̂(x)2 becomes φ̂(x)φ̂(x′ ),
whose expectation value is well defined. If one is interested in the low energy behavior captured
by the point-defined quantum field theory – as the effort in the 70’s was directed – one takes the

25
coincidence limit. Once the divergences present are identified, they may be removed (regularization)
or moved (by renormalizing the coupling constants), to produce a well-defined, finite stress tensor
at a single point.
Thus the first order of business is the construction of the stress tensor and then derive the
symmetric stress-energy tensor two point function, the noise kernel, in terms of the Wightman
Green function. In this section we will use the traditional notation for index tensors in the point-
separation context.

5.1.1 n-tensors and end-point expansions

An object like the Green function G(x, y) is an example of a bi-scalar: it transforms as scalar
at both points x and y. We can also define a bi-tensor Ta1 ···an b′1 ···b′m (x, y): upon a coordinate
transformation, this transforms as a rank n tensor at x and a rank m tensor at y. We will extend
this up to a quad-tensor Ta1 ···an1 b′1 ···b′n c′′1 ···c′′n d′′′ ···d′′′
n4
which has support at four points x, y, x′ , y ′ ,
2 3 1
transforming as rank n1 , n2 , n3 , n4 tensors at each of the four points. This also sets the notation
we will use: unprimed indices referring to the tangent space constructed above x, single primed
indices to y, double primed to x′ and triple primed to y ′ . For each point, there is the covariant
derivative ∇a at that point. Covariant derivatives at different points commute and the covariant
derivative at, say, point x′ , does not act on a bi-tensor defined at, say, x and y:

Tab′ ;c;d′ = Tab′ ;d′ ;c and Tab′ ;c′′ = 0. (5.1)

To simplify notation, henceforth we will eliminate the semicolons after the first one for multiple
covariant derivatives at multiple points.
Having objects defined at different points, the coincident limit is defined as evaluation “on the
diagonal”, in the sense of the spacetime support of the function or tensor, and the usual shorthand
[G(x, y)] ≡ G(x, x) is used. This extends to n-tensors as
h i
Ta1 ···an1 b′1 ···b′n c′′ ′′ ′′′ ′′′
1 ···cn3 d1 ···dn4
= Ta1 ···an1 b1 ···bn2 c1 ···cn3 d1 ···dn4 , (5.2)
2

i.e., this becomes a rank (n1 + n2 + n3 + n4 ) tensor at x. The multi-variable chain rule relates
covariant derivatives acting at different points, when we are interested in the coincident limit:
h i h i h i
Ta1 ···am b′1 ···b′n ;c = Ta1 ···am b′1 ···b′n ;c + Ta1 ···am b′1 ···b′n ;c′ . (5.3)

This result is referred to as Synge’s theorem in this context; we follow Fulling’s [33] discussion.
′
The bi-tensor of parallel transport ga b is defined such that when it acts on a vector vb′ at y, it
parallel transports the vector along the geodesics connecting x and y. This allows us to add vectors
and tensors defined at different points. We cannot directly add a vector va at x and vector wa′ at
b′ a b′
h But
y.
′
i by using ga , we can construct the sum v + ga wb′ . We will also need the obvious property
ga b = ga b .
The main bi-scalar we need is the world function σ(x, y). This is defined as a half of the square
of the geodesic distance between the points x and y. It satisfies the equation
1
σ = σ ;p σ;p (5.4)
2

26
Often in the literature, a covariant derivative is implied when the world function appears with
′
indices: σ a ≡ σ ;a , i.e.taking the covariant derivative at x, while σ a means the covariant derivative
at y. This is done since the vector −σ a is the tangent vector to the geodesic with length equal
the distance between x and y. As σ a records information about distance and direction for the two
points this makes it ideal for constructing a series expansion of a bi-scalar. The end point expansion
of a bi-scalar S(x, y) is of the form
S(x, y) = A(0) + σ p A(1) p q (2) p q r (3) p q r s (4)
p + σ σ Apq + σ σ σ Apqr + σ σ σ σ Apqrs + · · · (5.5)
(n)
where, following our convention, the expansion tensors Aa1 ···an with unprimed indices have support
at x and hence the name end point expansion. Only the symmetric part of these tensors contribute
to the expansion. For the purposes of multiplying series expansions it is convenient to separate the
distance √
dependence from the direction dependence. This is done by introducing the unit vector
a a
p = σ / 2σ. Then the series expansion can be written
1 3
S(x, y) = A(0) + σ 2 A(1) + σA(2) + σ 2 A(3) + σ 2 A(4) + · · · (5.6)
(n)
The expansion scalars are related, via A(n) = 2n/2 Ap1 ···pn pp1 · · · ppn , to the expansion tensors.
The last
object we need is the VanVleck-Morette determinant D(x, y), defined as D(x, y) ≡
− det −σ;ab′ . The related bi-scalar
!1
2
1/2 D(x, y)
∆ = p (5.7)
g(x)g(y)
satisfies the equation
∆1/2 (4 − σ;p p ) − 2∆1/2 ;p σ ;p = 0 (5.8)
h i
with the boundary condition ∆1/2 = 1.
Further details on these objects and discussions of the definitions and properties are contained
in [188, 189] and [259]. There it is shown how the defining equations for σ and ∆1/2 are used to
determine the coincident limit expression for the various covariant derivatives of the world function
([σ;a ], [σ;ab ], etc.) and how the defining differential equation for ∆1/2 can be used to determine
(n)
the series expansion of ∆1/2 . We show how the expansion tensors Aa1 ···an are determined in terms
of the coincident limits of covariant derivatives of the bi-scalar S(x, y). (Ref. [259] details how
point separation can be implemented on the computer to provide easy access to a wider range of
applications involving higher derivatives of the curvature tensors.)

5.2 Stress Energy Bi-Tensor Operator and Noise Kernel

Even though we believe that the point-separated results are more basic in the sense that it reflects
a deeper structure of the quantum theory of spacetime, we will nevertheless start with quantities
defined at one point because they are what enter in conventional quantum field theory. We will use
point separation to introduce the bi-quantities. The key issue here is thus the distinction between
point-defined (pt) and point-separated (bi) quantities.
For a free classical scalar field φ with the action Sm [g, φ] defined in Eq. (3.1), the classical
stress-energy tensor is
1
 
Tab = (1 − 2 ξ) φ;a φ;b + 2 ξ − φ;p φ;p gab + 2ξ φ (φ;p p − φ;ab g ab )
2

27
 1 1
 
+φ2 ξ Rab − R gab − m2 φ2 g ab , (5.9)
2 2
which is equivalent to the tensor of Eq. (3.3) but written in a slightly different form for convenience.
When we make the transition to quantum field theory, we promote the field φ(x) to a field operator
φ̂(x). The fundamental problem of defining a quantum operator for the stress tensor is immediately
visible: the field operator appears quadratically. Since φ̂(x) is an operator-valued distribution,
products at a single point are not well-defined. But if the product is point separated (φ̂2 (x) →
φ̂(x)φ̂(x′ )), they are finite and well-defined.
Let us first seek a point-separated extension of these classical quantities and then consider
the quantum field operators. Point separation is symmetrically extended to products of covariant
derivatives of the field according to
1  p′ ′

(φ;a ) (φ;b ) → ga ∇p′ ∇b + gb p ∇a ∇p′ φ(x)φ(x′ ),
2
1 ′ ′

φ (φ;ab ) → ∇a ∇b + ga p gb q ∇p′ ∇q′ φ(x)φ(x′ ).
2
′
The bi-vector of parallel displacement ga a (x, x′ ) is included so that we may have objects that are
rank 2 tensors at x and scalars at x′ .
To carry out point separation on (5.9), we first define the differential operator
1 1
 ′  
′ ′

Tab = (1 − 2ξ) ga a ∇a′ ∇b + gb b ∇a ∇b′ + 2ξ − gab g cd ∇c ∇d′
2  2
′ ′ ′
  
−ξ ∇a ∇b + ga a gb b ∇a′ ∇b′ + ξgab ∇c ∇c + ∇c′ ∇c
1 1
 
+ξ Rab − gab R − m2 gab (5.10)
2 2
from which we obtain the classical stress tensor as

Tab (x) = lim

′
Tab φ(x)φ(x′ ). (5.11)
x →x

That the classical tensor field no longer appears as a product of scalar fields at a single point allows
a smooth transition to the quantum tensor field. From the viewpoint of the stress tensor, the
separation of points is an artificial construct so when promoting the classical field to a quantum one,
neither point should be favored. The product of field configurations is taken to be the symmetrized
operator product, denoted by curly brackets:
1n o 1 
φ(x)φ(y) → φ̂(x), φ̂(y) = φ̂(x)φ̂(y) + φ̂(y)φ̂(x) (5.12)
2 2
With this, the point separated stress energy tensor operator is defined as
1 n o
T̂ab (x, x′ ) ≡ Tab φ̂(x), φ̂(x′ ) . (5.13)
2
While the classical stress tensor was defined at the coincidence limit x′ → x, we cannot attach any
physical meaning to the quantum stress tensor at one point until the issue of regularization is dealt
with, which will happen in the next section. For now, we will maintain point separation so as to
have a mathematically meaningful operator.

28
 The expectation value of the point-separated stress tensor can now be taken. This amounts
to replacing the field operators by their expectation value, which is given by the Hadamard (or
Schwinger) function n o
G(1) (x, x′ ) = h φ̂(x), φ̂(x′ ) i. (5.14)
and the point-separated stress tensor is defined as
1
hT̂ab (x, x′ )i = Tab G(1) (x, x′ ) (5.15)
2
where, since Tab is a differential operator, it can be taken “outside” the expectation value. The
expectation value of the point-separated quantum stress tensor for a free, massless (m = 0) con-
formally coupled (ξ = 1/6) scalar field on a four dimension spacetime with scalar curvature R
is
1  p′ (1) ′
 1 ′
hT̂ab (x, x′ )i = g b G ;p′ a + gp a G(1) ;p′ b − gp q G(1) ;p′ q gab
6 12
1  p′ q′ (1) 
− g a g b G ;p′ q′ + G(1) ;ab
12
1  (1) p′  
+ G ;p′ + G(1) ;p p g ab
12
1 1
 
+ G(1) Rab − R gab (5.16)
12 2

5.2.1 Finiteness of Noise Kernel

We now turn our attention to the noise kernel introduced in Eq. (3.11), which is the symmetrized
product of the (mean subtracted) stress tensor operator:
n o
8Nab,c′ d′ (x, y) = h T̂ab (x) − hT̂ab (x)i, T̂c′ d′ (y) − hT̂c′ d′ (y)i i
n o
= h T̂ab (x), T̂c′ d′ (y) i − 2hT̂ab (x)ihT̂c′ d′ (y)i (5.17)

Since T̂ab (x) defined at one point can be ill-behaved as it is generally divergent, one can question
the soundness of these quantities. But as will be shown later, the noise kernel is finite for y 6= x. All
field operator products present in the first expectation value that could be divergent are canceled
by similar products in the second term. We will replace each of the stress tensor operators in the
above expression for the noise kernel by their point separated versions, effectively separating the two
points (x, y) into the four points (x, x′ , y, y ′ ). This will allow us to express the noise kernel in terms
of a pair of differential operators acting on a combination of four and two point functions. Wick’s
theorem will allow the four point functions to be re-expressed in terms of two point functions. From
this we see that all possible divergences for y 6= x will cancel. When the coincidence limit is taken
divergences do occur. The above procedure will allow us to isolate the divergences and obtain a
finite result.
Taking the point-separated quantities as more basic, one should replace each of the stress tensor
operators in the above with the corresponding point separated version (5.13), with Tab acting at x
and x′ and Tc′ d′ acting at y and y ′ . In this framework the noise kernel is defined as

8Nab,c′ d′ (x, y) = lim

′
lim
′
Tab Tc′ d′ G(x, x′ , y, y ′ ) (5.18)
x →x y →y

29
where the four point function is
1 h nn o n oo
G(x, x′ , y, y ′ ) = h φ̂(x), φ̂(x′ ) , φ̂(y), φ̂(y ′ ) i
4
n o n o i
−2 h φ̂(x), φ̂(x′ ) ih φ̂(y), φ̂(y ′ ) i . (5.19)

We assume the pairs (x, x′ ) and (y, y ′ ) are each within their respective Riemann normal coordinate
neighborhoods so as to avoid problems that possible geodesic caustics might be present. When we
later turn our attention to computing the limit y → x, after issues of regularization are addressed, we
will want to assume all four points are within the same Riemann normal coordinate neighborhood.
Wick’s theorem, for the case of free fields which we are considering, gives the simple product
four point function in terms of a sum of products of Wightman functions (we use the shorthand
notation Gxy ≡ G+ (x, y) = hφ̂(x) φ̂(y)i):

hφ̂(x) φ̂(y) φ̂(x′ ) φ̂(y ′ )i = Gxy′ Gyx′ + Gxx′ Gyy′ + Gxy Gx′ y′ (5.20)

Expanding out the anti-commutators in (5.19) and applying Wick’s theorem, the four point function
becomes
G(x, x′ , y, y ′ ) = Gxy′ Gx′ y + Gxy Gx′ y′ + Gyx′ Gy′ x + Gyx Gy′ x′ . (5.21)
We can now easily see that the noise kernel defined via this function is indeed well defined for the
limit (x′ , y ′ ) → (x, y):  
G(x, x, y, y) = 2 G2xy + G2yx . (5.22)
From this we can see that the noise kernel is also well defined for y 6= x; any divergence present
in the first expectation value of (5.19) have been cancelled by those present in the pair of Green
functions in the second term, in agreement with the results of section 3.

5.2.2 Explicit Form of the Noise Kernel

We will let the points separated for a while so we can keep track of which covariant derivative
acts on which arguments of which Wightman function. As an example (the complete calculation is
quite long), consider the result of the first set of covariant derivative operators in the differential
operator (5.10), from both Tab and Tc′ d′ , acting on G(x, x′ , y, y ′ ):
1  ′′ ′′

(1 − 2ξ)2 ga p ∇p′′ ∇b + gb p ∇p′′ ∇a
4  ′′′ ′′′

× gc′ q ∇q′′′ ∇d′ + gd′ q ∇q′′′ ∇c′ G(x, x′ , y, y ′ ). (5.23)

(Our notation is that ∇a acts at x, ∇c′ at y, ∇b′′ at x′ and ∇d′′′ at y ′ ). Expanding out the
differential operator above, we can determine which derivatives act on which Wightman function:

(1 − 2 ξ)2 h ′′′ ′′
× g c′ p g q a Gxy′ ;bp′′′ Gx′ y;q′′ d′ + Gxy ;bd′ Gx′ y′ ;q′′ p′′′
4

+ Gyx′ ;q′′ d′ Gy′ x;bp′′′ + Gyx;bd′ Gy′ x′ ;q′′ p′′′
p′′′ q ′′
+gd′ g a Gxy′ ;bp′′′ Gx′ y;q′′ c′ + Gxy;bc′ Gx′ y′ ;q′′ p′′′

+ Gyx′ ;q′′ c′ Gy′ x;bp′′′ + Gyx;bc′ Gy′ x′ ;q′′ p′′′

30
 ′′′ ′′
+gc′ p gq Gxy′ ;ap′′′ Gx′ y ;q′′ d′ + Gxy;ad′ Gx′ y′ ;q′′ p′′′

b
+ Gyx′ ;q′′ d′ Gy′ x;ap′′′ + Gyx;ad′ Gy′ x′ ;q′′ p′′′
p′′′ q ′′
+gd′ g b Gxy′ ;ap′′′ Gx′ y;q′′ c′ + Gxy;ac′ Gx′ y′ ;q′′ p′′′

+ Gyx′ ;q′′ c′ Gy′ x;ap′′′ +Gyx;ac′ Gy′ x′ ;q′′ p′′′ . (5.24)
1
If we now let x′ → x and y ′ → y the contribution to the noise kernel is (including the factor of 8
present in the definition of the noise kernel):
1n
(1 − 2 ξ)2 (Gxy ;ad′ Gxy ;bc′ + Gxy;ac′ Gxy ;bd′ )
8 o
+(1 − 2 ξ)2 (Gyx;ad′ Gyx;bc′ + Gyx;ac′ Gyx;bd′ ) . (5.25)

That this term can be written as the sum of a part involving Gxy and one involving Gyx is a general
property of the entire noise kernel. It thus takes the form

Nabc′ d′ (x, y) = Nabc′ d′ [G+ (x, y)] + Nabc′ d′ [G+ (y, x)] . (5.26)

We will present the form of the functional Nabc′ d′ [G] shortly. First we note, for x and y time-like
separated, the above split of the noise kernel allows us to express it in terms of the Feynman (time
ordered) Green function GF (x, y) and the Dyson (anti-time ordered) Green function GD (x, y):

Nabc′ d′ (x, y) = Nabc′ d′ [GF (x, y)] + Nabc′ d′ [GD (x, y)] . (5.27)

This can be connected with the zeta function approach to this problem [199] as follows: Recall
when the quantum stress tensor fluctuations determined in the Euclidean section is analytically
continued back to Lorentzian signature (τ → it), the time ordered product results. On the other
hand, if the continuation is τ → −it, the anti-time ordered product results. With this in mind, the
noise kernel is seen to be related to the quantum stress tensor fluctuations derived via the effective
action as  
2 2
16Nabc′ d′ = ∆Tabc + ∆T . (5.28)
 
′ d′  ′ ′
abc d 
′ ′ t=−iτ,t =−iτ ′ ′ t=iτ,t =iτ

The complete form of the functional Nabc′ d′ [G] is

′
Nabc′ d′ [G] = Ñabc′ d′ [G] + gab Ñc′ d′ [G] + gc′ d′ Ñab [G] + gab gc′ d′ Ñ [G] . (5.29)

with

8Ñabc′ d′ [G] = (1−2ξ)2 (G;c′ b G;d′ a + G;c′ a G;d′ b )

+4ξ 2 (G;c′ d′ G;ab + G G;abc′ d′ )
−2ξ(1−2ξ)(G;b G;c′ ad′ +G;a G;c′ bd′ +G;d′ G;abc′ +G;c′ G;abd′)
+2ξ(1−2ξ)(G;a G;b Rc′ d′ + G;c′ G;d′ Rab )
−4ξ 2 (G;ab Rc′ d′ + G;c′ d′ Rab ) G + 2 ξ 2 Rc′ d′ Rab G2 , (5.30)

1
  
′ ′ ′
 
′
8Ñab [G] = 2(1−2ξ) G;p′ b G; p a +ξ G;b G;p′ a p +G;a G;p′ b p
2ξ −
2
1
 
′ ′ ′

−4ξ 2ξ − G; p G;abp′ + ξ G;p′ p G;ab + G G;abp′ p
2

31
 −(m2
+ ξR′ )[(1 − 2 ξ) G;a G;b − 2 G ξ G;ab ]
1 ′ ′
+2ξ 2ξ − G;p′ G; p + 2 G ξ G;p′ p Rab
2
−(m2 + ξR′ ) ξ Rab G2 , (5.31)

1 2
 
′ ′ ′
 
8Ñ [G] = 2 2ξ − G;p′ q G; p q + 4ξ 2 G;p′ p G;q q + G G;p p q′ q
 2
1 

′ ′

+4ξ 2ξ − G;p G;q′ pq + G; p G;q q p′
2
1 h 2

′
   i
− 2ξ − m + ξR G;p′ G; p + m2 + ξR′ G;p G;p
h 2  ′
  i
−2ξ m2 + ξR G;p′ p + m2 + ξR′ G;p p G
1  
+ m2 + ξR m2 + ξR′ G2 . (5.32)
2

5.2.3 Trace of the Noise Kernel

One of the most interesting and surprising results to come out of the investigations undertaken in
the 1970’s of the quantum stress tensor was the discovery of the trace anomaly [260, 261]. When
the trace of the stress tensor T = g ab Tab is evaluated for a field configuration that satisties the
field equation (3.2) the trace is seen to vanish for massless conformally coupled fields. When this
analysis is carried over to the renormalized expectation value of the quantum stress tensor, the trace
no longer vanishes. Wald [191] showed this was due to the failure of the renormalized Hadamard
function Gren (x, x′ ) to be symmetric in x and x′ , implying it does not necessarily satisfy the field
equation (3.2) in the variable x′ . The definition of Gren (x, x′ ) in the context of point separation
will come next.)
With this in mind, we can now determine the noise associated with the trace. Taking the trace
at both points x and y of the noise kernel functional (5.27):
′ ′
N [G] = gab gc d 
Nabc′ d′ [G] 
1 1
  
′
= −3 G ξ m2 + ξR G;p′ p + m2 + ξR′ G;p p
2 2
9 ξ2 h ′ ′
i
+ G;p′ p G;p p + G G;p p p′ p
2
1 1
  
+ m2 + ξR m2 + ξR′ G2
2   2
1 1

′ ′ ′
 
+3 −ξ 3 −ξ G;p′ p G; p p −3ξ G;p G;p′ pp +G;p′ G;p pp
6  6
2 1 2 1
   
p′ ′ ;p
+ m + ξR G;p′ G; + m + ξR G;p G . (5.33)
2 2
For the massless conformal case, this reduces to
1 n o
RR′ G2 − 6G R✷′ + R′ ✷ G + 18 (✷G) ✷′ G + ✷′ ✷G ,



N [G] = (5.34)
144
which holds for any function G(x, y). For G being the Green function, it satisfies the field equation
(3.2):
✷G = (m2 + ξR)G. (5.35)

32
We will only assume the Green function satisfies the field equation in its first variable. Using
the fact ✷′ R = 0 (because the covariant derivatives act at a different point than at which R is
supported), it follows that
✷′ ✷G = (m2 + ξR)✷′ G. (5.36)
With these results, the noise kernel trace becomes
1 1
   
N [G] = m2 (1 − 3 ξ) + 3 R −ξ ξ
2 6
′ ′
h   i
× G2 2 m2 + R′ ξ + (1 − 6 ξ) G;p′ G; p − 6 G ξ G;p′ p
1 1
 h 
′

+ − ξ 3 2 m2 + R′ ξ G;p G;p − 18 ξ G;p G;p′ pp
2 6
1
  
p′ p
+18 − ξ G;p′ p G; , (5.37)
6
which vanishes for the massless conformal case. We have thus shown, based solely on the definition
of the point separated noise kernel, there is no noise associated with the trace anomaly. This
result obtained in Ref. [8] is completely general since it is assumed that the Green function is only
satisfying the field equations in its first variable; an alternative proof of this result was given in Ref.
[6]. This condition holds not just for the classical field case, but also for the regularized quantum
case, where one does not expect the Green function to satisfy the field equation in both variables.
One can see this result from the simple observation used in section 3: since the trace anomaly
is known to be locally determined and quantum state independent, whereas the noise present in
the quantum field is non-local, it is hard to find a noise associated with it. This general result
is in agreement with previous findings [102, 99, 104], derived from the Feynman-Vernon influence
functional formalism [12, 13] for some particular cases.

6 Metric fluctuations in Minkowski spacetime

Although the Minkowski vacuum is an eigenstate of the total four-momentum operator of a field in
Minkowski spacetime, it is not an eigenstate of the stress-energy operator. Hence, even for those
solutions of semiclassical gravity such as the Minkowski metric, for which the expectation value of
the stress-energy operator can always be chosen to be zero, the fluctuations of this operator are
non-vanishing. This fact leads to consider the stochastic metric perturbations induced by these
fluctuations.
Here we derive the Einstein-Langevin equation for the metric perturbations in a Minkowski
background. We solve this equation for the linearized Einstein tensor and compute the associated
two-point correlation functions, as well as, the two-point correlation functions for the metric pertur-
bations. Even though, in this case, we expect to have negligibly small values for these correlation
functions for points separated by lengths larger than the Planck length, there are several reasons
why it is worth carrying out this calculation.
On the one hand, these are the first back-reaction solutions of the full Einstein-Langevin equa-
tion. There are analogous solutions to a “reduced” version of this equation inspired in a “mini-
superspace” model [105, 106], and there is also a previous attempt to obtain a solution to the
Einstein-Langevin equation in Ref. [104], but, there, the non-local terms in the Einstein-Langevin
equation were neglected.

33
 On the other hand, the results of this calculation, which confirm our expectations that gravita-
tional fluctuations are negligible at length scales larger than the Planck length, but also predict that
the fluctuations are strongly suppressed on small scales, can be considered a first test of stochastic
semiclassical gravity. These results reveal also an important connection between stochastic gravity
and the large N expansion of quantum gravity. In addition, they are used in 6.5 to study the stabil-
ity of the Minkowski metric as a solution of semiclassical gravity, which constitutes an application
of the validity criterion introduced in section 3.3. This calculation requires also a discussion of the
problems posed by the so called runaway solutions, which arise in the back-reaction equations of
semiclassical and stochastic gravity, and some of the methods to deal with them. As a result we
conclude that Minkowski spacetime is a stable and valid solution of semiclassical gravity.
We advise the reader that his section is rather technical since it deals with an explicit non trivial
back-reaction computation in stochastic gravity. We tried to make it reasonable self-contained and
detailed, however a more detailed exposition can be found in Ref. [109].

6.1 Perturbations around Minkowski spacetime

The Minkowski metric ηab , in a manifold M which is topologically IR4 , and the usual Minkowski
vacuum, denoted as |0i, are the class of simplest solutions to the semiclassical Einstein equation
(3.7), the so called trivial solutions of semiclassical gravity [229]. They constitute the ground state
of semiclassical gravity. In fact, we can always choose a renormalization scheme in which the
renormalized expectation value h0| T̂Rab [η]|0i = 0. Thus, Minkowski spacetime (IR4 , ηab ) and the
vacuum state |0i are a solution to the semiclassical Einstein equation with renormalized cosmological
constant Λ = 0. The fact that the vacuum expectation value of the renormalized stress-energy
operator in Minkowski spacetime should vanish was originally proposed by Wald [190] and it may
be understood as a renormalization convention [33, 35]. Note that other possible solutions of
semiclassical gravity with zero vacuum expectation value of the stress-energy tensor are the exact
gravitational plane waves, since they are known to be vacuum solutions of Einstein equations which
induce neither particle creation nor vacuum polarization [262, 263, 264].
As we have already mentioned the vacuum |0i is an eigenstate of the total four-momentum
operator in Minkowski spacetime, but not an eigenstate of T̂ab R [η]. Hence, even in the Minkowski

background, there are quantum fluctuations in the stress-energy tensor and, as a result, the noise
kernel does not vanish. This fact leads to consider the stochastic corrections to this class of trivial
solutions of semiclassical gravity. Since, in this case, the Wightman and Feynman functions (4.23),
their values in the two-point coincidence limit, and the products of derivatives of two of such
functions appearing in expressions (4.24) and (4.25) are known in dimensional regularization, we
can compute the Einstein-Langevin equation using the methods outlined in sections 3 and 4.
To perform explicit calculations it is convenient to work in a global inertial coordinate system
µ
{x } and in the associated basis, in which the components of the flat metric are simply ηµν =
diag(−1, 1, . . . , 1). In Minkowski spacetime, the components of the classical stress-energy tensor
(3.3) reduce to
1 µν ρ 1
T µν [η, φ] = ∂ µ φ∂ ν φ − η ∂ φ∂ρ φ − η µν m2 φ2 + ξ (η µν ✷ − ∂ µ ∂ ν ) φ2 , (6.1)
2 2
where ✷ ≡ ∂µ ∂ µ , and the formal expression for the components of the corresponding “operator” in

34
dimensional regularization, see Eq. (3.4), is
1 µ
T̂nµν [η] = {∂ φ̂n , ∂ ν φ̂n } + D µν φ̂2n , (6.2)
2
where D µν is the differential operator (3.5), with gµν = ηµν , Rµν = 0, and ∇µ = ∂µ . The field
φ̂n (x) is the field operator in the Heisenberg representation in a n-dimensional Minkowski spacetime,
which satisfies the Klein-Gordon equation (3.2). We use here a stress-energy tensor which differs
from the canonical one, which corresponds to ξ = 0, both tensors, however, define the same total
momentum.
The Wightman and Feynman functions (4.23) when gµν = ηµν , are well known:

G+ +
n (x, y) = i∆n (x − y), GF n(x, y) = ∆F n(x − y), (6.3)

with
dn k ikx
Z
∆+
n (x) = −2πi e δ(k2 + m2 ) θ(k0 ),
(2π)n
dn k eikx
Z
∆F n(x) = − , ǫ → 0+ , (6.4)
(2π)n k2 + m2 − iǫ

where k2 ≡ ηµν kµ kν and kx ≡ ηµν kµ xν . Note that the derivatives of these functions satisfy
∂µx ∆+ + y + +
n (x − y) = ∂µ ∆n (x − y) and ∂µ ∆n (x − y) = −∂µ ∆n (x − y), and similarly for the Feynman
propagator ∆F n(x − y).
To write down the semiclassical Einstein equation (3.7) in n-dimensions for this case, we need
to compute the vacuum expectation value of the stress-energy operator components (6.2). Since,
from (6.3), we have that h0|φ̂2n (x)|0i = i∆F n(0) = i∆+
n (0), which is a constant (independent of x),
we have simply
!n/2 
dn k kµ kν η µν m2 n
Z 
h0|T̂nµν [η]|0i = −i n 2 2
= Γ − , (6.5)
(2π) k + m − iǫ 2 4π 2

where the integrals in dimensional regularization have been computed in the standard way (see Ref.
[109]) and where Γ(z) is the Euler’s gamma function. The semiclassical Einstein equation (3.7) in
n-dimensions before renormalization reduces now to
ΛB µν
η = µ−(n−4) h0|T̂nµν [η]|0i. (6.6)
8πGB
This equation, thus, simply sets the value of the bare coupling constant ΛB /GB . Note, from
(6.5), that in order to have h0| T̂Rµν |0i[η] = 0, the renormalized and regularized stress-energy tensor
“operator” for a scalar field in Minkowski spacetime, see Eq. (3.6), has to be defined as
n−4
!
η µν m4 m2 n
 
T̂Rµν [η] −(n−4) 2
=µ T̂nµν [η] − Γ − , (6.7)
2 (4π)2 4πµ2 2

which corresponds to a renormalization of the cosmological constant

ΛB Λ 2 m4
= − κn + O(n − 4), (6.8)
GB G π n(n−2)

35
where
n−4
! !
1 eγ m2 2 1 1 eγ m2
κn ≡ = + ln + O(n − 4), (6.9)
(n−4) 4πµ2 n−4 2 4πµ2
being γ the Euler’s constant. In the case of a massless scalar field, m2 = 0, one simply has
ΛB /GB = Λ/G. Introducing this renormalized coupling constant into Eq. (6.6), we can take the
limit n → 4. We find that, for (IR4 , ηab , |0i) to satisfy the semiclassical Einstein equation, we must
take Λ = 0.
We can now write down the Einstein-Langevin equations for the components hµν of the stochas-
tic metric perturbation in dimensional regularization. In our case, using h0|φ̂2n (x)|0i = i∆F n(0) and
the explicit expression of Eq. (4.20) we obtain
" #
1 1 4

G(1) µν + ΛB hµν − η µν h (x) − αB D (1) µν (x) − 2βB B (1) µν (x)
8πGB 2 3
1 n −(n−4) µναβ
Z
−ξG(1) µν(x)µ−(n−4)i∆F n(0)+ d yµ Hn (x, y)hαβ (y) = ξ µν(x). (6.10)
2
The indices in hµν are raised with the Minkowski metric and h ≡ hρρ , and here a superindex (1)
denotes the components of a tensor linearized around the flat metric. Note that in n-dimensions
the two-point correlation functions for the field ξ µν is written as

hξ µν (x)ξ αβ (y)is = µ−2 (n−4) Nnµναβ (x, y), (6.11)

Explicit expressions for D (1) µν and B (1) µν are given by

1 µναβ
D (1) µν (x) = F hαβ (x), B (1) µν (x) = 2Fxµν Fxαβ hαβ (x), (6.12)
2 x
µ(α β)ν
with the differential operators Fxµν ≡ η µν ✷x − ∂xµ ∂xν and Fxµναβ ≡ 3Fx Fx − Fxµν Fxαβ .

6.2 The kernels in the Minkowski background

Since the two kernels (4.22) are free of ultraviolet divergences in the limit n → 4, we can deal
directly with the F µναβ (x − y) ≡ limn→4 µ−2 (n−4) Fnµναβ in Eq. (4.21). The kernels N µναβ (x, y) =
Re F µναβ (x − y) and HAµναβ (x, y) = Im F µναβ (x − y) are actually the components of the “physical”
noise and dissipation kernels that will appear in the Einstein-Langevin equations once the renor-
malization procedure has been carried out. The bi-tensor F µναβ can be expressed in terms of the
Wightman function in four spacetime dimensions, according to (4.24). The different terms in this
kernel can be easily computed using the integrals

d4 k
Z
I(p) ≡ δ(k2 + m2 ) θ(−k0 ) δ[(k − p)2 + m2 ] θ(k0 − p0 ), (6.13)
(2π)4

and I µ1 ...µr (p) which are defined as the previous one by inserting the momenta kµ1 . . . kµr with
r = 1, 2, 3, 4 in the integrand. All these integral can be expressed in terms of I(p); see Ref. [109] for
the explicit expressions. It is convenient to separate I(p) in its even and odd parts with respect to
the variables pµ as
I(p) = IS (p) + IA (p), (6.14)

36
where IS (−p) = IS (p) and IA (−p) = −IA (p). These two functions are explicitly given by
s
1 m2
IS (p) = θ(−p2 − 4m2 ) 1+4 ,
8 (2π)3 p2
s
−1 m2
IA (p) = sign p0 θ(−p2 − 4m2 ) 1+4 . (6.15)
8 (2π)3 p2

After some manipulations, we find

!2
π 2 µναβ d4 p −ipx m2
Z
µναβ
F (x) = F e 1 + 4 I(p)
45 x (2π)4 p2
!2
8π 2 µν αβ d4 p −ipx m2
Z
+ Fx Fx e 3 ∆ξ + I(p), (6.16)
9 (2π)4 p2

where ∆ξ ≡ ξ − 1/6. The real and imaginary parts of the last expression, which yield the noise
and dissipation kernels, are easily recognized as the terms containing IS (p) and IA (p), respectively.
To write them explicitly, it is useful to introduce the new kernels
s !2
1 d4 p ipx m2 m2
Z
NA (x; m2 ) ≡ e θ(−p2 − 4m2 ) 1+4 2 1+4 2 ,
480π (2π)4 p p
1 d4 p ipx
Z
2
NB (x; m , ∆ξ) ≡ e θ(−p2 − 4m2 )
72π (2π)4
s !2
m2 m2
× 1+4 2 3 ∆ξ + 2 ,
p p
−i d4 p ipx
Z
DA (x; m2 ) ≡ e sign p0 θ(−p2 − 4m2 )
480π (2π)4
s !2
m2 m2
× 1+4 2 1+4 2 ,
p p
−i d4 p ipx
Z
DB (x; m2 , ∆ξ) ≡ e sign p0 θ(−p2 − 4m2 )
72π (2π)4
s !2
m2 m2
× 1+4 2 3 ∆ξ + 2 , (6.17)
p p

and we finally get

1
N µναβ (x, y) = Fxµναβ NA (x−y; m2 ) + Fxµν Fxαβ NB (x−y; m2 , ∆ξ),
6
1
HAµναβ (x, y) = Fxµναβ DA (x−y; m2 ) + Fxµν Fxαβ DB (x−y; m2 , ∆ξ). (6.18)
6
Notice that the noise and dissipation kernels defined in (6.17) are actually real because, for the
noise kernels, only the cos px terms of the exponentials eipx contribute to the integrals, and, for the
dissipation kernels, the only contribution of such exponentials comes from the i sin px terms.

37
 The evaluation of the kernel HSµναβ
n (x, y) is a more involved task. Since this kernel contains
divergences in the limit n → 4, we use dimensional regularization. Using Eq. (4.25), this kernel can
be written in terms of the Feynman propagator (6.4) as

µ−(n−4) HSµναβ
n (x, y) = Im K µναβ (x − y), (6.19)

where

K µναβ (x) ≡ −µ−(n−4) 2∂ µ ∂ (α ∆F n(x) ∂ β) ∂ ν ∆F n(x)
   
+2D µν ∂ α ∆F n(x)∂ β ∆F n(x) + 2D αβ ∂ µ ∆F n(x) ∂ ν ∆F n(x)
  
µν αβ
+2D D ∆F n(x) + η µν ∂ (α ∆F n(x)∂ β) +η αβ ∂ (µ ∆F n(x)∂ ν)
2

1
     
µν αβ αβ µν
+∆F n(0) η D +η D + η µνη αβ ∆F n(x)✷−m2∆F n(0) δn (x) . (6.20)
4
Let us define the integrals
dn k 1
Z
Jn (p) ≡ µ−(n−4) , (6.21)
(2π) (k + m − iǫ) [(k − p)2 + m2 − iǫ]
n 2 2

and Jnµ1 ...µr (p) obtained by inserting the momenta kµ1 . . . kµr into the previous integral, together
with
dn k 1
Z
I0n ≡ µ−(n−4) , (6.22)
(2π) (k + m2 − iǫ)
n 2

and I0µn1 ...µr which are also obtained by inserting momenta in the integrand. Then, the different
terms in Eq. (6.20) can be computed. These integrals are explicitly given in Ref. [109]. It is found
that I0µn = 0 and the remaining integrals can be written in terms of I0n and Jn (p). It is useful to
introduce the projector P µν orthogonal to pµ and the tensor P µναβ as

p2 P µν ≡ η µν p2 − pµ pν , P µναβ ≡ 3P µ(α P β)ν − P µν P αβ , (6.23)

then the action of the operator Fxµν is simply written as Fxµν dn p eipx f (p) = − dn p eipx f (p) p2 P µν
R R

where f (p) is an arbitrary function of pµ .

Finally after a rather long but straightforward calculation, and after expanding around n = 4,
we get,
(
i 1 µναβ n

µναβ
K (x) = κn F δ (x) + 4∆ξ 2 Fxµν Fxαβ δn (x)
(4π)2 90 x
2 m2
+ (η µν η αβ ✷x − η µ(α η β)ν ✷x + η µ(α ∂xβ) ∂xν + η ν(α ∂xβ) ∂xµ
3 (n−2)
4m4

−η µν ∂xα ∂xβ − η αβ ∂xµ ∂xν )δn (x) + (2η µ(α η β)ν − η µν η αβ ) δn (x)
n(n−2)
!2
1 µναβ dn p ipx m2
Z
+ F e 1 + 4 φ̄(p2 )
180 x (2π)n p2

38
 !2
2 dn p ipx m2
Z
+ Fxµν Fxαβ e 3∆ξ + φ̄(p2 )
9 (2π)n p2
4 1
 
− Fxµναβ + (60ξ −11) Fxµν Fxαβ δn (x)
675 270
)
2 1 µν αβ
 
2 µναβ
−m F + F F ∆n (x) + O(n − 4), (6.24)
135 x 27 x x

where κn has been defined in (6.9), and φ̄(p2 ) and ∆n (x) are given by
s
1 p2 m2
Z  
φ̄(p2 ) ≡ dα ln 1+ 2 α(1−α)−iǫ = −iπθ(−p2 −4m2 ) 1+4 2 +ϕ(p2), (6.25)
0 m p

dn p ipx 1
Z
∆n (x) ≡ e , (6.26)
(2π)n p2
2
where ϕ(p2 ) ≡ 01 dα ln |1 + m
p
R
2 α(1−α)|. The imaginary part of (6.24) gives the kernel components
−(n−4) µναβ
µ HSn (x, y), according to (6.19). It can be easily obtained multiplying this expression by
−i and retaining only the real part, ϕ(p2 ), of the function φ̄(p2 ).

6.3 The Einstein-Langevin equation

With the previous results for the kernels we can now write the n-dimensional Einstein-Langevin
equation (6.10), previous to the renormalization. Taking also into account Eqs. (6.5) and (6.6), we
may finally write:
1 4
G(1) µν (x) − αB D (1) µν (x) − 2βB B (1) µν (x)
8πGB 3
" #
κn m2 (1) µν 1 (1) µν 2 (1) µν
+ −4∆ξ G + D ∆ξ B (x)
(4π)2 (n−2) 90
(
1 16 1
 
+ − D (1) µν (x) + −10∆ξ B (1) µν (x)
2880π 2 15 6
" !2
dn p ip(x−y) m2
Z Z
n
+ d y e ϕ(p2 ) 1+4 2 D (1)µν (y)
(2π)n p
!2 #
m2 (1) µν
+10 3∆ξ + 2 B (y)
p
)
m2
Z  
− dn y ∆n (x−y) 8D (1) µν + 5B (1) µν (y)
3
1
Z
+ dn y µ−(n−4) HAµναβ
n (x, y) hαβ (y) + O(n−4) = ξ µν (x). (6.27)
2
Notice that the terms containing the bare cosmological constant have cancelled. These equations
can now be renormalized, that is, we can now write the bare coupling constants as renormalized
coupling constants plus some suitably chosen counterterms and take the limit n → 4. In order to

39
carry out such a procedure, it is convenient to distinguish between massive and massless scalar
fields. The details of the calculation can be found in Ref. [109].
It is convenient to introduce the two new kernels
( !2
1 d4 p ipx m2
Z
2
HA (x; m ) ≡ e 1+4 2
480π 2 (2π)4 p
" s # )
m2 8 m2
× −iπsignp0 θ(−p2 −4m2 ) 1 + 4 2 + ϕ(p2 ) − ,
p 3 p2
( !2
1 d4 p ipx m2
Z
HB (x; m2 , ∆ξ) ≡ e 3∆ξ + 2
72π 2 (2π)4 p
" s # )
m2 1 m2
0 2
× −iπsign p θ(−p −4m ) 1+4 2 +ϕ(p2 ) − , 2
(6.28)
p 6 p2

where ϕ(p2 ) is given by the restriction to n = 4 of expression (6.25). The renormalized coupling
constants 1/G, α and β are easily computed as it was done in Eq. (6.8). Substituting their
expressions into Eq. (6.27), we can take the limit n → 4, using the fact that, for n = 4, D (1) µν (x) =
(3/2) A(1) µν (x), we obtain the corresponding semiclassical Einstein-Langevin equation.
For the massless case one needs the limit m → 0 of equation (6.27). In this case it is convenient
to separate κn in (6.9) as κn = κ̃n + 12 ln(m2 /µ2 ) + O(n−4), where

1 eγ n−4 1 1 eγ
   
κ̃n ≡ 2
= + ln + O(n − 4), (6.29)
(n−4) 4π n−4 2 4π
and use that, from Eq. (6.25), we have
 
h  p2 i
2 2 2
lim ϕ(p ) + ln(m /µ ) = −2 + ln  2  . (6.30)
 
m2 →0 µ 

The coupling constants are then easily renormalized. We note that in the massless limit, the
Newtonian gravitational constant is not renormalized and, in the conformal coupling case, ∆ξ = 0,
we have that βB = β. Note also that, by making m = 0 in (6.17), the noise and dissipation kernels
can be written as

NA (x; m2 = 0) = N (x), NB (x; m2 = 0, ∆ξ) = 60∆ξ 2 N (x),

DA (x; m2 = 0) = D(x), DB (x; m2 = 0, ∆ξ) = 60∆ξ 2 D(x), (6.31)

where
1 d4 p ipx −i d4 p ipx
Z Z
N (x) ≡ 4
e θ(−p2 ), D(x) ≡ e sign p0 θ(−p2 ). (6.32)
480π (2π) 480π (2π)4
It is also convenient to introduce the new kernel
"   #
1 d4 p ipx  p2 
Z
2
H(x; µ ) ≡ e ln  2  − iπ sign p0 θ(−p2 )
 
480π 2 (2π)4 µ 
!
1 d4 p ipx −(p0 + iǫ)2 + pi pi
Z
= lim e ln . (6.33)
480π 2 ǫ→0+ (2π)4 µ2

40
This kernel is real and can be written as the sum of an even part and an odd part in the variables
xµ , where the odd part is the dissipation kernel D(x). The Fourier transforms (6.32) and (6.33)
can actually be computed and, thus, in this case we have explicit expressions for the kernels in
position space; see, for instance, Refs. [265, 266, 110].
Finally, the Einstein-Langevin equation for the physical stochastic perturbations hµν can be
written in both cases, for m 6= 0 and for m = 0, as
1
G(1) µν (x) − 2 ᾱA(1) µν (x) + β̄B (1) µν (x)


8πG Z
1
d4 y HA (x−y)A(1) µν (y) + HB (x−y)B (1) µν (y) = ξ µν (x),
 
+ (6.34)
4

where in terms of the renormalized constants α and β the new constants are ᾱ = α + (3600π 2 )−1
and β̄ = β − (1/12 − 5∆ξ)(2880π 2 )−1 . The kernels HA (x) and HB (x) are given by Eqs. (6.28)
when m 6= 0, and HA (x) = H(x; µ2 ), HB (x) = 60∆ξ 2 H(x; µ2 ) when m = 0. In the massless
case, we can use the arbitrariness of the mass scale µ to eliminate one of the parameters ᾱ or β̄.
The components of the Gaussian stochastic source ξ µν have zero mean value and their two-point
correlation functions are given by hξ µν (x)ξ αβ (y)is = N µναβ (x, y), where the noise kernel is given
in (6.18), which in the massless case reduces to (6.31).
It is interesting to consider the massless conformally coupled scalar field, i.e., the case ∆ξ = 0, of
particular interest because of its similarities with the electromagnetic field, and also because of its
interest in cosmology: massive fields become conformally invariant when their masses are negligible
compared to the spacetime curvature. We have already mentioned that for a conformally coupled,
field, the stochastic source tensor must be traceless (up to first order in perturbation theory around
semiclassical gravity), in the sense that the stochastic variable ξµµ ≡ ηµν ξ µν behaves deterministically
as a vanishing scalar field. This can be directly checked by noticing, from Eqs. (6.18) and (6.31),
that, when ∆ξ = 0, one has hξµµ (x)ξ αβ (y)is = 0, since Fµµ = 3✷ and F µα Fµβ = ✷F αβ . The Einstein-
Langevin equations for this particular case (and generalized to a spatially flat Robertson-Walker
background) were first obtained in Ref. [104], where the coupling constant β was fixed to be zero.
See also Ref. [4] for a discussion of this result and its connection to the problem of structure
formation in the trace anomaly driven inflation [135, 136, 137].
Note that the expectation value of the renormalized stress-energy tensor for a scalar field can be
obtained by comparing Eq. (6.34) with the Einstein-Langevin equation (3.14), its explicit expression
is given in Ref. [109]. The results agree with the general form found by Horowitz [110, 111] using an
axiomatic approach and coincides with that given in Ref. [229]. The particular cases of conformal
coupling, ∆ξ = 0, and minimal coupling, ∆ξ = −1/6, are also in agreement with the results for this
cases given in Refs. [110, 111, 267, 96, 24], modulo local terms proportional to A(1) µν and B (1) µν
due to different choices of the renormalization scheme. For the case of a massive minimally coupled
scalar field, ∆ξ = −1/6, our result is equivalent to that of Ref. [268].

6.4 Correlation functions for gravitational perturbations

Here we solve the Einstein-Langevin equations (6.34) for the components G(1) µν of the linearized
Einstein tensor. Then we use these solutions to compute the corresponding two-point correlation
functions, which give a measure of the gravitational fluctuations predicted by the stochastic semi-
classical theory of gravity in the present case. Since the linearized Einstein tensor is invariant

41
under gauge transformations of the metric perturbations, these two-point correlation functions are
also gauge invariant. Once we have computed the two-point correlation functions for the linearized
Einstein tensor, we find the solutions for the metric perturbations and compute the associated
two-point correlation functions. The procedure used to solve the Einstein-Langevin equation is
similar to the one used by Horowitz [110], see also Ref. [229], to analyze the stability of Minkowski
spacetime in semiclassical gravity.
We first note that the tensors A(1) µν and B (1) µν can be written in terms of G(1) µν as
2
A(1) µν = (F µν G(1) αα − Fαα G(1) µν ), B (1) µν = 2F µν G(1) αα , (6.35)
3
where we have used that 3✷ = Fαα . Therefore, the Einstein-Langevin equation (6.34) can be seen as
a linear integro-differential stochastic equation for the components G(1) µν . In order toR find solutions
to Eq. (6.34), it is convenient to Fourier transform. With the convention f˜(p) = d4 xe−ipx f (x)
for a given field f (x), one finds, from (6.35),
2
Ã(1)µν (p) = 2p2 G̃(1)µν (p)− p2P µν G̃(1) αα (p),
3
B̃ (1)µν (p) = −2p2P µνG̃(1) αα (p). (6.36)
The Fourier transform of the Einstein-Langevin Eq. (6.34) now reads
F µναβ (p) G̃(1) αβ (p) = 8πG ξ˜µν (p), (6.37)
where
µ ν
F µναβ (p) ≡ F1 (p) δ(α δβ) + F2 (p) p2 P µν ηαβ , (6.38)
with
1
 
2
F1 (p) ≡ 1 + 16πG p H̃A (p) − 2ᾱ ,
4
16 1 3
 
F2 (p) ≡ − πG H̃A (p) + H̃B (p) − 2ᾱ − 6β̄ . (6.39)
3 4 4
In the Fourier transformed Einstein-Langevin Eq. (6.37), ξ˜µν (p), the Fourier transform of ξ µν (x),
is a Gaussian stochastic source of zero average and
hξ̃ µν (p)ξ̃ αβ (p′ )is = (2π)4 δ4 (p + p′ ) Ñ µναβ (p), (6.40)
where we have introduced the Fourier transform of the noise kernel. The explicit expression for
Ñ µναβ (p) is found from (6.18) and (6.17) to be
s  !2
µναβ 1 m2 1 m2
Ñ (p) = θ(−p2 −4m2 ) 1+4 2  1+4 2 (p2 )2 P µναβ
720π p 4 p
!2 
m2 2 2 µν αβ 
+10 3∆ξ + 2 (p ) P P , (6.41)
p

which in the massless case reduces to

1 1 2 2 µναβ
 
lim Ñ µναβ (p) = θ(−p2 ) (p ) P + 60∆ξ 2 (p2 )2 P µν P αβ . (6.42)
m→0 480π 6

42
6.4.1 Correlation functions for the linearized Einstein tensor
µν µν
In general, we can write G(1) µν = hG(1) µν is + G(1)f , where G(1)
f is a solution to Eqs. (6.34)
with zero average, or (6.37) in the Fourier transformed version. The averages hG(1) µν is must be
a solution of the linearized semiclassical Einstein equations obtained by averaging Eqs. (6.34), or
(6.37). Solutions to these equations (specially in the massless case, m = 0) have been studied by
several authors [110, 111, 269, 270, 271, 272, 273, 214, 228, 24, 229], particularly in connection with
the problem of the stability of the ground state of semiclassical gravity. The two-point correlation
functions for the linearized Einstein tensor are defined by
G µναβ (x, x′ ) ≡ hG(1) µν (x)G(1) αβ (x′ )is − hG(1) µν (x)is hG(1) αβ (x′ )is
µν αβ ′
= hG(1)
f (x)G(1)
f (x )is . (6.43)
Now we shall seek the family of solutions to the Einstein-Langevin equation which can be
written as a linear functional of the stochastic source and whose Fourier transform, G̃(1) µν (p),
depends locally on ξ̃ αβ (p). Each of such solutions is a Gaussian stochastic field and, thus, it can be
completely characterized by the averages hG(1) µν is and the two-point correlation functions (6.43).
µν
For such a family of solutions, G̃(1)
f (p) is the most general solution to Eq. (6.37) which is linear,
αβ
homogeneous and local in ξ̃ (p). It can be written as
µν
G̃(1)
f (p) = 8πG D µναβ (p) ξ˜αβ (p), (6.44)
where D µναβ (p) are the components of a Lorentz invariant tensor field distribution in Minkowski
spacetime (by “Lorentz invariant” we mean invariant under the transformations of the orthochronous
Lorentz subgroup; see Ref. [110] for more details on the definition and properties of these tensor
distributions), symmetric under the interchanges α ↔ β and µ ↔ ν, which is the most general
solution of
µ ν
F µνρσ (p) D ρσαβ (p) = δ(α δβ) . (6.45)
µν
In addition, we must impose the conservation condition: pν G̃(1) f (p) = 0, where this zero must be
understood as a stochastic variable which behaves deterministically as a zero vector field. We can
write D µναβ (p) = Dpµν αβ (p) + Dhµν αβ (p), where Dpµν αβ (p) is a particular solution to Eq. (6.45) and
Dhµν αβ (p) is the most general solution to the homogeneous equation. Consequently, see Eq. (6.44),
µν µν (p) + G̃(1) µν (p). To find the particular solution, we try an ansatz of
we can write G̃(1)f (p) = G̃(1)
p h
the form
Dpµναβ (p) = d1 (p) δ(α µ ν
δβ) + d2 (p) p2 P µν ηαβ . (6.46)
Substituting this ansatz into Eqs. (6.45), it is easy to see that it solves these equations if we take
1 F2 (p)
   
d1 (p) = , d2 (p) = − , (6.47)
F1 (p) r F1 (p)F3 (p) r
with
1
 
2 2
F3 (p) ≡ F1 (p) + 3p F2 (p) = 1 − 48πG p H̃B (p) − 2β̄ , (6.48)
4
and where the notation [ ]r means that the zeros of the denominators are regulated with appro-
priate prescriptions in such a way that d1 (p) and d2 (p) are well defined Lorentz invariant scalar
distributions. This yields a particular solution to the Einstein-Langevin equations:
G̃(1)
p
µν
(p) = 8πG Dpµν αβ (p) ξ˜αβ (p), (6.49)

43
which, since the stochastic source is conserved, satisfies the conservation condition. Note that, in
the case of a massless scalar field, m = 0, the above solution has a functional form analogous to
that of the solutions of linearized semiclassical gravity found in the Appendix of Ref. [229]. Notice
also that, for a massless conformally coupled field, m = 0 and ∆ξ = 0, the second term on the right
hand side of Eq. (6.46) will not contribute in the correlation functions (6.43), since in this case the
stochastic source is traceless.
µν
A detailed analysis given in Ref. [109] concludes that the homogeneous solution G̃(1) h (p) gives
no contribution to the correlation functions (6.43). Consequently G µναβ (x, x′ ) = hG(1)
p
µν (x)G(1) αβ (x′ )i ,
p s
where G(1) µν (x) is the inverse Fourier transform of (6.49), and the correlation functions (6.43) are
p

hG̃(1)
p
µν
(p)G̃(1)
p (p )is = 64(2π)6G2δ4(p+p′ )Dpµν ρσ (p)Dpαβλγ (−p)Ñ ρσλγ(p).
αβ ′
(6.50)
It is easy to see from the above analysis that the prescriptions [ ]r in the factors Dp are irrelevant
in the last expression and, thus, they can be suppressed. Taking into account that Fl (−p) = Fl∗ (p),
with l = 1, 2, 3, we get from Eqs. (6.46) and (6.47)
µν αβ ′
hG̃(1)
p (p) G̃(1)
p (p )is =
δ4 (p + p′ ) F2 (p) 2 µν αβρ

64(2π)6 G2 2 Ñ µναβ (p) − p P Ñ ρ (p)
|F1 (p)| F3 (p)
F ∗ (p) 2 αβ µνρ |F2 (p)|2 2 µν 2 αβ ρ σ
#
− 2∗ p P Ñ ρ (p)+ p P p P Ñ ρ σ (p) .
F3 (p) |F3 (p)|2
(6.51)
This last expression is well defined as a bi-distribution and can be easily evaluated using Eq. (6.41).
The final explicit result for the Fourier transformed correlation function for the Einstein tensor is
thus
µν αβ ′
hG̃(1)
p (p) G̃(1)
p (p )is =
s
2 δ4 (p + p′ ) m2
= (2π)5 G2 θ(−p2 −4m2 ) 1+4
45 |F1 (p)|2 p2
 !2
1 m2
× 1+4 2 (p2 )2 P µναβ
4 p
!2 
2
m2 F (p)

2
(p2 )2 P µν P αβ 1−3p2
 
+10 3∆ξ + 2  . (6.52)
p F (p) 3

To obtain the correlation functions in coordinate space, Eq. (6.43), we take the inverse Fourier
transform. The final result is:
π 2 µναβ 8π 2 µν αβ
G µναβ (x, x′ ) = G Fx GA (x − x′ ) + G Fx Fx GB (x − x′ ), (6.53)
45 9
with
s !2
m2 m2 1
G̃A (p) ≡ θ(−p2 − 4m2 ) 1+4 2 1+4 2 ,
p p |F1 (p)|2
s !2
m2 m2 1 2
1−3p2 F2 (p)  ,
 
G̃B (p) ≡ θ(−p2 −4m2 ) 1+4 2 3∆ξ + 2

(6.54)
p p |F1 (p)|2 F3 (p) 


44
where Fl (p), l = 1, 2, 3, are given in (6.39) and (6.48). Notice that, for a massless field (m = 0), we
have

F1 (p) = 1 + 4πGp2 H̃(p; µ̄2 ),

16 1
 
F2 (p) = − πG (1 + 180∆ξ 2 ) H̃(p; µ̄2 ) − 6Υ ,
3 4
h i
F3 (p) = 1 − 48πGp2 15∆ξ 2 H̃(p; µ̄2 ) − 2Υ , (6.55)

with µ̄ ≡ µ exp(1920π 2 ᾱ) and Υ ≡ β̄ − 60∆ξ 2 ᾱ, and where H̃(p; µ2 ) is the Fourier transform of
H(x; µ2 ) given in (6.33).

6.4.2 Correlation functions for the metric perturbations

Starting from the solutions found for the linearized Einstein tensor, which are characterized by
the two-point correlation functions (6.53) [or, in terms of Fourier transforms, (6.52)], we can now
solve the equations for the metric perturbations. Working in the harmonic gauge, ∂ν h̄µν = 0 (this
zero must be understood in a statistical sense) where h̄µν ≡ hµν −(1/2)ηµν hαα , the equations for the
metric perturbations in terms of the Einstein tensor are

✷h̄µν (x) = −2G(1) µν (x), (6.56)

˜µν (p) = 2G̃(1) µν (p). Similarly to the analysis of the equation
or, in terms of Fourier transforms, p2 h̄
for the Einstein tensor, we can write h̄µν = hh̄µν is + h̄µν µν
f , where h̄f is a solution to these equations
with zero average, and the two-point correlation functions are defined by

Hµναβ (x, x′ ) ≡ hh̄µν (x)h̄αβ (x′ )is − hh̄µν (x)is hh̄αβ (x′ )is
= hh̄µν αβ ′
f (x)h̄f (x )is . (6.57)

We seek solutions of the Fourier transform of Eq. (6.56) of the form h̄ ˜ µν (p) = 2D(p)G̃(1) µν (p),
f f
where D(p) is a Lorentz invariant scalar distribution in Minkowski spacetime, which is the most
general solution of p2 D(p) = 1. Note that, since the linearized Einstein tensor is conserved, solutions
of this form automatically satisfy the harmonic gauge condition. As in the previous subsection,
we can write D(p) = [1/p2 ]r + Dh (p), where Dh (p) is the most general solution to the associated
homogeneous equation and, correspondingly, we have h̄ ˜ µν (p) + h̄
˜ µν (p) = h̄ ˜ µν (p). However, since
f p h
Dh (p) has support on the set of points for which p2 = 0, it is easy to see from Eq. (6.52) (from the
factor θ(−p2 − 4m2 )) that hh̄ ˜ µν (p)G̃(1) αβ (p′ )i = 0 and, thus, the two-point correlation functions
h f s
˜ µν ˜ αβ ′
(6.57) can be computed from hh̄ (p)h̄ (p )i = hh̄ ˜ µν (p)h̄
˜ αβ (p′ )i . From Eq. (6.52) and due to the
f f s p p s
factor θ(−p2 − 4m2 ), it is also easy to see that the prescription [ ]r is irrelevant in this correlation
function and we obtain
˜ µν (p)h̄
˜ αβ (p′ )i = 4
hh̄ p p s hG̃(1) µν (p) G̃(1) αβ ′
(p )is , (6.58)
(p2 )2 p p

where hG̃(1) µν (p) G̃(1) αβ (p′ )i is given by Eq. (6.52). The right hand side of this equation is a well
p p s
defined bi-distribution, at least for m 6= 0 (the θ function provides the suitable cutoff). In the
massless field case, since the noise kernel is obtained as the limit m → 0 of the noise kernel for a

45
massive field, it seems that the natural prescription to avoid the divergences on the lightcone p2 = 0
is a Hadamard finite part, see Refs. [274, 275] for its definition. Taking this prescription, we also
get a well defined bi-distribution for the massless limit of the last expression.
The final result for the two-point correlation function for the field h̄µν is:
4π 2 µναβ 32π 2 µν αβ
Hµναβ (x, x′ ) = G Fx HA (x − x′ ) + G Fx Fx HB (x − x′ ), (6.59)
45 9
where H̃A (p) ≡ [1/(p2 )2 ] G̃A (p) and H̃B (p) ≡ [1/(p2 )2 ] G̃B (p), with G̃A (p) and G̃B (p) given by (6.54).
The two-point correlation functions for the metric perturbations can be easily obtained using hµν =
h̄µν −(1/2)ηµν h̄αα .

6.4.3 Conformally coupled field

For a conformally coupled field, i.e., when m = 0 and ∆ξ = 0, the previous correlation functions are
greatly simplified and can be approximated explicitly in terms of analytic functions. The detailed
results are given in Ref. [109], here we outline the main features.
When m = 0 and ∆ξ = 0 we have GB (x) = 0 and G̃A (p) = θ(−p2 ) |F1 (p)|−2 . Thus the two-point
correlations functions for the Einstein tensor is
′
π 2 µναβ d4 p eip(x−x ) θ(−p2 )
Z
µναβ ′
G (x, x ) = G Fx , (6.60)
45 (2π)4 |1 + 4πGp2 H̃(p; µ̄2 )|2

where H̃(p, µ2 ) = (480π 2 )−1 ln[−((p0 + iǫ)2 + pi pi )/µ2 ], see Eq. (6.33).
To estimate this integral let us consider spacelike separated points (x − x′ )µ = (0,x − x′ ), and
define y = x − x′ . We may now formally change the momentum variable pµ by the dimensionless
vector sµ : pµ = sµ /|y|, then the previous integral
√ denominator is |1 + 16π(lp /|y|)2 s2 H̃(s)|2 , where
we have introduced the Planck length lp = G. It is clear that we can consider two regimes: (a)
when lp ≪ |y|, and (b) when |y| ∼ lp . In case (a) the correlation function, for the 0000 component,
say, will be of the order
lp4
G 0000 (y) ∼ .
|y|8
In case (b) when the denominator has zeros a detailed calculation carried out in Ref. [109] shows
that: !
lp 1
G 0000 (y) ∼ e−|y|/lp + ... + 2 2
|y|5 lp |y|
which indicates an exponential decay at distances around the Planck scale. Thus short scale
fluctuations are strongly suppressed.
For the two-point metric correlation the results are similar. In this case we have
′
4π 2 µναβ d4 p eip(x−x ) θ(−p2 )
Z
µναβ ′
H (x, x ) = G Fx . (6.61)
45 (2π)4 (p2 )2 |1 + 4πGp2 H̃(p; µ̄2 )|2
The integrand has the same behavior of the correlation function of Eq. (6.60) thus matter fields
tends to suppress the short scale metric perturbations. In this case we find, as for the correlation
of the Einstein tensor, that for case (a) above we have,
lp4
H0000 (y) ∼ ,
|y|4

46
and for case (b) we have
lp
 
H0000 (y) ∼ e−|y|/lp + ... .
|y|
It is interesting to write expression (6.61) in an alternative way. If we use the dimensionless
tensor P µναβ introduced in Eq. (6.23), which accounts for the effect of the operator F µναβ
x , we can
write ′
4π 2 d4 p eip(x−x ) P µναβ θ(−p2 )
Z
µναβ ′
H (x, x ) = G . (6.62)
45 (2π)4 |1 + 4πGp2 H̃(p; µ̄2 )|2
This expression allows a direct comparison with the graviton propagator for linearized quantum
gravity in the 1/N expansion found by Tomboulis [215]. One can see that the imaginary part of
the graviton propagator leads, in fact, to Eq. (6.62). In Ref. [211] it is shown that, in fact, the
two-point correlation functions for the metric perturbations derived from the Einstein-Langevin
equation are equivalent to the symmetrized quantum two-point correlation functions for the metric
fluctuations in the large N expansion of quantum gravity interacting with N matter fields.
The main results of this section are the correlation functions (6.53) and (6.59). In the case of
a conformal field, the correlation functions of the linearized Einstein tensor have been explicitly
estimated. From the exponential factors e−|y|/lp in these results for scales near the Planck length,
we see that the correlation functions of the linearized Einstein tensor have the Planck length as
the correlation length. A similar behavior is found for the correlation functions of the metric
perturbations. Since these fluctuations are induced by the matter fluctuations we infer that the
effect of the matter fields is to suppress the fluctuations of the metric at very small scales. On the
other hand, at scales much larger than the Planck length the induced metric fluctuations are small
compared with the free Rgraviton propagator which goes like lp2 /|y|2 , since the action for the free
graviton goes like Sh ∼ d4 x lp−2 h✷h
For background solutions of semiclassical gravity with other scales present apart from the Planck
scales (for instance, for matter fields in a thermal state), stress-energy fluctuations may be impor-
tant at larger scales. For such backgrounds, stochastic semiclassical gravity might predict correla-
tion functions with characteristic correlation lengths larger than the Planck scales. It seems quite
plausible, nevertheless, that these correlation functions would remain non-analytic in their charac-
teristic correlation lengths. This would imply that these correlation functions could not be obtained
from a calculation involving a perturbative expansion in the characteristic correlation lengths. In
particular, if these correlation lengths are proportional to the Planck constant h̄, the gravitational
correlation functions could not be obtained from an expansion in h̄. Hence, stochastic semiclassical
gravity might predict a behavior for gravitational correlation functions different from that of the
analogous functions in perturbative quantum gravity [241, 242, 243, 244]. This is not necessarily
inconsistent with having neglected action terms of higher order in h̄ when considering semiclassical
gravity as an effective theory [229]. It is, in fact, consistent with the closed connection of stochastic
gravity with the large N expansion of quantum gravity interacting with N matter fields.

6.5 Stability of Minkowski spacetime

In this section we apply the validity criterion for semiclassical gravity introduced in section 3.3 to flat
spacetime. The Minkowski metric is a particularly simple and interesting solution of semiclassical
gravity. In fact, as we have seen in section 6.1, when the quantum fields are in the Minkowski
vacuum state one may take the renormalized expectation value of the stress tensor as hT̂ab R [η]i = 0

47
(this is equivalent to assuming that the cosmological constant is zero), then the Minkowski metric
ηab is a solution of the semiclassical Einstein equation (3.7). Thus, we can look for the stability of
Minkowski spacetime against quantum matter fields. According to the criteria we have established
we have to look for the behavior of the two-point quantum correlations for the metric perturbations
hab (x) over the Minkowski background which are given by Eqs. (3.15) and (3.16). As we have
emphasized before these metric fluctuations separate in two parts: the first term on the right hand
side of Eq. (3.16) which corresponds to the intrinsic metric fluctuations, and the second term which
corresponds to the induced metric fluctuations.

6.5.1 Intrinsic metric fluctuations

Let us first consider the intrinsic metric fluctuations,
hhab (x)hcd (y)iint = hh0ab (x)h0cd (y)is , (6.63)
where h0ab are the homogeneous solutions of the Einstein-Langevin equation (3.14), or equivalently
the linearly perturbed semiclassical equation, and where the statistical average is taken with respect
to the Wigner distribution that describes the initial quantum state of the metric perturbations.
Since these solutions are described by the linearized semiclassical equation around flat spacetime
we can make use of the results derived in Refs. [110, 229, 113, 205]. The solutions for the case of a
massless scalar field were first discussed in Ref. [110] and an exhaustive description can be found
in Appendix A of Ref. [229]. It is convenient to decompose the perturbation around Minkowski
spacetime into scalar, vectorial and tensorial parts, as
 
hab = φ̄ ηab + ∇(a ∇b) − ηab ∇c ∇c ψ + 2∇(a vb) + hTT
ab , (6.64)

where v a is a transverse vector and hTT a

ab is a transverse and traceless symmetric tensor, i.e., ∇a v =
a TT TT a a
0, ∇ hab = 0 and (h )a = 0. A vector field ζ characterizes the gauge freedom due to infinitesimal
diffeomorphisms hab → hab +∇a ζb +∇b ζa . We may use this freedom to choose a gauge, a convenient
election is the so called Lorentz or harmonic gauge defined as
1
 
a
∇ hab − ηab hcc = 0. (6.65)
2
When this gauge is imposed we have the following conditions on the metric perturbations ∇a ∇a v b =
0 and ∇b φ̄ = 0, which implies φ̄ = const. A remaining gauge freedom compatible with the Lorentz
gauge is still possible provided the vector field ζ a satisfies the condition ∇a ∇a ζ b = 0. One can
easily see [118] that the vectorial and scalar part φ̄ can be eliminated, as well as the contribution of
the scalar part ψ which corresponds to Fourier modes ψ̃(p) with p2 = 0. Thus we will assume that
we impose the Lorentz gauge with additional gauge transformations which leave only the tensorial
component and the modes of the scalar component ψ with p2 6= 0 in Fourier space.
(1)
Using the metric decomposition (6.64) we may compute the linearized Einstein tensor Gab . It is
found that the vectorial part of the metric perturbation gives no contribution to this tensor, and the
(1) (S)
scalar and tensorial components give rise, respectively, to scalar and tensorial components: Gab
(1) (T )
and Gab , respectively. Thus let us now write the Fourier transform of the homogeneous Einstein-
Langevin equation (6.37), which is equivalent to the linearized semiclassical Einstein equation,
F µναβ (p) G̃(1) αβ (p) = 0. (6.66)

48
Using the previous decomposition of the Einstein tensor this equation can be re-written in terms
of its scalar and tensorial parts as
h i
F1 (p) + 3p2 F2 (p) G̃(1)
µν
(S)
(p) = 0, (6.67)
F1 (p)G̃(1)
µν
(T)
(p) = 0. (6.68)

(1) (S) (1) (T)

where F1 (p) and F2 (p) are given by Eqs. (6.39), and G̃µν and G̃µν denote, respectively, the
Fourier transformed scalar and tensorial parts of the linearized Einstein tensor. To simplify the
problem and to illustrate, in particular, how the runaway solutions arise we will consider the case of
a massless and conformally coupled field; see Ref. [229] for the massless case with arbitrary coupling
and Refs. [109, 113] for the general massive case. Thus substituting m = 0 and ξ = 1/6 into the
functions F1 (p) and F2 (p), and using Eq. (6.33), the above equations become:
 
1 + 12κβ̄p2 G̃(1)
µν
(S)
(p) = 0, (6.69)
" !#
κp2 −(p0 + iǫ)2 + ~p 2
lim 1+ ln G̃(1)
µν
(T)
(p) = 0, (6.70)
ǫ→0+ 960π 2 µ2

where κ = 8πG. Let us consider these two equations separately.

(1) (S)
For the scalar component when β̄ = 0 the only solution is G̃µν (p) = 0. When β̄ > 0 the
solutions for the scalar component exhibit an oscillatory behavior in spacetime coordinates which
corresponds to a massive scalar field with m2 = (12κ|β̄|)−1 ; for β̄ < 0 the solutions correspond
to a tachyonic field with m2 = −(12κ|β̄|)−1 : in spacetime coordinates they exhibit an exponential
behavior in time, growing or decreasing, for wavelengths larger than 4π(3κ|β̄|)1/2 , and an oscillatory
(1) (S)
behavior for wavelengths smaller than 4π(3κ|β̄|)1/2 . On the other hand, the solution G̃µν (p) = 0
is completely trivial since any scalar metric perturbation h̃µν (p) giving rise to a vanishing linearized
Einstein tensor can be eliminated by a gauge transformation.
For the tensorial component, when µ ≤ µcrit = lp−1 (120π)1/2 eγ , where lp is the Planck length
(lp2 ≡ κ/8π) the first factor in Eq. (6.70) vanishes for four complex values of p0 of the form ±ω and
±ω ∗ , where ω is some complex value. This means that in the corresponding propagator, there are
two poles on the upper half plane of the complex p0 plane and two poles in the lower half plane. We
will consider here the case in which µ < µcrit ; a detailed description of the situation for µ ≥ µcrit
can be found in Appendix A of Ref. [229]. The two zeros on the upper half of the complex plane
correspond to solutions in spacetime coordinates which exponentially grow in time, whereas the
two on the lower half correspond to solutions exponentially decreasing in time. Strictly speaking,
these solutions only exist in spacetime coordinates, since their Fourier transform is not well defined.
They are commonly referred to as runaway solutions and for µ ∼ lp−1 they grow exponentially in
time scales comparable to the Planck time.
(1)
Consequently, in addition to the solutions with Gab (x) = 0, there are other solutions that in
(1)
Fourier space take the form G̃ab (p) ∝ δ(p2 − p20 ) for some particular values of p0 , but all of them
exhibit exponential instabilities with characteristic Planckian time scales. In order to deal with
those unstable solutions, one possibility is to make use of the order reduction prescription [230],
which we will briefly summarize in the last subsection. Note that the p2 terms in Eqs. (6.69)
and (6.70) come from two spacetime derivatives of the Einstein tensor, moreover, the p2 ln p2 term
comes from the nonlocal term of the expectation value of the stress tensor. The order reduction

49
prescription amounts here to neglecting these higher derivative terms. Thus, neglecting the terms
proportional to p2 in Eqs. (6.69) and (6.70), we are left only with the solutions which satisfy
(1)
G̃ab (p) = 0. The result for the metric perturbation in the gauge introduced above can be obtained
by solving for the Einstein tensor, which in the Lorentz gauge of Eq. (6.65) reads:

1 1
 
(1)
G̃ab (p) = p2 h̃ab (p) − ηab h̃cc (p) . (6.71)
2 2

These solutions for h̃ab (p) simply correspond to free linear gravitational waves propagating in
Minkowski spacetime expressed in the transverse and traceless (TT) gauge. When substituting
back into Eq. (6.63) and averaging over the initial conditions we simply get the symmetrized
quantum correlation function for free gravitons in the TT gauge for the state given by the Wigner
distribution. As far as the intrinsic fluctuations are concerned, it seems that the order reduction
prescription is too drastic, at least in the case of Minkowski spacetime, since no effects due to the
interaction with the quantum matter fields are left.
A second possibility, proposed by Hawking et al. [137, 276], is to impose boundary conditions
which discard the runaway solutions that grow unbounded in time. These boundary conditions
correspond to a special prescription for the integration contour when Fourier transforming back
to spacetime coordinates. As we will discuss in some more detail in the next subsection, this
prescription reduces here to integrating along the real axis in the p0 complex plane. Following that
procedure we get, for example, that for a massless conformally coupled matter field with β̄ > 0
the intrinsic contribution to the symmetrized quantum correlation function coincides with that
of free gravitons plus an extra contribution for the scalar part of the metric perturbations. This
extra massive scalar renders Minkowski spacetime stable, but also plays a crucial role in providing a
graceful exit in inflationary models driven by the vacuum polarization of a large number of conformal
fields. Such a massive scalar field would not be in conflict with present observations because, for the
range of parameters considered, the mass would be far too large to have observational consequences
[137].

6.5.2 Induced metric fluctuations

The induced metric fluctuations are described by the second term in Eq. (3.16). They are dependent
on the noise kernel that describes the stress tensor fluctuations of the matter fields,

κ̄2
Z q
hhab (x)hcd (y)iind = d4 x′ d4 y ′ g(x′ )g(y ′ )Gret ′
abef (x, x )N
ef gh ′ ′
(x , y )Gret ′
cdgh (y, y ), (6.72)
N
where here we have written the expression in the large N limit, so that κ̄ = N κ (κ = 8πG) and
N is the number of independent free scalar fields. The contribution corresponding to the induced
quantum fluctuations is equivalent to the stochastic correlation function obtained by considering
just the inhomogeneous part of the solution to the Einstein-Langevin equation. We can make use
of the results for the metric correlations obtained in sections 6.3 and 6.4 for solving the Einstein-
Langevin equation. In fact, one should simply take N = 1 to transform our expressions here to those
of sections 6.3 and 6.4 or, more precisely, one should multiply the noise kernel in the expressions of
sections 6.3 and 6.4 by N in order to use those expressions here, as follows from the fact that now
we have N independent matter fields.

50
 As we have seen in section 6.4, following Ref. [109], the Einstein-Langevin equation can be
entirely written in terms of the linearized Einstein tensor. The equation involves second spacetime
derivatives of that tensor and in terms of its Fourier components is given in Eq. (6.37) as

F µναβ (p) G̃(1) αβ (p) = κ̄ ξ̃ µν (p), (6.73)

where we have used now the rescaled coupling κ̄. The solution for the linearized Einstein tensor is
given in Eq. (6.49) in terms of the retarded propagator Dµνρσ (p) defined in Eq. (6.45). Now this
propagator, which is written in Eq. (6.46), exhibits two poles in the upper half complex p0 plane
and two poles in the lower half plane, as we have seen analyzing the zeros in Eqs. (6.69) and (6.70)
for the massless and conformally coupled case. The retarded propagator in spacetime coordinates
is obtained as usual by taking the appropriate integration contour in the p0 plane. It is convenient
in this case to deform the integration path along the real p0 axis so as to leave the two poles of the
upper half plane below that path. In this way when closing the contour by un upper half circle, in
order to compute the anti-causal part of the propagator, there will be no contribution. The problem
now is that when closing the contour on the lower half plane, in order to compute the causal part,
the contribution of the upper half plane poles gives an unbounded solution, a runaway instability.
If we adopt Hawking et al. [137, 276] criterion of imposing final boundary conditions which discard
solutions growing unboundedly in time this implies that we just need to take the integral along
the real axis, as was done in section 6.4.2. But now the propagator
√ is no longer strictly retarded,
there are causality violations in time scales of the order of N lp , which should have no observable
consequences. This propagator, however, has well defined Fourier transform.
Following the steps after Eq. (6.49) the Fourier transform of the two-point correlation for the
linearized Einstein tensor can be written in our case as,

(1) κ̄2
hG̃(1) ′
µν (p)G̃αβ (p )iind = (2π)4 δ4(p + p′ )Dµνρσ (p)Dαβλγ (−p)Ñ ρσλγ (p), (6.74)
N
where the noise kernel Ñ ρσλγ (p) is given by Eq. (6.41). Note that these correlation functions are
invariant under gauge transformations of the metric perturbations because the linearized Einstein
tensor is invariant under those transformations.
We may also take the order reduction prescription which amounts in this case to neglecting
terms in the propagator which are proportional to p2 , corresponding to two spacetime derivatives
of the Einstein tensor. The propagator then becomes a constant, and we have

(1) κ̄2
hG̃(1) ′
µν (p)G̃αβ (p )iind = (2π)4 δ4(p+p′ )ѵναβ (p). (6.75)
N
Finally we may derive the correlations for the metric perturbations from these equations (6.74)
or (6.75). In the Lorentz or harmonic gauge the linearized Einstein tensor takes the particularly
simple form of Eq. (6.71) in terms of the metric perturbation. One may derive the correlation
functions for h̃µν (p) as it was done in section 6.4.2 to get

˜ (p)h̄
˜ (p′ )i = 4
hh̄µν αβ ind hG̃µν (p)G̃αβ (p′ )iind . (6.76)
(p2 )2
There will be one possible expression for the two-point metric correlation which corresponds to the
Einstein tensor correlation of Eq. (6.74) and another expression corresponding to Eq. (6.74), when

51
the order reduction prescription is used. We should note that contrary to the correlation functions
for the Einstein tensor, the two-point metric correlation is not gauge invariant (it is given in the
Lorentz gauge). Moreover, when taking the Fourier transform to get the correlations in spacetime
coordinates there is an apparent infrared divergence when p2 = 0 in the massless case. This can
be seen from the expression for the noise kernel ѵναβ (p) defined in Eq. (6.41). For the massive
case no such divergence due to the factor θ(−p2 − 4m2 ) exists, but as one takes the limit m → 0 it
will show up. This infrared divergence, however, is a gauge artifact that has been enforced by the
use of the Lorentz gauge. A gauge different from the Lorentz gauge should be used in the massless
case; see Ref. [118] for a more detailed discussion of this point.
Let us now write the two-point metric correlation function in spacetime coordinates for the
massless and conformally coupled fields. In order to avoid runaway solutions we use the prescription
that the propagator should have a well defined Fourier transform, by integrating along the real axis
in the complex p0 plane. This was, in fact, done in section 6.4.3, see Eq. (6.62), which we may
write now as
˜ (y)i = κ̄
2 d4 p eip(x−y) Pµναβ θ(−p2 )
Z
˜ (x)h̄
hh̄ , (6.77)
µν αβ ind
720πN (2π)4 |1 + (κ̄/2)p2 H̃(p; µ̄2 )|2
where the projector Pµναβ is defined in Eq. (6.23). This correlation function for the metric per-
turbations is in agreement with the real part of the graviton propagator obtained by Tomboulis
in Ref. [215] using a large N expansion with Fermion fields. Note that when the order reduction
prescription is used the terms in the denominator of Eq. (6.77) which are proportional to p2 are ne-
glected. Thus, in contrast to the intrinsic metric fluctuations, there is still a nontrivial contribution
to the induced metric fluctuations due to the quantum matter fields in this case.
To estimate the above integral let us follow section 6.4.3 and consider space like separated
points x − y = (0, r) and introduce the Planck length lp . For space separations
√ |r| ≫ lp we have
that the two-point correlation (6.77) goes like ∼ N lp4 /|r|4 , and for |r| ∼ N lp we have that they
√
go like ∼ exp(−|r|/ N lp )lp /|r|. Since these metric fluctuations are induced by the matter stress
fluctuations we infer that the effect of the matter fields is to suppress metric fluctuations at small
scales. On the other hand, at large scales the induced metric fluctuations are small compared to
the free graviton propagator which goes like lp2 /|r|2 .
We thus conclude that, once the instabilities giving rise to the unphysical runaway solutions
have been discarded, the fluctuations of the metric perturbations around the Minkowski spacetime
induced by the interaction with quantum scalar fields are indeed stable (instabilities lead to di-
vergent results when Fourier transforming back to spacetime coordinates). We have found that,
indeed, both the intrinsic and the induced contributions to the quantum correlation functions of
metric perturbations are stable, and consequently Minkowski spacetime is stable.

6.5.3 Order reduction prescription and large N

Runaway solutions are a typical feature of equations describing back-reaction effects, such is in
classical electrodynamics, and are due to higher than two time derivatives in the dynamical equa-
tions. Here we will give a qualitative analysis of this problem in semiclassical gravity. In a very
schematic way the semiclassical Einstein equations have the form

Gh + lp2 G̈h = 0, (6.78)

52
where Gh stands for the linearized Einstein tensor over the Minkowski background, say, and we have
simplified the equation as much as possible. The second term of the equation is due to the vacuum
polarization of matter fields and contains four time derivatives of the metric perturbation. Some
specific examples of such an equation are, in momentum space, Eqs. (6.69) and (6.70). The order
reduction procedure is based on treating perturbatively the terms involving higher order derivatives,
differentiating the equation under consideration and substituting back the higher derivative terms in
the original equation keeping only terms up to the required order in the perturbative parameter. In
the case of the semiclassical Einstein equation, the perturbative parameter is lp2 . If we differentiate
twice Eq. (6.78) with respect to time it is clear that the second order derivatives of the Einstein
tensor are of order lp2 . Substituting back into the original equation, we get the following equation
up to order lp4 : Gh = 0 + O(lp4 ). Now, there are certainly no runaway solutions but also no effect
due to the vacuum polarization of matter fields. Note that the result is not so trivial when there is
an inhomogeneous term on the right hand side of Eq. (6.78), this is what happens with the induced
fluctuations predicted by the Einstein-Langevin equation.
Semiclassical gravity is expected to provide reliable results as long as the characteristic length
scales under consideration, say L, satisfy that L ≫ lp [229]. This can be qualitatively argued by
estimating the magnitude of the different contributions to the effective action for the gravitational
field, considering the relevant Feynman diagrams and using dimensional arguments. Let us write
the effective gravitational action, again in a very schematic way, as
!
√ 1
Z
Seff = d4 x −g R + αR2 + lp2 R3 + . . . , (6.79)
lp2

where R is the Ricci scalar. The first term is the usual classical Einstein-Hilbert term, the second
stands for terms quadratic in the curvature (square of Ricci and Weyl tensors) these terms appear
as radiative corrections due to vacuum polarization of matter fields, here α is a dimensionless
parameter presumably of order 1, the R3 terms are higher order corrections which appear for
instance when one considers internal graviton propagators inside matter loops. Let us assume that
R ∼ L−2 then the different terms in the action are of the order of R2 ∼ L−4 and lp2 R3 ∼ lp2 L−6 .
Consequently when L ≫ lp2 , the term due to matter loops is a small correction to the Einstein-
Hilbert term (1/lp2 )R ≫ R2 , and this term can be treated as a perturbation. The justification of the
order reduction prescription is actually based on this fact. Therefore, significant effects from the
vacuum polarization of the matter fields are only expected when their small corrections accumulate
in time, as would be the case, for instance, for an evaporating macroscopic black hole all the way
before reaching Planckian scales (see section 8.3).
However if we have a large number N of matter fields the estimates for the different terms
change in a remarkable way. This is interesting because the large N expansion seems, as we have
argued in section 3.3.1, the best justification for semiclassical gravity. In fact, now the N vacuum
polarization terms involving loops of matter are of order N R2 ∼ N L−4 . For this reason the
contribution of the graviton loops, which is just of order R2 as any other loop of matter, can be
neglected in front of the matter loops; this justifies the semiclassical limit. Similarly
√ higher order
corrections are of order N lp2 R3 ∼ N lp2 L−6 . Now there is a regime, when L ∼ N lp , where the
Einstein-Hilbert term is comparable to the vacuum polarization of matter fields, (1/lp2 )R ∼ N R2 ,
and yet the higher correction terms can be neglected because we still have L ≫ lp , provided N ≫ 1.
This is the kind of situation considered in trace anomaly driven inflationary models [137], such as
that originally proposed by Starobinsky [135], see also Ref. [136], where exponential inflation is

53
driven by a large number of massless conformal fields. The order reduction prescription would
completely discard the effect from the vacuum polarization of the matter fields even though it is
comparable to the Einstein-Hilbert term. In contrast, the procedure proposed by Hawking et al.
keeps the contribution from the matter fields. Note that here the actual physical Planck length lp
is considered, not the rescaled one, ¯lp2 = κ̄/8π, which is related to lp by lp2 = κ/8π = ¯lp2 /N .

6.5.4 Summary
An analysis of the stability of any solution of semiclassical gravity with respect to small quantum
perturbations should include not only the evolution of the expectation value of the metric per-
turbations around that solution, but also their fluctuations, encoded in the quantum correlation
functions. Making use of the equivalence (to leading order in 1/N , where N is the number of mat-
ter fields) between the stochastic correlation functions obtained in stochastic semiclassical gravity
and the quantum correlation functions for metric perturbations around a solution of semiclassical
gravity, the symmetrized two-point quantum correlation function for the metric perturbations can
be decomposed into two different parts: the intrinsic metric fluctuations due to the fluctuations of
the initial state of the metric perturbations itself, and the fluctuations induced by their interaction
with the matter fields. From the linearized perturbations of the semiclassical Einstein equation, in-
formation on the intrinsic metric fluctuations can be retrieved. On the other hand, the information
on the induced metric fluctuations naturally follows from the solutions of the Einstein-Langevin
equation.
We have analyzed the symmetrized two-point quantum correlation function for the metric
perturbations around the Minkowski spacetime interacting with N scalar fields initially in the
Minkowski vacuum state. Once the instabilities that arise in semiclassical gravity which are com-
monly regarded as unphysical, have been properly dealt with by using the order reduction prescrip-
tion or the procedure proposed by Hawking et al. [137, 276], both the intrinsic and the induced
contributions to the quantum correlation function for the metric perturbations are found to be sta-
ble [118]. Thus, we conclude that Minkowski spacetime is a valid solution of semiclassical gravity.

7 Structure formation in the early universe

Structure formation in the early universe is a key problem in modern cosmology. It is believed that
galaxies, clusters or any large scale structure observed today originated from random sources of
primordial inhomogeneities (density contrast) in the early universe amplified by the expansion of the
universe. Theories of structure formation based on general relativity theory have been in existence
for over 60 years [277, 278, 193] (see e.g., [279, 280, 281]) long before the advent of inflationary
cosmology [282, 283, 284] but the inflation paradigm [195, 196, 197, 198] provided at least two
major improvements in the modern theory of cosmological structure formation [285, 286, 287, 288]:

• The sources: instead of a classical white noise source arbitrarily specified, the seeds of struc-
tures of the new theory are from quantum fluctuations which obey equations derivable from
the dynamics of the inflaton field, the quantum field which is responsible for driving inflation.

• The spectrum: The almost scale- invariant spectrum (masses of galaxies as a function of
their scales) has a more natural explanation from the almost exponential expansion of the

54
 inflationary universe than from the power law expansion of the FRW universe in the traditional
theory.

Stochastic gravity provides a sound and natural formalism for the derivation of the cosmological
perturbations generated during inflation. In Ref. [134] it was shown that the correlation functions
that follow from the Einstein-Langevin equation which emerges in the framework of stochastic
gravity coincide with that obtained with the usual quantization procedures [288] when both the
metric perturbations and the inflaton fluctuations are linearized. Stochastic gravity, however, can
naturally deal with the fluctuations of the inflaton field even beyond the linear approximation. In
the last subsection we will enumerate possible advantages of the stochastic gravity treatment of this
problem over the usual methods based on the quantization of the linear cosmological and inflaton
perturbations.
We should point out that the equivalence at the linearized level is proved in Ref. [134] directly
from the field equations of the perturbations and by showing that the stochastic and the quantum
correlations are both given by identical expressions. Within the stochastic gravity framework an
explicit computation of the curvature perturbation correlations was performed by Urakawa and
Maeda [289]. A convenient approximation for that computation, used by these authors, leads only
to a small discrepancy with the usual approach for the observationally relevant part of the spectrum.
We think the deviation from the standard result found for superhorizon modes would not arise if
an exact calculation were used.
Here, we illustrate the equivalence with the conventional approach with one of the simplest
chaotic inflationary models in which the background spacetime is a quasi de Sitter universe [217,
134].

7.1 The model

In this chaotic inflationary model [283] the inflaton field φ of mass m is described by the following
Lagrangian density,
1 1
L(φ) = gab ∇a φ∇b φ + m2 φ2 . (7.1)
2 2
The conditions for the existence of an inflationary period, which is characterized by an accelerated
cosmological expansion, is that the value of the field over a region with the typical size of the Hubble
radius is higher than the Planck mass mP . This is because in order to solve the cosmological horizon
and flatness problem more than 60 e-folds of expansion are needed, to achieve this the scalar field
should begin with a value higher than 3mP . The inflaton mass is small: as we will see, the large
scale anisotropies measured in the cosmic background radiation [290] restrict the inflaton mass to
be of the order of 10−6 mP . We will not discuss the naturalness of this inflationary model but simply
assume that if one such region is found (inside a much larger universe) it will inflate to become our
observable universe.
We want to study the metric perturbations produced by the stress-energy tensor fluctuations
of the inflaton field on the homogeneous background of a flat Friedmann-Robertson-Walker model,
described by the cosmological scale factor a(η), where η is the conformal time, which is driven by
the homogeneous inflaton field φ(η) = hφ̂i. Thus we write the inflaton field in the following form

φ̂ = φ(η) + ϕ̂(x), (7.2)

55
where ϕ̂(x) corresponds to a free massive quantum scalar field with zero expectation value on
the homogeneous background metric: hϕ̂i = 0. We will restrict ourselves to scalar-type metric
perturbations because these are the ones that couple to the inflaton fluctuations in the linear
theory. We note that this is not so if we were to consider inflaton fluctuations beyond the linear
approximation, then tensorial and vectorial metric perturbations would also arise. The perturbed
metric g̃ab = gab + hab can be written in the longitudinal gauge as,

ds2 = a2 (η)[−(1 + 2Φ(x))dη 2 + (1 − 2Ψ(x))δij dxi dxj ], (7.3)

where the scalar metric perturbations Φ(x) and Ψ(x) correspond to Bardeen’s gauge invariant
variables [280].

7.2 Einstein-Langevin equation for scalar metric perturbations

The Einstein-Langevin equation as described in Sec. 3 is gauge invariant, and thus we can work in
a desired gauge and then extract the gauge invariant quantities. The Einstein-Langevin equation
(3.14) now reads:
(0) (0) (1) (1)
Gab − 8πGhT̂ab i + Gab (h) − 8πGhTˆab (h)i = 8πGξab , (7.4)
(0) (0)
Note that the two first terms cancel, that is, Gab − 8πGhT̂ab i = 0, if the background metric is a
solution of the semiclassical Einstein equations. Here the subscripts (0) and (1) refer to functions
in the background metric gab and linear in the metric perturbation hab , respectively. The stress
tensor operator T̂ab for the minimally coupled inflaton field in the perturbed metric is:

T̂ab = ∇ ˜ b φ̂ + 1 g̃ab (∇
˜ a φ̂∇ ˜ c φ̂∇
˜ c φ̂ + m2 φ̂2 ). (7.5)
2
Using the decomposition of the scalar field into its homogeneous and inhomogeneous part, see
Eq. (7.2), and the metric g̃ab into its homogeneous background gab and its perturbation hab , the
renormalized expectation value for the stress-energy tensor operator can be written as
R R
hT̂ab [g̃]i = hT̂ab [g̃]iφφ + hT̂ab [g̃]iφϕ + hT̂ab [g̃]iϕϕ , (7.6)

where the subindices indicate the degree of dependence on the homogeneous field φ and its pertur-
bation ϕ. The first term in this equation depends only on the homogeneous field and it is given by
the classical expression. The second term is proportional to hϕ̂[g̃]i which is not zero because the
field dynamics is considered on the perturbed spacetime, i.e., this term includes the coupling of
the field with hab and may be obtained from the expectation value of the linearized Klein-Gordon
equation,  
✷g+h − m2 ϕ̂ = 0. (7.7)
The last term in Eq. (7.6) corresponds to the expectation value of the stress tensor for a free scalar
field on the spacetime of the perturbed metric.
After using the previous decomposition, the noise kernel Nabcd [g; x, y) defined in Eq. (3.11) can
be written as

h{t̂ab [g; x), t̂cd [g; y)}i = h{t̂ab [g; x), t̂cd [g; y)}i(φϕ)2
+h{t̂ab [g; x), tˆcd [g; y)}i(ϕϕ)2 , (7.8)

56
where we have used the fact that hϕ̂i = 0 = hϕ̂ϕ̂ϕ̂i for Gaussian states on the background geometry.
We consider the vacuum state to be the Euclidean vacuum which is preferred in the de Sitter
background, and this state is Gaussian. In the above equation the first term is quadratic in ϕ̂
whereas the second one is quartic, both contributions to the noise kernel are separately conserved
since both φ(η) and ϕ̂ satisfy the Klein-Gordon field equations on the background spacetime.
Consequently, the two terms can be considered separately. On the other hand, if one treats ϕ̂ as a
small perturbation the second term in (7.8) is of lower order than the first and may be consistently
neglected, this corresponds to neglecting the last term of Eq. (7.6). The stress tensor fluctuations
due to a term of that kind were considered in Ref. [291].
We can now write down the Einstein-Langevin equations (7.4) to linear order in the inflaton
fluctuations. It is easy to check [134] that the space-space components coming from the stress tensor
expectation value terms and the stochastic tensor are diagonal, i.e. hT̂ij i = 0 = ξij for i 6= j. This,
in turn, implies that the two functions characterizing the scalar metric perturbations are equal:
Φ = Ψ in agreement with ref. [288]. The equation for Φ can be obtained from the 0i-component of
the Einstein-Langevin equation which, neglecting a nonlocal term, reads in Fourier space as

2iki (HΦk + Φ′k ) = 8πG(ξ0i )k , (7.9)

where ki is the comoving momentum component associated to the comoving coordinate xi , and we
have used the definition Φk (η) = d x exp(−i~k · ~x)Φ(η, ~x). Here primes denote derivatives with
R 3

respect to the conformal time η and H = a′ /a. A nonlocal term of dissipative character which
comes from the second term in Eq. (7.6) should also appear on the left hand side of Eq. (7.9), but
we have neglected it to simplify the forthcoming expressions (the large scale spectrum does not
change in a substantial way). We must emphasize, however, that the proof of the equivalence of
the stochastic approach to linear order in ϕ̂ to the usual linear cosmological perturbations approach
does not assume that simplification [134]. To solve Eq. (7.9), whose left-hand side comes from the
linearized Einstein tensor for the perturbed metric [288], we need the retarded propagator for the
gravitational potential Φk ,

4π a(η ′ )
 
Gk (η, η ′ ) = −i θ(η − η ′ ) + f (η, η ′ ) , (7.10)
ki m2P a(η)

where f is a homogeneous solution of Eq. (7.9) related to the initial conditions chosen and m2P =
1/G. For instance, if we take f (η, η ′ ) = −θ(η0 − η ′ )a(η ′ )/a(η) the solution would correspond to
“turning on” the stochastic source at η0 . With the solution of the Einstein-Langevin equation (7.9)
for the scalar metric perturbations we are in the position to compute the two-point correlation
functions for these perturbations.

7.3 Correlation functions for scalar metric perturbations

The two-point correlation function for the scalar metric perturbations induced by the inflaton
fluctuations is thus given by

hΦk (η)Φk′ (η ′ )is = (2π)2 δ(~k + ~k′ )

Z η Z η′
× dη1 dη2 Gk (η, η1 )Gk′ (η ′ , η2 )h(ξ0i )k (η1 )(ξ0i )k′ (η2 )is . (7.11)

57
Here the two-point correlation function for the stochastic source, which is connected to the stress-
energy tensor fluctuations through the noise kernel, is given by,
1
h(ξ0i )k (η1 )(ξ0i )−k (η2 )is =h{(t̂0i )k (η1 ), (t̂0i )−k (η2 )}iφϕ
2
1 (1)
= ki ki φ′ (η1 )φ′ (η2 )Gk (η1 , η2 ), (7.12)
2
(1)
where Gk (η1 , η2 ) = h{ϕ̂k (η1 ), ϕ̂−k (η2 )}i is the kth-mode Hadamard function for a free minimally
coupled scalar field in the appropriate vacuum state on the Friedmann-Robertson-Walker back-
ground.
In practice, to make the explicit computation of the Hadamard function we will assume that the
field state is in the Euclidean vacuum and the background spacetime is de Sitter. Furthermore we
will compute the Hadamard function for a massless field, and will make a perturbative expansion
in terms of the dimensionless parameter m/mP . Thus we consider
(1)
Ḡk (η1 , η2 ) = h0|{ŷk (η1 ), ŷ−k (η2 )}|0i = 2R (uk (η1 )u∗k (η2 )) ,

with
ŷk (η) = a(η)ϕ̂k (η) = âk uk (η) + â†−k u∗−k (η),
and where
uk = (2k)−1/2 eikη (1 − i/η)
are the positive frequency k-modes for a massless minimally coupled scalar field on a de Sitter
background, which define the Euclidean vacuum state, âk |0i = 0 [32].
The assumption of a massless field for the computation of the Hadamard function is made
because massless modes in de Sitter are much simpler to deal with than massive modes. We
can see that this is nonetheless a reasonable approximation as follows: For a given mode the
m = 0 approximation is reasonable when its wavelength λ is shorter than the Compton wavelength,
λc = 1/m. In our case we have a very small mass m and the horizon size H −1 , where H is the
Hubble constant H = ȧ/a (here a(t) with t the physical time dt = adη), satisfies that H −1 < λc .
Thus, for modes inside the horizon λ < λc and m = 0 is a good approximation. Outside the horizon
massive modes decay in amplitude as ∼ exp(−m2 t/3H) whereas massless modes remain constant,
thus when modes leave the horizon the approximation will eventually break down. However, we
only need to ensure that the approximation is still valid after 60 e-folds, i.e. H∆t ∼ 60 (∆t being
the time between horizon exit and the end of inflation). But this is the case provided 3H 2 /m2 > 60,
since the decay factor ∼ exp[−(m2 /3H 2 )H∆t] will not be too different from unity for those modes
that left the horizon during the last sixty e-folds of inflation. This condition is indeed satisfied
given that m ≪ H in most slow-roll inflationary models [282, 281], and in particular for the model
considered here where m ∼ 10−6 mP .
We note that the background geometry is not exactly that of de Sitter spacetime, for which
a(η) = −(Hη)−1 with −∞ < η < 0. One can expand in terms of the “slow-roll” parameters and
assume that to first order φ̇(t) ≃ m2P (m/mP ), where t is the physical time. The correlation function
for the metric perturbation (7.11) can then be easily computed; see Ref. [217, 134] for details. The
final result, however, is very weakly dependent on the initial conditions as one may understand
from the fact that the accelerated expansion of the quasi-de Sitter spacetime during inflation erases

58
the information about the initial conditions. Thus one may take the initial time to be η0 = −∞
and obtain to lowest order in m/mP the expression
2
m

hΦk (η)Φk′ (η ′ )is ≃ 8π 2 k−3 (2π)3 δ(~k + ~k′ ) cos k(η − η ′ ). (7.13)
mP
From this result two main conclusions can be derived. First, the prediction of an almost scale-
invariant Harrison-Zel’dovich spectrum for large scales, i.e. small values of k. Second, since the
correlation function is of the order of (m/mP )2 a severe bound to the mass m is imposed by
the gravitational fluctuations derived from the small values of the Cosmic Microwave Background
(CMB) anisotropies detected by COBE. This bound is of the order of (m/mP ) ∼ 10−6 [290, 288].
We now comment on some differences with Refs. [130, 131, 132, 133] which used a self-interacting
scalar field or a scalar field interacting nonlinearly with other fields. In those works an important
relaxation of the ratio m/mP was found. The long wavelength modes of the inflaton field were
regarded as an open system in an environment made out of the shorter wavelength modes. Then,
Langevin type equations were used to compute the correlations of the long wavelength modes driven
by the fluctuations of the shorter wavelength modes. In order to get a significant relaxation on the
above ratio, however, one had to assume that the correlations of the free long wavelength modes,
which correspond to the dispersion of the system’s initial state, had to be very small. Otherwise
they dominate by several orders of magnitude those fluctuations that come from the noise of the
environment. This would require a great amount of fine-tuning for the initial quantum state of
each mode [134].
We should remark that in the linear model discussed here there is no environment for the
inflaton fluctuations. When one linearizes with respect to both the scalar metric perturbations
and the inflaton perturbations, the system cannot be regarded as a true open quantum system.
The reason is that Fourier modes decouple and the dynamical constraints due to diffeomorphism
invariance link the metric perturbations of scalar type with the perturbations of the inflaton field
so that only one true dynamical degree of freedom is left for each Fourier mode. Nevertheless, the
inflaton fluctuations are responsible for the noise that induces the metric perturbations.

7.4 Summary and outlook

Stochastic gravity provides an alternative framework to study the generation of primordial inho-
mogeneities in inflationary models. Besides the interest of the problem in its own right, there are
also other reasons that make this problem worth discussing from the point of view of stochastic
gravity. The Einstein-Langevin equation is not restricted by the use of linearized perturbations of
the inflaton field. In practice this may not be very important for inflationary models which are
driven by an inflaton field which takes a non-zero expectation value, because the linear perturba-
tions will give the leading contribution. However, the importance of considering corrections due
to one-loop contributions from the inflaton field perturbations, beyond the tree level of the linear
cosmological perturbation theory, has recently been emphasized by Weinberg [292, 293] as a means
to understand the consequences of the theory even where in practice its predictions may not be
observed. Note also Refs. [294, 295] for the effect of the quantum gravitational loop corrections on
the dynamics of the inflaton field.
In the stochastic gravity approach some insights on the exact treatment of the inflaton scalar
field perturbations have been discussed in Refs. [134, 289, 296]. The main features that would char-
acterize an exact treatment of the inflaton perturbations are the following. First, the three types

59
of metric perturbations (scalar, vectorial and tensorial perturbations) couple to the perturbations
of the inflaton field. Second, the corresponding Einstein-Langevin equation for the linear metric
perturbations will explicitly couple the scalar and tensorial metric perturbations. Furthermore,
although the Fourier modes (with respect to the spatial coordinates) for the metric perturbations
will still decouple in the Einstein-Langevin equation, any given mode of the noise and dissipa-
tion kernels will get contributions from an infinite number of Fourier modes of the inflaton field
perturbations. This fact will imply, in addition, the need to properly renormalize the ultraviolet
divergences arising in the dissipation kernel, which actually correspond to the divergences associ-
ated with the expectation value of the stress tensor operator of the quantum matter field evolving
on the perturbed geometry.
We should remark that although the gravitational fluctuations are here assumed to be classi-
cal, the correlation functions obtained correspond to the expectation values of the symmetrized
quantum metric perturbations [218, 134]. This means that even in the absence of decoherence,
the fluctuations predicted by the Einstein-Langevin equation still give the correct symmetrized
quantum two-point correlation functions. In Ref. [218] it was explained how a stochastic descrip-
tion based on a Langevin-type equation could be introduced to gain information on fully quantum
properties of simple linear open systems. In a forthcoming paper [211] it will be shown that, by
carefully dealing with the gauge freedom and the consequent dynamical constraints, this result can
be extended to the case of N free quantum matter fields interacting with the metric perturbations
around a given background. In particular, the correlation functions for the metric perturbations
obtained using the Einstein-Langevin equation are equivalent to the correlation functions that
would follow from a purely quantum field theory calculation up to the leading order contribution
in the large N limit. This will generalize the results already obtained on a Minkowski background
[118, 119].
These results have important implications on the use of the Einstein-Langevin equation to ad-
dress situations in which the background configuration for the scalar field vanishes. This includes
not only the case of a Minkowski background spacetime, but also the remarkably interesting case of
the trace anomaly induced inflation. That is, inflationary models driven by the vacuum polarization
of a large number of conformal fields [135, 136, 137], where the usual approaches based on the lin-
earization of both the metric perturbations and the scalar field perturbations and their subsequent
quantization can no longer be applied. More specifically, the semiclassical Einstein equations (3.7)
for massless quantum fields conformally coupled to the gravitational field admit an inflationary
solution that begins in an almost de Sitter like regime and ends up in a matter-dominated like
regime [135, 136]. In these models the standard approach based on the quantization of the gravi-
tational and the matter fields to linear order cannot be used because the calculation of the metric
perturbations correspond to having only the last term in the noise kernel in Eq. (7.8), since there
is no homogeneous field φ(η) as the expectation value hφ̂i = 0, and linearization becomes trivial.
In the trace anomaly induced inflation model Hawking et al. [137] were able to compute the
two-point quantum correlation function for scalar and tensorial metric perturbations in a spatially
closed de Sitter universe, making use of the anti-de Sitter / conformal field theory correspondence.
They find that short scale metric perturbations are strongly suppressed by the conformal matter
fields. This is similar to what we obtained in Sec. 6 for the induced metric fluctuations in Minkowski
spacetime. In the stochastic gravity context, the noise kernel in a spatially closed de Sitter back-
ground was derived in Ref. [291], and in a spatially-flat arbitrary Friedmann-Robertson-Walker
model the Einstein-Langevin equations describing the metric perturbations were first obtained in

60
Ref. [104]. The computation of the corresponding two-point correlation functions for the metric
perturbations is now work in progress.

8 Black Hole Backreaction and Fluctuations

As another illustration of the application of stochastic gravity we now consider the backreaction and
fluctuations in black hole spacetimes. Backreaction refers to the quantum effects of matter fields
such as vacuum polarization, quantum fluctuations and particle creation on the spacetime structure
and dynamics. Studying the dynamics of quantum fields in a fixed background spacetime, Hawking
found that black holes emit thermal radiation with a temperature inversely proportional to their
mass [184, 185, 186, 187]. When the backreaction of the quantum fields on the spacetime dynamics
is included, one expects that the mass of the black hole decreases as thermal radiation at higher and
higher temperatures is emitted. The reduction of the mass of a black hole due to particle creation
is often referred to as the black hole ‘evaporation’ process. Backreaction of Hawking radiation
[297, 298, 233, 234, 235, 299, 300, 301] could alter the evolution of the background spacetime and
change the nature of its end state, more drastically so for Planck size black holes.
Backreaction is a technically challenging but conceptually rewarding problem. Progress is slow
in this long standing problem but it cannot be ignored because existing results from test field
approximations or semiclassical analysis are not trustworthy when backreaction becomes strong
enough as to alter the structure and dynamics of the background spacetime. At the least one needs
to know how strong the backreaction effects are under what circumstances the existing predictions
make sense. Without an exact quantum solution of the black hole plus quantum field system
or at least a full backreaction consideration including the intrinsic and induced effects of metric
fluctuations, much of the long speculations on the end-state of black hole collapse – remnants,
naked singularity, baby universe formation or complete evaporation (see e.g., [302, 303, 304, 305,
306, 307, 308, 309, 310, 311, 312].) – and the information loss issue [313, 314, 315, 316] (see, e.g.,
[317, 318] for an overview, and recent results from quantum information [319, 320]) will remain
to be speculations and puzzles. This issue also enters in the extension of the well-known black
hole thermodynamics [321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334] to
nonequilibrium conditions [161] and can lead to new inferences on the microscopic structure of
spacetime and the true nature of Einstein’s equations [160] along the viewpoint regarding general
relativity as geometro-hydrodynamics and gravity as emergent phenomena. (See the nontraditional
views of Volovik [159, 162], Hu [157, 169, 168] on spacetime structure, Wen [335, 166, 336, 167] on
quantum order, Seiberg [164], Horowitz and Polchinsky [165] on emergent gravity, Herzog on the
hydrodynamics of M-theory [163] and the seminal work of Unruh, Jacobson [337, 338] leading to
analog gravity [49, 50, 51].)

8.1 General Issues on Backreaction

Backreaction studies of quantum field processes in cosmological spacetimes have progressed further
than the corresponding black hole problems, partly because of relative technical simplicity associ-
ated with the higher symmetry of relevant cosmological background geometries. (For a summary
of the cosmological backreaction problem treated in the stochastic gravity theory, see [4]). The way
how the problem is set up and approached (e.g. via effective action) in these previously studied
models can be carried over to black hole problems. In fact since the interior of a black hole can be

61
described by a cosmological model (e.g. the Kantowski-Sachs universe for a spherically symmetric
black hole), some aspects even convey directly. The latest important work on this problem is that
of Hiscock, Larson and Anderson [47] on backreaction in the interior of a black hole, where one can
find a concise summary of earlier work.

8.1.1 Regularized Energy Momentum Tensor

The first step in a backreaction problem is to find a regularized energy momentum tensor of the
quantum fields using reasonable techniques since the expectation value of which serves as the source
in the semiclassical Einstein equation. For this, much work started in the 80’s (and still ongoing
sparingly) are concerned with finding the right approximations for the regularized energy momen-
tum tensor [339, 340, 341, 342, 343, 344, 47]. Even in the simplest spherically symmetric spacetime,
including the important Schwarzschild metric, is technically quite involved. To name a few of the
important landmarks in this endeavor (this is adopted from [47]), Howard and Candelas [345, 346]
have computed the stress-energy of a conformally invariant scalar field in the Schwarzschild ge-
ometry. Jensen and Ottewill [347] have computed the vacuum stress-energy of a massless vector
field in Schwarzschild. Approximation methods have been developed by Page, Brown, and Ottewill
[348, 349, 350] for conformally invariant fields in Schwarzschild spacetime, Frolov and Zel’nikov
[351] for conformally invariant fields in a general static spacetime, Anderson, Hiscock and Samuel
[342, 343] for massless arbitrarily coupled scalar fields in a general static spherically symmetric
spacetime. Furthermore the DeWitt-Schwinger approximation has been derived by Frolov and
Zel’nikov [352, 353] for massive fields in Kerr spacetime, Anderson Hiscock and Samuel [342, 343]
for a general (arbitrary curvature coupling and mass) scalar field in a general static spherically
symmetric spacetime and have applied their method to the Reissner-Nordström geometry [344].
Though arduous and demanding, the effort continues on because of its importance in finding the
backreaction effects of Hawking radiation on the evolution of black holes and the quantum structure
of spacetime.
Here we wish to address the black hole backreaction problem with new insights and methods
provided by stochastic gravity. (For the latest developments see reviews, e.g., [101, 2, 3, 4]). It is
not our intention to seek better approximations for the regularized energy momentum tensor, but
to point out new ingredients lacking in the existing semiclassical gravity framework. In particular
one needs to consider both the dissipation and the fluctuations aspects in the back reaction of
particle creation or vacuum polarization.
In a short note Hu, Raval and Sinha [140] first used the stochastic gravity formalism to address
the backreaction of evaporating black holes. A more detailed analysis is given by the recent work
of Hu and Roura [143, 354] For the class of quasi-static black holes, the formulation of the problem
in this new light was sketched out by Sinha Raval and Hu [142]. We follow these two latter works
in the stochastic gravity theory approach to the black hole fluctuations and backreaction problems.

8.1.2 Backreaction and Fluctuation-Dissipation relation

From the statistical field theory perspective provided by stochastic gravity one can understand that
backreaction effect is the manifestation of a fluctuation-dissipation relation [355, 356, 357, 358, 359,
360]. This was first conjectured by Candelas and Sciama [97, 138, 139] for a dynamic Kerr black hole
emitting Hawking radiation, and Mottola [98] for a static black hole (in a box) in quasi-equilibrium
with its radiation via linear response theory [361, 362, 363, 364, 365]. This postulate was shown to

62
hold for fully dynamical spacetimes. From cosmological backreaction problem Hu and Sinha [99]
derived a generalized fluctuation-dissipation relation relating dissipation (of anisotropy in Bianchi
Type I universes) and fluctuations (measured by particle numbers created in neighboring histories).
While the fluctuation-dissipation relation in a linear response theory captures the response of the
system (e.g., dissipation of the black hole) to the environment (in these cases the quantum matter
field) linear response theory (in the way it is commonly presented in statistical thermodynamics)
cannot provide a full description of self-consistent backreaction on at least two counts:
First, because it is usually based on the assumption of a specified background spacetime (static
in this case) and state (thermal) of the matter field(s) (e.g., [98]). The spacetime and the state of
matter should be determined in a self-consistent manner by their dynamics and mutual influence.
Second, the fluctuation part represented by the noise kernel is amiss (e.g., [205, 113]) This is also
a problem in the fluctuation-dissipation relation proposed by Candelas and Sciama [97, 138, 139]
(see below). As demonstrated by many authors [99, 104] backreaction is intrinsically a dynamic
process. The Einstein-Langevin equation in stochastic gravity overcomes both of these deficiencies.
For Candelas and Sciama [97, 138, 139], the classical formula they showed relating the dissipation
in area linearly to the squared absolute value of the shear amplitude is suggestive of a fluctuation-
dissipation relation. When the gravitational perturbations are quantized (they choose the quantum
state to be the Unruh vacuum) they argue that it approximates a flux of radiation from the hole
at large radii. Thus the dissipation in area due to the Hawking flux of gravitational radiation is
allegedly related to the quantum fluctuations of gravitons. The criticism in Ref. [140] is that their’s
is not a fluctuation-dissipation relation in the truly statistical mechanical sense because it does not
relate dissipation of a certain quantity (in this case, horizon area) to the fluctuations of the same
quantity. To do so would require one to compute the two point function of the area, which, being a
four-point function of the graviton field, is related to a two-point function of the stress tensor. The
stress tensor is the true “generalized force” acting on the spacetime via the equations of motion,
and the dissipation in the metric must eventually be related to the fluctuations of this generalized
force for the relation to qualify as a fluctuation-dissipation relation.

8.1.3 Noise and Fluctuations – the missing ingredient in older treatment

From this reasoning, we see that the vacuum expectation value of the stress energy bi-tensor,
known as the noise kernel, is the necessary new ingredient in addition to the dissipation kernel, and
stochastic gravity as an extension of semiclasical gravity is the appropriate framework for backre-
action considerations. The noise kernel for quantum fields in Minkowski and de Sitter spacetime
has been carried out by Martin, Roura and Verdaguer [5, 109, 134], for thermal fields in black hole
spacetime and scalar fields in general spacetimes by Campos, Hu and Phillips [100, 141, 7, 8].

8.2 Backreaction on black holes under quasi-static conditions

As an illustration of the application of stochastic gravity theory we outline the essential steps in a
black hole backreaction calculation, focusing on a more manageable quasi-static class. We adopt
the Hartle-Hawking picture [366] where the black hole is bathed eternally – actually in quasi-
thermal equilibrium – in the Hawking radiance it emits. It is described here by a massless scalar
quantum field at the Hawking temperature. As is well-known, this quasi-equilibrium condition is
possible only if the black hole is enclosed in a box of size suitably larger than the event horizon.
We can divide our consideration into the far field case and the near horizon case. Campos and Hu

63
[100, 141] have treated a relativistic thermal plasma in a weak gravitational field. Since the far
field limit of a Schwarzschild metric is just the perturbed Minkowski spacetime, one can perform a
perturbation expansion off hot flat space using the thermal Green functions [367]. Strictly speaking
the location of the box holding the black hole in equilibrium with its thermal radiation is as far
as one can go, thus the metric may not reach the perturbed Minkowski form. But one can also
put the black hole and its radiation in an anti-de Sitter space [368], which contains such a region.
Hot flat space has been studied before for various purposes. See e.g., [369, 370, 371, 372, 373].
Campos and Hu derived a stochastic CTP effective action and from it an equation of motion, the
Einstein Langevin equation, for the dynamical effect of a scalar quantum field on a background
spacetime. To perform calculations leading to the Einstein-Langevin equation one needs to begin
with a self-consistent solution of the semiclassical Einstein equation for the thermal field and the
perturbed background spacetime. For a black hole background, a semiclassical gravity solution is
provided by York [233, 234, 235]. For a Robertson-Walker background with thermal fields it is
given by Hu [374].
We follow the strategy outlined by Sinha, Raval and Hu [142] for treating the near horizon
case, following the same scheme of Campos and Hu. In both cases two new terms appear which are
absent in semiclassical gravity considerations: a nonlocal dissipation and a (generally colored) noise
kernel. When one takes the noise average one recovers York’s [233, 234, 235] semiclassical equations
for radially perturbed quasi-static black holes. For the near horizon case one cannot obtain the full
details yet, because the Green function for a scalar field in the Schwarzschild metric comes only in
an approximate form (e.g. Page approximation [348]), which, though reasonably accurate for the
stress tensor, fails poorly for the noise kernel [8]. In addition a formula is derived in [142] expressing
the CTP effective action in terms of the Bogoliubov coefficients. Since it measures not only the
number of particles created, but also the difference of particle creation in alternative histories, this
provides a useful avenue to explore the wider set of issues in black hole physics related to noise and
fluctuations.
Since backreaction calculations in semiclassical gravity has been under study for a much longer
time than in stochastic gravity we will concentrate on explaining how the new stochastic features
arise from the framework of semiclassical gravity, i.e., noise and fluctuations and their conse-
quences. Technically the goal is to obtain an influence action for this model of a black hole coupled
to a scalar field and to derive an Einstein-Langevin equation from it. As a by-product, from the
fluctuation-dissipation relation, one can derive the vacuum susceptibility function and the isother-
mal compressibility function for black holes, two quantities of fundamental interest in characterizing
the nonequilibrium thermodynamic properties of black holes.

8.2.1 The model

In this model the black hole spacetime is described by a spherically symmetric static metric with line
element of the following general form written in advanced time Eddington-Finkelstein coordinates
2m
 
ds2 = gµν dxµ dxν = −e2ψ 1 − dv 2 + 2e2ψ dvdr + r 2 dΩ2 , (8.1)
r
r
− 1 and dΩ2 is the line element on the two


where ψ = ψ(r) and m = m(r) , v = t + r + 2M ln 2M
sphere. Hawking radiation is described by a massless, conformally coupled quantum scalar field φ

64
with the classical action
1 √
Z
Sm [φ, gµν ] = − dn x −g[g µν ∂µ φ∂ν φ + ξ(n)Rφ2 ], (8.2)
2
(n−2)
where ξ(n) = 4(n−1) (n is the dimension of spacetime) and R is the curvature scalar of the spacetime
it lives in.
(0)
Let us consider linear perturbations hµν off a background Schwarzschild metric gµν
(0)
gµν = gµν + hµν , (8.3)

with standard line element

2M
 
(ds2 )0 = 1 − dv 2 + 2dvdr + r 2 dΩ2 . (8.4)
r
We look for this class of perturbed metrics in the form given by (8.1), (thus restricting our consid-
eration only to spherically symmetric perturbations):

eψ ≃ 1 + ǫρ(r), (8.5)

and
m ≃ M [1 + ǫµ(r)], (8.6)
ǫ 1 π 2
where λM 4 4 2
2 = 3 aTH ; a = 30 ; λ = 90(8 )π . TH is the Hawking temperature. This particular
parametrization of the perturbation is chosen following York’s [233, 234, 235] notation. Thus the
only non-zero components of hµν are

2M 2M ǫµ(r)
 
hvv = − (1 − )2ǫρ(r) + , (8.7)
r r
and
hvr = ǫρ(r). (8.8)
So this represents a metric with small static and radial perturbations about a Schwarzschild black
hole. The initial quantum state of the scalar field is taken to be the Hartle Hawking vacuum, which
is essentially a thermal state at the Hawking temperature and it represents a black hole in (unstable)
thermal equilibrium with its own Hawking radiation. In the far field limit, the gravitational field
is described by a linear perturbation from Minkowski spacetime. In equilibrium the thermal bath
can be characterized by a relativistic fluid with a four-velocity (time-like normalized vector field)
uµ , and temperature in its own rest frame β −1 .
To facilitate later comparisons with our program we briefly recall York’s work [233, 234, 235].
See also work by Hochberg and Kephart [299] for a massless vector field, Hochberg, Kephart and
York [300] for a massless spinor field, and Anderson, Hiscock, Whitesell, and York [301] for a
quantized massless scalar field with arbitrary coupling to spacetime curvature. York considered the
semiclassical Einstein equation
Gµν (gαβ ) = κhTµν i, (8.9)
(0) (0)
with Gµν ≃ Gµν +δGµν where Gµν is the Einstein tensor for the background spacetime. The zeroth
order solution gives a background metric in empty space, i.e, the Schwarzschild metric. δGµν is

65
the linear correction to the Einstein tensor in the perturbed metric. The semiclassical Einstein
equation in this approximation therefore reduces to
δGµν (g(0) , h) = κhTµν i. (8.10)
York solved this equation to first order by using the expectation value of the energy momentum
tensor for a conformally coupled scalar field in the Hartle-Hawking vacuum in the unperturbed
(Schwarzschild) spacetime on the right hand side and using (8.7) and (8.8) to calculate δGµν on
the left hand side. Unfortunately, no exact analytical expression is available for the hTµν i in a
Schwarzschild metric with the quantum field in the Hartle-Hawking vacuum that goes on the right
hand side. York therefore uses the approximate expression given by Page [348] which is known
to give excellent agreement with numerical results. Page’s approximate expression for hTµν i was
constructed using a thermal Feynman Green’s function obtained by a conformal transformation of
a WKB approximated Green’s function for an optical Schwarzschild metric. York then solves the
semiclassical Einstein equation (8.10) self consistently to obtain the corrections to the background
metric induced by the backreaction encoded in the functions µ(r) and ρ(r). There was no mention of
fluctuations or its effects. As we shall see, in the language of the previous section, the semiclassical
gravity procedure which York followed working at the equation of motion level is equivalent to
looking at the noise-averaged backreaction effects.

8.2.2 CTP Effective Action for the Black Hole

We first derive the CTP effective action for the model described in the previous section. Using the
metric (8.4) (and neglecting the surface terms that appear in an integration by parts) we have the
action for the scalar field written perturbatively as
1
Z q h i
Sm [φ, hµν ] = dn x −g(0) φ ✷(0) + V (1) + V (2) + · · · φ, (8.11)
2
where the first and second order perturbative operators V (1) and V (2) are given by
1
 q  
V (1)≡− q ∂µ −g(0) h̄µν ∂ν + h̄µν ∂µ ∂ν +ξ(n)R(1) ,
−g (0)
1 1
 q  
V (2)≡−q ∂µ −g(0) ĥµν ∂ν + ĥµν ∂µ ∂ν −ξ(n)(R(2) + hR(1) ) . (8.12)
−g (0) 2

In the above expressions, R(k) is the k-order term in the perturbation hµν (x) of the scalar curvature
R and h̄µν and ĥµν denote a linear and a quadratic combination of the perturbation, respectively,
1 (0)
h̄µν ≡ hµν − hgµν ,
2
1 1 1
ĥµν ≡ hµα hαν − hhµν + h2 gµν
(0)
− hαβ hαβ gµν
(0)
. (8.13)
2 8 4
From quantum field theory in curved spacetime considerations discussed above we take the following
action for the gravitational field:
1 αµ̄n−4
Z q Z q
Sg [gµν ]= n−2 dn x −g(x)R(x) + dn x −g(x)
(16πG) 2 4(n − 4)

66
 ( " 2 # )
1

µναβ 2
× 3Rµναβ (x)R (x)− 1−360 ξ(n)− R (x) . (8.14)
6

The first term is the classical Einstein-Hilbert action and the second term is the counterterm in
four dimensions used to renormalize the divergent effective action. In this action ℓ2P = 16πGN ,
α = (2880π 2 )−1 and µ̄ is an arbitrary mass scale.
We are interested in computing the CTP effective action (8.11) for the matter action and when
the field φ is initially in the Hartle-Hawking vacuum. This is equivalent to saying that the initial
state of the field is described by a thermal density matrix at a finite temperature T = TH . The CTP
effective action at finite temperature T ≡ 1/β for this model is given by (for details see [100, 141])

β ± + − i β ±
Srmef f [hµν ] = Sg [hµν ] − Sg [hµν ] − T r{ln Ḡab [hµν ]}, (8.15)
2

where ± denote the forward and backward time path of the CTP formalism and Ḡβab [h±
µν ] is the
complete 2 × 2 matrix propagator (a and b take ± values: G++ , G+− and G−− correspond to
the Feynman, Wightman and Schwinger q Greens functions respectively) with thermal boundary
conditions for the differential operator −g(0) (✷ + V (1) + V (2) + · · ·). The actual form of Ḡβab
cannot be explicitly given. However, it is easy to obtain a perturbative expansion in terms of
(k) (k) (k)
Vab , the k-order matrix version of the complete differential operator defined by V±± ≡ ±V±
(k)
and V±∓ ≡ 0, and Gβab , the thermal matrix propagator for a massless scalar field in Schwarzschild
spacetime . To second order Ḡβab reads,
(1) (2) (1) (1)
Ḡβab = Gβab − Gβac Vcd Gβdb − Gβac Vcd Gβdb + Gβac Vcd Gβde Vef Gβfb + · · · (8.16)

Expanding the logarithm and dropping one term independent of the perturbation h±
µν (x), the CTP
effective action may be perturbatively written as,
β
Seff [h± + −
µν ]=Sg [hµν ]−Sg [hµν ]
i (1) (1) (2) (2)
+ T r[V+ Gβ++ −V− Gβ−− +V+ Gβ++ −V− Gβ−− ]
2
i (1) (1) (1) (1)
− T r[V+ Gβ++ V+ Gβ++ +V− Gβ−− V− Gβ−−
4
(1) (1)
−2V+ Gβ+− V− Gβ−+ ]. (8.17)

In computing the traces, some terms containing divergences are canceled using counterterms intro-
duced in the classical gravitational action after dimensional regularization.

8.2.3 Near Flat Case

At this point we divide our considerations into two cases. In the far field limit hµν represent
(0)
perturbations about flat space, i.e., gµν = ηµν . The exact “unperturbed” thermal propagators
for scalar fields are known, i.e., the Euclidean propagator with periodicity β. Using the Fourier
transformed form (those quantities are denoted with a tilde) of the thermal propagators G̃βab (k) ,

67
 (1) (1)
the trace terms of the form T r[Va Gβmn Vb Gβrs ] can be written as [100, 141],
Z
(1)
T r[Va(1) Gβmn Vb Gβrs ] = dn xdn x′ haµν (x)hbαβ (x′ )
dn k dn q ik·(x−x′ ) β
Z
× e G̃mn (k + q)G̃βrs (q)Tµν,αβ (q, k), (8.18)
(2π)n (2π)n

where the tensor Tµν,αβ (q, k) is defined in [100, 141] after an expansion in terms of a basis of
14 tensors [370, 371]. In particular, the last trace of (8.17) may be split in two different kernels
Nµν,αβ (x − x′ ) and Dµν,αβ (x − x′ ),
i (1) (1)
T r[V+ Gβ+− V− Gβ−+ ] =
2 Z
− d4 xd4 x′ h+ − ′
µν (x)hαβ (x )[D
µν,αβ
(x − x′ ) + iNµν,αβ (x − x′ )]. (8.19)

One can express the Fourier transforms of these kernels, respectively, as

d4 q
Z
µν,αβ
Ñ (k)=π 2 {θ(ko + q o )θ(−q o ) + θ(−ko − q o )θ(q o )
(2π)4
+nβ (|q o |) + nβ (|ko + q o |)
+2nβ (|q o |)nβ (|ko + q o |)} δ(q 2 )δ[(k + q)2 ]Tµν,αβ (q, k), (8.20)

d4 q
Z
µν,αβ 2
D̃ (k)=−iπ {θ(ko + q o )θ(−q o ) − θ(−ko − q o )θ(q o )
(2π)4
+sg(ko + q o )nβ (|q o |)
−sg(q o )nβ (|ko + q o |)} δ(q 2 )δ[(k + q)2 ]Tµν,αβ (q, k). (8.21)

Using the property Tµν,αβ (q, k) = Tµν,αβ (−q, −k), it is easy to see that the kernel Nµν,αβ (x − x′ ) is
symmetric and Dµν,αβ (x − x′ ) antisymmetric in their arguments; that is, Nµν,αβ (x) = Nµν,αβ (−x)
and Dµν,αβ (x) = −Dµν,αβ (−x).
The physical meanings of these kernels can be extracted if we write the renormalized CTP
effective action at finite temperature (8.17) in an influence functional form [375, 376, 377, 378]. N,
the imaginary part of the CTP effective action can be identified with the noise kernel and D, the
antisymmetric piece of the real part, with the dissipation kernel. Campos and Hu [100, 141] have
shown that these kernels identified as such indeed satisfy a thermal fluctuation-dissipation relation.
If we denote the difference and the sum of the perturbations h± µν defined along each branch C±
of the complex time path of integration C by [hµν ] ≡ h+ µν − h− and {h } ≡ h+ + h− , respectively,
µν µν µν µν
the influence functional form of the thermal CTP effective action may be written to second order
in hµν as,
1
Z
β
Seff [h±
µν ] ≃ d4 x d4 x′ [hµν ](x)Lµν,αβ
(o) (x − x′ ){hαβ }(x′ )
2(16πGN )
1
Z
µν
+ d4 x [hµν ](x)T(β)
2
1
Z
+ d4 x d4 x′ [hµν ](x)Hµν,αβ (x − x′ ){hαβ }(x′ )
2

68
 1
Z
− d4 x d4 x′ [hµν ](x)Dµν,αβ (x − x′ ){hαβ }(x′ )
2
i
Z
+ d4 x d4 x′ [hµν ](x)Nµν,αβ (x − x′ )[hαβ ](x′ ). (8.22)
2
µν,αβ
The first line is the Einstein-Hilbert action to second order in the perturbation h±
µν (x). L(o) (x)
is a symmetric kernel (i.e. Lµν,αβ
(o) (x) = Lµν,αβ
(o) (−x)). In the near flat case its Fourier transform is
given by
µν,αβ 1 h 2 µν,αβ
L̃(o) (k) = −k T1 (q, k) + 2k2 Tµν,αβ
4 (q, k)
4 i
+Tµν,αβ
8 (q, k) − 2T µν,αβ
13 (q, k) . (8.23)

The fourteen elements of the tensor basis Tµν,αβ

i (q, k) (i = 1, · · · , 14) are defined in [370, 371]. The
±
second is a local term linear in hµν (x). Only when far away from the hole that it takes the form of
the stress tensor of massless scalar particles at temperature β −1 , which has the form of a perfect
fluid stress-energy tensor
π2 1
 
µν
T(β) = uµ uν + (η µν + uµ uν ) , (8.24)
30β 4 3
π 2
where uµ is the four-velocity of the plasma and the factor 30β 4 is the familiar thermal energy
−1
density for massless scalar particles at temperature β . In the far field limit, taking into account
the four-velocity uµ of the fluid, a manifestly Lorentz-covariant approach to thermal field theory
may be used [379]. However, in order to simplify the involved tensorial structure we work in the
co-moving coordinate system of the fluid where uµ = (1, 0, 0, 0). In the third line, the Fourier
transform of the symmetric kernel Hµν,αβ (x) can be expressed as
( )
µν,αβ αk 4 1 |k2 | µν,αβ 1 µν,αβ
H̃ (k) =− ln 2 Q (k) + Q̄ (k)
4 2 µ 3
π2 n
+ 4
−Tµν,αβ
1 (u, k) − 2Tµν,αβ
2 (u, k)
180β
o
+Tµν,αβ
4 (u, k) + 2Tµν,αβ
5 (u, k)
ξ n 2 µν,αβ
+ k T1 (u, k) − 2k2 Tµν,αβ
4 (u, k)
96β 2
o
−Tµν,αβ
8 (u, k) + 2Tµν,αβ
13 (u, k)
d4 q 1
Z   
+π δ(q 2 )nβ (|q o |)P
(2π)4 (k + q)2
1
 
2 o o
+δ[(k+q) ]nβ (|k +q |)P 2 ×Tµν,αβ (q, k), (8.25)
q
23
where µ is a simple redefinition of the renormalization parameter µ̄ given by µ ≡ µ̄ exp( 15 + 21 ln 4π−
1 µν,αβ µν,αβ
2 γ), and the tensors Q (k) and Q̄ (k) are defined, respectively, by
3 1 2
 
Qµν,αβ (k)= Tµν,αβ
1 (q, k) − 2 Tµν,αβ (q, k) + 4 Tµν,αβ (q, k)
2 k 8 k 12

69
 1 1 1
 
−[1−360(ξ − )2 ] Tµν,αβ
4 (q, k)+ 4 Tµν,αβ (q, k)− 2 Tµν,αβ (q, k) , (8.26)
6 k 12 k 13

µν,αβ 1 1 n
Q̄ (k)=[1 + 576(ξ − )2 − 60(ξ − )(1 − 36ξ ′ )] Tµν,αβ
4 (q, k)
6 6
1 1

+ 4 Tµν,αβ
12 (q, k) − 2 Tµν,αβ (q, k) . (8.27)
k k 13
In the above and subsequent equations, we denote the coupling parameter in four dimensions ξ(4) by
µν,αβ
ξ and consequently ξ ′ means dξ(n)/dn evaluated at n = 4. H̃ (k) is the complete contribution of
a free massless quantum scalar field to the thermal graviton polarization tensor[370, 371, 372, 373]
and it is responsible for the instabilities found in flat spacetime at finite temperature [369, 370,
371, 372, 373]. Note that the addition of the contribution of other kinds of matter fields to the
effective action, even graviton contributions, does not change the tensor structure of these kernels
and only the overall factors are different to leading order [370, 371]. Eq. (8.25) reflects the fact
µν,αβ
that the kernel H̃ (k) has thermal as well as non-thermal contributions. Note that it reduces
to the first term in the zero temperature limit (β → ∞)
( )
µν,αβ αk 4 1 |k2 | µν,αβ 1 µν,αβ
H̃ (k) ≃ − ln 2 Q (k) + Q̄ (k) . (8.28)
4 2 µ 3

and at high temperatures the leading term (β −4 ) may be written as

14
µν,αβ π2 X
H̃ (k) ≃ Hi (r)Tµν,αβ
i (u, K), (8.29)
30β 4 i=1

where we have introduced the dimensionless external momentum K µ ≡ kµ /|~k| ≡ (r, k̂). The Hi (r)
coefficients were first given in [370, 371] and generalized to the next-to-leading order (β −2 ) in
[372, 373]. (They are given with the MTW sign convention in [100, 141].)
Finally, as defined above, Nµν,αβ (x) is the noise kernel representing the random fluctuations of
the thermal radiance and Dµν,αβ (x) is the dissipation kernel, describing the dissipation of energy
of the gravitational field.

8.2.4 Near Horizon Case

In this case, since the perturbation is taken around the Schwarzschild spacetime, exact expressions
for the corresponding unperturbed propagators Gβab [h± µν ] are not known. Therefore apart from the
approximation of computing the CTP effective action to certain order in perturbation theory, an
appropriate approximation scheme for the unperturbed Green’s functions is also required. This
feature manifested itself in York’s calculation of backreaction as well, where, in writing the hTµν i
on the right hand side of the semiclassical Einstein equation in the unperturbed Schwarzschild
metric, he had to use an approximate expression for hTµν i in the Schwarzschild metric given by
Page [348]. The additional complication here is that while to obtain hTµν i as in York’s calculation,
the knowledge of only the thermal Feynman Green’s function is required, to calculate the CTP
effective action one needs the knowledge of the full matrix propagator, which involves the Feynman,
Schwinger and Wightman functions.

70
 It is indeed possible to construct the full thermal matrix propagator Gβab [h± µν ] based on Page’s
approximate Feynman Green’s function by using identities relating the Feynman Green’s function
with the other Green’s functions with different boundary conditions. One can then proceed to
explicitly compute a CTP effective action and hence the influence functional based on this approx-
imation. However, we desist from delving into such a calculation for the following reason. Our
main interest in performing such a calculation is to identify and analyze the noise term which is
the new ingredient in the backreaction. We have mentioned that the noise term gives a stochastic
contribution ξ µν to the Einstein-Langevin equation (3.14). We had also stated that this term is
2 i. However, a calculation of
related to the variance of fluctuations in Tµν , i.e, schematically, to hTµν
2
hTµν i in the Hartle-Hawking state in a Schwarzschild background using the Page approximation was
performed by Phillips and Hu [7, 8] and it was shown that though the approximation is excellent as
2 i at the horizon. In fact, similar
far as hTµν i is concerned, it gives unacceptably large errors for hTµν
errors will be propagated in the non-local dissipation term as well, because both terms originate
from the same source, that is, they come from the last trace term in (8.17) which contains terms
quadratic in the Green’s function. However, the Influence Functional or CTP formalism itself does
not depend on the nature of the approximation, so we will attempt to exhibit the general structure
of the calculation without resorting to a specific form for the Greens function and conjecture on
what is to be expected. A more accurate computation can be performed using this formal structure
once a better approximation becomes available.
The general structure of the CTP effective action arising from the calculation of the traces in
equation (8.17) remains the same. But to write down explicit expressions for the non-local kernels
one requires the input of the explicit form of Gβab [h± µν ] in the Schwarzschild metric, which is not
available in closed form. We can make some general observations about the terms in there. The
first line containing L does not have an explicit Fourier representation as given in the far field
µν
case, neither will T(β) in the second line representing the zeroth order contribution to hTµν i have
a perfect fluid form. The third and fourth terms containing the remaining quadratic component
of the real part of the effective action will not have any simple or even complicated analytic form.
The symmetry properties of the kernels H µν,αβ (x, x′ ) and D µν,αβ (x, x′ ) remain intact, i.e., they are
respectively even and odd in x, x′ . The last term in the CTP effective action gives the imaginary
part of the effective action and the kernel N (x, x′ ) is symmetric.
Continuing our general observations from this CTP effective action, using the connection be-
tween this thermal CTP effective action to the influence functional [18, 102] via an equation in the
schematic form (4.3). We see that the nonlocal imaginary term containing the kernel N µν,αβ (x, x′ )
is responsible for the generation of the stochastic noise term in the Einstein-Langevin equation
and the real non-local term containing kernel D µν,αβ (x, x′ ) is responsible for the non-local dissi-
pation term. To derive the Einstein-Langevin equation we first construct the stochastic effective
action (4.13). We then derive the equation of motion, as shown earlier in (4.15), by taking its
functional derivative with respect to [hµν ] and equating it to zero. With the identification of noise
and dissipation kernels, one can write down a linear, non-local relation of the form,
Z
N (t − t′ ) = d(s − s′ )K(t − t′ , s − s′ )γ(s − s′ ), (8.30)

where D(t, t′ ) = −∂t′ γ(t, t′ ). This is the general functional form of a fluctuation-dissipation re-
lation and K(t, s) is called the fluctuation-dissipation kernel [375, 376, 377, 378]. In the present
context this relation depicts the backreaction of thermal Hawking radiance for a black hole in

71
quasi-equilibrium.

8.2.5 Einstein-Langevin equation

In this section we show how a semiclassical Einstein-Langevin equation can be derived from the
previous thermal CTP effective action. This equation depicts the stochastic evolution of the per-
turbations of the black hole under the influence of the fluctuations of the thermal scalar field.
The influence functional FIF ≡ exp(iSIF ) previously introduced in Eq. (4.2) can be written
β
in terms of the the CTP effective action Seff [h±
µν ] derived in equation (8.22) using Eq. (4.3). The
Einstein-Langevin equation follows from taking the functional derivative of the stochastic effective
action (4.13) with respect to [hµν ](x) and imposing [hµν ](x) = 0. This leads to

1 1 µν
Z Z 
d x 4 ′
Lµν,αβ
(o) (x ′ ′
− x )hαβ (x ) + T(β) + d4 x′ Hµν,αβ (x − x′ )
16πGN 2

−Dµν,αβ (x − x′ ) hαβ (x′ ) + ξ µν (x) = 0. (8.31)

where
hξ µν (x)ξ αβ (x′ )ij = Nµν,αβ (x − x′ ), (8.32)
In the far field limit this equation should reduce to that obtained by Campos and Hu [100, 141]:
For gravitational perturbations hµν defined in (8.13) under the harmonic gauge h̄µν,ν = 0, their
Einstein-Langevin equation is given by
1
 Z 
µν µν
✷h̄ (x) + T(β) + 2Pρσ,αβ d4 x′ Hµν,αβ (x − x′ )
16πG2N
 o
−Dµν,αβ (x − x′ ) h̄ρσ (x′ ) + 2ξ µν (x) = 0, (8.33)

where the tensor Pρσ,αβ is given by

1
Pρσ,αβ = (ηρα ησβ + ηρβ ησα − ηρσ ηαβ ) . (8.34)
2
The expression for Pρσ,αβ in the near horizon limit of course cannot be expressed in such a simple
form. Note that this differential stochastic equation includes a non-local term responsible for the
dissipation of the gravitational field and a noise source term which accounts for the fluctuations
of the quantum field . Note also that this equation in combination with the correlation for the
stochastic variable (8.32) determine the two-point correlation for the stochastic metric fluctuations
hh̄µν (x)h̄αβ (x′ )iξ self-consistently.
As we have seen before and here, the Einstein-Langevin equation is a dynamical equation
governing the dissipative evolution of the gravitational field under the influence of the fluctuations
of the quantum field, which, in the case of black holes, takes the form of thermal radiance. From
its form we can see that even for the quasi-static case under study the back reaction of Hawking
radiation on the black hole spacetime has an innate dynamical nature.
For the far field case making use of the explicit forms available for the noise and dissipation
kernels Campos and Hu [100, 141] formally proved the existence of a fluctuation-dissipation relation
at all temperatures between the quantum fluctuations of the thermal radiance and the dissipation of
the gravitational field. They also showed the formal equivalence of this method with linear response

72
theory for lowest order perturbations of a near-equilibrium system, and how the response functions
such as the contribution of the quantum scalar field to the thermal graviton polarization tensor
can be derived. An important quantity not usually obtained in linear response theory, but of equal
importance, manifest in the CTP stochastic approach is the noise term arising from the quantum
and statistical fluctuations in the thermal field. The example given in this section shows that the
back reaction is intrinsically a dynamic process described (at this level of sophistication) by the
Einstein-Langevin equation. By comparison, traditional linear response theory calculations cannot
capture the dynamics as fully and thus cannot provide a complete description of the backreaction
problem.

8.2.6 Comments
As remarked earlier, except for the near-flat case, an analytic form of the Green function is not
available. Even the Page approximation [348] which gives unexpectedly good results for the stress
energy tensor has been shown to fail in the fluctuations of the energy density [8]. Thus using such an
approximation for the noise kernel will give unreliable results for the Einstein-Langevin equation.
If we confine ourselves to Page’s approximation and derive the equation of motion without the
stochastic term, we expect to recover York’s semiclassical Einstein’s equation if one retains only
the zeroth order contribution, i.e, the first two terms in the expression for the CTP effective action
in Eq. (8.22). Thus, this offers a new route to arrive at York’s semiclassical Einstein’s equations.
(Not only is it a derivation of York’s result from a different point of view, but it also shows how
his result arises as an appropriate limit of a more complete framework, i.e, it arises when one
averages over the noise.) Another point worth noting is that a non-local dissipation term arises
from the fourth term in Eq. (8.22) in the CTP effective action which is absent in York’s treatment.
This difference exists primarily due to the difference in the way backreaction is treated, at the
level of iterative approximations on the equation of motion as in York, versus the treatment at
the effective action level as in the influence functional approach. In York’s treatment, the Einstein
tensor is computed to first order in perturbation theory, while hTµν i on the right hand side of the
semiclassical Einstein equation is replaced by the zeroth order term. In the influence functional
treatment the full effective action is computed to second order in perturbation, and hence includes
the higher order non-local terms.
The other important conceptual point that comes to light from this new approach is that
related to the Fluctuation-Dissipation Relation. In the quantum Brownian motion analog (e.g.,
[375, 376, 377, 378] and references therein), the dissipation of the energy of the Brownian particle
as it approaches equilibrium and the fluctuations at equilibrium are connected by the Fluctuation-
Dissipation relation. Here the backreaction of quantum fields on black holes also consists of two
forms – dissipation and fluctuation or noise – corresponding to the real and imaginary parts of
the influence functional as embodied in the dissipation and noise kernels. A fluctuation-dissipation
relation has been shown to exist for the near flat case by Campos and Hu [100, 141] and is expected
to exist between the noise and dissipation kernels for the general case, as it is a categorical relation
[375, 376, 377, 378, 101]. Martin and Verdaguer have also proved the existence of a fluctuation-
dissipation relation when the semiclassical background is a stationary spacetime and the quantum
field is in thermal equilibrium. Their result was then extended to a conformal field in a conformally
stationary background [5]. As discussed earlier the existence of a fluctuation-dissipation relation for
the black hole case has been suggested by some authors previously [97, 138, 139, 98]. This relation

73
and the relevant physical quantities contained therein, such as the black hole susceptibility function
which characterizes the statistical mechanical and dynamical responses of a black hole interacting
with its quantum field environment, will allow us to study the nonequilibrium thermodynamic
properties of the black hole, and through it perhaps the microscopic structure of spacetime.
There are limitations of a technical nature in the quasi-static case studied, as mentioned above,
i.e., there is no reliable approximation to the Schwarzschild thermal Green’s function to explicitly
compute the noise and dissipation kernels. Another technical limitation of this example is the
following. Although we have allowed for backreaction effects to modify the initial state in the sense
that the temperature of the Hartle-Hawking state gets affected by the backreaction, our analysis is
essentially confined to a Hartle-Hawking thermal state of the field. It does not directly extend to
a more general class of states, for example to the case where the initial state of the field is in the
Unruh vacuum. To study the dynamics of a radiating black hole under the influence of a quantum
field and its fluctuations a different model and approach are needed which we now discuss.

8.3 Metric Fluctuations of an Evaporating Black Hole

At the semiclassical gravity level of description, black hole evaporation results from the backreaction
of particle production in Hawking effect. This is believed to be valid at least before the Planckian
scale is reached [298, 341]. However, as is explained above, semiclassical gravity [32, 34, 229] is
a mean field description that neglects the fluctuations of the spacetime geometry. A number of
studies have suggested the existence of large fluctuations near black hole horizons [380, 123, 381,
382] (and even instabilities [383]) with characteristic time-scales much shorter than the black hole
evaporation time. For example, Casher et al [381] and Sorkin [123, 124] have concentrated on the
issue of fluctuations of the horizon induced by a fluctuating metric. Casher et al [381] considers
the fluctuations of the horizon induced by the “atmosphere” of high angular momentum particles
near the horizon, while Sorkin [123, 124] calculates fluctuations of the shape of the horizon induced
by the quantum field fluctuations under a Newtonian approximations. A relativistic generalization
of this vein is given by Marolf [382]. Both groups of authors came to the conclusion that horizon
fluctuations become large at scales much larger than the Planck scale (note Ford and Svaiter [121]
later presented results contrary to this claim). Though these works do deal with backreaction, the
fluctuations considered do not arise as an explicit stochastic noise term as in stochastic gravity.
Either states which are singular on the horizon (such as the Boulware vacuum for Schwarzschild
spacetime) were explicitly considered, or fluctuations were computed with respect to those states
and found to be large near the horizon. Whether these huge fluctuations are of a generic nature or
an artifact from the consideration of states singular on the horizon is an issue that deserves further
investigation. By contrast, the fluctuations for states regular on the horizon were estimated in
Ref. [122] and found to be small even when integrated over a time of the order of the evaporation
time. These apparently contradictory claims and the fact that most claims on black hole horizon
fluctuations were based on qualitative arguments and/or semi-quantitative estimates indicate that
a more quantitative and self-consistent description is needed. This is what stochastic gravity theory
can provide.
Such a program of research has been pursued rigorously by Hu and Roura (HR) [144, 354] In
contrast to the claims made before, they find that even for states regular on the horizon the accu-
mulated fluctuations become significant by the time the black hole mass has changed substantially,
but well before reaching the Planckian regime. This result is different from those obtained in prior

74
studies, but in agreement with earlier work by Bekenstein [384]. The apparent difference from
the conclusions drawn in the earlier work of Hu, Raval and Sinha [140], which was also based on
stochastic gravity, will be explained later. We begin with the evolution of the mean geometry.

8.3.1 Evolution of Mean Geometry of an Evaporating Black Hole

Backreaction of the Hawking radiation emitted by the black hole on the dynamics of spacetime
geometry has been studied in some detail for spherically symmetric black holes [298, 341]. For a
general spherically-symmetric metric there always exists a system of coordinates in which it takes
the form
 
ds2 = −e2ψ(v,r) (1 − 2m(v, r)/r)dv 2 + 2eψ(v,r) dvdr + r 2 dθ 2 + sin2 θdϕ2 . (8.35)

In general this metric exhibits an apparent horizon, where the expansion of the outgoing radial
null geodesics vanishes and which separates regions with positive and negative expansion for those
geodesics, at those radii that correspond to (odd degree) zeroes of the vv metric component. We
mark the location of the apparent horizon by rAH (v) = 2M (v), where M (v) satisfies the equation
2m(2M (v), v) = 2M (v).
The non-zero components of the semiclassical Einstein equation associated with the metric in
Eq. (8.35) become

∂m
= 4πr 2 hTvr i, (8.36)
∂v
∂m
= −4πr 2 hTvv i, (8.37)
∂r
∂ψ
= 4πrhTrr i, (8.38)
∂r
where in the above and henceforth we use hTµν i to denote the renormalized or regularized vacuum
expectation value of the stress energy tensor hT̂µν [g]iren and employ Planckian units (with m2p = 1).
Solving Eqs. (8.36)-(8.38) is not easy. However, one can introduce a useful adiabatic approxi-
mation in the regime where the mass of the black hole is much larger than the Planck mass, which
is in any case a necessary condition for the semiclassical treatment to be valid. What this entails
is that when M ≫ 1 (remember that we are using Planckian units) for each value of v one can
simply substitute hTµν i by its “parametric value” – by this we mean the expectation value of the
stress energy tensor of the quantum field in a Schwarzschild black hole with a mass corresponding
to M (v) evaluated at that value of v. This is in contrast to its dynamical value, which should be
determined by solving self-consistently the semiclassical Einstein equation for the spacetime metric
and the equations of motion for the quantum matter fields. This kind of approximation introduces
errors of higher order in LH ≡ B/M 2 (B is a dimensionless parameter that depends on the number
of massless fields and their spins and accounts for their corresponding grey-body factors; it has
been estimated to be of order 10−4 [385]), which are very small for black holes well above Planckian
scales. These errors are due to the fact that M (v) is not constant and that, even for a constant
M (v), the resulting static geometry is not exactly Schwarzschild because the vacuum polarization
of the quantum fields gives rise to a non-vanishing hT̂ab [g]iren [234].
The expectation value of the stress tensor for Schwarzschild spacetime has been found to corre-
spond to a thermal flux of radiation (with hTvr i = LH /(4πr 2 )) for large radii and of order LH near

75
the horizon [386, 348, 345, 346, 343]. This shows the consistency of the adiabatic approximation
for LH ≪ 1: the right-hand side of Eqs. (8.36)-(8.38) contains terms of order LH and higher, so
that the derivatives of m(v, r) and ψ(v, r) are indeed small. We note that the natural quantum
state for a black hole formed by gravitational collapse is the Unruh vacuum, which corresponds to
the absence of incoming radiation far from the horizon. The expectation value of the stress tensor
operator for that state is finite on the future horizon of Schwarzschild, which is the relevant one
when identifying a region of the Schwarzschild geometry with the spacetime outside the collapsing
matter for a black hole formed by gravitational collapse.
One can use the v component of the stress-energy conservation equation

∂ r 2 hTvr i ∂hTvv i


+ r2 = 0, (8.39)
∂r ∂v
to relate the hTvr i components on the horizon and far from it. Integrating Eq. (8.39) radially, one
gets
(r 2 hTvr i)(r = 2M (v), v) = (r 2 hTvr i)(r ≈ 6M (v), v) + O(L2H ), (8.40)
where we considered a radius sufficiently far from the horizon, but not arbitrarily far (i.e. 2M (v) ≪
r ≪ M (v)/LH ).Hence, Eq. (8.40) relates the positive energy flux radiated away far from the
horizon and the negative energy flux crossing the horizon. Taking into account this connection
between energy fluxes and evaluating Eq. (8.36) on the apparent horizon, we finally get the equation
governing the evolution of its size:
dM B
= − 2. (8.41)
dv M
Unless M (v) is constant, the event horizon and the apparent horizon do not coincide. However,
in the adiabatic regime their radii are related, differing by a quantity of higher order in LH :
rEH (v) = rAH (v) (1 + O(LH )).

8.3.2 Spherically-symmetric induced fluctuations

We now consider metric fluctuations around a background metric gab that corresponds to a given
solution of semiclassical gravity. Their dynamics is governed by the Einstein-Langevin equation
[103, 99, 104, 107] D E
(1) (1)
Gab [g + h] = κ T̂ab [g + h] + κ ξab [g] , (8.42)
ren
(The superindex (1) indicates that only the terms linear in the metric perturbations should be
considered) Here ξab is a Gaussian stochastic source with vanishing expectation value and correlation
function hξab (x)ξcd (x′ )iξ = (1/2)h{t̂ab (x), t̂cd (x′ )}i (with t̂ab ≡ T̂ab −hT̂ab i), known as the noise kernel
and denoted by Nabcd (x, x′ ).
As explained earlier the symmetrized two-point function consists of two contributions: intrinsic
and induced fluctuations. The intrinsic fluctuations are a consequence of the quantum width of the
initial state of the metric perturbations, and they are obtained in stochastic gravity by averaging
over the initial conditions for the solutions of the homogeneous part of Eq. (8.42) distributed
according to the reduced Wigner function associated with the initial quantum state of the metric
perturbations. On the other hand, the induced fluctuations are due to the quantum fluctuations of
the matter fields interacting with the metric perturbations, and they are obtained by solving the
Einstein-Langevin equation using a retarded propagator with vanishing initial conditions.

76
 In this section we study the spherically-symmetric sector [i.e., the monopole contribution, which
corresponds to l = 0, in a multipole expansion in terms of spherical harmonics Ylm (θ, φ)] of metric
fluctuations for an evaporating black hole. Restricting one’s attention to the spherically-symmetric
sector of metric fluctuations necessarily implies a partial description of the fluctuations because,
contrary to the case for semiclassical gravity solutions, even if one starts with spherically-symmetric
initial conditions, the stress tensor fluctuations will induce fluctuations involving higher multipoles.
Thus, the multipole structure of the fluctuations is far richer than that of spherically-symmetric
semiclassical gravity solutions, but this also means that obtaining a complete solution (including
all multipoles) for fluctuations rather than the mean value is much more difficult. For spherically
symmetric fluctuations only induced fluctuations are possible. The fact that intrinsic fluctuations
cannot exist can be clearly seen if one neglects vacuum polarization effects, since Birkhoff’s theorem
forbids the existence of spherically-symmetric free metric perturbations in the exterior vacuum
region of a spherically-symmetric black hole that keep the ADM mass constant. (This fact rings
an alarm in the approach taken in Ref. [387] to the black hole fluctuation problem. The degrees of
freedom corresponding to spherically-symmetric perturbations are constrained by the Hamiltonian
and momentum constraints both at the classical and quantum level. Therefore, they will not
exhibit quantum fluctuations unless they are coupled to a quantum matter field.) Even when
vacuum polarization effects are included, spherically-symmetric perturbations, characterized by
m(v, r) and ψ(v, r), are not independent degrees of freedom. This follows from Eqs. (8.36)-(8.38),
which can be regarded as constraint equations.
Considering only spherical symmetry fluctuations is a simplification but it should be emphasized
that it gives more accurate results than two-dimensional dilaton-gravity models resulting from
simple dimensional reduction [388, 389, 390]. This is because we project the solutions of the
Einstein-Langevin equation just at the end, rather than considering only the contribution of the
s-wave modes to the classical action for both the metric and the matter fields from the very
beginning. Hence, an infinite number of modes for the matter fields with l 6= 0 contribute to the
l = 0 projection of the noise kernel, whereas only the s-wave modes for each matter field would
contribute to the noise kernel if dimensional reduction had been imposed right from the start, as
done in Refs. [128, 391, 392] as well as in studies of two-dimensional dilaton-gravity models.
The Einstein-Langevin equation for the spherically-symmetric sector of metric perturbations can
be obtained by considering linear perturbations of m(v, r) and ψ(v, r), projecting the stochastic
source that accounts for the stress tensor fluctuations to the l = 0 sector, and adding it to the
right-hand side of Eqs. (8.36)-(8.38). We will focus our attention on the equation for the evolution
of η(v, r), the perturbation of m(v, r):

∂(m + η) B 2 r

2

=− + 4πr ξ v + O L H , (8.43)
∂v (m + η)2

which reduces, after neglecting terms of order L2H or higher, to the following equation to linear
order in η:
∂η 2B
= 3 η + 4πr 2 ξvr . (8.44)
∂v m
It is important to emphasize that in Eq. (8.43) we assumed that the change in time of η(v, r) is
sufficiently slow so that the adiabatic approximation employed in the previous section to obtain
the mean evolution of m(v, r) can also be applied to the perturbed quantity m(v, r) + η(v, r). This

77
is guaranteed as long as the term corresponding to the stochastic source is of order LH or higher,
a point which will be discussed below.
A more serious issue raised by HR is that in most previous investigations [384, 122] of the
problem of metric fluctuations driven by quantum matter field fluctuations of states regular on the
horizon (as far as the expectation value of the stress tensor is concerned) most authors assumed the
existence of correlations between the outgoing energy flux far from the horizon and a negative energy
flux crossing the horizon. (See, however, Refs. [391, 392], where those correlators were shown to
vanish in an effectively two-dimensional model.) In semiclassical gravity, using energy conservation
arguments, such correlations have been confirmed for the expectation value of the energy fluxes,
provided that the mass of the black hole is much larger than the Planck mass. However, a more
careful analysis by HR shows that no such simple connection exists for energy flux fluctuations. It
also reveals that the fluctuations on the horizon are in fact divergent. This requires one modifies
the classical picture of the event horizon from a sharply defined three-dimensional hypersurface to
that possessing a finite width, i.e., a fluctuating geometry. One needs to find an appropriate way of
probing the metric fluctuations near the horizon and extracting physically meaningful information.
It also testifies to the necessity of a complete reexamination of all cases afresh and that an evaluation
of the noise kernel near the horizon seems unavoidable for the consideration of fluctuations and
back-reaction issues.
Having registered this cautionary note, Hu and Roura [354] first make the assumption that
a relation between the fluctuations of the fluxes exists, so as to be able to compare with earlier
work. They then show that this relation does not hold and discuss the essential elements required
in understanding not only the mathematical theory but also the operational meaning of metric
fluctuations.

Assuming that there is a relation between fluctuations Since the generation of Hawking
radiation is especially sensitive to what happens near the horizon, from now on we will concentrate
on the metric perturbations near the horizon and consider η(v) = η(v, 2M (v)). This means that
possible effects on the Hawking radiation due to the fluctuations of the potential barrier for the
radial mode functions will be missed by our analysis. Assuming that the fluctuations of the energy
flux crossing the horizon and those far from it are exactly correlated, from Eq. (8.44) we have

dη(v) 2B
= 3 η(v) + ξ(v), (8.45)
dv M (v)

where ξ(v) ≡ (4πr 2 ξvr )(v, r ≈ 6M (v)). The correlation function for the spherically-symmetric
fluctuation ξ(v) is determined by the integral over the whole solid angle of the N rv rv component
of the noise kernel, which is given by (1/2)h{t̂rv (x), t̂rv (x′ )}i. The l = 0 fluctuations of the energy
flux of Hawking radiation far from a black hole formed by gravitational collapse, characterized
also by (1/2)h{t̂rv (x), t̂rv (x′ )}i averaged over the whole solid angle, have been studied in Ref. [122].
Its main features are a correlation time of order M and a characteristic fluctuation amplitude of
order ǫ0 /M 4 (this is the result of smearing the stress tensor two-point function, which diverges
in the coincidence limit, over a period of time of the order of the correlation time). The order
of magnitude of ǫ0 has been estimated to lie between 0.1B and B [384, 122]. For simplicity, we
will consider quantities smeared over a time of order M . We can then introduce the Markovian
approximation (ǫ0 /M 3 (v))δ(v − v ′ ), which coarse-grains the information on features corresponding

78
to time-scales shorter than the correlation time M . Under those conditions r 2 ξvr is of order 1/M 2
and thus the adiabatic approximation made when deriving Eq. (8.43) is justified.
The stochastic equation (8.45) for η can be solved in the usual way and the correlation function
for η(v) can then be computed. Alternatively one can obtain an equation for hη 2 (v)iξ by first
multiplying Eq. (8.45) by η(v) and then taking the expectation value. This brings out a term
hη(v)ξ(v)iξ on the right hand side. For delta-correlated noise (the Stratonovich prescription is the
appropriate one here), it is equal to one half the time-dependent coefficient multiplying the delta
function δ(v − v ′ ) in the correlator hξ(v)ξ(v ′ )iξ , which is given by ǫ0 /M 3 (v) in our case. Finally,
changing from the v coordinate to the mass function M (v) for the background solution, we obtain
d 4 (ǫ0 /B)
hη 2 (M )iξ = − hη 2 (M )iξ − . (8.46)
dM M M
The solutions of this equation are given by
4 " 4 #
M0 ǫ0 M0

2 2
hη (M )iξ = hη (M0 )iξ + −1 . (8.47)
M 4B M

Provided thatpthe fluctuations at the initial time corresponding to M = M0 are negligible (much
smaller than ǫ0 /4B ∼ 1), the fluctuations become comparable to the background solution when
2/3
M ∼ M0 . Note that fluctuations of the horizon radius of order one in Planckian units do not
correspond to Planck scale
√ physics because near the horizon ∆R = r − 2M corresponds to a
physical distance L ∼ M ∆R, as can be seen from the line element for Schwarzschild, ds2 =
−(1 − 2M/r)dt2 + (1 − 2M/r)−1 dr 2√+ r 2 (dθ 2 + sin2 θdϕ2 ), by considering pairs of points at constant
t. So ∆R ∼ 1 corresponds to L ∼ M , whereas a physical distance of order one is associated with
∆R ∼ 1/M , which corresponds to an area change of order one for spheres with those radii. One
can, therefore, have initial fluctuations of the horizon radius of order one for physical distances
well above the Planck length for a black hole with a mass much larger than the Planck mass.
One expects that the fluctuations for states that are regular on the horizon correspond to physical
distances not much larger than the Planck length, so that the horizon radius fluctuations would
be much smaller than one for sufficiently large black hole masses. Nevertheless, that may not be
the case when dealing√ with states which are singular on the horizon, with estimated fluctuations of
order M 1/3 or even M [381, 382, 383].
The result of HR for the growth of the fluctuations of the size of the black hole horizon agrees
with the result obtained by Bekenstein in Ref. [384] and implies that, for a sufficiently massive black
hole (with a few solar masses or a supermassive black hole), the fluctuations become important
before the Planckian regime is reached.
This growth of the fluctuations which was found by Bekenstein and confirmed here via the
Einstein-Langevin equation seems to be in conflict with the estimate given by Wu and Ford in
Ref. [122]. According to their estimate, the accumulated mass fluctuations over a period of the
order of the black hole evaporation time (∆t ∼ M03 ) would be of the order of the Planck mass.
The discrepancy is due to the fact that the first term on the right-hand side of Eq. (8.45), which
(1)
corresponds to the perturbed expectation value hT̂ab [g + h]iren in Eq. (8.42), was not taken into
account in Ref. [122]. The larger growth obtained here is a consequence of the secular effect of that
term, which builds up in time (slowly at first, during most of the evaporation time, and becoming
more significant at late times when the mass has changed substantially) and reflects the unstable
nature of the background solution for an evaporating black hole.

79
 As for the relation between HR’s results reported here and earlier results of Hu, Raval and
Sinha in Ref. [140], there should not be any discrepancy since both adopted the stochastic gravity
framework and performed their analysis based on the Einstein-Langevin equation. The claim in
Ref. [140] was based on a qualitative argument that focused on the dynamics of the stochastic
source alone. If one adds in the consideration that the perturbations around the mean are unstable
for an evaporating black hole, their results agree.
All this can be qualitatively understood as follows. Consider an evaporating black hole with
initial mass M0 and suppose that the initial mass is perturbed by an amount δM0 = 1. The mean
evolution for the perturbed black hole (without taking into account any fluctuations) leads to a
mass perturbation that grows like δM = (M0 /M )2 δM0 = (M0 /M )2 , so that it becomes comparable
2/3
to the unperturbed mass M when M ∼ M0 , which coincides with the result obtained above. Such
a coincidence has a simple explanation: the fluctuations of the Hawking flux slowly accumulated
during most of the evaporating time, which are of the order of the Planck mass, as found by Wu
and Ford, give a dispersion of that order for the mass distribution at the time when the instability
of the small perturbations around the background solution start to become significant.

When no such relation exists and consequences For conformal fields in two dimensional
spacetimes, HR showed that the correlations between the energy flux crossing the horizon and the
flux far from it vanish. The correlation function for the outgoing and ingoing null energy fluxes in
an effectively two-dimensional model was explicitly computed in Refs. [391, 392] and it was also
found to vanish. On the other hand, in four dimensions the correlation function does not vanish
in general and correlations between outgoing and ingoing fluxes do exist near the horizon (at least
partially).
For black hole masses much larger than the Planck mass, one can use the adiabatic approx-
imation for the background mean evolution. Therefore, to lowest order in LH one can compute
the fluctuations of the stress tensor in Schwarzschild spacetime. In Schwarzschild, the amplitude
of the fluctuations of r 2 hTvr i far from the horizon is of order 1/M 2 (= M 2 /M 4 ) when smearing
over a correlation time of order M , which one can estimate for a hot thermal plasma in flat space
[100, 141] (see also Ref. [122] for a computation of the fluctuations of r 2 hTvr i far from the horizon).
The amplitude of the fluctuations of r 2 hTvr i is thus of the same order as its expectation value.
However, their derivatives with respect to v are rather different: since the characteristic variation
times for the expectation value and the fluctuations are M 3 and M respectively, ∂(r 2 hTvr i)/∂v is
of order 1/M 5 whereas ∂(r 2 ξvr )/∂v is of order 1/M 3 . This implies an additional contribution of
order LH due to the second term in Eq. (8.39) if one radially integrates the same equation applied
to stress tensor fluctuations (the stochastic source in the Einstein-Langevin equation). Hence, in
contrast to the case of the mean value, the contribution from the second term in Eq. (8.39) cannot
be neglected when radially integrating since it is of the same order as the contributions from the
first term, and one can no longer obtain a simple relation between the outgoing energy flux far
from the horizon and the energy flux crossing the horizon.
What then? Without this convenience (which almost all earlier researchers had taken for
granted) to get a more precise depiction we need to compute the noise kernel near the horizon.
However, as shown by Hu and Phillips earlier [8] when they examine the coincidence limit of the
noise kernel, and confirmed by the careful analysis of Hu and Roura using smearing functions [354],
the noise kernel smeared over the horizon is divergent, and so are the induced metric fluctuations.
Hence, one cannot study the fluctuations of the horizon as a three-dimensional hypersurface for

80
each realization of the stochastic source because the amplitude of the fluctuations is infinite, even
when restricting one’s attention to the l = 0 sector. Instead one should regard the horizon as
possessing a finite effective width due to quantum fluctuations. In order to characterize its width
one must find a sensible way of probing the metric fluctuations near the horizon and extracting
physically meaningful information, such as their effect on the Hawking radiation emitted by the
black hole. How to probe metric fluctuations is an issue at the root base which needs be dealt with
in all discussions of metric fluctuations.

8.3.3 Summary and prospects

The work of Hu and Roura [354] based on the stochastic gravity program found that the spherically-
symmetric fluctuations of the horizon size of an evaporating black hole become important at late
2/3
times, and even comparable to its mean value when M ∼ M0 , where M0 is the mass of the
black hole at some initial time when the fluctuations of the horizon radius are much smaller than
the Planck length (remember that for large black hole masses this can still correspond to physical
distances much larger than the Planck length, as explained in Sec. 4). This is consistent with the
result previously obtained by Bekenstein in Ref. [384].
It is important to realize that for a sufficiently massive black hole, the fluctuations become
significant well before the Planckian regime is reached. More specifically, for a solar mass black
hole they become comparable to the mean value when the black hole radius is of the order of 10nm,
whereas for a supermassive black hole with M ∼ 107 M⊙ , that happens when the radius reaches a
size of the order of 1mm. One expects that in those circumstances the low-energy effective field
theory approach of stochastic gravity should provide a reliable description.
Due to the nonlinear nature of the back-reaction equations, such as Eq. (8.43), the fact that
the fluctuations of the horizon size can grow and become comparable to the mean value implies
non-negligible corrections to the dynamics of the mean value itself. This can be seen by expanding
Eq. (8.43) (evaluated on the horizon) in powers of η and taking the expectation value. Through
order η 2 we get

d(M (v) + hη(v)iξ ) B

 
= −
dv (M (v) + η(v))2 ξ
" !#
B 2 3 η3
= − 2 1− hη(v)iξ + 2 hη 2 (v)iξ + O . (8.48)
M (v) M (v) M (v) M3

When the fluctuations become comparable to the mass itself, the third term (and higher order
terms) on the right-hand side is no longer negligible and we get non-trivial corrections to Eq. (8.41)
for the dynamics of the mean value. These corrections can be interpreted as higher order radiative
corrections to semiclassical gravity that include the effects of metric fluctuations on the evolution
of the mean value.
Finally we remark on the relation of this finding with earlier well-known results. Does the
existence of the significant deviations for the mean evolution mentioned above invalidate the earlier
results by Bardeen and Massar based on semiclassical gravity in Refs. [298, 341]? First, those devi-
ations start to become significant only after a period of the order of the evaporation time when the
mass of the black hole has decreased substantially. Second, since fluctuations were not considered
in those references, a direct comparison cannot be established. Nevertheless,we can compare the

81
average of the fluctuating ensemble. Doing so exhibits an evolution that deviates significantly when
the fluctuations become important. However, if one considers a single member of the ensemble at
that time, its evolution will be accurately described by the corresponding semiclassical gravity so-
lution until the fluctuations around that particular solution become important again, after a period
of the order of the evaporation time associated with the new initial value of the mass at that time.

8.4 Other work on metric fluctuations but without backreaction

In closing we mention some work on metric fluctuations where no backreaction is considered.
Barrabes et al [125, 126] have considered the propagation of null rays and massless fields in a
black hole fluctuating geometry and have shown that the stochastic nature of the metric leads to a
modified dispersion relation and helps to confront the trans-Planckian frequency problem. However,
in this case the stochastic noise is put in by hand and does not naturally arise from coarse graining
as in a quantum open systems approach, in terms of which stochastic gravity can be interpreted.
It also does not take backreaction into account. It will be interesting to explore how a stochastic
black hole metric, arising as a solution to the Einstein-Langevin equation, hence fully incorporating
backreaction, would affect the trans-Planckian problem.
As mentioned earlier, Ford and his collaborators [121, 204, 122] have also explored the issue of
metric fluctuations in detail and in particular have studied the fluctuations of the black hole horizon
induced by metric fluctuations. However, the fluctuations they considered are in the context of a
fixed background and do not relate to the backreaction.
Another work on metric fluctuations with no backreaction is that of Hu and Shiokawa [120], who
study effects associated with electromagnetic wave propagation in a Robertson-Walker universe
and the Schwarzschild spacetime with a small amount of given metric stochasticity. They find
that time-independent randomness can decrease the total luminosity of Hawking radiation due to
multiple scattering of waves outside the black hole and gives rise to event horizon fluctuations and
fluctuations in the Hawking temperature. The stochasticity in a background metric in their work
is assumed rather than derived (as induced by quantum field fluctuations). But it is interesting to
compare their results with that obtained in stochastic gravity with backreaction, so one can begin
to get a sense of the different sources of stochasticity and their weights (see, e.g., [2] for a list of
possible sources of stochasticity.)
In a subsequent paper Shiokawa [156] showed that the scalar and spinor waves in a stochastic
spacetime behave similarly to the electrons in a disordered system. Viewing this as a quantum
transport problem, he expressed the conductance and its fluctuations in terms of a nonlinear sigma
model in the closed time path formalism and showed that the conductance fluctuations are universal,
independent of the volume of the stochastic region and the amount of stochasticity. This result
can have significant importance in characterizing the mesoscopic behavior of spacetimes resting
between the semiclassical and the quantum regimes.

9 Concluding Remarks
In the first part of this review on the fundamentals of the theory we have given two routes to the
establishment of stochastic gravity and derived a general (finite) expression for the noise kernel. In
the second part we gave three applications, the correlation functions of gravitons in a perturbed
Minkowski metric, structure formation in stochastic gravity theory and the outline of a program

82
for the study of black hole fluctuations and backreaction. We have also discussed the problem of
the validity of semiclassical gravity, a central issue which stochastic gravity is in a unique position
to address.
We have also pointed out a number of ongoing research related to the topics discussed in this
review. Such as the equivalence of the correlation functions for the metric perturbations obtained
using the Einstein-Langevin equations and the quantum correlation functions that follow from a
purely quantum field theory calculation up to leading order in the large N limit. Or the calculation
of the spectrum of the metric fluctuations in inflationary models driven by the trace anomaly due
to conformally coupled fields. The related problem of the runaway solutions in the backreaction
equations. The issue of the coincidence limit in the noise kernel for the black hole fluctuations.
Theoretically, stochastic gravity is at the frontline of the ‘bottom-up’ approach to quantum
gravity [157, 2, 170, 169]. pathway or angle starting from the well-defined and well-understood
theory of semiclassical gravity. Structurally, as can be seen from the issues discussed and the appli-
cations given, stochastic gravity has a very rich constituency because it is based on quantum field
theory and nonequilibrium statistical mechanics in a curved spacetime context. The open systems
concepts and the closed-time-path / influence functional methods constitute an extended framework
suitable for treating the backreaction and fluctuations problems of dynamical spacetimes interact-
ing with quantum fields. We have seen applications to cosmological structure formation and black
hole backreaction from particle creation. A more complete understanding of the backreaction of
Hawking radiation in a fully dynamical black hole situation will enable one to address fundamental
issues such as the black hole end state and information loss puzzles. The main reason why this
program has not progressed as swiftly as desired is due more to technical rather than programatic
difficulties (such as finding reasonable analytic approximations for the Green function or numeri-
cal evaluation of mode-sums near the black hole horizon). Finally, the multiplex structure of this
theory could be used to explore new lines of inquiry and launch new programs of research, such as
nonequilibrium black hole thermodynamics and the microscopic structures of spacetime.

10 Acknowledgements
The materials presented here originated from research work of BLH with Antonio Campos, Nicholas
Phillips, Alpan Raval, Albert Roura and Sukanya Sinha, and of EV with Rosario Martin and Albert
Roura. We thank them as well as Daniel Arteaga, Andrew Matacz, Tom Shiokawa, and Yuhong
Zhang for fruitful collaboration and their cordial friendship since their Ph. D. days. We enjoy
lively discussions with our friends and colleagues Esteban Calzetta, Diego Mazzitelli and Juan
Pablo Paz whose work in the early years contributed toward the establishment of this field. We
acknowledge useful discussions with Paul Anderson, Larry Ford, Ted Jacobson, Emil Mottola,
Renaud Parentani, Raphael Sorkin and Richard Woodard. This work is supported in part by NSF
grant PHY06-01550, the MEC Research projects FPA-2004-04582C02 and FPA-2007-66665C02 and
by DURSI 2005SGR00082.

References
[1] B. L. Hu, Physica A, 158, 399 (1989).

[2] B. L. Hu, Int. J. Theor. Phys. 38, 2987 (1999).

83
 [3] B.L. Hu and E. Verdaguer, “Recent advances in stochastic gravity: theory and issues” in
Advances in the interplay between quantum and gravity physics, 2001 Erice Lectures, edited
by V. De Sabbata (Kluwer Academic Publishers, Dordrecht, 2002), [arXiv:gr-qc/0110092].

[4] B. L. Hu and E. Verdaguer, “Stochastic gravity: A primer with applications”, Class. Quantum
Grav. 20, R1 (2003).

[5] R. Martin and E. Verdaguer, Phys. Lett. B 465, 113 (1999).

[6] R. Martin and E. Verdaguer, Phys. Rev. D 60, 084008 (1999).

[7] N. G. Phillips and B. L. Hu, Phys. Rev. D 63, 104001 (2001).

[8] N. G. Phillips and B. L. Hu, Phys. Rev. D 67, 104002 (2003).

[9] E. B. Davies, The quantum theory of open systems (Academic Press, London, 1976).

[10] K. Lindenberg and B. J. West, The nonequilibrium statistical mechanics of open and closed
systems (VCH Press, New York, 1990).

[11] U. Weiss, Quantum dissipative systems (World Scientific, Singapore, 1993).

[12] R. P. Feynman and F. L. Vernon, Ann. Phys. 24, 118 (1963).

[13] R. P. Feynman and A. R. Hibbs, Quantum mechanics and path integrals (McGraw-Hill, New
York, 1965).

[14] J. Schwinger, J. Math. Phys. 2, 407 (1961).

[15] P. M. Bakshi and K. T. Mahanthappa, J. Math. Phys. 4, 1 (1963).

[16] L.V. Keldish, Zh. Eksp. Teor. Fiz. 47 , 1515 (1964) [Engl. trans. Sov. Phys. JEPT 20, 1018
(1965)].

[17] K. Chou, Z. Su, B. Hao and L. Yu, Phys. Rep. 118, 1 (1985).

[18] Z. Su, L. Y. Chen, X. Yu and K. Chou, Phys. Rev. B 37, 9810 (1988).

[19] E. Calzetta and B. L. Hu, Phys. Rev. D 40, 656 (1989).

[20] F. Cooper, S. Habib, Y. Kluger, E. Mottola, J.P. Paz and P.R. Anderson, Phys. Rev. D 50,
2848 (1994).

[21] B. S. DeWitt, in Quantum Concepts in Space and Time edited by R. Penrose and C. J. Isham
(Claredon Press, Oxford, 1986).

[22] R. D. Jordan, Phys. Rev. D 33, 444 (1986).

[23] E. Calzetta and B. L. Hu, Phys. Rev. D 35, 495 (1987).

[24] R. D. Jordan, Phys. Rev. D 36, 3593 (1987)

[25] J. P. Paz, Phys. Rev. D 41, 1054 (1990).

84
[26] E. Calzetta and B. L. Hu, Phys. Rev. D 37, 2878 (1988).

[27] E. Calzetta and B. L. Hu, “Decoherence of Correlation Histories” in Directions in General

Relativity, Vol II: Brill Festschrift, eds B. L. Hu and T. A. Jacobson (Cambridge University
Press, Cambridge, 1993).

[28] E. Calzetta and B. L. Hu, “Correlations, Decoherence, Disspation and Noise in Quantum
Field Theory”, in Heat Kernel Techniques and Quantum Gravity, edited by S. Fulling (Texas
A and M Press, College Station, 1995), [arXiv:hep-th/9501040].

[29] E. Calzetta and B. L. Hu, Phys. Rev. D 61, 025012 (2000).

[30] C. W. Misner, K. S. Thorne and J. A. Wheeler, Gravitation (Freeman, San Francisco, 1973).

[31] R. M. Wald, General Relativity (The University of Chicago Press, Chicago, 1984).

[32] N. D. Birrell and P. C. W. Davies, Quantum fields in curved space (Cambridge University
Press, Cambridge, 1982).

[33] S. A. Fulling, Aspects of quantum field theory in curved spacetime (Cambridge University
Press, Cambridge, 1989).

[34] R. M. Wald, Quantum field theory in curved spacetime and black hole thermodynamics (Cam-
bridge University Press, Cambridge, 1994).

[35] A. A. Grib, S. G. Mamayev and V. M. Mostepanenko, Vacuum quantum effects in strong

fields (Friedmann Laboratory Publishing, St. Petersburg, 1994).

[36] V.N. Lukash and A.A. Starobinsky, Sov. Phys. JETP 39, 742 (1974).

[37] L. Grishchuk, Ann. N. Y. Acad. Sci. 302, 439 (1976).

[38] J. B. Hartle, Phys. Rev. Lett. 39, 1373 (1977).

[39] B. L. Hu and L. Parker, Phys. Lett. A 63, 217 (1977).

[40] B. L. Hu and L. Parker, Phys. Rev. D 17, 933 (1978).

[41] M. Fischetti, J. B. Hartle and B. L. Hu, Phys. Rev. D 20, 1757 (1979).

[42] J. B. Hartle and B. L. Hu, Phys. Rev. D 20, 1772 (1979).

[43] J. B. Hartle and B. L. Hu, Phys. Rev. 21, 2756 (1980).

[44] J. B. Hartle, Phys. Rev. D 23, 2121 (1981).

[45] P. A. Anderson, Phys. Rev. D 28, 271 (1983).

[46] P. A. Anderson, Phys. Rev. D 29, 615 (1984).

[47] W. A. Hiscock, S. L. Larson and P. A. Anderson, Phys. Rev. D 56, 3571 (1997).

85
[48] Paul R. Anderson, Mathew Binkley, Hector Calderon, William A. Hiscock, Emil Mottola,
Ruslan Vaulin, “Effects of quantized fields on the spacetime geometries of static spherically
symmetric black holes” in Proceedings of the Eleventh Marcel Grossmann Meeting on General
Relativity (Berlin, Germany, 2006), arXiv:0709.4457 [gr-qc].

[49] C. Barceló, S. Liberati, and M. Visser, Living Rev. Relativity 8, 12 (2005)

[50] R. Balbinot, A. Fabbri, S. Fagnocchi, R. Parentani. Riv. Nuovo Cim. 28, 1 (2005).

[51] R. Schutzhold, Lect. Notes Phys. 718, 5 (2007).

[52] Ralf Schutzhold, Michael Uhlmann, Lutz Petersen, Hector Schmitz, Axel Friedenauer, and
Tobias Schutz, Phys. Rev. Lett. 99, 201301 (2007).

[53] Clovis Maia and Ralf Schutzhold, Phys. Rev. D 76, 101502 (2007).

[54] Esteban A. Calzetta, Bei-Lok B. Hu, Nonequilibrium Quantum Field Theory (Cambridge
University Press, Cambridge, 2008) (ISBN-13:9780521641685).

[55] W. H. Zurek, Phys. Rev. D 24, 1516 (1981).

[56] W. H. Zurek, Phys. Rev. D 26, 1862 (1982).

[57] W. H. Zurek, Frontiers of Nonequilibrium Statistical Physics, edited by G. T. Moore and M.

O. Scully (Plenum, N. Y., 1986).

[58] W. H. Zurek, Physics Today 44, 36 (1991).

[59] E. Joos and H. D. Zeh, Z. Phys. B 59, 223 (1985).

[60] A. O. Caldeira and A. J. Leggett, Phys. Rev. A 31, 1059 (1985).

[61] W. G. Unruh and W. H. Zurek, Phys. Rev. D 40, 1071 (1989).

[62] W. H. Zurek, Prog. Theor. Phys. 89, 281 (1993).

[63] D. Giulini et al, Decoherence and the Appearance of a Classical World in Quantum Theory
(Springer Verlag, Berlin, 1996).

[64] R. B. Griffiths, J. Stat. Phys. 36, 219 (1984).

[65] R. Omnés, J. Stat Phys. 53, 893 (1988).

[66] R. Omnés, J. Stat Phys. 53, 933 (1988).

[67] R. Omnés, J. Stat Phys. 53, 957 (1988).

[68] R. Omnés, Ann. Phys. (NY) 201, 354 (1990).

[69] R. Omnés, Rev. Mod. Phys. 64, 339 (1992).

[70] R. Omnés, The Interpretation of Quantum Mechanics (Princeton University Press, Princeton,
1994).

86
[71] M. Gell-Mann and J. B. Hartle, in Complexity, Entropy and the Physics of Information,
edited by W. H. Zurek (Addison-Wesley, Reading, 1990).

[72] J. B. Hartle, “Quantum Mechanics of Closed Systems” in Directions in General Relativity Vol.
1, eds B. L. Hu, M. P. Ryan and C. V. Vishveswara (Cambridge University Press, Cambridge,
1993).

[73] H. F. Dowker and J. J. Halliwell, Phys. Rev. D 46, 1580 (1992).

[74] J. J. Halliwell, Phys. Rev. D 48, 4785 (1993).

[75] J. J. Halliwell, Phys. Rev. D 57, 2337 (1998).

[76] T. Brun, Phys. Rev. D 47, 3383 (1993).

[77] J. P. Paz and W. H. Zurek, Phys. Rev. D 48 2728 (1993).

[78] J. Twamley, Phys. Rev. D 48, 5730 (1993).

[79] C. J. Isham, J. Math. Phys. 35, 2157 (1994).

[80] C. J. Isham and N. Linden, J. Math. Phys. 35, 5452 (1994).

[81] C. J. Isham and N. Linden, J. Math. Phys. 36, 5392 (1994).

[82] J. J. Halliwell, Ann. N.Y. Acad. Sc. 755, 726 (1995).

[83] F. Dowker and A. Kent, Phys. Rev. Lett. 75, 3038 (1995).

[84] F. Dowker and A. Kent, J. Stat. Phys. 82, 1575 (1996).

[85] A. Kent, Phys. Rev. A 54, 4670 (1996).

[86] A. Kent, Phys. Rev. Lett. 78, 2874 (1997).

[87] A. Kent, Phys. Rev. Lett. 81, 1982 (1998).

[88] C. J. Isham, N. Linden, K. Savvidou and S. Schreckenberg, J. Math. Phys. 39, 1818 (1998).

[89] C. Kiefer, Class. Quant. Grav. 4, 1369 (1987).

[90] J. J. Halliwell, Phys. Rev. D 39, 2912 (1989).

[91] T. Padmanabhan, Phys. Rev. D 39 2924 (1989).

[92] B. L. Hu “Quantum and Statistical Effects in Superspace Cosmology” in Quantum mechanics

in curved spacetime, edited by J. Audretsch and V. de Sabbata (Plenum, London, 1990).

[93] E. Calzetta, Class. Quant. Grav. 6, L227 (1989).

[94] E. Calzetta, Phys. Rev. D 43, 2498 (1991).

87
 [95] B. L. Hu, J. P. Paz and S. Sinha, “Minisuperspace as a Quantum Open System” in Direc-
tions in General Relativity Vol. 1, (Misner Festschrift) eds B. L. Hu, M. P. Ryan and C.V.
Vishveswara (Cambridge University Press, Cambridge, 1993).

[96] A. Campos and E. Verdaguer, Phys. Rev. D 49, 1861 (1994).

[97] P. Candelas and D. W. Sciama, Phys. Rev. Lett. 38, 1372 (1977).

[98] E. Mottola, Phys. Rev. D 33, 2136 (1986).

[99] B.L. Hu and S. Sinha, Phys. Rev. D 51, 1587 (1995).

[100] A. Campos and B. L. Hu, Phys. Rev. D 58, 125021 (1998).

[101] B. L. Hu, ‘Quantum statistical fields in gravitation and cosmology” in Proc. Third Inter-
national Workshop on Thermal Field Theory and Applications, edited by R. Kobes and G.
Kunstatter (World Scientific, Singapore, 1994), [arXiv:gr-qc/9403061].

[102] E. Calzetta and B.L. Hu, Phys. Rev. D 49, 6636 (1994).

[103] B. L. Hu and A. Matacz, Phys. Rev. D51, 1577 (1995).

[104] A. Campos and E. Verdaguer, Phys. Rev. D 53, 1927 (1996).

[105] A. Campos and E. Verdaguer, Int. J. Theor. Phys. 36, 2525 (1997).

[106] E. Calzetta, A. Campos and E. Verdaguer, Phys. Rev. D 56, 2163 (1997).

[107] F. C. Lombardo and F. D. Mazzitelli, Phys. Rev. D 55, 3889 (1997).

[108] R. Martin and E. Verdaguer, Int. J. Theor. Phys. 38, 3049 (1999).

[109] R. Martin and E. Verdaguer, Phys. Rev. D 61, 124024 (2000).

[110] G. T. Horowitz, Phys. Rev. D 21, 1445 (1980).

[111] G. T. Horowitz, in Quantum Gravity 2: A Second Oxford Symposium, edited by C. J. Isham,

R. Penrose and D. W. Sciama (Clarendon Press, Oxford, 1981).

[112] J. Z. Simon, Phys. Rev. D 41, 3720 (1990).

[113] P.R. Anderson, C. Molina-Paris and E. Mottola, Phys. Rev. D 67, 024026 (2003).

[114] L. H. Ford, Ann. Phys. (N.Y.) 144, 238 (1982).

[115] C. I. Kuo and L. H. Ford, Phys. Rev. D 47, 4510 (1993).

[116] B. L. Hu and N. G. Phillips, Int. J. Theor. Phys. 39, 1817 (2000).

[117] N. G. Phillips and B. L. Hu, Phys. Rev. D 62, 084017 (2000).

[118] B. L. Hu, A. Roura and E. Verdaguer, Phys. Rev. D 70, 044002 (2004).

[119] B. L. Hu, A. Roura and E. Verdaguer, Int. J. Theor. Phys. 43, 749 (2004).

88
[120] B. L. Hu and K. Shiokawa, Phys. Rev. D 57, 3474 (1998).
[121] L. H. Ford and N. F. Svaiter, Phys. Rev. D 56, 2226 (1997).
[122] C. H. Wu and L. H. Ford, Phys. Rev. D 60, 104013 (1999)

[123] R. D. Sorkin, “How wrinkled is the surface of the black hole?” in Proceedings of the First
Australasian Conference on General Relativity and Gravitation February 1996, Adelaide,
Australia, edited by David Wiltshire, pp. 163-174 (University of Adelaide, 1996), [arXiv:gr-
qc/9701056].
[124] R. D. Sorkin and D. Sudarsky, Class. Quantum Grav. 16, 3835 (1999).

[125] C. Barrabes, V. Frolov and R. Parentani, Phys. Rev. D 59, 124010 (1999).
[126] C. Barrabes, V. Frolov and R. Parentani, Phys. Rev. D 62, 044020 (2000).

[127] S. Massar and R. Parentani, Nucl. Phys. B 575, 353 (2000).

[128] R. Parentani, Phys. Rev. D 63, 041503 (2001).

[129] E. Calzetta and E. Verdaguer, Phys. Rev. D 59, 083513 (1999).

[130] E. Calzetta and B. L. Hu, Phys. Rev. D 52, 6770 (1995).
[131] A. Matacz, Phys. Rev. D 55, 1860 (1997).

[132] A. Matacz, Phys. Rev. D 56, 1836 (1997).

[133] E. Calzetta and S. Gonorazky, Phys. Rev. D 55, 1812 (1997).

[134] A. Roura and E. Verdaguer, “Cosmological perturbations from stochastic gravity”,

arXiv:0709.1940 [gr-qc].
[135] A. A. Starobinsky, Phys. Lett. B 91, 99 (1980).

[136] A. Vilenkin, Phys. Rev. D 32, 2511 (1985).

[137] S. W. Hawking, T. Hertog and H. S. Reall, Phys. Rev. D 63, 083504 (2001).

[138] D. W. Sciama, in Relativity, quanta and cosmology – Centenario di Einstein edited by F.

DeFinis (Editrici Giunti Barbera Universitaria, Firenze, Italy, 1979).
[139] D. Sciama, P. Candelas and D. Deutsch, Adv. Phys. 30, 327 (1981).

[140] B. L. Hu, A. Raval and S. Sinha, “Notes on black hole fluctuations and backreaction, in Black
Holes, Gravitational Radiation and the Universe edited by B. R. Iyer and B. Bhawal (Kluwer
Academic Publishers, Dordtrecht, 1999).

[141] A. Campos and B. L. Hu, Int. J. Theor. Phys. 38, 1253 (1999).
[142] S. Sinha, A. Raval and B. L. Hu, “Black hole fluctuations and backreaction in stochastic
gravity” in Bekenstein issue of Foundations of Physics Thirty years of black hole physics
edited by L. Horwitz (2003).

89
[143] B. L Hu and A. Roura, Int. J. Theor. Phys. 46, 2204 (2007).

[144] B. L Hu and A. Roura, Proc. 7th Asia-Pacific International Conference on Gravitation and As-
trophysics, published in Chungli 2005: Gravitation and Astrophysics p. 236 (2006), [arXiv:gr-
qc/0610066].

[145] J. C. Niemeyer and R. Parentani, Phys.Rev. D 64, 101301 (2001).

[146] S. Carlip, Phys. Rev. Lett. 79, 4071 (1997).

[147] S. Carlip, Class. Quantum Grav. 15, 2629 (1998).

[148] L. J. Garay, Phys. Rev. Lett. 80, 2508 (1998).

[149] L. J. Garay, Phys. Rev. D 58, 124015 (1998).

[150] L. J. Garay, Int. J. Mod. Phys. A 14, 4079 (1999).

[151] Y. Jack Ng, Mod. Phys. Lett. A 18, 1073 (2003).

[152] W. A. Christiansen, Y. Jack Ng, H. van Dam, Phys. Rev. Lett. 96, 051301 (2006).

[153] S. Bernadotte and F. R. Klinkhamer, Phys. Rev. D 75, 024028 (2007).

[154] Hermann Nicolai, Kasper Peeters, “Loop and spin foam quantum gravity: a brief guide for
beginners” in An assessment of current paradigms in the physics of fundamental interactions,
edited by I. Stamatescu [arXiv:hep-th/0601129].

[155] Laurent Freidel, Kirill Krasnov, “A New Spin Foam Model for 4d Gravity”, arXiv:0708.1595
[hep-th].

[156] K. Shiokawa, Phys. Rev. D 62, 024002 (2000).

[157] B. L. Hu, “General Relativity as Geometro-Hydrodynamics” Invited talk at the Second

Sakharov Conference, Moscow (1996), [arXiv:gr-qc/9607070].

[158] B. L. Hu, Int. J. Theor. Phys. 44, 1785 (2005).

[159] G. E. Volovik: The Universe in a Helium Droplet (Clarendon Press, Oxford, 2003),
http://boojum.hut.fi/personnel/THEORY/volovik.html.

[160] T. Jacobson, Phys. Rev. Lett. 75, 1260 (1995).

[161] C. Eling, R. Guedens and T. Jacobson, Phys. Rev. Lett. 96, 121301 (2006).

[162] G.E. Volovik, “Fermi-point scenario for emergent gravity”, Proceedings of From Quantum to
Emergent Gravity: Theory and Phenomenology, Trieste June 11-15 (2007), arXiv:0709.1258
[hep-th].

[163] Christopher P. Herzog, J. High Energy Phys. 12, 026 (2002).

[164] Nathan Seiberg, “Emergent Spacetime” Rapporteur talk at the 23rd Solvay Conference in
Physics, (2005), [arXiv:hep-th/0601234].

90
[165] Gary T. Horowitz and Joseph Polchinski, “Gauge/gravity duality” (2006), [arXiv:gr-
qc/0602037].

[166] Michael Levin, Xiao-Gang Wen, Rev. Mod. Phys. 77, 871 (2005).

[167] Zheng-Cheng Gu, Xiao-Gang Wen, “A lattice bosonic model as a quantum theory of gravity”
(2006), [arXiv:gr-qc/0606100].

[168] B. L. Hu, “Semiclassical Gravity and Mesoscopic Physics” in Quantum Classical Correspon-
dence edited by D. S. Feng and B. L. Hu (International Press, Boston, 1997), [arXiv:gr-
qc/9511077].

[169] B. L. Hu, “New View on Quantum Gravity and the Origin of the Universe” in Where Do We
Come From? – on the Origin of the Universe (in Chinese) (Commercial Press, Hong Kong
2007), [arXiv:gr-qc/0611058].

[170] B. L. Hu, Int. J. Theor. Phys. 41, 2091 (2002).

[171] D. Oriti, Approaches to Quantum Gravity - toward a new understanding of space, time, and
matter (Cambridge University Press, Cambridge, 2008).

[172] B. S. DeWitt, Phys. Rep. 19C, 295 (1975).

[173] L. Parker, Phys. Rev. 183, 1057 (1969).

[174] R. U. Sexl and H. K. Urbantke, Phys. Rev. 179, 1247 (1969).

[175] Ya. B. Zeldovich, Zh. Eksp. Teor. Fiz. Pis’ma Red. 12, 443 (1970) [JETP Lett.12, 307 (1970)].

[176] Y. B. Zel’dovich and A. A. Starobinsky, Zh. Eksp. Teor. Fiz. 61, 2161 (1971) [Sov. Phys. –
JETP 34, 1159 (1971)].

[177] B. L. Hu, Phys. Rev. D 9, 3263 (1974).

[178] B. Berger, Ann. Phys. (N. Y. ) 83, 458 (1974).

[179] B. Berger, Phys. Rev. D 11, 2770 (1975).

[180] B. Berger, Phys. Rev. D 12, 368 (1975).

[181] J. A. Frieman, Phys. Rev. D 39, 389 (1989).

[182] J. Cespedes and E. Verdaguer, Phys. Rev. D 41, 1022 (1990).

[183] S. W. Hawking, Nature 248, 30 (1974).

[184] S. W. Hawking, Comm. Math. Phys. 43, 199 (1975).

[185] W. Israel, Phys. Lett. A 57, 107 (1975).

[186] L. Parker, Phys. Rev. D 12, 1519 (1975).

[187] R. M. Wald, Comm. Math. Phys. 45, 9 (1975).

91
[188] S. M. Christensen, Phys. Rev. D 14, 2490 (1976).

[189] S. M. Christensen, Phys. Rev. D 17, 946 (1978).

[190] R. M. Wald, Commun. Math. Phys. 54, 1 (1977).

[191] R. M. Wald, Phys. Rev. D 17, 1477 (1977).

[192] C. W. Misner, Phys. Rev. Lett. 22,, 1071 (1969).

[193] V. A. Belinsky, I. M. Khalatnikov and E. M. Lifshitz, Adv. Phys. 19, 525 (1970).

[194] V. A. Belinsky, I. M. Khalatnikov and E. M. Lifshitz, Adv. Phys. 31, 639 (1982).

[195] A. H. Guth, Phys. Rev. D 23, 347 (1981).

[196] A. Albrecht and P. J. Steinhardt, Phys. Rev. Lett. 48, 1220 (1982).

[197] A. D. Linde, Phys. Lett. B 114, 431 (1982).

[198] A. D. Linde, Phys. Lett. B 162, 281 (1985).

[199] N. G. Phillips and B. L. Hu, Phys. Rev. D 55, 6123 (1997).

[200] A. Roura and E. Verdaguer, Phys. Rev. D 60, 107503 (1999).

[201] H. Osborn and G. M. Shore, Nucl. Phys. B 57 287 (2000).

[202] G. Cognola, E. Elizalde, S. Zerbini Phys. Rev. D 65, 085031 (2002).

[203] L.H. Ford, Int. Jour. Theor. Phys. 39, 1803 (2000).

[204] L. H. Ford and C. H. Wu, Int. J. Theor. Phys. 42, 15 (2003).

[205] P.R. Anderson, C. Molina-Paris and E. Mottola, “Linear response and the validity of the
semi-classical approximation in gravity” (2002), [arXiv:gr-qc/0204083].

[206] H. Yu and L. H. Ford, Phys. Rev. D 60, 084023 (1999).

[207] H. Yu and L. H. Ford, Phys. Lett. B 496, 107 (2000).

[208] M. Gell-Mann and J. B. Hartle, Phys. Rev. D 47, 3345 (1993).

[209] J. P. Paz, Phys. Rev. D 44, 1038 (1991).

[210] J. P. Paz and S. Sinha, Phys. Rev. D 45, 2823 (1992).

[211] A. Roura and E. Verdaguer, in preparation (2008).

[212] G. T. Horowitz and R. M. Wald, Phys. Rev. D 21, 1462 (1980).

[213] G. T. Horowitz and R. M. Wald, Phys. Rev. D 25, 3408 (1982).

[214] J.B. Hartle and G.T. Horowitz, Phys. Rev. D 24, 257 (1981).

92
[215] E. Tomboulis, Phys. Lett. B 70, 361 (1977).

[216] T. S. Bunch, J. Phys. A 12, 517 (1979).

[217] A. Roura and E. Verdaguer, Int. J. Theor. Phys. 39, 1831 (2000).

[218] E. Calzetta, A. Roura and E. Verdaguer, Physica A 319, 188 (2003).

[219] J. B. Hartle, in Gravitation and Quantizations, proceedings of the 1992 Les Houches Summer
School, edited by B. Julia and J. Zinn-Justin (North-Holland, Amsterdam, 1995), [arXiv:gr-
qc/9304006].

[220] J. T. Whelan, Phys. Rev. D 57, 768 (1998).

[221] P. R. Johnson and B. L. Hu, Phys. Rev. D 65 065015 (2002).

[222] E. Calzetta, A. Roura and E. Verdaguer, Phys. Rev. D 64, 105008 (2001).

[223] E. Calzetta, A. Roura and E. Verdaguer, Phys. Rev. Lett. 88, 010403 (2002).

[224] L. H. Ford, Int. J. Theor. Phys. 38, 2941 (1999).

[225] J. Borgman and L. H. Ford, Phys. Rev. D 70, 064032 (2004).

[226] J. Borgman and L. H. Ford, Phys. Rev. D 70, 127701 (2004).

[227] J. Borgman and L. H. Ford, Int. J. Mod. Phys. A 20, 2364 (2005).

[228] J. Z. Simon, Phys. Rev. D 43, 3308 (1991).

[229] É. É. Flanagan and R. M. Wald, Phys. Rev. D 54, 6233 (1996).

[230] L. Parker and J. Z. Simon, Phys. Rev. D 47, 1339 (1993).

[231] A. Ashtekar, Phys. Rev. Lett. 77, 4864 (1996).

[232] C. Beetle, Adv. Theor. Math. Phys. 2, 471 (1998).

[233] J. W. York, Jr., Phys. Rev. D 28, 2929 (1983).

[234] J. W. York, Jr., Phys. Rev. D 31, 775 (1985).

[235] J. W. York, Jr., Phys. Rev. D 33, 2092 (1986).

[236] F. Lombardo and F. D. Mazzitelli, Phys. Rev. D 53, 2001 (1996).

[237] C. Greiner and B. Müller, Phys. Rev. D 55, 1026 (1997).

[238] M. Morikawa, Phys. Rev. D 33, 3607 (1986).

[239] D.-S. Lee and D. Boyanovsky, Nucl. Phys. B 406, 631 (1993).

[240] B. L. Hu and A. Matacz, Phys. Rev. D 49, 6612 (1994).

93
[241] J. F. Donoghue, Phys. Rev. Lett. 72, 2996 (1994).

[242] J. F. Donoghue, Phys. Rev. D 50, 3874 (1994).

[243] J. F. Donoghue, Helv. Phys. Acta 69, 269 (1996).

[244] J. F. Donoghue, in Advanced School on Effective Theories, edited by F. Cornet and M. J.

Herrero (World Scientific, Singapore, 1996) [arXiv:gr-qc/9512024; gr-qc/9712070].

[245] S. Sinha and B. L. Hu, Phys. Rev. D44, 1028 (1991).

[246] S. Weinberg, The Quantum Theory of Fields, vol. I (Cambridge University Press, Cambridge,
1995).

[247] S. Weinberg, The Quantum Theory of Fields, vol. II (Cambridge University Press, Cambridge,
1996).

[248] E. Calzetta and A. Kandus, Phys. Rev. D 55, 1795 (1997).

[249] M. Gleiser and R. O. Ramos, Phys. Rev. D 50, 2441 (1994).

[250] D. Boyanovsky, H. J. de Vega, R. Holman, D.S-Lee and A. Singh, Phys. Rev. D 51, 4419
(1995).

[251] M. Yamaguchi and J. Yokoyama, Phys. Rev. D 56, 4544 (1997).

[252] S. A. Ramsey, B. L. Hu and A. M. Stylianopoulos, Phys. Rev. D 57, 6003 (1998).

[253] Chun-Hsien Wu, L. H. Ford, Phys. Rev. D 64, 045010 (2001).

[254] E. Elizalde, S.D. Odintsov, A. Romeo, A.A. Bytsenko and S. Zerbini, Zeta regularization
techniques with applications (World Scientific, Singapore, 1994).

[255] Klaus Kirsten, Spectral functions in mathematics and physics (Chapman and Hall/CRC, Boca
Raton, 2001).

[256] R. Camporesi, Phys. Rep. 196, 1 (1990).

[257] S. L. Adler, J. Lieberman and Y. J. Ng, Ann. Phys. (N.Y.) 106, 279 (1977).

[258] B. S. DeWitt, Dynamical Theory of Groups and Fields (Gordon and Breach, 1965).

[259] N. G. Phillips, “Symbolic Computation of Higher Order Correlation Functions of Quantum

Fields in Curved Spacetimes” (in preparation).

[260] D. M. Capper and M. J. Duff, Nouvo Cimento A 23, 173 (1974).

[261] M. J. Duff, in Quantum Gravity: An Oxford symposium, edited by C. J. Isham, R. Penrose

and D. W. Sciama (Oxford University Press, Oxford, 1975).

[262] G. W. Gibbons, Commun. Math. Phys. 45, 191 (1975).

[263] S. Deser, J. Phys. A 8, 1972 (1975).

94
[264] J. Garriga and E. Verdaguer, Phys. Rev. D 43, 391 (1991).

[265] D. S. Jones, Generalized functions (McGraw-Hill, New York, 1966).

[266] A. Campos, R. Martin and E. Verdaguer, Phys. Rev. D 52, 4319 (1995).

[267] A. A. Starobinsky, Pis’ma Zh. Eksp. Teor. Fiz. 34, 460 (1981) [JEPT Lett. 34, 438, (1981)].

[268] W. Tichy and É. É. Flanagan, Phys. Rev. D 58, 124007 (1998).

[269] G. T. Horowitz and R. M. Wald, Phys. Rev. D 17, 414 (1978).

[270] S. Randjbar-Daemi, J. Phys. A 14, L229 (1981).

[271] S. Randjbar-Daemi, J. Phys. A 15, 2209 (1982).

[272] W.-M. Suen, Phys. Rev. Lett. 62, 2217 (1989).

[273] W.-M. Suen, Phys. Rev. D 40, 315 (1989).

[274] L. Schwartz, Théorie des distributions, Tomes I et II (Hermann, Paris, 1957 and 1959).

[275] A. H. Zemanian, Distribution Theory and Transform Analysis (Dover, New York, 1987).

[276] S. W. Hawking and T. Hertog, Phys. Rev. D 65, 103515 (2002).

[277] E.M. Lifshitz, Zh. Eksp. Teor. Phys. 16, 587 (1946) [J. Phys. USSR 10, 116 (1946)].

[278] E.M. Lifshitz and I. Khalatnikov, Adv. Phys. 12, 185 (1963).

[279] P.J.E. Peebles, Large Scale Structure of the Universe (Princeton University Press, Princeton,
1980).

[280] J. M. Bardeen, Phys. Rev. D 22, 1882 (1980).

[281] T. Padmanabhan, Structure formation (Cambridge University Press, Cambridge, 1993).

[282] E.W. Kolb and M. Turner, The early Universe (Addison-Wesley, Reading, Massachusetts,
1990).

[283] A. Linde, Particle physics and inflationary cosmology (Harwood Academic Publishers,
Switzerland, 1990).

[284] V. Mukhanov, Physical Foundations of Cosmology (Cambridge University Press, Cambridge,

2005)

[285] S.W. Hawking, Phys. Lett. B 115, 295 (1982).

[286] A.H. Guth and S.Y. Pi, Phys. Rev. Lett. 49, 1110 (1982).

[287] J.M. Bardeen, P.J. Steinhardt and M. S. Turner, Phys. Rev. D 28, 679 (1983).

[288] V.F. Mukhanov, H.A. Feldman and R.H. Brandenberger, Phys. Rep. 215, 203 (1992).

95
[289] Y. Urakawa and K. Maeda, “Cosmological density fluctuations in stochastic gravity. Formal-
ism and linear analysis”, arXiv:0710.5342 [hep-th].

[290] G.F. Smoot et al., Astrophys. J. Lett. 396, L1 (1992).

[291] A. Roura and E. Verdaguer, Int. J. Theor. Phys. 38, 3123 (1999).

[292] S. Weinberg, Phys. Rev. D 72, 043514 (2005).

[293] S. Weinberg, Phys. Rev. D 74, 023508 (2006).

[294] E.O. Kahya and R.P. Woodard, Phys.Rev. D 76, 124005 (2007).

[295] E.O. Kahya and R.P. Woodard, Scalar field equations from quantum gravity during inflation,
(2007), arXiv:0710.5282 [gr-qc].

[296] Y. Urakawa and K. Maeda, “ One-loop Corrections to Scalar and Tensor Perturbations during
Inflation in Stochastic Gravity”, arXiv:0801.0126 [hep-th].

[297] P. Hajicek and W. Israel, Phys. Lett. A 80, 9 (1980).

[298] J. M. Bardeen, Phys. Rev. Lett. 46, 382 (1981).

[299] D. Hochberg and T. W. Kephart, Phys. Rev. D 47, 1465 (1993).

[300] D. Hochberg, T. W. Kephart, and J. W. York, Jr., Phys. Rev. D 48, 479 (1993).

[301] P. R. Anderson, W. A. Hiscock, J. Whitesell, and J. W. York, Jr., Phys. Rev. D 50, 6427
(1994).

[302] E. Witten, Phys. Rev. D 44, 314 (1991).

[303] C. G. Callan, S. B. Giddings, J. A. Harvey, and A. Strominger, Phys. Rev. D 45, R1005
(1992).

[304] J. G. Russo, L. Susskind, and L. Thorlacius, Phys. Lett. B 292, 13 (1992).

[305] J. G. Russo, L. Susskind, and L. Thorlacius, Phys. Rev. D 46, 3444 (1992).

[306] C. Holzhey and F. Wilczek, Nucl. Phys. B 380, 447 (1992).

[307] F. Wilczek, in Black Holes, Membranes, Wormholes and Superstrings, edited by K. Kalara
and D. V. Nanopoulos (World Scientific, Singapore, 1993), [arXiv:hep-th/9302096].

[308] J. Polchinski and A. Strominger, Phys. Rev. D 50, 7403 (1994).

[309] J. R. Anglin, R. Laflamme, W. H. Zurek and J. P. Paz, Phys. Rev. D 52, 2221 (1995).

[310] B. L. Hu, Erice Lectures, Sept. 1995, in String Gravity and Physics at the Planck Energy Scale,
edited by N. Sanchez and A. Zichichi (Kluwer, Dordrecht, 1996), [arXiv:gr-qc/9511075].

[311] D. A. Lowe and L. Thorlacius, Phys. Rev. D 73, 104027 (2006).

96
[312] S. B. Giddings, Phys. Rev. D 74, 106005 (2006).

[313] S. W. Hawking, Phys. Rev. D 13, 191 (1976).

[314] S. W. Hawking, Phs. Rev. D 14, 2460 (1976).

[315] S. W. Hawking, Phys. Rev. D 72, 084013 (2005).

[316] D. N. Page , Phys. Rev. Lett. 44, 301 (1980).

[317] J. Preskill, in Black Holes, Membranes, Wormholes and Superstrings, edited by K. Kalara
and D. V. Nanopoulos (World Scientific, Singapore, 1993), [arXiv:hep-th/9209058].

[318] D. N. Page, in proceedings of the 5th Canadian Conference on General Relativity and Rela-
tivistic Astrophysics, edited by R. B. Mann and R. G. McLenaghan (World Scientific, Singa-
pore, 1994), [arXiv:hep-th/9305040].

[319] J. A. Smolin and J. Oppenheim, Phys. Rev. Lett. 96, 081302 (2006).

[320] S. L. Braunstein and A. K. Pati, Phys. Rev. Lett. 98, 080502 (2007).

[321] J. D. Bekenstein, Phys. Rev. D 7, 2333 (1973).

[322] J. D. Bekenstein, “Do we understand black hole entropy?” Proc. Seventh Marcel Grossmann
Meeting (Stanford University, 1994), [arXiv:gr-qc/9409015].

[323] J. D. Bekenstein and V. F. Mukhanov, Phys. Lett. B 360, 7 (1995).

[324] R. D. Sorkin, “The statistical mechanics of black hole thermodynamics,” in Black holes and
relativistic stars, edited by R. M. Wald (The University of Chicago Press, Chicago, 1998),
[arXiv:gr-qc/9705006].

[325] T. Jacobson, “On the nature of black hole entropy,” in General Relativity and Relativistic
Astrophysics: Eighth Canadian Conference, edited by C. Burgess and R.C. Myers (AIP Press,
1999), [arXiv:gr-qc/9908031].

[326] R. M. Wald, in Advances in the interplay between quantum and gravity physics, edited by P.
Bergmann and V. De Sabbata, (Kluwer, Dortrecht, 2002), [arXiv:gr-qc/9912119].

[327] R.M. Wald, Living Rev. Rel. 4, 6 (2001).

[328] L. Susskind and J. Uglam, Phys. Rev. D 50, 2700 (1994).

[329] D. Kabat, S. H. Shenker, and M. J. Strassler, Phys. Rev. D 52, 7027 (1995).

[330] A. Strominger and C. Vafa, Phys. Lett. B 379, 99 (1996).

[331] G. T. Horowitz, “The Origin of Black Hole Entropy in String Theory” in the Proceedings
of the Pacific Conference on Gravitation and Cosmology (Seoul, Korea, 1996), [arXiv:gr-
qc/9604051].

[332] G. Horowitz and J. Polchinski, Phys. Rev. D 55, 6189 (1997).

97
[333] J. M. Maldacena, A. Strominger and E. Witten, J. High Energy Phys. 12, 002 (1997).

[334] J. M. Maldacena, Nucl. Phys. Proc. Suppl. A 61, 111 (1998).

[335] Xiao-Gang Wen, Quantum field Theory of Many-Body Systems-From the Origin of Sound to
an Origin of Light and Electrons (Oxford University Press, Oxford, 2004).

[336] Xiao-Gang Wen, Ann. Phys. 316, 1 (2005), [arXiv:cond-mat/0406441].

[337] W. G. Unruh, Phys. Rev. Lett. 46, 1351 (1980).

[338] T. Jacobson, Phys. Rev. D 48, 728 (1993).

[339] B. P. Jenson, J. G. McLaughlin and A. C. Ottewill, Phys. Rev. D 51, 5676 (1995).

[340] R. Parentani and T. Piran, Phys. Rev. Lett. 73, 2805 (1994).

[341] S. Massar, Phys. Rev. D 52, 5857 (1995).

[342] P. R. Anderson, W. A. Hiscock and D. A. Samuel, Phys. Rev. Lett. 70, 1739 (1993).

[343] R. Anderson, W. A. Hiscock and D. A. Samuel, Phys. Rev. D 51, 4337 (1995).

[344] P. R. Anderson, W. A. Hiscock and D. J. Loranz, Phys. Rev. Lett. 74, 4365 (1995).

[345] K. W. Howard and P. Candelas, Phys. Rev. Lett. 53, 403 (1984).

[346] K. W. Howard, Phys. Rev. D 30, 2532 (1984).

[347] B. P. Jensen and A. Ottewill, Phys. Rev. D 39, 1130 (1989).

[348] D. N. Page, Phys. Rev. D 25, 1499 (1982).

[349] M. R. Brown and A. C. Ottewill, Phys. Rev. D 31, 2514 (1985).

[350] M. R. Brown, A. C. Ottewill, and D. N. Page, Phys. Rev. D 33, 2840 (1986).

[351] V. P. Frolov and A. I. Zel’nikov, Phys. Rev. D 35, 3031 (1987).

[352] V. P. Frolov and A. I. Zel’nikov, Phys. Lett. B 115, 372 (1982).

[353] V. P. Frolov and A. I. Zel’nikov, Phys. Rev. D 29, 1057 (1984).

[354] B. L Hu and A. Roura, Phys. Rev. D 76, 1240189 (2007).

[355] A. Einstein, Ann. Physik 17, 549 (1905).

[356] A. Einstein, Ann. Physik 19, 371 (1906).

[357] H. Nyquist, Phys. Rev. 32, 110 (1928).

[358] H. B. Callen and T. A. Welton, Phys. Rev. 83, 34 (1951).

[359] H. B. Callen and R. F. Greene, Phys. Rev. 86, 702 (1952).

98
[360] J. Weber, Phys. Rev. 101, 1620 (1956).

[361] R. Kubo, J. Phys. Soc. Japan 12, 570 (1957).

[362] W. Bernard and H. B. Callen, Rev. Mod. Phys. 31, 1017 (1959).

[363] R. Kubo, Rep. Prog. Phys. 29, 255 (1966).

[364] L. Landau, E. Lifshitz and L. Pitaevsky, Statistical Physics, Vol. 1 (Pergamon, London, 1980).

[365] R. Kubo, M. Toda and N. Hashitsume, Statistical Physics II, (Springer-Verlag, Berlin, 1985).

[366] J. B. Hartle and S. W. Hawking, Phys. Rev. D 13, 2188 (1976).

[367] G. Gibbons and M. J. Perry, Proc. Roy. Soc. Lon. A 358, 467 (1978).

[368] S. W. Hawking and D. Page, Comm. Math. Phys. 87, 577 (1983).

[369] D. J. Gross, M. J. Perry and L. G. Yaffe, Phys. Rev. D 25, 330 (1982).

[370] A. Rebhan, Nucl. Phys. B 351, 706 (1991).

[371] A. Rebhan, Nucl. Phys. B 368, 479 (1992).

[372] A. P. de Almeida, F. T. Brandt and J. Frenkel, Phys. Rev. D 49, 4196 (1994).

[373] F. T. Brandt and J. Frenkel, Phys. Rev. D 58, 085012 (1998).

[374] B. L. Hu, Phys. Lett. B 103, 331 (1981).

[375] A. O. Caldeira and A. J. Leggett, Physica A 121, 587 (1983).

[376] H. Grabert, P. Schramm and G. L. Ingold, Phys. Rep. 168, 115 (1988).

[377] B. L. Hu, J. P. Paz and Y. Zhang, Phys. Rev. D 45, 2843 (1992).

[378] B. L. Hu, J. P. Paz and Y. Zhang, Phys. Rev. D 47, 1576 (1993).

[379] H. A. Weldon, Phys. Rev. D 26, 1394 (1982).

[380] R. D. Sorkin. “Two topics concerning black holes: Extremality of the energy, fractality of the
horizon”, in proceedings of the Conference on Heat Kernel Techniques and Quantum Gravity,
Discourses in Mathematics and its Applications vol. 4, aditor S. A. Fulling (College Station,
Texas, University of Texas Press, 1995).

[381] A. Casher, F. Englert, N. Itzhaki, S. Massar, R. Parentani, Nucl. Phys. B 484, 419 (1997).

[382] D. Marolf, “On the quantum width of a black hole horizon”, in Particle physics and the
universe, Springer Proceedings in Physics, Vol. 98, edited by J. Trampetic and J. Wess,
(Springer-Verlag, 2005).

[383] P. O. Mazur and E. Mottola, Proc. Nat. Acad. Sci. 111, 9545, (2004).

99
[384] J. D. Bekenstein, in Quantum Theory of Gravity edited by S. M. Christensen , (Adam Hilger,
Bristol, 1984).

[385] D. N. Page, Phys. Rev. D 13, 198 (1976).

[386] P. Candelas, Phys. Rev. D 21, 2185 (1980).

[387] James W. York and Bjoern S. Schmekel, Phys. Rev. D 72, 024022 (2005).

[388] S. P. Trivedi, Phys. Rev. D 47, 4233 (1993).

[389] A. Strominger and S. P. Trivedi, Phys. Rev. D 48, 5778 (1993).

[390] F. C. Lombardo, F. D. Mazzitelli, and J. G. Russo, Phys. Rev. D 59, 064007 (1999).

[391] R. Parentani, Int. J. Theor. Phys. 40, 2201 (2001).

[392] R. Parentani, Int. J. Theor. Phys. 41, 2175 (2002).

100