PHYSICAL REVIEW D 111, 066001 (2025)

Gravity from entropy

*
Ginestra Bianconi
School of Mathematical Sciences, Queen Mary University of London, London, E1 4NS, United Kingdom

(Received 26 August 2024; accepted 27 January 2025; published 3 March 2025)

Gravity is derived from an entropic action coupling matter fields with geometry. The fundamental idea is
to relate the metric of Lorentzian spacetime to a quantum operator, playing the role of an renormalizable
effective density matrix and to describe the matter fields topologically, according to a Dirac-Kähler
formalism, as the direct sum of a 0-form, a 1-form and a 2-form. While the geometry of spacetime is
defined by its metric, the matter fields can be used to define an alternative metric, the metric induced by the
matter fields, which geometrically describes the interplay between spacetime and matter. The proposed
entropic action is the quantum relative entropy between the metric of spacetime and the metric induced by
the matter fields. The modified Einstein equations obtained from this action reduce to the Einstein
equations with zero cosmological constant in the regime of low coupling. By introducing the G-field, which
acts as a set of Lagrangian multipliers, the proposed entropic action reduces to a dressed Einstein-Hilbert
action with an emergent small and positive cosmological constant only dependent on the G-field. The
obtained equations of modified gravity remain second order in the metric and in the G-field. A canonical
quantization of this field theory could bring new insights into quantum gravity while further research might
clarify the role that the G-field could have for dark matter.

DOI: 10.1103/PhysRevD.111.066001

I. INTRODUCTION In this work, a continuum modified theory of gravity

The relation between general relativity, statistical based on a statistical mechanics action is considered. This
mechanics and information theory is a central research theory treats the metric at each point of spacetime as a
topic in theoretical physics. The interest in the subject has “renomalizable” density matrix, or more precisely a local
its roots in the discovery that black holes have an quantum operator. This central idea is relating geometry
entropy [1,2] and emit Hawking radiation [3]. Recently, with the mathematical foundation of quantum field
important results have been obtained relating information theory [33] and is inspired by the used of von Neumann
theory, entanglement entropy [4–6] and gravity [7–13] algebras in explaining entanglement in field theories [17,18]
involving the holographic principle [14–16], the entangle- and quantum gravity [34–36]. While the geometry of
ment properties of quantum field theory and the theory of spacetime is defined through its associated metric, the
von Neumann algebras [17,18]. interplay between matter fields and geometry is captured by
These results define a very active research direction [19] the metric induced by the matter fields which describes how
indicating that the quest for an ultimate gravitational the matter fields effectively curve spacetime. Embracing a
theory based on information theory and statistical mechan- statistical mechanics approach to gravity, this work inter-
ics is ongoing. A comprehensive statistical mechanics prets these metric tensors as quantum operators and
approach to gravity is expected to give rise to modified postulates an action for gravity given by the quantum
Einstein equations [20–22] that on the one side can be relative entropy between the metric of the manifold and the
testable experimentally [23,24] while on the other side metric induced by the matter fields. From the mathematical
can bring important conceptual insights into the ultimate point of view, the quantum relative entropy proposed in this
work is strictly related to the Araki quantum relative
theory for black holes [25], dark matter [26] and quantum
entropy for von Neumann algebras [17,37–39]. From the
gravity [27–32].
physics point of view, the proposed action fully describes
how matter curves geometry and how geometry affects the
*
Contact author: [email protected] matter fields.
A crucial aspect of the proposed theory is the adoption of
Published by the American Physical Society under the terms of a topological (Dirac-Kähler like [40,41]) description
the Creative Commons Attribution 4.0 International license.
Further distribution of this work must maintain attribution to of bosonic matter fields. Note that the extension of
the author(s) and the published article’s title, journal citation, Dirac-Kähler and staggered fermions formalism to bosonic
and DOI. particles is gaining increasing interest in lattice gauge

2470-0010=2025=111(6)=066001(17) 066001-1 Published by the American Physical Society

GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)

theory [42,43] and in network theory as well [44]. These postulate the entropic action for gravity and we derive the
bosonic matter fields are described as the direct sum of a corresponding modified Einstein equations. In Sec. IV we
0-form, a 1-form, and a 2-form defined on the Riemannian provide the concluding remarks. The paper also includes
manifold describing spacetime. Moreover the metric three Appendixes discussing possible extensions of the
induced by the topological matter fields is expressed in proposed theoretical framework, providing the mathemati-
terms of the Hodge-Dirac operator [45]. From this stat- cal background and all the necessary details regarding the
istical mechanics approach to gravity we derive the notation used in Sec. III, and establishing the connection of
modified Einstein equations by introducing an auxiliary the present theory with the theory of local quantum
field associated to gravity which we call G-field. The operators and the Araki entropy.
introduction of this new field is justified as it acts as a set of
Lagragian multipliers enforcing linear constraints on metric II. MOTIVATION OF THE THEORY AND
induced by the matter fields. In this way the G-field extends PRELIMINARY CONSIDERATIONS
the popular use of the Legendre transformation in fðRÞ A. Eigenvalues and logarithm of rank 2-tensors
theories [20,22]. Given the particular entropic structure of
the action, the modified Einstein equations take a very Spacetime is described by a torsion-free, d-dimensional
simplified expression. The gravitational part of the action Riemannian manifold K associated with a Lorentzian
takes the form of a dressed Einstein-Hilbert action in which metric gμν of signature f−1; 1; ; 1…; 1g and a metric
we observe an emergent positive cosmological constant compatible Levi-Civita connection Γσνμ determining the
that depends exclusively on the G-field. covariant derivative ∇μ . In order to formulate our statistical
This work greatly expands on previous results [44] mechanics and information theory action for gravity, we
obtained in the discrete setting by the same author. On need first to define the eigenvalues the logarithm of rank-2
one side, here a continuum and fully Lorentz invariant tensors Ĝ. To this end we first define the eigenvalues λ and
theory is proposed. This progress is based on the develop- ðλÞ
eigenvectors V ν of the covariant tensor Ĝ of elements Ĝμν
ment of the suitable mathematical framework to define the in a Lorentz invariant way. These satisfy the eigenvalue
Lorentz invariant entropy and cross-entropy between problem
the metric of the spacetime and the metric induced by
the matter fields. On the other side, here the relation of this ðλÞ
Ĝμν ½V ðλÞ ν ¼ λV μ : ð1Þ
statistical mechanics/information theory action with the
Einstein-Hilbert action [46,47] is established defining a
We say that a rank-2 tensor is positively defined if all its
clear connection to gravity. Two fundamental aspects of
eigenvalues are positive. We notice that this definition of
this work that are not present in Ref. [44] are relevant to
the eigenvalue of a rank-2 tensor reduces to the definition
clearly relate this approach to gravity. First, this work
considers a local theory, defining the entropy of the metric of the eigenvalue of the matrix Ĝg−1 as Eq. (1) can be
at each point of spacetime, while the previous work rewritten as
considers only the entropy associated to the full metric ðλÞ
of the higher-order network. The present local theory Ĝμσ gσν ½V ðλÞ ν ¼ λV μ : ð2Þ
allows a closer connection to gravity and constitutes a
step forward to establish the connection between this One striking consequence of this definition is that the
approach and the quantum theory of entanglement [17]. eigenvalues of the metric gμν are all identically equal to one.
Secondly, by adopting the continuum limit, in this work the Assuming that the tensor Ĝμν is positively defined, we
intrinsic difficulty related to the definition of the curvature define the logarithm of this tensor as
of networks, simplicial and cell complexes is avoided. ðλÞ ðλÞ
To keep the discussion concise the focus is here mostly ½lnðĜÞμν ¼ V μ V ν lnðλÞ: ð3Þ
on scalar (bosonic) matter fields, and their topological
generalizations with a brief mention in Appendix A of the and the inverse of a tensor as
natural extension of this framework to Abelian gauge
fields, while in Ref. [44] the theory covers also fer- ½Ĝ−1 μν ¼ ½V ðλÞ μ ½V ðλÞ ν λ−1 : ð4Þ
mionic matter fields. Further extensions of the proposed
local framework to Dirac and non-Abelian gauge fields It follows that if the tensor Ĝ is invertible, the logarithm of
[44,48,49] in the continuum or in the discrete setting are left the inverse of a positively defined tensor is given by
for future investigations.
This work is structured as follows. In Sec. II we provide ½lnðĜ−1 Þμν ¼ −½V ðλÞ μ ½V ðλÞ ν lnðλÞ: ð5Þ
the motivation of the proposed theory and we discuss
preliminary results on an instructive warm-up scenario. In Finally we define the trace of a rank two tensor as the sum
Sec. III we outline the proposed theoretical framework, we of its eigenvalues, i.e.,

066001-2
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)
X
TrĜ ¼ λ: ð6Þ H ¼ Trg ln g−1 ¼ 0: ð9Þ
λ
The fundamental assumption of the present theory is that
Thus the trace of a rank-2 tensor can be also calculated as spacetime is endowed with two metrics: the metric g fully
usual in tensor calculus, as the trace of the matrix Ĝg−1 , i.e., determining the spacetime geometry and the metric induced
by the matter fields G that fully capture the interplay
TrĜ ¼ TrM Ĝg−1 ¼ Ĝμν gμν : ð7Þ between matter fields and geometry. Leaving the detailed
discussion about the metric induced by the matter fields to
In this first section we are interested exclusively on rank- the next paragraph, we postulate that the action should
two tensors that are metrics between vectors (and 1-forms). explicitly express the relation between these two metrics
In the subsequent paragraphs and in Appendix B we will and their reciprocal coupling. By embracing a statistical
extended the notion of eigenvalues also to metric matrices mechanics approach we consider the Lagrangian L given
between bivectors (and 2-forms) represented by rank-4 by the quantum relative entropy between the metric g and
tensors. Such an approach will be shown in Appendix B to the metric G induced by the matter field defined as
be general and applicable to metric tensors between two
n-vectors (an n-forms) with n of any order.
L ¼ −Trg ln g−1 þ Trg ln G−1 : ð10Þ
B. Warm-up scenario: Entropic action
Using H ¼ 0 we thus obtain
Before we develop our theory, let us consider an
instructive warm-up scenario that will help justify the L ¼ Trg ln G−1 : ð11Þ
theory that we will present in the following. Having defined
the logarithm of positively defined rank-2 tensors, we are
We observe that since g has all the eigenvalues equal to the
now in the position to define their Lorentz invariant
quantum entropy H for metric associated to 1-forms. identity, and Eq. (2) holds for Ĝ ¼ G, the Lagrangian L can
This is inspired by the expression of the von Neumann be also be expressed as
entropy albeit we do not require the tensor to have trace X
one at every point in spacetime. Thus strictly speaking L ≡ −TrM ln Gg−1 ¼ − lnðλ0 Þ; ð12Þ
we interpret the generic metric tensor as a quantum λ0
operator [33], which has the physical interpretation of a
renormalizable effective density matrix [18]. We consider where with TrM ln Gg−1 we indicate the usual trace of the
positive definite, invertible metric for 1-forms, represented logarithm of the matrix ln Gg−1 and with λ0 the eigenvalues
by the covariant tensor Ĝ of rank two. We define the of the 2-tensor G as defined in Eq. (2).
entropy H of Ĝ as An important point here is that the Lorentz invariant
definition of the quantum relative entropy given by Eq. (12)
X
H ¼ TrĜ ln Ĝ−1 ¼ Ĝμν ½ln Ĝ−1 νμ ¼ − λ ln λ: ð8Þ requires that the matrix G has to be invertible as well as the
matrix g. Thus this is another difference with respect to the
Note that the problem of defining eigenvalues and entropy strictly speaking density matrices that can be semidefinite
of tensors is a topic of intense research and similar positive. We are now in the position to consider an action S
definitions have been provided for instance in the theory associated to the Lagrangian L, as given by
of elasticity [50] and in applied tensor analysis [51] Z pffiffiffiffiffiffiffiffiffiffiffi
(although not defined in a Lorentz invariant way). In the 1
S¼ d j − gjLdr; ð13Þ
present warm-up scenario we will define the entropy and lP
the quantum relation entropy of different metrics taking the
form of rank-2 tensors. As a reference for the notation used, where j − gj indicates the absolute value of the determinant
the reader can refer to Table I. Recalling that the metric g of g and lP indicates the Planck length.
has all the eigenvalues equal to one, it follows that the We note that the quantum relative entropy [17,37,38] is a
entropy of the metric g is null, i.e., central quantity in quantum information [39], in the theory

TABLE I. Notation used to indicate the different metrics in the warm-up-scenario.

Covariant metrics Elements Definition

g gμν Default covariant rank-2 metric tensor between vectors associated to the manifold K
G Gμν Covariant rank-2 metric between vectors induced by the matter fields
Ĝ Ĝμν General covariant rank-2 metric tensor including both g and G

066001-3
GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)

of local quantum operators [33] and the mathematical where here and in the following ¯· indicates complex
foundations of quantum gravity [34–36]. conjugation. Since G should be adimensional it is conven-
Although we are not aware of previous interpretations of ient to work in the units ℏ ¼ c ¼ 1 and to put α ¼ α0 ldP
the metrics as quantum operators, it is well known that the where lP is the Planck length and α0 is adimensional.
quantum relative entropy can also be defined among We observe that in the limit in which the field ϕ is real,
quantum operators that generalize density matrices. In i.e., ϕðxÞ ∈ R and the metric flat and Euclidean, i.e.,
particular, in the fundamental theory of quantum operators gμν ¼ δμν , the metric induced by the matter fields G given
algebras [37,38], quantum operators generalize the notion by Eq. (14) reduces to the first fundamental form of Gauss
of density matrices in an analogous way of the metric for the d manifold immersed in Rdþ1 defined by the set of
tensors adopted here. In particular in this theory, quantum points ðx; ϕðxÞÞ. For the underlying mathematical treat-
operators, like our metric matrices, might admit a finite ment of the metric induced by real scalar fields we refer the
trace at each point of the manifold but this trace might be interested reader to Ref. [52].
not unitary and might be a function of the considered Let us define the scalar product j∇ϕj2 as
point of the manifold. For these quantum operators the
Araki quantum relative entropy [37,38] is defined. This j∇ϕj2 ¼ ∇μ ϕ̄gμν ∇ν ϕ: ð16Þ
entropy reduces to the von Neumann entropy when the
quantum operators reduce to density matrices of unitary We observe that the inverse G−1 of the induced metric G
trace but is defined also among quantum operators of has metric given by
nonunitary trace.
While we leave the discussion of the relation of our Mμν
entropic action to the Araki quantum relative entropy to ½G−1 μν ¼ gμν − α : ð17Þ
1 þ αj∇ϕj2
Appendix C, a series of physical observations are here in
place to justify our interpretation of the metrics matrices as
Adopting this notation we observe that the logarithm of
quantum operators. As remarked previously we treat the
the induced metric ln G and the logarithm of ln G−1 have
metric tensors as quantum operators or effective density
elements
matrices. The main differences between the metric tensor
and the density operators include the fact that we require
that the metric matrices are invertible and we do not require ½ln Gμν ¼ fðj∇ϕj2 ÞMμν ;
that they have unitary trace at each point of the manifold. ½ln G−1 μν ¼ −fðj∇ϕj2 ÞMμν ; ð18Þ
The requirement of treating exclusively invertible metrics g
and G is dictated by our desire to define the entropy in a where
Lorentz invariant way. In fact here we desire to treat the
metrics g and G on the same footing of their inverse g−1 and lnð1 þ αwÞ
G−1 . It might be argued that also our relaxation of the fðwÞ ¼ : ð19Þ
w
requirement to have density matrices of trace one has
similar roots. In fact if we require g and G to have unitary where fðwÞ → α for jαwj ≪ 1. With this choice of the
trace at each point of the manifold their inverse as well as metric induced by the matter field we obtain that the
their dual (defined in Appendix C) in general will not have Lagrangian (11) reads
a unitary trace.
L ¼ − lnð1 þ αj∇ϕj2 Þ: ð20Þ
C. Warm-up scenario: Scalar matter fields
We now apply this warm-up scenario in presence of We observe that in the limit αj∇ϕj → 0 and then we have
scalar matter fields. We consider the complex valued scalar L → −αj∇ϕj2 , i.e., we recover the Lagrangian correspond-
matter field ϕðxÞ ∈ C with x ∈ K and we consider the ing to the massless Klein-Gordon equation.
d-dimensional manifold immersed in K ⊗ C defined by the By minimizing the action S with respect to the field ϕ
points ðx; ϕðxÞÞ. The metric G induced on K by this matter and to the metric g we obtain the Euler-Lagrange equations
fields is given by of motion,

G ¼ g þ αM; ð14Þ ∇μ hðj∇ϕj2 Þgμν ∇ν ϕ ¼ 0; ð21Þ

where α is a real positive parameter, and M is the rank-2 where hðwÞ is given by
tensor of elements
α
Mμν ¼ ð∇μ ϕ̄Þð∇ν ϕÞ; ð15Þ hðwÞ ¼ : ð22Þ
1 þ αw

066001-4
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)

In the limit αw → 0 we have hðwÞ → α and the equation for and its conjugate topological field hΦj as
the scalar field reduces to the simple massless Klein-
Gordon equation. By putting equal to zero the variation hΦj ¼ ϕ̄ ⊕ ω̄μ dxμ ⊕ ζ̄μν dxμ ∧ dxν : ð25Þ
of the action δS ¼ 0 with respect to the metric g, we get
The considered covariant metric g̃ is defined as the direct
1 sum of the metric among scalars (the identity), the metric g
δS ¼ −hðj∇ϕj2 ÞM μν − Lgμν ¼ 0: ð23Þ
2 among vectors introduced previously, and the metric gð2Þ
among bivectors given by
In empty spacetime M ¼ 0 and L ¼ 0, thus this equation is
automatically satisfied independently on the value of g. Thus gð2Þ ¼ gμρ gνσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ; ð26Þ
in this limit, the metric g is not determined by the action.
Let us now make some remarks about this warm-up or, exploiting the antisymmetry of the 2-forms,
derivation. We proposed a statistical mechanics framework
that is very inspiring as we get the massless Klein-Gordon gð2Þ ¼ ½gð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ; ð27Þ
equation as the outcome of the minimization of a quantum
entropy action for low coupling, i.e., 0 < αj∇ϕj2j ≪ 1.
However this approach has two important limitations. The with
first limitation is that the Klein-Gordon equation does not
1
contain the mass term. The second limitation is that in ½gð2Þ μνρσ ¼ ðgμρ gνσ − gμσ gνρ Þ: ð28Þ
absence of matter fields the metric is not determined. In the 2
following we provide a more comprehensive framework to It then follows that g̃ is given by
address these two limitations. This framework can be related
to gravity as it gives rise to modified Einstein equations that g̃ ¼ 1 ⊕ gμν dxν ⊗ dxν
reduce to the Einstein equations for low coupling.
⊕ ½gð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ: ð29Þ
III. THE ENTROPIC THEORY OF MATTER
FIELDS COUPLED TO GEOMETRY The local scalar product among topological fields is
defined as
A. Topological matter fields and their associated
metrics hΦjΦi ¼ jϕj2 þ ω̄μ ωμ þ ζ̄ μν ζμν ; ð30Þ
In order to derive gravity from our entropic action we
need to consider the topological bosonic matter field. The where ωμ ωμ ¼ ων gνμ ωμ , while the local outer product is
topological bosonic matter field is a type of Dirac- given by
Kähler [40,41] boson given by the direct sum of a 0-form,
a 1-form and a 2-form. Topological bosonic fields are jΦihΦj ¼ ϕ̄ϕ ⊕ ðω̄μ ων dxμ ⊗ dxν Þ
receiving increasing attention in discrete theories devel- ⊕ ζ̄ μν ζρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ: ð31Þ
oped in network [44] and lattice gauge theories [42,43].
Taking into consideration topological bosonic fields will Indicating here with d the differential operator and with δ
allow us to introduce in the metric induced by the matter the codifferential operator, we define the Dirac operator D
fields terms depending on the mass of the bosonic field. as D ¼ δ þ d, which is restricted to the space of topologi-
Thus, in this way, we address the first limitation of the cal bosons. Thus we define the action of D over jΦi as
warm-up scenario that we have presented above.
In order to address the second limitation of the warm-up DjΦi ¼ −∇μ ωμ ⊕ ð∇μ ϕ − ∇ρ ζρμ Þdxμ
scenario discussed previously, we include in the expression
of the metric induced by the matter field also terms ⊕ ∇μ ων dxμ ∧ dxν : ð32Þ
depending directly on the curvature of the manifold.
These terms will be expressed in terms of the Ricci scalar The metric G̃ induced by the topological matter field will
R, the Ricci (covariant) tensor R of elements Rμν and the have structure similar to g̃ that we can write in full
Riemann tensor of elements Rμνρσ . generality as
We define the topological fields jΦi as the direct sum
between a 0- (complex valued) form ϕ and a (complex G̃ ¼ Gð0Þ ⊕ ½Gð1Þ μν dxμ ⊗ dxν
valued) 1-form ωμ dxμ , and a (complex valued) 2-form ⊕ ½Gð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ: ð33Þ
ζμν dxμ ∧ dxν , with ζμν ¼ −ζνμ i.e.,
Proceeding as in the warm-up scenario we might wish to
jΦi ¼ ϕ ⊕ ωμ dxμ ⊕ ζμν dxμ ∧ dxν ; ð24Þ define the metric induced by the topological field as

066001-5
GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)

TABLE II. Covariant topological metrics used in the general scenario.

Covariant topological metrics Interpretation

ν ν μ ν ρ σ
g̃ ¼ 1 ⊕ gμν dx ⊗ dx ⊕ ½gð2Þ μνρσ ðdx ∧ dx Þ ⊗ ðdx ∧ dx Þ Default covariant topological metric between
the topological fields associated to the manifold K
G̃ ¼ Gð0Þ ⊕ ½Gð1Þ μν dxμ ⊗ dxν ⊕ ½Gð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ Covariant topological metric induced by the matter fields

TABLE III. Covariant metric tensors between n-vectors that are included in the covariant topological metrics.

Covariant metric tensors Elements Interpretation

1 1 Default metric tensor between scalars
g gμν Default metric tensor between vectors
gð2Þ ½gð2Þ μνρσ Default metric tensor between bivectors
Gð0Þ Gð0Þ Metric tensor between scalars induced by the matter fields
Gð1Þ ½Gð1Þ μν Metric tensor between vectors induced by the matter fields
Gð2Þ ½Gð2Þ μνρσ Metric tensor between bivectors induced by the matter fields

G̃ ¼ g̃ þ αðDjΦihΦjDÞ: ð34Þ From these considerations, it follows that in the induced

metric G̃, we will substitute the term DjΦihΦjD with M̃
defined as
Thus in this general scenario we consider the two covariant
topological metrics g̃ and G̃ each given by the direct sum of
M̃ ¼ DjΦihΦjD þ ðm2 þ ξRÞjΦihΦj: ð36Þ
metrics among scalars, vectors and bivectors. As a refer-
ence on our notation for this general scenario we refer the
As we will see in the following paragraph this choice will
reader to the Table II and the Table III.
allow us to effectively overcome the first limitation of the
In our interpretation of the metric as a quantum density
warm-up scenario and to recover the Klein-Gordon equa-
matrix, this would correspond to density matrix corre-
tion in curved spacetime in full. Furthermore we consider
sponding to a pure state DjΦi. However, if the induced
also the additional term proportional to R̃ and we postulate
metric G̃ is interpreted as a density matrix, it is natural to
that the metric G̃ induced by the geometry and the matter
add further terms in the metric G̃. These terms will depend
fields is given by
on the topological field and the geometry of the space
which will allow us to describe mixed states. First, we
G̃ ¼ g̃ þ αM̃ − βR̃: ð37Þ
introduce in G̃ a term proportional to jΦihΦj. Specifically
we introduce a term ðm2 þ ξRÞjΦihΦj where R is the Ricci
scalar and ξ might include the case of conformal coupling As we will see in the following the addition of the term
ξ ¼ ðd − 2Þ=ð4ðd − 1ÞÞ. This term can be also interpreted proportional to R̃ will allow also us to overcome the
as a projector. Secondly, we introduce a term depending second limitation of the warm-up scenario. Note that here
explicitly on the curvature of the manifold. Since G̃ α, β are positive constants. In particular, since we require G̃
involves the metric for scalars, vectors and bivectors on to be adimensional, we need to consider α ¼ α0 ldP and
equal footing, it is natural to consider a further term β ¼ β0 l2P where lP is the Planck length and α0 ; β0 are
involving R̃ given by the direct sum of the Ricci scalar adimensional in the units ℏ ¼ c ¼ 1.
R, the Ricci tensor of elements Rμν and the Riemann tensor In the main body of this paper we will investigate only
of elements Rμνρσ , i.e., (bosonic) Dirac-Kähler matter fields. However gauge fields
and fermionic Dirac fields can be included as well. For a
discussion of the inclusion of Abelian gauge field see
R̃ ¼ R ⊕ ðRμν dxμ ⊗ dxν Þ Appendix A.
⊕ Rμνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ: ð35Þ Possibly this approach could be extended to include also
higher-forms. However, for simplicity, we consider here only
topological matter fields formed by the direct sum between a
Including R̃ into the metric G̃ will allow us to describe 0-form, a 1-form and a 2-form as this the minimal choice that
more general metric matrices that are not decomposable will allow us to include in the action the Ricci scalar, the Ricci
into the sum of projectors. and the Riemann tensor explicitly.

066001-6
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)

B. Entropic topological and geometrical action L ¼ Trg̃ ln G̃−1 : ð41Þ

We propose a statistical mechanics action formulated in
terms of the quantum relative entropy between the metric g̃ By treating separately the contributions of Gð0Þ , Gð1Þ and
and the metric G̃ induced by the topological matter fields. Gð2Þ we get the explicit expression for the Lagrangian L
In order to define our action we need to extend the notion of given by
eigenvalues to metric tensors between bivectors. This will
allow us to define the entropy and the quantum relative L ≔ ln½Gð0Þ −1 þ Trg ln½Gð1Þ −1 þ Trgð2Þ ln½Gð2Þ −1 : ð42Þ
entropy in this novel framework. For a detailed discussion
of this mathematical background see Appendix B that Since the metric matrices gðnÞ have all their eigenvalues
constitutes also the reference for our notation. First of all equal to one, using the notation developed in Appendix B
we observe that the entropy associated to the metric g̃ we can express this Lagrangian as well as
remains zero. Indeed we define H̃ as
L ≔ −TrF ln G̃g̃−1 ;
−1
H̃ ¼ Trg̃ ln g̃ ¼ − ln½Gð0Þ  − TrF ln½Gð1Þ g−1
−1
¼ 1 ln 1 þ Trg ln g þ Trgð2Þ ln g−1
ð2Þ ¼ 0; ð38Þ − TrF ln½Gð2Þ −1 ½gð2Þ −1 : ð43Þ

where we have used Eq. (9) and the identity derived in the The resulting statistical mechanics action S associated to
Appendix B [in Eq. (B39)], the Lagrangian L is strictly related to the Araki quantum
relative entropy [17,37,38] and can be formally derived
Trgð2Þ ln g−1
ð2Þ ¼ 0: ð39Þ from a mathematical theory of quantum operators (see
Appendix C). Specifically, the considered action is given by
This equation is a direct consequence of the fact that not Z pffiffiffiffiffiffiffiffiffiffiffi
1
only g but also gð2Þ has all the eigenvalues equal to one. S¼ d j − gjLdr: ð44Þ
We are now in the position to consider the Lagrangian lP
given by the quantum relative entropy between g̃ and G̃,
This action defines a modified theory of gravity. In the
linearised limit α0 ≪ 1, β0 ≪ 1 this action reduces to the
L ≔ −Trg̃ ln g̃−1 þ Trg̃ ln G̃−1 ð40Þ Einstein-Hilbert action with zero cosmological constant
[46,47] coupled with the scalar topological field. Indeed
(see Fig. 1 for an illustration of the physical model beyond in this limit L reduces to
the choice of this Lagrangian). Since we have H̃ ¼ 0 we
obtain L ¼ 3βR − αhΦjDg̃−1 DjΦi
− αðm2 þ ξRÞðjϕj2 þ ω̄μ ωμ þ ζ̄ μν ζμν Þ: ð45Þ

where

hΦjDg̃−1 DjΦi ¼ j∇ϕj2 þ j∇μ ωμ j2 þ j∇ρ ζρμ j2

þ jϵμνρ ∇μ ων j2 : ð46Þ

In the interesting limit in which ωμ ¼ 0 and ζ μν ¼ 0 and the

topological matter field is only scalar, Eq. (45) reduces to the
widely studied Einstein-Hilbert action with zero cosmologi-
cal constant coupled with the scalar field [53],
FIG. 1. Schematic representation of this theoretical framework.
The metric induced by the matter field G̃ affects the metric of the L ¼ 3βR − αj∇ϕj2 − αðm2 þ ξRÞjϕj2 : ð47Þ
manifold g̃ and vice versa the metric of the manifold affects the
metric induced by the matter field. The considered Lagrangian As already anticipated, this implies, among other things, that
is given by the quantum relative entropy between the topologi-
this framework overcomes the first and the second limitations
cal metric g̃ of spacetime and the topological metric induced
of the warm-up scenario. In fact in this way we recover the
by the topological matter fields G̃. Since, by definition, the
entropy of the metric vanishes identically, the quantum relative
Klein-Gordon Lagrangian in curved spacetime in full.
entropy in the Lagrangian L reduces to a single term: the Moreover, by adding to G̃ the term proportional to the
quantum cross-entropy. curvature R̃ we solve the second limitation of the warm-up

066001-7
GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)

scenario, and we obtain that the entropic action allows for the Lagrangian multipliers that are measurable. Here we would
full determination of the metric in the vacuum. like to embrace this point of view and give a physical
meaning to the introduced Lagrangian multipliers as
C. Introduction of G-field G̃ and the field Θ̃ physical fields defined on our manifold K. Interpreting
In this section our goal is to investigate the property of in this way the Lagrangian multipliers changes the physics
the modified gravity emerging from the entropic action. To of the problem as the phase space associated to the
this end we introduce the auxiliary G-field G̃ and the equations of motion acquires new dimensions and the
fixed points as well as the initial value problem will involve
auxiliary field Θ̃. In this way we will be able to show in the
the newly introduced fields. In this latter interpretation,
following that the proposed entropic action can be inter-
introducing the auxiliary fields and giving a physical
preted as the sum of a dressed Einstein-Hilbert action with
meaning to them as encoding for new fields, we have
an emergent positive cosmological constant only depend-
reduced our theory to a theory linear in R̃. In particular, as
ing on the G-field G̃ and a matter action.
we will see below, the equations of motion will involve at
As mentioned before we can express the Lagrangian L
most second order derivatives of the fields, thus the model
given by Eq. (41) as
might avoid the Ostrogradsky instability. A definitive
answer to the stability question would follow from further
L ¼ −TrF ln G̃g̃−1 : ð48Þ
analysis (for example a Hamiltonian analysis).
The fields Θ̃ and G̃ are given by
This Lagragian is nonlinear in G̃ and thus in R̃.
Several modified gravity actions are questioned because
Θ̃ ¼ Θð0Þ ⊕ Θð1Þ ⊕ Θð2Þ ;
they give rise to theories with derivatives of the metric
higher than two. Such theories can be affected by the G̃ ¼ Gð0Þ ⊕ Gð1Þ ⊕ Gð2Þ ; ð51Þ
Ostrogradsky instability, however nonlinear theories not
suffering from this pathology are also known, most notably with Θð1Þ having elements ½Θð1Þ μ ν , Θð2Þ having elements
the fðRÞ theories [20,22]. Thus an important question is ½Θð2Þ μν ρσ and Gð1Þ having elements ½Gð1Þ μ ν , Gð2Þ having
whether the proposed theory is also affected by the
elements ½Gð2Þ μν ρσ . The term W, of the Lagrangian L̃ that
Ostrogradsky instability or rather, it generalizes and
extends the realm of viable modified gravity theories is linear in G̃g̃−1 , can be expressed as
beyond the fðRÞ theories. In particular the Ostrogradsky
instability can arise when the Lagrangian depends on X
2

derivatives of order higher than two of the fields, and thus, W ¼ TrF G̃ðG̃g̃−1 − Θ̃Þ ¼ W n; ð52Þ
n¼0
in the gravitational setting on higher powers of the
curvature R̃. with
In order to tackle this question we observe that the
Lagrangian L has an easy expression as it is only dependent W 0 ¼ Gð0Þ ðGð0Þ − Θð0Þ Þ;
on the product G̃g̃−1 . Thus, by introducing an auxiliary
field Θ̃ and imposing the constraints W 1 ¼ ½Gð1Þ ρ μ ð½Gð1Þ μν gνρ − ½Θð1Þ μ ρ Þ;
W 2 ¼ ½Gð2Þ ηθ μν ð½Gð2Þ μνρσ ½gð2Þ ρσηθ − ½Θð2Þ μν ηθ Þ: ð53Þ
G̃g̃−1 ¼ Θ̃ ð49Þ
Note that the use of Lagrange multipliers in the present
with Lagrangian multipliers G̃ (constituting another aux- theory extends the equivalence between the fðRÞ theories
iliary field) that we will call the G-field, we can reduce our and Brans-Dicke theory achieved through a Legendre
theory to a theory driven by the Lagrangian L̃ given by transformation [20,22]. However using an extensive num-
ber of Lagrange multipliers might be considered a danger-
L̃ ¼ −TrF ln Θ̃ − TrF G̃ðG̃g̃−1 − Θ̃Þ: ð50Þ ous mathematical operation. For the sake of simplicity here
we work under the assumption that L̃ is equivalent to L.
This Lagrangian is now linear in G̃g̃−1 . Note that a Whereas it might be proven that L is not equivalent to L̃ we
conservative interpretation of this transformation will
assume that our true theory is given by L̃ which we might
not give a physical interpretation to the auxiliary fields.
always consider as a “canonical” version of L.
Thus, according to this point of view, the auxiliary fields
The resulting statistical mechanics action S̃ associated to
are not changing the physics of the original Lagrangian and
this theory might be prone to Ostrogradsky instability. the Lagrangian L̃ is given by
However, we know from statistical physics that Lagrangian Z pffiffiffiffiffiffiffiffiffiffiffi
1
multipliers can also acquire a rather physical meaning. For S̃ ¼ d j − gjL̃dr: ð54Þ
instance the temperature and the chemical potential are lP

066001-8
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)
 
We are now in the position to derive the equations for the d
− TrF ΘðnÞ þ αTrF MðnÞ g−1
ðnÞ ¼ βR; ð58Þ
matter fields Φ, the metric gμν and the fields Θ̃ and G̃ by n
considering the variation of L̃ with respect to these fields.
for any n ∈ f0; 1; 2g.
D. Equation of motion for the matter fields
Here we consider the action L̃ defined in Eq. (50) 2. Variation with respect to Θ̃
depending linearly on G̃ which is encoding for the matter Thanks to the logarithmic nonlinearity in L̃ the variation
fields. By using the definition of G̃ given by Eq. (A1), we of the action S̃ with respect to Θ̃ simply reads
can derive the equation of motion of the matter fields by
considering the variation with respect to hΦj, getting, Θ̃−1 ¼ G̃: ð59Þ
Dg̃G −1 DjΦi þ g̃G −1 ðm2 þ ξRÞjΦi ¼ 0: ð55Þ
Thus we obtain that the field Θ̃ corresponds to the inverse
Here g̃G , given by of the dressing G-field G̃. Using Eq. (57) we find that the
equation of motion for the G-field G̃ is given by
g̃G ¼ G̃−1 g; ð56Þ
G̃−1 ¼ Ĩ þ αM̃g̃−1 − βR̃g̃−1 ; ð60Þ
can be interpreted as the dressed metric which affects the
matter fields. Specifically, g can be seen as a bare metric and we can simply eliminate the field Θ̃ from the action. In
and the G-field G̃−1 can be seen as a dressing of this metric this way the considered action S̃ can be decomposed into
that gives rise to the dressed metric g̃G given by Eq. (56). two terms as

E. Modified gravity S̃ ¼ βS G þ αS M ; ð61Þ

We now turn to the equations for modified gravity
considering variation of our action S̃ with respect to the with
metric g and to the fields Θ̃ and G̃. Z pffiffiffiffiffiffiffiffiffiffiffi
1
SG ¼ d j − gjLG dr;
lP
1. Variation with respect to G̃ Z pffiffiffiffiffiffiffiffiffiffiffi
1
The variation with respect to the fields G̃ enforces the SM ¼ d j − gjLM dr: ð62Þ
lP
constraints in Eq. (49) which, by using the expression for G̃
given by Eq. (A1), can be expressed explicitly as
The Lagrangians LG and LM appearing in the above
definition of S G and S M can be expressed as
Θ̃ ¼ Ĩ þ αM̃g̃−1 − βR̃g̃−1 ; ð57Þ
LG ¼ ðRG − 2ΛG Þ; LM ¼ −MG : ð63Þ
where Ĩ is the (topological) identity. Since α ¼ α0 l4P and
β ¼ β0 l2P in the regime of low energies and low curvatures, where RG is the dressed Ricci scalar and MG includes all
we have that Θ̃ is a small perturbation of the identity Ĩ. dependence with the matter fields jΦi while ΛG is the
Note that these are the equations that determine the emergent positive cosmological constant. Specifically we
relation between the fields Θ̃ the curvature R̃ and the have
matter field determining M̃. We can write these equations
by separating the contribution corresponding to the metric RG ¼ TrF g̃−1
G R̃; MG ¼ TrF g̃−1
G M̃;
for 0-forms, 1-forms and 2-forms obtaining
1
ΛG ¼ Tr ðG̃ − Ĩ − ln G̃Þ: ð64Þ
Θð0Þ ¼ 1 þ αMð0Þ − βR; 2β F
½Θð1Þ μ ν ¼ δμ ν þ α½Mð1Þ μρ gρν − βRμρ gρν ; From this reformulation of the action S̃ we see that the
ν⃗ ρ⃗
½Θð2Þ μ⃗ ¼ δμ⃗ þ α½Mð2Þ μ⃗ ρ⃗ ½gð2Þ  ρ⃗ ν⃗ ρ⃗ ν⃗
− βRμ⃗ ρ⃗ ½gð2Þ  ; Lagrangian LG has a the structure of a dressed Einstein-
Hilbert action and depends on the metric and the G-field G̃.
where we have indicated a pair of indices with a vector In particular the Ricci scalar R is substituted to the dressed
symbol, e.g., μ⃗ . Since the trace of the Ricci tensor as well as Ricci scalar RG and the role of the cosmological constant is
the trace of the Riemann tensor are both equal to the Ricci played by ΛG that is nonnegative and only dependent on the
scalar, upon performing the trace of these expressions, we G-field G̃. Whenever this field is close to the identity, e.g.,
get some constraints on the trace of the fields ΘðnÞ , i.e., G̃ ≃ Ĩ þ ϵ̃ we obtain that ΛG is positive but very small,

066001-9
GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)

i.e., its first nontrivial term is quadratic in ϵ̃. Thus the and the gravity dressed Lagrangians Eq. (63) respectively.
emergent positive cosmological constant ΛG is small in this The gravitational dressed action LG displays an emergent
limit and only determined by the G-field. Finally we positive cosmological constant ΛG only dependent on the
recover as already noticed, that the action S̃ reduces to G-field and given by Eq. (64). The equations of motion
the Einstein-Hilbert with zero cosmological constant in the associated to this action involve solving Eq. (55) for the
low coupling limit. matter fields, and Eqs. (60) and (66) for the metric and the
G-field. All these equations involve at most second
3. Variation with respect to g derivatives of the fields.

The modified Einstein equations are obtained by per- IV. CONCLUSIONS

forming the variation of the action S̃ with respect to g. Let
us define the stress-energy tensor T in the usual way with This work proposes a modified theory of gravity
elements T μν given by emerging from statistical mechanics and information theory
action. The fundamental idea of this theory is to associate
1 δS the metric to a quantum operator, playing the role of a
− pffiffiffiffiffiffiffiffiffiffiffi M ¼ T μν : ð65Þ renormalizable and effective density matrix. In particular
j − gj dgμν
two metrics are discussed: the metric of spacetime that fully
With this notation we can expressed the modified Einstein defines its geometry and the metric induced by matter fields
equations as that are effectively curving the space. The interplay
between geometry of spacetime and matter fields is
1 explicitly captured by the proposed action given by the
RGðμνÞ − gμν ðRG − 2ΛG Þ þ DðμνÞ ¼ κT ðμνÞ ; ð66Þ quantum relative entropy between the metric and the metric
2
induced by the matter fields.
where κ ¼ α=β, ðμνÞ indicates symmetrization of the Here a Lorentzian theory consistent with this fundamen-
indices, RGμν are the elements or the dressed Ricci tensor tal physical interpretation of gravity is formulated. In order
given by to do this we have built the necessary mathematical
background to define the entropy and the cross-entropy
RGμν ¼ Gð0Þ Rμν þ ½Gð1Þ μ ρ Rρν − ½Gð2Þ ρ ρ μη Rν ηρ1 ρ2 of metric tensors in a Lorentzian well defined way. A
1 2
modified gravity is emerging from this framework when the
ηρ1 ρ2
þ 2½Gð2Þ μ Rρ1 ρ2 νη ; ð67Þ matter fields are described by topological Dirac-Kähler
bosons formed by the direct sum between a 0-form, a
and Dμν are the elements depending on second derivatives 1-form and a 2-form and the induced metric also depends
of the G-field G̃ given by on the curvature of spacetime. The modified Einstein
equations reduce to the Einstein equations in the regime
Dμν ¼ ð∇ρ ∇ρ gμν − ∇μ ∇ν ÞGð0Þ − ∇ρ ∇ν ½Gð1Þ ðρμÞ of low coupling. By introducing the G-field we obtain the
modified Einstein equations and the equations of motion
1 1
þ ∇ρ ∇ρ ½Gð1Þ μν þ ∇ρ ∇η ½Gð1Þ ρη gμν for the matter and the G-field. From this theory it emerges
2 2 that the proposed entropic action takes the form of a sum
þ ∇η ∇ρ ½Gð2Þ μρνη þ ∇ρ ∇η ½Gð2Þ ημρν between a dressed Einstein-Hilbert action and a matter
1 action. Interestingly, the dressed Einstein-Hilbert action
þ ½∇ρ ; ∇η ½Gð2Þ ρημν : ð68Þ displays an emergent positive cosmological constant that
2
only depends on the G-field. Moreover, thanks to the
It follows that the modified Einstein equations involve only introduction of the G-field, the equations of modified
second derivatives of the metric and second derivatives of gravity remain second order in the metric, in the matter
the field G̃. However these equations might have more and in the G-field.
solutions than the Einstein equations. A detailed study of The interpretation of the topological metrics tensors as
the solutions and the viability of these equations is beyond quantum operators, or effective density matrix where we
the scope of this work and will be the subject of future have relaxed the constraints of having unitary trace, and we
investigations. have required the existence of the inverse, is shown to be
very useful. These choices are motivated here by the
necessity to have a Lorentzian invariant theory. Here we
4. Discussion have established the relation of the adopted quantum
In summary the introduction of the G-field turns the relative action with Araki quantum relative entropy [37],
proposed entropy action Eq. (41) into the action Eq. (61) opening the perspective of using the theory of quantum von
involving a dressed Einstein-Hilbert action and a dressed Neumann algebras and the theory of entanglement [17] to
matter action given by Eq. (62) depending on the matter investigate further the properties of the proposed theory.

066001-10
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)

In conclusion, we hope that this approach can help 1. Eigenvalues of metric matrices GðnÞ
identify the deep connections among gravity, quantum A 1-form can be written as
mechanics and statistical physics. Given the interpretation
of the metrics as a quantum operators our hope is that ω ¼ ωρ dxρ : ðB1Þ
this approach will also be instrumental to formulating
approaches to quantum gravity in second quantization. The metric Gð1Þ can be interpreted as a quantum operator
Finally future investigation might explore the role of the
among two of such 1-forms and can be written as
G-field in dark matter. Future directions in this line of research
involve also the investigation of the proposed entropic action
Gð1Þ ¼ ½Gð1Þ μν dxμ ⊗ dxν ; ðB2Þ
under the renormalization group, and possible connections
with phenomenology and experimental results.
where ½Gð1Þ μν is Hermitian. The · product between the
ACKNOWLEDGMENTS metric Gð1Þ and the generic 1-form ω is defined as

This work was partially supported by a grant from the Gð1Þ · ω ¼ ω̂ ¼ ω̂μ dxμ ; ðB3Þ
Simons Foundation. The author would like to thank the
Isaac Newton Institute for Mathematical Sciences, where, by definition
Cambridge, for support and hospitality during the program
Hypergraphs: Theory and Applications, where work on this ω̂μ ≔ ½Gð1Þ μν ωρ hdxν ; dxρ i; ðB4Þ
paper was undertaken. This work was supported by EPSRC
Grant No. EP/V521929/1.
where here and in the following h; i indicates the inner
product between n-forms. Thus for canonical 1-forms
APPENDIX A: INCLUSION we have
OF AN ABELIAN GAUGE FIELD
An Abelian gauge field Aμ can be easily included in the hdxν ; dxρ i ¼ gνρ : ðB5Þ
general framework introduced in this work. We consider
the operator F̃ defined as Inserting this expression in Eq. (B4) we get

F̃ ¼ 0 ⊕ 0μ dxμ ⊕ Fμν Fρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ; ω̂μ ¼ ½Gð1Þ μν gνρ ωρ : ðB6Þ

where Fμν ¼ ∇μ Aν − ∇ν Aμ and we consider the following Summarizing, we have shown that according to our
expression for the metric induced by matter fields: definition we have
G̃ ¼ g̃ þ αðM̃ þ F̃Þ − βR̃: ðA1Þ
Gð1Þ · ω ¼ ½Gð1Þ μν gνρ ωρ dxμ : ðB7Þ
Note that in this expression we modify as well the
definition of M̃ as the Dirac operator will be defined as The eigenvalue problem of a metric Gð1Þ is then defined as
in Eq. (32) by performing the minimal substitution in the
covariant derivative Gð1Þ · ω ≡ λω ðB8Þ
ðAÞ
∇μ → ∇μ ¼ ∇μ − ieAμ : ðA2Þ or equivalently in matrix form as

½Gð1Þ μν gνρ ωρ ¼ λωμ : ðB9Þ

APPENDIX B: EIGENVALUES AND ENTROPY
FOR METRICS Gð1Þ AND FOR METRICS GðnÞ Note that ½Gð1Þ μν admits also an interpretation as metric
In this Appendix our goal is to use algebraic geometry between vectors. We are now in the position to extend the
[54] to define the eigenvalues of the metrics GðnÞ consid- formalism to the metrics Gð2Þ considered as quantum
ered as quantum operators between n-forms and their operators among two forms of order two. The generic
associated entropy. In order to establish the foundation 2-form ζ is defined as
of this treatment let us revisit the treatment outlined in
Sec. II for defining the eigenvectors and the associated ζ ¼ ζ ηθ dxη ∧ dxθ ; ðB10Þ
entropy of the metrics Gð1Þ viewed as quantum operators
between 1-forms and treat on the same footing metrics where ζ ηθ ¼ −ζ θη , i.e., where ζηθ is antisymmetric.
tensor GðnÞ viewed as quantum operators between n-forms The metric Gð2Þ considered as a quantum operator
with arbitrary values of n. between two two-forms is defined as

066001-11
GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)

Gð2Þ ¼ ½Gð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ: ðB11Þ for any arbitrary 2-form associated to the antisymmetric
tensor ζ μν. Thus the degeneracy of the eigenvalue is given
where Gð2Þ is antisymmetric in the first two indices and in by dðd − 1Þ=2.
the last two indices and Hermitian under the exchange of The treatment of the eigenvalues associated to the
the first two indices with the second two indices. Following metrics tensors GðnÞ considered as quantum operators
a similar line of reasoning used to define the eigenvalue of between two n-forms can be performed in a straightforward
the metric Gð1Þ we define the contraction way, following similar steps. Let us observe that the generic
n-form is given by
Gð2Þ · ζ ¼ ζ̂ ¼ ζ̂μν dxμ ∧ dxν ; ðB12Þ
ζ ¼ ζ ν1 ν2 …νn dxν1 ∧ dxν2 ; ∧ … ∧ dxνn ; ðB20Þ
where, by definition, ζ̂μν is defined as
where ζν1 ν2 …νn is antisymmetric, while the generic expres-
1
ζ̂ μν ≔ ½Gð2Þ μνρσ ζ ηθ hdxρ ∧ dxσ ; dxη ∧ dxθ i: ðB13Þ sion for GðnÞ is given by
2
Performing the inner products we thus get GðnÞ ¼ ½GðnÞ μ μ …μ ðdxμ1 ∧ dxμ2 ∧ …dxμn Þ
1 2 n ν1 ν2 …νn

1 ⊗ ðdxν1 ∧ dxν2 ; ∧ … ∧ dxνn Þ; ðB21Þ

ζ̂ μν ¼ ½Gð2Þ μνρσ ζ ηθ ðgρη gσθ − gρθ gση Þ
2
¼ ½Gð2Þ μνρσ ½gð2Þ ρσηθ ζ ηθ ; ðB14Þ where GðnÞ is antisymmetric in the first n indices and in the
last n indices and Hermitian under the exchange of the first
where we have used the definition of gð2Þ given in Eq. (28) n indices with the second n indices.
that we rewrite here for convenience, For instance, it is instructive to discuss explicitly the
form of the metric gðnÞ induced on n-forms by g. This can be
1 immediately shown to be given by
½gð2Þ μνρσ ¼ ðgμρ gνσ − gμσ gνρ Þ: ðB15Þ
2
Y
n
Summarizing, we have shown that according to our gðnÞ ¼ gμi ρi ðdxμ1 ∧ dxμ2 ∧ …dxμn Þ
definition we have i¼1

⊗ ðdxν1 ∧ dxν2 ; ∧ … ∧ dxνn Þ; ðB22Þ

Gð2Þ · ζ ¼ ½Gð2Þ μνρσ ½gð2Þ ρσηθ ζ ηθ dxμ ∧ dxν : ðB16Þ
or, more explicitly, using the antisymmetric properties of
In total analogy with the previous case, the eigenvalue the n forms, as
problem associated to the metric Gð2Þ viewed as quantum
operator among two 2-forms is defined as gðnÞ ¼ ½gðnÞ μ μ …μ ðdxμ1 ∧ dxμ2 ∧ …dxμn Þ
1 2 n ν1 ν2 …νn

Gð2Þ · ζ ≡ λζ; ðB17Þ ⊗ ðdxν1 ∧ dxν2 ; ∧ … ∧ dxνn Þ; ðB23Þ

or equivalently, with

½Gð2Þ μνρσ ½gð2Þ ρσηθ ζ ηθ ¼ λζ μν : ðB18Þ 1 ρ1 ρ2 …ρn Y n

½gðnÞ μ ¼ δν1 ;ν2 …νn gμi ρi : ðB24Þ
1 μ2 …μn ν1 ν2 …νn n! i¼1
Note that ½Gð2Þ μνρσ admits also an interpretation as metric
between bivectors. Let us now consider the particular case
Here and in the following we will use the notation ½gðnÞ μ⃗ ν⃗
in which Gð2Þ ¼ gð2Þ . In this case the eigenvalue problem
Eq. (B18) reads of indicating the element of this tensor where μ⃗ ¼
ðμ1 μ2 …μn Þ and ν⃗ ¼ ðν1 ν2 …νn Þ indicate n-tuples of indi-
ces. Using an analogous notation, we define the · product
1
ðg g − gμσ gνρ Þðgρη gσθ − gρθ gση Þζηθ ¼ λζμν ; between the generic metric GðnÞ and the n-form ζ as
4 μρ νσ

which has solution GðnÞ · ζ ¼ ζ̂; ðB25Þ

λ ¼ 1; ðB19Þ with

066001-12
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)

1 Using these flattened matrices we then construct the matrix

ζ̂μ⃗ ≔ ½G  ζ ðdxμ1 ∧ dxμ2 ∧ …dxμn Þ
n! ðnÞ μ⃗ ν⃗ ρ⃗ Nð2Þ of matrix elements
× hdxν1 ∧ dxν2 ; ∧ … ∧ dxνn ; dxρ1 ∧ dxρ2 ∧ …dxρn i:
½Nð2Þ ηθ ≔ ½GFð2Þ  ½gFð2Þ ρσηθ ; ðB30Þ
ðB26Þ μν μνρσ

Thus the eigenvalue problem for GðnÞ can be written as where μ < ν and η < θ. Note that now the usual matrix
multiplication of Nð2Þ with itself preserves Lorentz invari-
½GðnÞ μ⃗ ν⃗ ½gðnÞ ν⃗ ρ⃗ ζ ρ⃗ ¼ λζμ⃗ ; ðB27Þ ance. The eigenvalues of this matrix Nð2Þ are equal to the
eigenvalue of Gð2Þ defined in Eq. (B18).
where here and in the following we have indicated with μ⃗ , ρ⃗ Similarly metric tensors GðnÞ considered as quantum
and η⃗ tuples of n indices. From this definition of the operators between n-forms can be flattened by considering
eigenvector of a metric between n-forms it follows immedi- a basis of independent n-forms. For instance a basis for
ately that the metric tensor gðnÞ has all eigenvalues λ equal these n-forms can be given by the canonical n-forms
to one, i.e.,
dxμ1 ∧ dxμ2 ∧ … ∧ dxμn ; ðB31Þ
λ ¼ 1; ðB28Þ
with μ1 < μ2 < … < μn . This canonical basis is formed by
for any arbitrary value of n, with 0 ≤ n ≤ d. m ¼ ðdnÞ n-forms.
On this basis we can define the m × m flattened matrices
2. Flattening of the metric tensors GðnÞ and reduction
to a matrix eigenvalue problem ½GFð2Þ  ¼ n!½Gð2Þ μ⃗ ;⃗ν ; ½gFð2Þ  ¼ n!½gð2Þ μ⃗ ;⃗ν :
μ⃗ ;⃗ν μ⃗ ;⃗ν
To define the eigenvalues associated to the metric tensor
Gð1Þ between 1-forms, we can use a matrix formalism and Starting from these flattened matrices we define the matrix
define NðnÞ of elements

½Nð1Þ μ ν ≔ ½Gð1Þ μρ gρν : ðB29Þ

½NðnÞ ημ⃗⃗ ≔ ½GFðnÞ  ½gFðnÞ ρ⃗ η⃗ : ðB32Þ
μ⃗ ρ⃗
The eigenvalues of the tensor Gð1Þ are defined as the
eigenvalues of the matrix Nð1Þ . Note that now the multi- As we have seen for n ¼ 1 and n ¼ 2 now these matrices
plication of this latter matrix with itself preserves Lorentz can be multiplied preserving Lorentz invariance, and the
invariance. eigenvalues of these matrices correspond to the eigenvalues
The eigenvalue problem for the metric Gð2Þ Eq. (B18) of the metric tensor GðnÞ.
can be also written in matrix form, by considering the
m×m flattened matrices GFð2Þ and gFð2Þ with m ¼ dðd−1Þ=2 3. Trace of the metric tensors GðnÞ
representing the dimension of the basis for the 2-forms The trace of the metric tensors GðnÞ is defined as the trace
defined on the d dimensional manifold K. For instance, of the matrix NðnÞ or equivalently as
taking the basis fdxμ ∧ dxν g with μ < ν the flattened
matrices GFð2Þ and gM ð2Þ will have, respectively, matrix TrGðnÞ ¼ TrM NðnÞ : ðB33Þ
elements
For ease of notation in the following we will also define
½GFð2Þ  ¼ 2½Gð2Þ μνρσ ; ½gFð2Þ  ¼ 2½gð2Þ μνρσ :
μν;ρσ μν;ρσ indicate this trace with

A practical example might be helpful to illustrate this TrF GðnÞ g−1

ðnÞ ¼ TrGðnÞ ¼ TrM NðnÞ : ðB34Þ
construction. Assuming that the manifold K is d ¼ 3
dimensional and has three coordinates 0,1,2; the flattened
matrix gFð2Þ associated to gð2Þ , defined in the basis dx0 ∧ From the above definition of NðnÞ it follows that our
definition of the trace is the usual definition of the trace
dxð1Þ ; dxð0Þ ∧ dxð2Þ and dxð1Þ ∧ dxð2Þ is given by of a tensor, i.e.,
0 1
g00 g11 − g01 g10 g00 g12 − g02 g10 g01 g12 − g02 g11 ρ⃗ μ⃗
B C TrF GðnÞ g−1
ðnÞ ¼ ½GðnÞ μ⃗ ρ⃗ ½gðnÞ  : ðB35Þ
gFð2Þ ¼ @ g00 g21 − g01 g20 g00 g22 − g02 g20 g01 g22 − g02 g21 A:
g10 g21 − g11 g20 g10 g22 − g12 g20 g11 g22 − g12 g21 Specifically, for n ∈ f1; 2g we obtain

066001-13
GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)
Z pffiffiffiffiffiffiffiffiffi
TrGð1Þ ¼ ½Gð1Þ μρ gρμ ;
hhΨ; Φii ¼ −jgjðϕ̄ ϕ̂ þω̄μ ω̂μ þ ζ̄ μν ζ̂ μν Þdr; ðC2Þ
TrGð2Þ ¼ ½Gð2Þ μνρσ ½gð2Þ ρσμν ; ðB36Þ
where ω̂μ ¼ gμρ ω̂ρ and ζ̂ μν ¼ gμρ gνσ ζ̂ ρσ ¼ ½gð2Þ μνρσ ζ̂ ρσ .
and for general n it follows that the trace of gðnÞ is given by Thus the Hilbert space H is endowed with a scalar product
  defined through the default metric tensor g̃−1 given by
d
TrgðnÞ ¼ ; ðB37Þ
n g̃−1 ¼ 1 ⊕ gμν dxμ ⊗ dxν

for every 0 ≤ n ≤ d. ⊕ ½gð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ ðC3Þ

This scalar product has the following properties:

4. Entropy for metric tensors GðnÞ (i) Linearity on the second term
The entropy H of metric tensor GðnÞ is defined as
X hhΨ; c1 Φ1 þ c2 Φ2 ii
H ¼ TrGðnÞ ln G−1
ðnÞ ≔ − λ ln λ ðB38Þ ¼ c1 hhΨ; Φ1 ii þ c2 hhΨ; Φ2 ii; ðC4Þ
λ

where λ indicates the generic eigenvalue of GðnÞ . Since gðnÞ for arbitrary c1 ; c2 ∈ C.
(ii) Antilinearity on the first term
has all eigenvalues equal to one, it follows
hhc1 Ψ1 þ c2 Ψ2 Φii ¼ c̄1 hhΨ1 jΦii þ c̄2 hhΨ2 jΦii;
HðnÞ ¼ TrgðnÞ ln g−1
ðnÞ ¼ 0; ðB39Þ
ðC5Þ
for any value of 0 ≤ n ≤ d. The cross entropy between gðnÞ
and GðnÞ is defined as for arbitrary c1 ; c2 ∈ C.
(iii) Scalar product of a topological field with itself
The scalar product of a topological field with
TrgðnÞ ln G−1 −1
ðnÞ ≔ −TrF ln GðnÞ gðnÞ
X itself is real,
≔ −TrM NðnÞ ¼ − lnðλ0 Þ; ðB40Þ
μ hhΨ; Φii ∈ R; ðC6Þ

where λ0 indicates the generic eigenvalue of NðnÞ . unless it is infinite.

The Hilbert space H is formed by all topological fields jΦi
such that
APPENDIX C: THE TOPOLOGICAL METRIC
AS A QUANTUM OPERATOR hhΦjΦii < ∞: ðC7Þ
In this Appendix we outline the underlying quantum
theory of the proposed general scenario and the relation of The metric induced by the matter field G̃
the proposed entropy action with the Araki quantum
relative entropy [37,38]. G̃ ¼ ½Gð0Þ  ⊕ ½Gð1Þ μν dxμ ⊗ dxν
Our treatment will consider quantum operators in first ⊕ ½Gð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ ðC8Þ
quantization. We will provide the foundation of the theory
of topological metric tensors interpreted as quantum
endowed with the · product defined in Appendix B
operators [33] where the vectors of our Hilbert space are
[Eqs. (B3) and (B12)] can be viewed as a quantum operator
the topological fields formed by the direct sum of a 0-form,
a 1-form and a 2-form. G̃∶H → H where G̃ · jΦi ∈ H,
We consider the two generic topological fields jΨi and
jΦi given by G̃ · jΦi ¼ ϕ ⊕ ½Gð1Þ μν ων dxμ
⊕ ½Gð2Þ μνρσ ζ ρσ dxμ ∧ dxν : ðC9Þ
jΨi ¼ ϕ ⊕ ωμ dxμ ⊕ ζ μν dxμ ∧ dxν ;
jΦi ¼ ϕ̂ ⊕ ω̂μ dxμ ⊕ ζ̂ μν dxμ ∧ dxν : ðC1Þ The metric g̃ of the manifold K can be used to define a
dual Hilbert space H⋆ given by the dual topological field.
The scalar product among these two topological fields is For instance jΨ⋆ i is the dual of jΨi and jΦ⋆ i is the dual of
given by jΦi with jΨ⋆ i; jΦ⋆ i given by

066001-14
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)

jΨ⋆ i ¼ ϕ ⊕ ωμ dxμ ⊕ ζμν dxμ ∧ dxν ; Interestingly, the metric tensors are such that the dual of the
dual coincides with the metric tensor itself, indeed we have
jΦ⋆ i ¼ ϕ̂ ⊕ ω̂μ dxμ ⊕ ζ̂μν dxμ ∧ dxν ; ðC10Þ
G̃ ¼ G̃⋆⋆ ¼ g̃G̃⋆ g̃: ðC18Þ
where ωμ ¼ gμρ ωρ ; ω̂μ ¼ gμρ ω̂ρ and ζμν ¼ ½gð2Þ μνρσ ζ ρσ ,
ζ̂μν ¼ ½gð2Þ μνρσ ζ̂ρσ . While the Hilbert space H is endowed The topological metrics G̃ and G̃⋆ that we consider in
with a scalar product defined through the default metric this work are respectively elements of the algebras and
tensor g̃−1 the Hilbert space H⋆ is endowed with a scalar that generalize the C algebra [33] and have the following
product hhΨ⋆ ; Φ⋆ ii⋆ mediated by g̃ such that properties:
(1) and are algebras with complex numbers as the
hhΨ; Φii ¼ hhΨ⋆ ; Φ⋆ ii⋆ : ðC11Þ coefficient field.
(2) The product between given by
The dual operator G̃⋆ ∶ H⋆ → H⋆ of G̃ is given by
G̃n ¼ Gð0Þ;n ⊕ ½Gð1Þ;n μν dxμ ⊗ dxν
G̃⋆ ¼ G⋆ð0Þ ⊕ ½G⋆ð1Þ μν dxμ ⊗ dxν
⊕ ½G½2;n μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ;
⊕ ½G⋆ð2Þ μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ ðC12Þ
ðC19Þ
and satisfies
for n ∈ f1; 2g, is mediated by the metric g̃−1 , while
the product between is mediated by
hhΨ; G̃ · Φii ¼ hhG̃⋆ · Ψ⋆ ; Φ⋆ ii⋆ ; ðC13Þ
the metric g̃. Specifically we have,
for any arbitrary choice of jΨi and jΦi. Here the · product
of G̃⋆ with jΦ⋆ i is given by G̃1 g̃−1 G̃2 ¼ ½Gð0Þ;12  ⊕ ½Gð1Þ;12 μν dxμ ⊗ dxν
⊕ ½Gð2Þ;12 μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ;
⋆
G̃ · jΦ⋆ i ¼ ½G⋆ð0Þ ϕ ⊕ ½G⋆ð1Þ μν ων dxμ
with
⊕ ½G⋆ð2Þ μνρσ ζρσ dxμ ∧ dxν : ðC14Þ
½Gð0Þ;12  ¼ ½Gð0Þ;1 ½Gð0Þ;2 ;
Due to the fact the metrics GðnÞ that constitute G̃ are
½Gð1Þ;12 μν ¼ ½Gð1Þ;1 μη gη1 η2 ½Gð1Þ;2 η ν ;
Hermitian under the exchange of the first n and the second 1 2

n indices, we have that G̃⋆ is related to G̃ by ½Gð2Þ;12 μ⃗ ρ⃗ ¼ ½Gð2Þ;1 μ⃗ η⃗ ½gð2Þ  η⃗ 1 ;⃗η2
½Gð2Þ;2 η⃗ ρ⃗ : ðC20Þ
1 2

G⋆ð0Þ ¼ Gð0Þ ; Similarly we have

½G⋆ð1Þ μν μν
¼ ½Gð1Þ  ;
G̃⋆1 g̃G̃⋆2 ¼ ½G⋆ð0Þ;12  ⊕ ½G⋆ð1Þ;12 μν dxμ ⊗ dxν
½G⋆ð2Þ μνρσ μνρσ
¼ ½Gð1Þ  ; ðC15Þ
⊕ ½G⋆ð2Þ;12 μνρσ ðdxμ ∧ dxν Þ ⊗ ðdxρ ∧ dxσ Þ;
where
with
½Gð1Þ μν ¼ gμρ ½Gð1Þ ρσ gνσ ;
½G⋆ð0Þ;12  ¼ ½G⋆ð0Þ;1 ½G⋆ð0Þ;2 ;
½Gð2Þ μνρσ ¼ ½gð2Þ μνη1 η2 ½Gð2Þ η η θ ½gð2Þ θ1 θ2 ρσ :
1 2 1 θ2
½G⋆ð1Þ;12 μν ¼ ½G⋆ð1Þ;1 μη1 gη1 η2 ½G⋆ð1Þ;2 η2 ν ;
Thus we will indicate the relation between G̃⋆ and G̃ for ½G⋆ð2Þ;12 μ⃗ ρ⃗ ¼ ½G⋆ð2Þ;1  ½gð2Þ η⃗ ;⃗η ½G⋆ð2Þ;2 η⃗ 2 ρ⃗ : ðC21Þ
μ⃗ η⃗ 1 1 2
short as
(3) A bijection and a bijection
G̃⋆ ¼ g̃−1 G̃g̃−1 : ðC16Þ are defined and satisfy the
following properties:
Form this relation it follows that the dual of the default (a) The dual of the dual coincides with the metric
metric g̃ is the inverse metric g̃−1 , i.e., tensor itself,

g̃⋆ ¼ g̃−1 : ðC17Þ ðG̃⋆ Þ⋆ ¼ G̃⋆⋆ ¼ G̃: ðC22Þ

066001-15
GINESTRA BIANCONI PHYS. REV. D 111, 066001 (2025)
pffiffiffiffiffiffiffiffiffiffi
(b) The product between metrics in maps, under Δ1=2
G̃;g
¼ G̃G̃⋆ ¼ G̃g̃−1 ; ðC27Þ
the bijection, to the product between dual
metrics in and vice versa according to
where the last identity is derived under the assumption that
ðG̃1 g̃−1 G̃2 Þ⋆ ¼ G̃⋆2 g̃G̃⋆1 ; G̃ is positively definite, i.e., it has only positive eigenval-
ues. In this assumption the action of the square root of the
ðG̃⋆1 g̃G̃⋆2 Þ⋆ ¼ G̃2 g̃−1 G̃1 : ðC23Þ
modular operator Δ1=2G̃;g
on the topological field jΦi given by
(c) The dual of a linear combination of metrics is Eq. (C1) can be explicitly expressed as
and the dual of a linear combination of metrics in
obeys Δ1=2
G̃;g
jΦi ¼ Gð0Þ ϕ ⊕ ½Gð1Þ μρ gρν ων dxμ
η1 η2 ρσ
ðc1 G̃1 þ c2 G̃2 Þ⋆ ¼ c̄1 G̃⋆1 þ c̄2 G̃⋆2 ; ⊕ ½Gð2Þ μνη η gð2Þ ζρσ ðdxμ ∧ dxν Þ: ðC28Þ
1 2

ðc1 G̃⋆1 þ c2 G̃⋆2 Þ⋆ ¼ c̄1 G̃1 þ c̄2 G̃2 : ðC24Þ

In terms of the modular operator the considered entropic
(4) The norm associated to the topological metric action is defined similarly to the usual Araki quantum
is equal to the norm associated to the dual and given relative entropy [37,38] as
by kG̃k ¼ kG̃⋆ k defined as Z pffiffiffiffiffiffiffiffiffiffiffi
Z pffiffiffiffiffiffiffiffiffiffiffi 1
S¼ j − gjLdr; ðC29Þ
⋆
kG̃k ¼ kG̃ k ¼ j − gjTrF ðG̃G̃⋆ Þdr; ðC25Þ lP

where
where TrF ðG̃G̃⋆ Þ is given by
L ¼ −TrF ln Δ1=2 ¼ −TrF ln G̃g̃−1 : ðC30Þ
TrF ðG̃G̃⋆ Þ ¼ G2ð0Þ þ ½Gð1Þ μν ½Gð1Þ νμ G̃;g

þ ½Gð2Þ μνρσ ½Gð2Þ ρσμν : ðC26Þ Thus the entropic action adopted in this work is strictly
related to the Araki quantum relative entropy [37,38], albeit
the definition needs to take into account the structure of the
For a topological metric G̃ interpreted as a quantum Hilbert space H that is here formed by topological fields
operator in we define the square root of the modular given by the direct sum of a 0-form, with a 1-form and a
operator Δ1=2
G̃;g
∶ H → H as 2-form.

[1] Jacob D. Bekenstein, Black holes and entropy, Phys. Rev. D [8] Sean M. Carroll and Grant N. Remmen, What is the entropy
7, 2333 (1973). in entropic gravity?, Phys. Rev. D 93, 124052 (2016).
[2] Jacob D. Bekenstein, Generalized second law of thermo- [9] Ted Jacobson, Entanglement equilibrium and the Einstein
dynamics in black-hole physics, Phys. Rev. D 9, 3292 equation, Phys. Rev. Lett. 116, 201101 (2016).
(1974). [10] Curtis Callan and Frank Wilczek, On geometric entropy,
[3] Stephen W Hawking, Particle creation by black holes, Phys. Lett. B 333, 55 (1994).
Commun. Math. Phys. 43, 199 (1975). [11] Goffredo Chirco, Hal M. Haggard, Aldo Riello, and Carlo
[4] Shinsei Ryu and Tadashi Takayanagi, Aspects of holo- Rovelli, Spacetime thermodynamics without hidden degrees
graphic entanglement entropy, J. High Energy Phys. 08 of freedom, Phys. Rev. D 90, 044044 (2014).
(2006) 045. [12] Goffredo Chirco and Stefano Liberati, Nonequilibrium
[5] Tatsuma Nishioka, Shinsei Ryu, and Tadashi Takayanagi, thermodynamics of spacetime: The role of gravitational
Holographic entanglement entropy: An overview, J. Phys. A dissipation, Phys. Rev. D 81, 024016 (2010).
42, 504008 (2009). [13] Erik Verlinde, On the origin of gravity and the laws of
[6] Thomas Faulkner, Aitor Lewkowycz, and Juan Maldacena, Newton, J. High Energy Phys. 04 (2011) 029.
Quantum corrections to holographic entanglement entropy, [14] Gerard’t Hooft, The holographic principle, in Basics
J. High Energy Phys. 11 (2013) 074. and Highlights in Fundamental Physics (World Scientific,
[7] Ted Jacobson, Thermodynamics of spacetime: The Singapore, 2001), pp. 72–100.
Einstein equation of state, Phys. Rev. Lett. 75, 1260 [15] Leonard Susskind, The world as a hologram, J. Math. Phys.
(1995). (N.Y.) 36, 6377 (1995).

066001-16
GRAVITY FROM ENTROPY PHYS. REV. D 111, 066001 (2025)

[16] Brian Swingle, Entanglement renormalization and holog- [33] Huzihiro Araki, Mathematical Theory of Quantum Fields
raphy, Phys. Rev. D 86, 065007 (2012). (Oxford University Press, USA, 1999).
[17] Edward Witten, APS medal for exceptional achievement in [34] Fabio Ciolli, Roberto Longo, and Giuseppe Ruzzi, The
research: Invited article on entanglement properties of information in a wave, Commun. Math. Phys. 379, 979
quantum field theory, Rev. Mod. Phys. 90, 045003 (2018). (2020).
[18] Jonathan Sorce, Notes on the type classification of von [35] Roberto Longo and Feng Xu, Von neumann entropy in QFT,
Neumann algebras, Rev. Math. Phys. 36, 2430002 (2024). Commun. Math. Phys. 381, 1031 (2021).
[19] Thanu Padmanabhan, Thermodynamical aspects of gravity: [36] Edward Witten, Gravity and the crossed product, J. High
New insights, Rep. Prog. Phys. 73, 046901 (2010). Energy Phys. 10 (2022) 008.
[20] Thomas P. Sotiriou and Valerio Faraoni, f(R) theories of [37] Huzihiro Araki, Relative entropy of states of von Neumann
gravity, Rev. Mod. Phys. 82, 451 (2010). algebras, Publ. RIMS 11, 809 (1975).
[21] Thomas P. Sotiriou and Stefano Liberati, Metric-affine f(R) [38] Masanori Ohya and Dénes Petz, Quantum Entropy and its
theories of gravity, Ann. Phys. (Amsterdam) 322, 935 (2007). Use (Springer Science & Business Media, 2004).
[22] Richard Woodard, Avoiding dark energy with 1/r modifi- [39] Vlatko Vedral, The role of relative entropy in quantum
cations of gravity, in The Invisible Universe: Dark Matter information theory, Rev. Mod. Phys. 74, 197 (2002).
and Dark Energy (Springer, New York, 2007), pp. 403–433. [40] SI Kruglov, Dirac–Kähler equation, Int. J. Theor. Phys. 41,
[23] Emanuele Berti, Enrico Barausse, Vitor Cardoso, Leonardo 653 (2002).
Gualtieri, Paolo Pani, Ulrich Sperhake, Leo C Stein, Norbert [41] Peter Becher and Hans Joos, The Dirac-Kähler equation and
Wex, Kent Yagi, Tessa Baker et al., Testing general fermions on the lattice, Z. Phys. C 15, 343 (1982).
relativity with present and future astrophysical observations, [42] David Berenstein, Simon Catterall, and P. N. Lloyd,
Classical Quantum Gravity 32, 243001 (2015). Staggered bosons and Kahler-Dirac bosons, Proc. Sci.
[24] Enrico Barausse, Emanuele Berti, Thomas Hertog, Scott A. CORFU2023 (2024) 280 ,
Hughes, Philippe Jetzer, Paolo Pani, Thomas P. Sotiriou, [43] David Berenstein, Staggered bosons, Phys. Rev. D 108,
Nicola Tamanini, Helvi Witek, Kent Yagi et al., Prospects 074509 (2023).
for fundamental physics with LISA, Gen. Relativ. Gravit. [44] Ginestra Bianconi, Quantum entropy couples matter with
52, 1 (2020). geometry, J. Phys. A 57, 365002 (2024).
[25] Leor Barack, Vitor Cardoso, Samaya Nissanke, Thomas P. [45] Mikio Nakahara, Geometry, Topology and Physics (CRC
Sotiriou, Abbas Askar, Chris Belczynski, Gianfranco Press, 2018).
Bertone, Edi Bon, Diego Blas, Richard Brito et al., Black [46] Albert Einstein, Die feldgleichungen der gravitation,
holes, gravitational waves and fundamental physics: A Sitzungsberichte der Königlich Preußischen Akademie
roadmap, Classical Quantum Gravity 36, 143001 (2019). der Wissenschaften (1915), pp. 844–847.
[26] Gianfranco Bertone, Dan Hooper, and Joseph Silk, Particle [47] Sean M. Carroll, Spacetime and Geometry (Cambridge
dark matter: Evidence, candidates and constraints, Phys. University Press, Cambridge, England, 2019).
Rep. 405, 279 (2005). [48] Ginestra Bianconi, The mass of simple and higher-order
[27] Carlo Rovelli and Francesca Vidotto, Covariant Loop networks, J. Phys. A 57, 015001 (2023).
Quantum Gravity: An Elementary Introduction to Quantum [49] Nicolas Delporte, Saswato Sen, and Reiko Toriumi,
Gravity and Spinfoam Theory (Cambridge University Press, Dirac walks on regular trees, J. Phys. A 57, 275002
Cambridge, England, 2015). (2023).
[28] Jan Ambjørn, Jerzy Jurkiewicz, and Renate Loll, Emergence [50] Eduardo A. de Souza Neto, Djordje Peric, and David R. J.
of a 4D world from causal quantum gravity, Phys. Rev. Lett. Owen, Computational Methods for Plasticity: Theory and
93, 131301 (2004). Applications (John Wiley & Sons, New York, 2011).
[29] Alessandro Codello, Roberto Percacci, and Christoph [51] Liqun Qi and Ziyan Luo, Tensor Analysis: Spectral Theory
Rahmede, Investigating the ultraviolet properties of gravity and Special Tensors (SIAM, 2017).
with a Wilsonian renormalization group equation, Ann. [52] Esko Keski-Vakkuri, Claus Montonen, and Marco Panero,
Phys. (Amsterdam) 324, 414 (2009). Mathematical Methods for Physics: An Introduction to
[30] Astrid Eichhorn, An asymptotically safe guide to quantum Group Theory, Topology and Geometry (Cambridge
gravity and matter, Front. Astron. Space Sci. 5, 47 (2019). University Press, Cambridge, England, 2022).
[31] Daniele Oriti, Approaches to Quantum Gravity: Toward a [53] Viatcheslav Mukhanov and Sergei Winitzki, Introduction to
New Understanding of Space, Time and Matter (Cambridge Quantum Effects in Gravity (Cambridge University Press,
University Press, Cambridge, England, 2009). Cambridge, England, 2007).
[32] Carlos Barcelo, Stefano Liberati, and Matt Visser, Analogue [54] Paul Renteln, Manifolds, Tensors and Forms (Cambridge
gravity, Living Rev. Relativity 14, 1 (2011). University Press, Cambridge, England, 2014).

066001-17