de Sitter special relativity

R Aldrovandi, J P Beltrán Almeida and J G Pereira‡

Instituto de Fı́sica Teórica, Universidade Estadual Paulista, Rua Pamplona 145,
01405-900 São Paulo, Brazil

Abstract. A special relativity based on the de Sitter group is introduced, which is the
theory that might hold up in the presence of a non-vanishing cosmological constant.
Like ordinary special relativity, it retains the quotient character of spacetime, and
arXiv:gr-qc/0606122v2 8 Feb 2007

a notion of homogeneity. As a consequence, the underlying spacetime will be a de

Sitter spacetime, whose associated kinematics will differ from that of ordinary special
relativity. The corresponding modified notions of energy and momentum are obtained,
and the exact relationship between them, which is invariant under a re-scaling of the
involved quantities, explicitly exhibited. Since the de Sitter group can be considered
a particular deformation of the Poincaré group, this theory turns out to be a specific
kind of deformed (or doubly) special relativity. Some experimental consequences, as
well as the causal structure of spacetime—modified by the presence of the de Sitter
horizon—are briefly discussed.

1. Introduction

When the cosmological constant Λ vanishes, absence of gravitation is represented by

Minkowski spacetime, a solution of the sourceless Einstein’s equation. Its isometry
transformations are determined by the Poincaré group, which is also the group governing
the kinematics of special relativity. For a non-vanishing Λ, however, Minkowski is no
longer a solution of the corresponding Einstein’s equation and becomes, in this sense,
physically meaningless. In this case, absence of gravitation turns out to be represented
by the de Sitter spacetime. Now, the group governing the kinematics in a de Sitter
spacetime is not the Poincaré, but the de Sitter group. This means essentially that, for
a non-vanishing Λ, ordinary Poincaré special relativity will no longer be valid, and must
be replaced by a de Sitter special relativity. Since the local symmetry of spacetime will
be represented by the de Sitter group, the tangent space at every point of spacetime
must also be replaced by an osculating de Sitter space. This is the geometrical setting
of a de Sitter special relativity. In this theory, due to the change in the kinematic group,
the ordinary notions of energy and momentum, as well as the relationship between them,
will change [1]. Furthermore, the causal structure of spacetime will also be modified by
the presence of the de Sitter horizon.
‡ jpereira@[Link]
de Sitter special relativity 2

At the Planck scale, gravitation and quantum mechanics are somehow expected
to meet. That scale is, in consequence, believed to be the threshold of a new physics.
In particular, consistency arguments related to quantum gravity seem to indicate that
Lorentz symmetry must be broken and ordinary special relativity might be no longer
true [2]. To comply with that violation without producing significant changes in special
relativity far from that scale, the idea of a deformed (or doubly, as it has been called)
special relativity (DSR) has been put forward recently [3]. In this kind of theory, Lorentz
symmetry is deformed through the agency of a dimensional parameter κ, proportional
to the Planck length.§ Such a deformation implies that, in the high energy limit, a
quantum theory of gravitation must be invariant, not under the Poincaré group, but
under a “κ-deformed” Poincaré group which reduces to the standard Poincaré group in
the low energy limit.
Now, the de Sitter group naturally involves a length parameter, which is related
to the cosmological constant by Einstein’s equations. In addition, since it has the
Lorentz group as a subgroup, it can also be interpreted as a particular deformation of
the Poincaré group. In fact, it is related to the Poincaré group through the contraction
limit of a vanishing cosmological constant, in the very same way the Galilei group is
related to the Poincaré group through the contraction limit of an infinite velocity of
light. A special relativity based on the de Sitter group, therefore, gives rise to a kind of
DSR.k The fundamental difference in relation to the usual DSR models is that, in this
case, the equivalence between frames is ruled by the de Sitter group. As a consequence,
the energy and momentum definitions will change, and will satisfy a generalized relation.
Furthermore, since the Lorentz group is a sub-group of de Sitter, the Lorentz symmetry
will remain as a sub-symmetry in the theory. The presence of the de Sitter length-
parameter, however, in addition to modifying the symmetry group, modifies also the
usual Lorentz causal structure of spacetime, defined by the light cone. In fact, the causal
domain of any observer will be further restricted by the presence of an event horizon:
the de Sitter horizon.
To get some insight on how a de Sitter special relativity might be thought of, let
us briefly recall the relationship between the de Sitter and the Galilei groups, which
comes from the Wigner–Inönü processes of group contraction and expansion [6, 7].
Ordinary Poincaré special relativity can be viewed as describing the implications to
Galilei’s relativity of introducing a fundamental velocity-scale into the Galilei group.
Conversely, the latter can be obtained from the special-relativistic Poincaré group by
taking the formal limit of the velocity scale going to infinity (non-relativistic limit).
We can, in an analogous way, say that de Sitter relativity describes the implications to
Galilei’s relativity of introducing both a velocity and a length scales in the Galilei group.
In the formal limit of the length-scale going to infinity, the de Sitter groups contract
to the Poincaré group, in which only the velocity scale is present. It is interesting to
observe that the order of the group expansions (or contractions) is not important. If
§ For some reviews, as well as for the relevant literature, see Ref. [4].
k Similar ideas have already been explored in Ref. [5].
de Sitter special relativity 3

we introduce in the Galilei group a fundamental length parameter, we end up with the
Newton-Hooke group [8], which describes a (non-relativistic) relativity in the presence of
a cosmological constant [9]. Adding to this group a fundamental velocity scale, we end
up again with the de Sitter group, whose underlying relativity involves both a velocity
and a length scales. Conversely, the low-velocity limit of the de Sitter group yields the
Newton-Hooke group, which contracts to the Galilei group in the limit of a vanishing
cosmological constant.
A crucial property of the de Sitter relativity is that it retains the quotient character
of spacetime and, consequently, a notion of homogeneity. As in special relativity,
whose underlying Minkowski spacetime is the quotient space of the Poincaré by the
Lorentz groups, the underlying spacetime of the de Sitter relativity will be the quotient
space of the de Sitter and the Lorentz groups. In other words, it will be a de Sitter
spacetime. Now, a space is said to be transitive under a set of transformations — or
homogeneous under them — when any two points of it can be attained from each other
by a transformation belonging to the set. For example, the Minkowski spacetime is
transitive under spacetime translations. The de Sitter spacetime, on the other hand,
is found to be transitive under a combination of translations and proper conformal
transformations, the relative importance of these contributions being determined by the
value of the cosmological constant. We are here taking advantage of a common abuse of
language, talking rather freely of the de Sitter group, while allowing its length parameter
to vary. Of course, what is meant is the family of all such groups, each one the group
of motions of a de Sitter space with a different scalar curvature.
Due to its quotient character, spacetime will respond concomitantly to any
deformation occurring in the symmetry group. For small values¶ of Λ, for example,
the underlying spacetime will approach Minkowski spacetime. In the contraction limit
Λ → 0, it is reduced to the flat Minkowski spacetime, which is transitive under ordinary
translations. In the contraction limit Λ → ∞, on the other hand, the underlying
spacetime is reduced to a new maximally-symmetric cone spacetime [10], which is
homogeneous under proper conformal transformations. It is important to remark that
the Λ → ∞ limit must be understood as purely formal. In fact, since a large Λ means a
small length parameter, quantum effects should necessarily be taken into account. Such
effects, as is well known, provides a cut-off value for Λ, which prevents the limit to be
physically achieved.
Motivated by the above arguments, the basic purpose of this paper is to develop a
special relativity based on the de Sitter group. We will proceed as follows. Section 2
is a review of the fundamental properties of the de Sitter groups and spaces. Section
3 describes, for the sake of completeness, the main geometrical properties of the cone
spacetime that emerges in the limit of an infinite cosmological constant. In section
4, the fundamentals of a de Sitter special relativity are presented and discussed. In
¶ The reference value for defining small and large Λ is the Planck cosmological constant ΛP = 3/lP2 ,
with lP the Planck length. A small Λ will then be characterized by Λ lP2 → 0. A large Λ, on the other
hand, will be characterized by Λ lP2 → 1.
de Sitter special relativity 4

particular, an analysis of the deformed group generators acting on the de Sitter space
is made, which allows us to understand how a de Sitter relativity can give rise to an
algebraically well defined theory. The modified notions of energy and momentum are
obtained in section 5, and the new relationship between them explicitly exhibited. A
discussion on the possible phenomenological implications is then presented. Finally,
section 6 discusses the results obtained.

2. de Sitter spacetimes and groups

2.1. The de Sitter spacetimes

Spacetimes with constant scalar curvature R are maximally symmetric: they can lodge
the highest possible number of Killing vectors. Given a metric signature, this spacetime
is unique [11] for each value of R. Minkowski spacetime M, with vanishing scalar
curvature, is the simplest one. Its group of motion is the Poincaré group P = L ⊘ T ,
the semi-direct product of the Lorentz L = SO(3, 1) and the translation group T . The
latter acts transitively on M and its group manifold can be identified with M. Indeed,
Minkowski spacetime is a homogeneous space under P, actually the quotient
M = P/L.
Amongst curved spacetimes, the de Sitter and anti-de Sitter spaces are the only
possibilities [12]. One of them has negative, and the other has positive scalar curvature.
They can be defined as hyper-surfaces in the “host” pseudo-Euclidean spaces E4,1 and
E3,2 , inclusions whose points in Cartesian coordinates (χA ) = (χ0 , χ1 , χ2 , χ3 , χ4 ) satisfy,
respectively,
ηAB χA χB ≡ (χ0 )2 − (χ1 )2 − (χ2 )2 − (χ3 )2 − (χ4 )2 = − l2
and
ηAB χA χB ≡ (χ0 )2 − (χ1 )2 − (χ2 )2 − (χ3 )2 + (χ4 )2 = l2 ,
with l the de Sitter length parameter. The Latin alphabet (a, b, c . . . = 0, 1, 2, 3) will
be used to denote the four-dimensional algebra and tangent space indices. Using then
ηab for the Lorentz metric η = diag (1, −1, −1, −1), and the sign notation s = η44 , the
above conditions can be put together as
ηab χa χb + s (χ4 )2 = s l2 . (1)
Defining the dimensionless coordinate χ′4 = χ4 /l, it becomes
1
ηab χa χb + s (χ′4 )2 = s. (2)
l2
For s = −1, we have the de Sitter space dS(4, 1), whose metric is induced from the
pseudo-Euclidean metric ηAB = (+1, −1, −1, −1, −1). It has the pseudo-orthogonal
group SO(4, 1) as group of motions. Sign s = +1 corresponds to anti-de Sitter space,
de Sitter special relativity 5

denoted by dS(3, 2). It comes from ηAB = (+1, −1, −1, −1, +1), and has SO(3, 2) as its
group of motions. Both spaces are homogeneous [13]:
dS(4, 1) = SO(4, 1)/L and dS(3, 2) = SO(3, 2)/L.
In addition, each group manifold is a bundle with the corresponding de Sitter or anti-
de Sitter space as base space and the Lorentz group L as fiber [14]. These spaces are
solutions of the sourceless Einstein’s equation, provided the cosmological constant Λ
and the length parameter l are related by
3s
Λ=− 2. (3)
l

2.2. Stereographic coordinates

For definiteness, as well as to comply with observational data [15], we consider from
now on the SO(4, 1) positive cosmological constant case. The de Sitter space is then
defined by
1
− 2 ηab χa χb + (χ′4 )2 = 1, (4)
l
and the four-dimensional stereographic coordinates {xa } are obtained through a
projection from the de Sitter hyper-surfaces into a target Minkowski spacetime. They
are given by [16]
χa = Ω(x) xa (5)
and
σ2
 
′4
χ = − Ω(x) 1 + 2 , (6)
4l
where
1
Ω(x) = , (7)
1 − σ 2 /4l2
with σ 2 = ηab xa xb . The {xa } take values on the Minkowski spacetime on which the
stereographic projection is made.

2.3. Kinematic groups: transitivity

In terms of the host-space Cartesian coordinates χA , the generators of the infinitesimal
de Sitter transformations are
∂ ∂
LAB = ηAC χC B
− ηBC χC . (8)
∂χ ∂χA
They satisfy the commutation relations
[LAB , LCD ] = ηBC LAD + ηAD LBC − ηBD LAC − ηAC LBD . (9)
In terms of the stereographic coordinates {xa }, these generators are written as
Lab = ηac xc Pb − ηbc xc Pa (10)
de Sitter special relativity 6

and
La4 = lPa − (4l)−1 Ka , (11)
where
Pa = ∂/∂xa (12)
are the translation generators (with dimension of length−1 ), and
Ka = 2ηab xb xc − σ 2 δa c Pc


(13)
are the generators of proper conformal transformations (with dimension of length).
Whereas Lab refer to the Lorentz subgroup of de Sitter, La4 define transitivity on
the corresponding de Sitter space. As implied by the generators (11), the de Sitter
spacetime is found to be transitive under a combination of translations and proper
conformal transformations. The relative importance of each one of these transformations
is determined by the value of the length parameter l or, equivalently, by the value of
the cosmological constant. It is important to remark once more that the generators Lab
and La4 provide a realization of the de Sitter transformations on Minkowski spacetime,
the target space of the stereographic projection. Observe that the indices a, b, c, . . . are
raised and lowered with the Minkowski metric ηab .

2.4. Contraction limits

The art of group contraction involves the ability of placing beforehand the contraction
parameters in appropriate positions. This is usually achieved by performing a similarity
transformation in the original generators [17]. Furthermore, since these appropriate
positions are different for different limits, the vanishing and the infinite limits of
the cosmological constant must be considered separately as they require different
parameterization.

2.4.1. Vanishing cosmological constant limit To study the limit of a vanishing

cosmological constant (l → ∞), it is convenient to write the de Sitter generators in
the form
Lab = ηac xc Pb − ηbc xc Pa (14)
and
La4 1
Πa ≡ = P a − 2 Ka . (15)
l 4l
The generators Lab give rise to the usual Lorentz transformation in Minkowski spacetime,
and satisfy the commutation relation
[Lab , Lcd ] = ηbc Lad + ηad Lbc − ηbd Lac − ηac Lbd . (16)
For l → ∞, the generators Πa reduce to ordinary translations, and the de Sitter group
contracts to the Poincaré group P = L ⊘ T . Concomitant with the algebra and group
deformations, the de Sitter space dS(4, 1) reduces to the Minkowski spacetime
M = P/L,
de Sitter special relativity 7

which is transitive under ordinary translations only.

2.4.2. Infinite cosmological constant limit To begin with we recall that, in this limit,
the de Sitter space tends to the conic spacetime, denoted N, which is related to
Minkowski through the spacetime inversion [10]
xa
xa → − 2 . (17)
σ
In fact, under the spacetime inversion (17), the points at infinity of M are led to the
vertex of the cone-space N, and those on the light-cone of M become the infinity of
N. Using this relation, by applying the duality transformation (17) to the Minkowski
interval
ds2 = ηab dxa dxb , (18)
we see that+
ds2 → ds̄2 = η̄ab dxa dxb , (19)
where
η̄ab = σ −4 ηab , η̄ ab = σ 4 η ab (20)
is the metric on the cone-space N. It is important to recall also that the spacetime
inversion (17) is well known to relate translations with proper conformal transformations
[18]:
P a → Ka . (21)
The Lorentz generators, on the other hand, are found not to change:
Lab → Lab . (22)
The above results imply that, to study the limit of an infinite cosmological constant
(l → 0), it is convenient to write the de Sitter generators in the form
L̄ab ≡ σ −4 Lab = η̄ac xc Pb − η̄bc xc Pa (23)
and
Π̄a ≡ 4l La4 = 4l2 Pa − Ka . (24)
The generators L̄ab satisfy the commutation relation
 
L̄ab , L̄cd = η̄bc L̄ad + η̄ad L̄bc − η̄bd L̄ac − η̄ac L̄bd . (25)
Since the interval ds̄2 is conformally invariant, and since L̄ab satisfy a Lorentz-like
commutation relation, the latter can be interpreted as the generators of a conformal
Lorentz transformation. For l → 0, Π̄a reduce to (minus) the proper conformal
generators, and the de Sitter group contracts to the conformal Poincaré group P̄ =
L̄ ⊘ T̄ , the semi-direct product of the conformal Lorentz L̄ and the proper conformal
+
In addition to denoting the indices of the Minkowski spacetime M , the Latin alphabet (a, b, c . . . =
0, 1, 2, 3) will also be used to denote the algebra and space indices of the cone-spacetime N .
de Sitter special relativity 8

T̄ groups [19]. Metric (20) is actually invariant under P̄. Concomitant with the above
group contraction, the de Sitter spacetime reduces to the conic spacetime
N = P̄/L̄.
It is a new maximally symmetric spacetime, transitive under proper conformal
transformations [10].

3. The cone-spacetime

Before proceeding further, we present a glimpse of the general properties of the cone
spacetime N. It represents an empty spacetime, in which all energy is in the form
of dark energy [20]. It is what a purely classical physics would lead to, and can be
interpreted as the fundamental spacetime around which quantum fluctuations changing
l = 0 to l = lP (or equivalently, changing Λ ∼ ∞ to Λ = ΛP = 3/lP2 ) would take place.

3.1. Geometry
The metric (20) of the cone spacetime leads to the Christoffel components
Γc ab = 2σ −2 xd (ηad δ c b + ηbd δ c a − ηab δ c d ). (26)
In terms of η̄ab , it is written as
Γc ab ≡ Γ̄c ab = 2σ̄ −2 xd (η̄ad δ c b + η̄bd δ c a − η̄ab δ c d ), (27)
where σ̄ 2 = η̄ab xa xb . As an easy calculation shows, the corresponding Riemann and
Ricci curvatures vanish. In consequence, also the scalar curvature vanishes. Except at
the origin, therefore, where the metric tensor is singular and the Riemann tensor cannot
be defined, the cone N is a flat spacetime.

3.2. Killing vectors

We are going now to solve the Killing equation for the conformal invariant metric η̄ab .
The resulting vectors ξa will be referred to as the conformal Killing vectors.∗ The Killing
equation Lξ η̄ab = 0, as usual, can be written in the form
¯ a ξb + ∇
∇ ¯ b ξa = 0, (28)
where ∇¯ a is the covariant derivative in the connection Γ̄c ab . Using Eq. (27), it can be
rewritten as
η̄ac ∂b ξ c + η̄bc ∂a ξ c + η̄ab ∂c (ln σ̄ −4 ) ξ c = 0. (29)
The corresponding solution is
ξ a (x) = αc (σ̄ 2 δc a − 2η̄cd xd xa ) + β ac xc , (30)
∗ Not to be confused with the vectors solving the conformal Killing equation L g = Ω2 g .
ξ µν µν
de Sitter special relativity 9

with αc and β ac = − β ca integration constants. We can then choose a set of ten Killing
vectors as follows:
a
ξ(c) (x) = σ̄ 2 δc a − 2η̄cd xd xa (31)
and
a
ξ(cd) (x) = δ a c xd − δ a d xc . (32)
a
The four vectors ξ(c) (x) represent proper conformal transformations, whereas the six
a
vectors ξ(cd) (x) represent spacetime rotations. The general Killing vector, therefore, is
given by
ξ a (x) = αc ξ(c)
a
(x) + β cd ξ(cd)
a
(x). (33)
The existence of ten independent Killing vectors shows that the cone spacetime N is,
in fact, maximally symmetric.

3.3. Casimir invariants

Ordinary relativistic fields, and the particles which turn up as their quanta, are classified
by representations of the Poincaré group P = L ⊘ T , which is of rank two. Each
representation is, consequently, fixed by the values of two Casimir invariants. As any
functions of two invariants are also invariant, it is possible to choose two which have a
clear relationship with simple physical characteristics: mass (m) and spin (s). Of all
the families of representations of the Poincaré group [21], Nature seems to have given
preference to one of the so-called discrete series, whose representations are fixed by the
two invariants
C2 = γab P a P b = ≡ − m2 c2 (34)
and
C4 = γab W a W b ≡ − m2 c2 s(s + 1), (35)
with W a the Pauli-Lubanski vector
W a = 21 ǫabcd Pb Scd . (36)
Any metric γab invariant under the group action would provide invariants, but to arrive
at the above mentioned physical choice, the Lorentzian metric ηab must be chosen. The
first, involving only translation generators, fixes the mass. It defines the 4-dimensional
Laplacian operator and, in particular, the Klein-Gordon equation
( + m2 c2 )φ = 0, (37)
which all relativistic fields satisfy. The second invariant is the square of the Pauli-
Lubanski operator, used to fix the spin.
de Sitter special relativity 10

Analogously to the ordinary Poincaré group, the Casimir invariants of the conformal
Poincaré group P̄ = L̄ ⊘ T̄ can be constructed in terms of the metric γab = ηab , and of
the generators Sab and Ka .♯ The first Casimir invariant is the norm of Ka ,
C̄2 = ηab K a K b = ¯ = − m̄2 c2 , (38)
where m̄ is the conformal equivalent of the mass. If we identify ∂ a ∂a ≡ m2 , we find that
m̄2 = σ 4 m2 . (39)
The conformal Klein-Gordon equation is consequently
(¯ + m̄2 c2 )φ = 0. (40)
The second Casimir invariant, on the other hand, is defined as
C̄4 = ηab W̄ a W̄ b = −m̄2 c2 s(s + 1), (41)
where W̄ a is the Pauli-Lubanski conformal-vector
W̄ a = 21 ǫabcd Kb Scd . (42)

4. The de Sitter special relativity

We construct now a relativity theory based on the de Sitter group. In ordinary special
relativity, the underlying Minkowski spacetime appears as the quotient space between
the Poincaré and the Lorentz groups. Similarly, in a de Sitter relativity, the underlying
spacetime will be the quotient space between de Sitter and the Lorentz groups. This
aspect is crucial, as it ensures the permanence of a notion of homogeneity. Instead of
Minkowski space, however, the homogeneous spacetime will be, for positive Λ, the de
Sitter spacetime dS(4, 1) = SO(4, 1)/L.
The Greek alphabet (µ, ν, ρ, . . . = 0, 1, 2, 3) will be used to denote indices related
to the de Sitter spacetime. For example, its coordinates will be denoted by {xµ }. We
recall that the Latin alphabet (a, b, c . . . = 0, 1, 2, 3) relate to the de Sitter algebra, as
well as to the spacetime indices of both limits of the de Sitter spacetime: Minkowski,
which appears in the limit of a vanishing Λ, and the cone spacetime, which appears
in the limit of an infinite Λ. This allows the introduction of the holonomic tetrad δ a µ ,
which satisfies
ηµν = δ a µ δ b ν ηab and η̄µν = δ a µ δ b ν η̄ab . (43)
Consequently, we can also write
σ 2 = ηab xa xb = ηµν xµ xν and σ̄ 2 = η̄ab xa xb = η̄µν xµ xν , (44)
where we have identified xa = δ a µ xµ .
♯ Alternatively, they can be obtained from the de Sitter Casimir invariants by taking the contraction
limit l → 0.
de Sitter special relativity 11

4.1. Transitivity and the notion of distance

The two concurrent, but different types of transformations appearing in the generators
defining transitivity on the de Sitter spacetime give rise to two different notions of
distance: one which is related to translations, and another which is related to proper
conformal transformations. This means that it is possible to define two different metrics
in the de Sitter spacetime, one invariant under translations, and another invariant under
proper conformal transformations. As a consequence, there will be two different family
of geodesics, one joining all points equivalent under translations, and another joining
all points equivalent under proper conformal transformations. If one considers only one
of these families, therefore, there will be points in the de Sitter spacetime which cannot
be joined to each other by any geodesic. This is a well known property of the de Sitter
spacetime [12].

4.1.1. Translational distance The first notion of distance is that related to translations.
This notion will be important for small values of Λ, for which translations become
the dominant part of the de Sitter transitivity generators. To study its properties,
therefore, it is necessary to use a parameterization appropriate for the limit Λ → 0.
This parameterization is naturally provided by Eq. (2),
KG Ω2 (x) σ 2 + (χ′4 )2 = 1, (45)
where
KG = − 1/l2 (46)
represents the Gaussian curvature of the de Sitter spacetime. We introduce now the
anholonomic tetrad field
ha µ = Ω δ a µ . (47)
If ηab denotes the Minkowski metric, the de Sitter metric can, in this case, be written as
gµν ≡ ha µ hb ν ηab = Ω2 (x) ηµν . (48)
It defines the “translational distance”, with squared interval
dτ 2 = gµν dxµ dxν ≡ Ω2 (x) ηµν dxµ dxν . (49)
For l → ∞ (Λ → 0), it reduces to the Lorentz-invariant Minkowski interval
dτ 2 → ds2 = ηµν dxµ dxν . (50)

4.1.2. Conformal distance The second notion of distance is that related to the proper
conformal transformation. Since this transformation is the most important part of
the transitivity generators for large values of Λ, its study requires a parameterization
appropriate for the limit Λ → ∞. This can be achieved by rewriting Eq. (2) in the form
K̄G Ω̄2 (x) σ̄ 2 + (χ′4 )2 = 1, (51)
de Sitter special relativity 12

where
σ2 1
Ω̄(x) ≡ 2
Ω(x) = − (52)
4l (1 − 4l2 /σ 2 )
is the new conformal factor, and
K̄G = − 16 l2 (53)
is the conformal Gaussian curvature. We introduce now the anholonomic tetrad field
h̄a µ = Ω̄(x) δ a µ . (54)
If η̄ab denotes the cone spacetime metric, the corresponding de Sitter metric can, in this
case, be written as
ḡµν ≡ h̄a µ h̄b ν η̄ab = Ω̄2 (x) η̄µν . (55)
It defines the “conformal distance” on de Sitter spacetime, whose quadratic interval has
the form
dτ̄ 2 ≡ ḡµν dxµ dxν = Ω̄2 (x) η̄µν dxµ dxν . (56)
For l → 0 (Λ → ∞), de Sitter contracts to the cone spacetime N, and dτ̄ 2 reduces to
the conformal invariant interval on N:
dτ̄ 2 → ds̄2 = η̄µν dxµ dxν . (57)
On account of the conformal transitivity of this spacetime, this is the only notion of
distance that can be defined on N.

4.1.3. Two family of geodesics The Christoffel connection of the de Sitter spacetime
metric gµν is
Γλ µν = δ λ µ δ σ ν + δ λ ν δ σ µ − ηµν η λσ ∂σ [ln Ω(x)] .
 
(58)
The corresponding Riemann tensor is
1
Rµ νρσ = − 2 [δ µ ρ gνσ − δ µ σ gνρ ] . (59)
l
If we consider the family of geodesics defined by the Christoffel connection (58), there
will be points in the de Sitter spacetime which cannot be connected by anyone of
these geodesics. The reason for this fact is that the metric gµν defines a “translational
distance” only, whereas the de Sitter spacetime is homogeneous under a combination of
translation and proper conformal transformations.
On the other hand, the Christoffel connection of the de Sitter spacetime ḡµν is
Γ̄λ µν = δ λ µ δ σ ν + δ λ ν δ σ µ − η̄µν η̄ λσ ∂σ ln Ω̄(x) .
   
(60)
Similarly, the corresponding Riemann tensor is
R̄µ νρσ = −16l2 [δ µ ρ ḡνσ − δ µ σ ḡνρ ] . (61)
Since the metric ḡµν defines only a “conformal distance”, and since the de Sitter
spacetime is homogeneous under a combination of translation and proper conformal
de Sitter special relativity 13

transformations, there will again be points in the de Sitter spacetime which cannot
be connected by anyone of the geodesics belonging to the family of the Christoffel
connection (60). However, the two families of geodesics are complementary in the sense
that the points that cannot be connected by one family of geodesics can be connected
by the other family.
It is important to remark that both Riemann tensors Rµ νρσ and R̄µ νρσ represent
the curvature of the de Sitter spacetime. The difference is that, whereas Rµ νρσ
represents the curvature tensor in a parameterization appropriate for studying the
limit of a vanishing cosmological constant, R̄µ νρσ represents the curvature tensor in a
parameterization appropriate for studying the limit of an infinite cosmological constant.
As a straightforward calculation shows, both limits yield a spacetime with vanishing
curvature. This means that Minkowski and the cone spacetimes are both flat.

4.2. The de Sitter transformations

The de Sitter transformations can be thought of as rotations in a five-dimensional
pseudo-Euclidian spacetime,
C
χ′ = ΛC D χD , (62)
with ΛC D the group element in the vector representation. Since these transformations
leave invariant the quadratic form
− ηAB χA χB = l2 , (63)
they also leave invariant the length parameter l. Their infinitesimal form is
δχC = 21 E AB LAB χC , (64)
where E AB are the parameters and LAB the generators.

4.2.1. Small cosmological constant For Λ small, analogously to the identifications (14)
and (15), we define the parameters
ǫab = E ab and ǫa = l E a4 . (65)
In this case, in terms of the stereographic coordinates, the infinitesimal de Sitter
transformation assumes the form
δxc = 12 ǫab Lab xc + ǫa Πa xc , (66)
or equivalently
ǫb
δxc = ǫc a xa + ǫa − 2xb xc − σ 2 δb c .


2
(67)
4l
In the limit of a vanishing Λ, it reduces to the ordinary Poincaré transformation.
de Sitter special relativity 14

4.2.2. Large cosmological constant For Λ large, analogously to the identifications (23)
and (24), we define the parameters
ǭab = σ 4 E ab and ǭa = E a4 /4l. (68)
In this case, in terms of the stereographic coordinates, the de Sitter transformation
assumes the form
δxc = 21 ǭab L̄ab xc + ǭa Π̄a xc , (69)
or equivalently
δxc = ǭc a xa − ǭb 2xb xc − σ 2 δb c + 4l2 ǭa ,


(70)
where ǭc a = ǭcb η̄ba ≡ ǫc a . In the limit of an infinite Λ, it reduces to the a conformal
Poincaré transformation.

4.3. The Lorentz generators

Up to now, we have studied the de Sitter transformations in a Minkowski spacetime.
In what follows we are going to study the form of the corresponding generators in a de
Sitter spacetime, which is the spacetime of a de Sitter special relativity. This will be
done by contracting the generators acting in Minkowski spacetime with the appropriate
tetrads. We begin by considering the Lorentz generators.

4.3.1. Small cosmological constant For small Λ, the generators of an infinitesimal

Lorentz transformation are (see section 4.2.1)
Lab = ηac xc Pb − ηbc xc Pa . (71)
The corresponding generators acting on a de Sitter spacetime can be obtained by
contracting Lab with the tetrad ha µ , given by Eq. (47):
Lµν ≡ ha µ hb ν Lab = gµρ xρ Pν − gνρ xρ Pµ . (72)
The corresponding matrix vector representation is easily found to be
(Sµν )λ ρ = gµλ δν ρ − gνλ δµ ρ . (73)
The spinor representation, on the other hand, is
i
(Sµν ) = [γµ , γν ], (74)
4
where γµ = ha µ γa are the point-dependent Dirac matrices. For l → ∞, the de Sitter
spacetime reduces to Minkowski, and the corresponding Lorentz generators reduce to
the generators of the usual, Minkowski spacetime Lorentz transformation.
Now, the generators Lµν satisfy the commutation relation
[Lµν , Lρλ ] = gνρ Lµλ + gµλ Lνρ − gνλ Lµρ − gµρ Lνλ . (75)
Even when acting on de Sitter spacetime, therefore, these generators still present a well-
defined algebraic structure, isomorphic to the usual Lie algebra of the Lorentz group.
de Sitter special relativity 15

This is a fundamental property in the sense that it allows the construction, on the de
Sitter spacetime, of an algebraically well defined special relativity. This possibility is
related to the mentioned fact that, like the Minkowski spacetime, the (conformally-flat)
de Sitter spacetime is homogeneous and isotropic [22].

4.3.2. Large cosmological constant For Λ large, the generators of infinitesimal Lorentz
transformations are (see section 4.2.2)
L̄ab = η̄ac xc Pb − η̄bc xc Pa . (76)
On a de Sitter spacetime, their explicit form can be obtained by contracting (76) with
the tetrad h̄a µ , given by Eq. (54):
L̄µν ≡ h̄a µ h̄b ν L̄ab = ḡµρ xρ Pν − ḡνρ xρ Pµ (77)
These generators are easily found to satisfy the commutation relation
 
L̄µν , L̄ρλ = ḡνρ L̄µλ + ḡµλ L̄νρ − ḡνλ L̄µρ − ḡµρ L̄νλ . (78)
Like Lµν , therefore, they present a Lorentz-like algebraic structure. The corresponding
matrix vector representation is, in this case, given by
(S̄µν )λ ρ = ḡµλ δν ρ − ḡνλ δµ ρ , (79)
whereas the spinor representation is
i
(S̄µν ) = [γ̄µ , γ̄ν ], (80)
4
with γ̄µ = h̄a µ γa the point-dependent Dirac matrices. For l → 0, the de Sitter spacetime
reduces to the conic space N, and the corresponding Lorentz generators reduce to the
generators of a conformal Lorentz transformation.

4.4. The de Sitter “translation” generators

Like in the case of the Lorentz generators, the form of the generators Πa and Π̄a acting
in the de Sitter spacetime can be obtained through contractions with the appropriate
tetrad. For Λ small, they are given by
Πµ ≡ ha µ Πa = Ω Pµ − (1/4l2 )−1 Kµ ,
 
(81)
where
Pµ = ∂/∂xµ and Kµ = 2ηµρ xρ xν − σ 2 δµ ν Pν .


(82)
For Λ large, on the other hand, they are
Π̄µ ≡ h̄a µ Π̄a = Ω̄ Pµ − (1/4l2 )−1 Kµ .


(83)
We see from these expressions that the de Sitter spacetime is transitive under a
combination of of the translation and proper conformal generators. For Λ → 0, Πµ
reduce to the usual translation generators of Minkowski spacetime. For Λ → ∞, Π̄µ
reduce to the proper conformal generators, which define the transitivity on the cone
spacetime.
de Sitter special relativity 16

4.5. Understanding the de Sitter relativity

The de Sitter special relativity can be viewed as made up of two different relativities:
the usual one, related to translations, and a conformal one, related to proper conformal
transformations. It is a single relativity interpolating these two extreme limiting cases.
For small values of Λ, for example, usual special relativity will prevail over the conformal
one, and the Poincaré symmetry will be weakly deformed. In the contraction limit of a
vanishing cosmological constant, de Sitter relativity reduces to usual special relativity.
The underlying spacetime in this case reduces to the Minkowski space M, which is
transitive under translations only. For large values of Λ, on the other hand, conformal
relativity will prevail over the usual one, and the Poincaré symmetry will be strongly
deformed. In the contraction limit of an infinite Λ, de Sitter special relativity reduces
to conformal relativity. The underlying spacetime, in this case, will be the cone-space
N, which is transitive under proper conformal transformations only.

4.5.1. Conformal relativity Conformal relativity is, therefore, the limit of de Sitter
special relativity for an infinite cosmological constant. It is the special relativity
governing the equivalence of frames in the cone spacetime N. Notice that this
equivalence must be understood in the conformal sense. In fact, remember that two
points of this spacetime cannot be related by a translation, but only by a proper
conformal transformation. Accordingly, kinematics will be governed by the so called
conformal Lorentz group, whose generators are
L̄ab = η̄ac xc Pa − η̄bc xc Pa . (84)
The corresponding conformal vector and spinor matrix representations are the limiting
cases of (79) and (80),
(S̄ab )d c = η̄ad δb c − η̄bd δa c (85)
and
i
S̄ab = [γ̄a , γ̄b ], (86)
4
where γ̄a = − σ −2 γa is a kind of conformal Dirac matrix. Observe that the anti-
commutator of the γ̄a ’s yields the cone spacetime metric:
{γ̄a , γ̄b} = 2 η̄ab . (87)
Of course, like the cone spacetime N, this limiting theory has to be interpreted as
purely formal. It is what a classical physics would lead to, that is to say, it is the
classical relativity behind the quantum physics at the Planck scale.

5. Energy-momentum relations

5.1. Noether current

Let us consider now the mechanics of point particles on de Sitter spacetime. The
conserved Noether current associated to a particle of mass m is, in this case, the five-
de Sitter special relativity 17

dimensional angular momentum [16]

B A
 
AB A dχ B dχ
λ = mc χ −χ , (88)
dτ dτ
with dτ the de Sitter line element (49). In order to make contact with the usual
definitions of energy and momentum, we rewrite it in terms of the stereographic
coordinates {xa } and the Minkowski interval ds. The result is
λab = xa pb − xb pa (89)
and
λa4 = lpa − (4l)−1 k a , (90)
where
dxa
pa = m c Ω (91)
ds
is the momentum, and
k a = (2ηcb xc xa − σ 2 δb a ) pb (92)
is the so called conformal momentum [18]. Their form on the de Sitter spacetime can
be obtained through a contraction with appropriate tetrads.

5.2. Small cosmological constant limit

For Λ lP2 → 0, analogously to the generators, we define the de Sitter momentum
λa4 ka
πa ≡ = pa − 2 . (93)
l 4l
The corresponding spacetime version is
kµ
π µ ≡ ha µ π a = pµ − 2 , (94)
4l
where
dxµ
pµ = m c (95)
ds
is the Poincaré momentum, and
k µ = (2ηλρ xρ xµ − σ 2 δλ µ ) pλ (96)
is the corresponding conformal Poincaré momentum.†† We remark that π µ is the
conserved Noether momentum related to the transformations generated by Πa . Its
zero component,
0 0 k0
π ≡ p − 2, (97)
4l
represents the energy, whereas the space components (i, j, . . . = 1, 2, 3)
ki
π i ≡ pi − (98)
4l2
†† Analogously to the identification pµ = T µ0 , with T µν the energy-momentum current, the conformal
momentum k µ is defined by k µ = K µ0 , with K µν the conformal current [18].
de Sitter special relativity 18

represent the momentum. The presence of a cosmological constant, therefore, changes

the usual definitions of energy and momentum [1]. As a consequence, the energy-
momentum relation will also be changed [23].
In fact, the energy-momentum relation in de Sitter relativity is given by
 
µ ν 2 µ ν 1 µ ν 1 µ ν
gµν π π = Ω ηµν p p − 2 p k + k k . (99)
2l 16l4
The components of the Poincaré momentum pµ are
ε 
µ p i
p = ,p , (100)
c
where εp and pi are the usual Poincaré energy and momentum, respectively. As
is well known, they satisfy the relation ηµν pµ pν = m2 c2 , where m2 c2 is the first
Casimir invariant of the Poincaré group. Analogously, the components of the conformal
momentum k µ can be written in the form
ε 
k
kµ = , ki , (101)
c
with εk the conformal notion of energy, and k i the space components of the conformal
momentum. The conformal momentum satisfies ηµν k µ k ν = m̄2 c2 , where m̄2 c2 is the
first Casimir invariant of the conformal Poincaré group. Using the expressions above,
the relation (99) becomes
ε2p 2
  2 
1 εp εk 1 ε
2 2
−p = m c + 2 − p~ · ~k − mm̄c − 2
2 k 2 2
− k − m̄ c 2
.(102)
c2 2l c2 8l c2
For small values of Λ, the de Sitter length parameter l is large, and the modifications
in the energy-momentum relation will be small. In the limit of a vanishing cosmological
constant, the ordinary notions of energy and momentum are recovered, and the de Sitter
relativity reduces to the ordinary special relativity, in which the Poincaré symmetry is
exact. The energy-momentum relation, in this case, reduces to the usual expression
ε2p
− p2 = m2 c2 . (103)
c2

5.3. High cosmological constant limit

For Λ lP2 → 1, analogously to the generators, we define the de Sitter momentum
π a ≡ 4l λa4 = 4l2 pa − k a . (104)
The corresponding spacetime version is
4l2
π̄ µ ≡ h̄a µ π̄ a = 4l2 pµ − k µ .


2
(105)
σ
We remark that π̄ µ is the conserved Noether momentum related to the transformations
generated by Π̄a . Its zero component,
4l2 2 0
π̄ 0 = (4l p − k 0 ), (106)
σ2
de Sitter special relativity 19

represents the conformal energy, whereas the space components

4l2 2 i
π̄ i = (4l p − k i ) (107)
σ2
represent the conformal momentum.
The energy-momentum relation is now given by
ḡµν π̄ µ π̄ ν = 16l4 Ω̄2 σ −8 ηµν 16l4 pµ pν − 8l2 pµ k ν + k µ k ν .
 
(108)
In terms of the energy and momentum components, it becomes
 2
ε2k 2 εp
 
2 2 2 εp εk ~ 2 2 2 2 2
−k = m̄ c +8l − p~ · k − mm̄c − 2l −p −m c .(109)
c2 c2 c2
For large values of the cosmological constant, the de Sitter length parameter l is small.
In the formal limit Λ lP2 → ∞, only the conformal notions of energy and momentum will
remain, and de Sitter relativity will reduce to the pure conformal special relativity. In
this case, the energy-momentum relation will be
ε2k
− k 2 = m̄2 c2 . (110)
c2

5.4. Some conceptual issues

In the standard formulations of deformed special relativity, which are based on a κ-
deformed Poincaré group, the momentum space of the particles is identified with a de
Sitter space [24]. In these theories, although energy and momentum keep their special
relativistic notions, they satisfy a deformed dispersion relation. As a consequence, a
consistent notion of total energy and momentum, as well as a conservation law for
them, is lacking [24]. In the de Sitter special relativity, on the other hand, a precise
notion of momentum and energy is provided for the particles: they are the Noether
currents associated to de Sitter symmetry. This means that there is a clear relation
between the symmetry generators (10-11) and the conserved currents (89-90). The
resulting deformed dispersion relations for the particle’s energy and momentum, given
by Eqs. (102) and (109), are consequently relations between conserved quantities. Since
the de Sitter current is a linear combination of the momentum pµ and the conformal
current k µ , the dispersion relation turns out to depend on these two four-vectors. Notice,
however, that neither pµ nor k µ is conserved: only the de Sitter momentum is conserved.
Another relevant feature of the de Sitter modified dispersion relations concerns their
properties under a re-scaling of the fundamental quantities. In the very same way as
it happens with the ordinary special relativity dispersion relation (103), the de Sitter
dispersion relations (102) and (109) are invariant under a simultaneous re-scaling of
mass, energy and momentum. On the other hand, because it includes non-quadratic
terms in the momentum, the dispersion relations of the usual formulations of DSR are
not invariant under such a re-scaling. This invariance is an important issue because
it prevents the so called “soccer-ball problem” [30]. This means that the dispersion
relation of the de Sitter DSR is true for elementary particles (as it is argued since they
de Sitter special relativity 20

must be relevant for discussing elementary particle processes), as well as for macroscopic
objects, like for example a soccer-ball.

5.5. Experimental consequences and speculations

The change from Minkowski to a de Sitter spacetime implies a change in the symmetry
group of spacetime from Poincaré to de Sitter. Since, algebraically speaking, the only
difference between these groups is the replacement of Pa by a certain combination of
Pa and Ka , the net result of this change is ultimately the breakdown of translational
invariance. From the experimental point of view, therefore, a de Sitter special relativity
may be probed by looking for possible violations of translational invariance in high
energy processes. This can be done by applying the same procedure used in the search
for possible violations of Lorentz and CPT symmetries in high energy processes [25].
For small values of Λ, as we have seen, the homogeneity of spacetime is preponderantly
given by the translation generators, which means that the violation of the translation
invariance will be very small. Only when Λ is large this violation is expected to be
relevant.
Now, relying on our current theories of particle physics based on spontaneously
broken symmetry and phase transitions, there must have been some periods in the
history of the universe in which the value of Λ, and hence of the scale energy EΛ , were
large. For example, in the electroweak epoch characterized by ΛEW , the kinematics
of a typical electroweak process with energy EΛEW , according to the de Sitter special
relativity, must have been strongly influenced by ΛEW . In fact, as we have seen, a large
cosmological constant would produce significant changes in the definitions of energy and
momentum, as well as in the kinematic relations satisfied by them. These changes could
modify significantly the physics that should be applied in the study the early universe.
On the other hand, if we take the phase transitions associated to the spontaneously
broken symmetries as the primary source of a non-vanishing Λ, it is conceivable that a
high energy experiment could modify the local structure of space-time for a short period
of time, in such a way that the immediate neighborhood of a high energy collision would
depart from the Minkowski space and become a de Sitter spacetime. According to this
point of view, there would be a connection between the energy scale of the experiment
and the local value of Λ [26]. It is interesting to note that such a connection yields a
thermodynamical cutoff for the cosmological constant. To see it, observe first that the
area of the de Sitter horizon is AdS ∼ l2 . Since the entropy associated to this surface is
proportional to the logarithm of the number of states
n = AdS /lP2 ∼ l2 /lP2 ,
with lP the Planck length, and since the minimum allowed value for the entropy is
achieved for n = 1, we see that the minimum allowed value for l is of the order of the
Planck length. This relation provides a contact between de Sitter special relativity and
quantum gravity [27].
de Sitter special relativity 21

Assuming the above described connection between the energy scale of the exper-
iment and the local value of Λ, it is possible to envisage some potential experimental
consequences of the dispersion relations (102) and (109). Of course, at the cosmological
level, where [28]
∼ 1028 cm,
−1/2
lΛ0 ∼ Λ0
which corresponds to EΛ0 ∼ 10−33 eV, the deviation from the dispersion relations
are very small, and there is no hope for any experimental detection in the existing
colliders. However, for energies of the order of 200 GeV, corresponding to the electroweak
phase transition, the de Sitter parameter is lΛEW ∼ (1/4) cm, which is equivalent to
EΛEW ∼ 10−4 eV. For high energy experiments of order 20 TeV, one finds EΛT eV ∼ 1 eV.
And for energies of order 1000 TeV, we have EΛ ∼ 2500 eV. For particles of small mass,
such as neutrinos, there would be significant changes in the kinematics at very high
energies, which could eventually be tested in a foreseeable future [26].

6. Final remarks

If the cosmological constant Λ has a non-vanishing value, ordinary special relativity

breaks down and must necessarily be replaced by a de Sitter special relativity. A crucial
point of this theory is that it preserves the notion of spacetime homogeneity. In fact,
like Minkowski, the de Sitter spacetime is a quotient space: dS(4, 1) = SO(4, 1)/L.
As a consequence, any deformation occurring in the symmetry group will produce
concomitant deformations in the quotient space. In particular, different values of the
cosmological constant will give rise to different spacetimes. For small Λ, the de Sitter
group approaches the Poincaré group, and the de Sitter spacetime will approach the
Minkowski spacetime. For large Λ, on the other hand, the de Sitter group approaches
the conformal Poincaré group, and the underlying spacetime will approach a flat cone
space. If we consider a possible Λ-dependence of high energy processes [26] close to
the Planck scale, not only the symmetry group will change, but also the geometric
nature of spacetime will change. Transitivity properties, in special, will be completely
different. Accordingly, the energy and momentum definitions will change, and will
satisfy a generalized relation. Experimentally, these changes would appear as a violation
of the translational symmetry, and could eventually be tested through the application
of the same techniques already used in the search for possible violations of the Lorentz
and CPT symmetries.
Another important point is that, due to the homogeneous character of the de Sitter
spacetime, the Lorentz generators in this spacetime still present a well defined algebraic
structure, isomorphic to the usual Lie algebra of the Lorentz group. This means that the
Lorentz symmetry remains a sub-symmetry in a de Sitter relativity, and consequently
the velocity of light c is left unchanged by a de Sitter transformation. Since it also leaves
unchanged the length parameter l, a de Sitter transformation leaves unchanged both c
and l. This property has important consequences for causality. As is well known, the
de Sitter special relativity 22

constancy of c introduces a causal structure in spacetime, defined by the light cones.

Analogously, the presence of the de Sitter length parameter l adds to that structure some
further restrictions on the causal structure of spacetime. To see that, let us remember
that the de Sitter spacetime has a horizon, which restricts the causal region of each
observer. In terms of the stereographic coordinates, this horizon is identified by
x2 + y 2 + z 2 = l2 /Ω2 and (x0 )2 = l2 (2 − 1/Ω)2 . (111)
For small Λ, the horizon tends to infinity, and there are no significant causal changes.
For large values of Λ, however, the causal domain of each observer — restricted by the
horizon — becomes small. Considering again a possible Λ-dependence of high energy
processes, at the Planck scale this region would be of the order of the Planck length.
At this scale, therefore, the large value of Λ would introduce deep changes in the causal
structure of spacetime. This mechanism could eventually be an explanation for the
causal modifications expected to occur at the Planck scale.
Finally, it is worth mentioning a topic of special importance, which concerns
relativistic fields. If relativity changes, the concept of relativistic field must change
accordingly. For example, in the context of the de Sitter relativity, a scalar field should
be interpreted as a singlet representation, not of the Lorentz, but of the de Sitter group.
Among other consequences, the Klein-Gordon equation will have a different form. For
general values of Λ, it is [1]
R
φ + m2 c2 φ − φ = 0, (112)
6
with the Laplace-Beltrami operator in the de Sitter metric (48), and R = −12/l2 .
Notice in passing that this could be the solution to the famous controversy on the R/6
factor [29]. In fact, this factor appears naturally if, instead of a Lorentz scalar, field φ is
assumed to be a de Sitter scalar, in which case the corresponding Klein-Gordon equation
becomes naturally conformal invariant in the limit of a vanishing mass. The ordinary
Klein-Gordon equation (37) for a Lorentz scalar is recovered in the limit Λ → 0. For
large values of Λ, on the other hand, the Klein-Gordon equation becomes
R̄
¯ φ + m̄2 c2 φ −φ = 0, (113)
6
with ¯ the Laplace-Beltrami operator in the metric (55). In the limit Λ → ∞, it reduces
to the conformal Klein-Gordon equation (40).

Acknowledgments

The authors would like to thank FAPESP, CAPES and CNPq for financial support.

References

[1] Aldrovandi R, Beltrán Almeida J P and Pereira J G 2004 Int. J. Mod. Phys. D 13, 2241 (Preprint
gr-qc/0405104)
de Sitter special relativity 23

[2] Amelino-Camelia G 2000 Lect. Not. Phys. 541, 1 (Preprint gr-qc/9910089); Protheroe R J
and Meyer H 2000 Phys. Lett. B 493, 1; Sarkar S 2002 Mod. Phys. Lett. A 17, 1025
(Preprint gr-qc/0204092); Ahluwalia D V 2002 Mod. Phys. Lett. A 17, 1135 (Preprint
gr-qc/0205121); Jacobson T, Liberati S and Mattingly D 2002 Phys. Rev. D 66, 081302 (Preprint
hep-ph/0112207); Myers R C and Pospelov M 2003 Phys. Rev. Lett. 90, 211601 (Preprint
hep-ph/0301124); Brandenberger R H and Martin J 2002 Int. J. Mod. Phys. A 17, 3663 (Preprint
hep-th/0202142)
[3] Amelino-Camelia G 2001 Phys. Lett. B 510, 255 (Preprint hep-th/0012238); Amelino-Camelia G
2002 Int. J. Mod. Phys. D 11, 35 (Preprint gr-qc/0012051)
[4] Kowalski-Glikman J 2005 in Planck Scale Effects in Astrophysics and Cosmology, ed. by
G. Amelino-Camelia and J. Kowalski-Glikman, Lect. Not. Phys. 669, 131 (Preprint
hep-th/0405101); Kowalski-Glikman J 2006 in Approaches to quantum gravity - toward a new
understanding of space, time, and matter, ed. by D. Oriti (Preprint gr-qc/0603022)
[5] Guo H Y, Huang C G, Xu Z and Zhuo B 2004 Phys. Lett. A 331, 1 (Preprint hep-th/0403171)
[6] Inönü E and Wigner E P 1953 Proc. Natl. Acad. Scien. 39, 510
[7] Gilmore R 1974 Lie Groups, Lie Algebras, and Some of Their Applications (New York: Wiley)
[8] Bacry H and Lévy-Leblond J-M 1968 J. Math. Phys. 9, 1605; Duval C, Burdet G, Künsle H P K
and Perrin M 1985 Phys. Rev. D 31, 1841; Aldrovandi R, Barbosa A L, Crispino L C B and
Pereira J G 1999 Class. Quant. Grav. 16, 495
[9] Gibbons G W and Patricot C E 2003 Class. Quant. Grav. 20, 5225 (Preprint hep-th/0308200);
Tian Y, Guo H Y, Huang C G, Xu Z and Zhou B 2005 Phys. Rev. D 71 044030 (Preprint
hep-th/0411004)
[10] Aldrovandi R, Beltrán Almeida J P and Pereira J G 2006 J. Geom. Phys 56, 1042 (Preprint
gr-qc/0403099)
[11] Weinberg S 1972 Gravitation and Cosmology (New York: Wiley)
[12] Hawking S W and Ellis G F R 1973 The Large Scale Structure of Space-Time (Cambridge:
Cambridge University Press)
[13] Aldrovandi R and Pereira J G 1995 An Introduction to Geometrical Physics (Singapore: World
Scientific)
[14] Kobayashi S and Nomizu K 1963 Foundations of Differential Geometry (New York: Interscience)
[15] Riess A G et al 1998 Ap. J. 116, 1009; Perlmutter S et al 1999 Ap. J. 517, 565; de Bernardis P
et al 2000 Nature 404, 955; Hanany S et al 2000 Ap. J. Letters 545, 5
[16] Gürsey F 1962 in Group Theoretical Concepts and Methods in Elementary Particle Physics, ed. by
F. Gürsey, Istanbul Summer School of Theoretical Physics (New York: Gordon and Breach)
[17] Inönü E in Group Theoretical Concepts and Methods in Elementary Particle Physics, ed. by F.
Gürsey, Istanbul Summer School of Theoretical Physics (New York: Gordon and Breach)
[18] Coleman S 1985 Aspects of Symmetry (Cambridge: Cambridge University Press)
[19] Aldrovandi R and Pereira J G 1998 A Second Poincaré Group, in Topics in Theoretical Physics:
Festschrift for A. H. Zimerman, ed. by H. Aratyn et al (São Paulo: Fundação IFT) (Preprint
gr-qc/9809061)
[20] Aldrovandi R, Beltrán Almeida J P and Pereira J G 2005 Grav. Cosm. 11, 277 (Preprint
gr-qc/0312017)
[21] Wigner E 1939 Ann. Math. 40, 39
[22] Jackson J D 1999 Classical Electrodynamics (New York: Wiley)
[23] Hossenfelder S 2006 Class. Quant. Grav. 23 1815 (Preprint hep-th/0510245)
[24] Kowalski-Glikman J and Nowak S 2003 Class. Quant. Grav. 20, 4799 (Preprint hep-th/0304101)
[25] Kostelecky V A (ed.) 1999 CPT and Lorentz Symmetry (Singapore: World Scientific)
[26] Mansouri F 2002 Phys. Lett. B 538 239 (Preprint hep-th/0203150)
[27] Amelino-Camelia G, Smolin L and Starodubtsev A 2004 Class. Quant. Grav. 21, 3095 (Preprint
hep-th/0306134)
[28] Carrol S 2001 Living Rev. Rel. 4, 1 (Preprint astro-ph/0004075)
de Sitter special relativity 24

[29] Accioly A J, Aldrovandi R, Novaes S F, Spehler D and Mukai H 1997 Prog. Theor. Phys. 97, 121
[30] Girelli F and Livine E R 2006 Some comments on the universal constant in DSR, Proceedings of
DICE2006, Piombino, Italy (Preprint gr-qc/0612111)