Found Phys

DOI 10.1007/s10701-010-9496-1

Geometrizing Relativistic Quantum Mechanics

F.T. Falciano · M. Novello · J.M. Salim

Received: 2 March 2010 / Accepted: 22 July 2010

© Springer Science+Business Media, LLC 2010

Abstract We propose a new approach to describe quantum mechanics as a manifes-

tation of non-Euclidean geometry. In particular, we construct a new geometrical space
that we shall call Qwist. A Qwist space has a extra scalar degree of freedom that ul-
timately will be identified with quantum effects. The geometrical properties of Qwist
allow us to formulate a geometrical version of the uncertainty principle. This relativis-
tic uncertainty relation unifies the position-momentum and time-energy uncertainty
principles in a unique relation that recover both of them in the non-relativistic limit.

Keywords Foundations of quantum mechanics · Bohm-de Broglie interpretation ·

Weyl integrable space · Non-Euclidean geometry

1 Introduction

Non-relativistic theories were formulated to describe natural phenomena as a collec-

tion of events occurring on space using time as their external parameter. For that it
seemed reasonable to formalize the physical arena in an abstract language as a flat,
homogeneous and isotropic space. In fact, there was no other but one geometrical the-
ory available. Hence, Euclidean geometry was immediately identified with physical
space.

F.T. Falciano () · M. Novello · J.M. Salim

Instituto de Cosmologia Relatividade Astrofisica ICRA—CBPF, Rua Dr. Xavier Sigaud, 150,
CEP 22290-180, Rio de Janeiro, Brazil
e-mail: [email protected]
M. Novello
e-mail: [email protected]
J.M. Salim
e-mail: [email protected]
 Found Phys

Apart from its peculiarities, quantum mechanics promptly inherited Euclidean

geometry from classical mechanics. Even though some of its formulations do not
even have a well defined notion of trajectory, quantum mechanics is defined using
the flat Euclidean metric or at most the flat Minkowskian metric when we are dealing
with relativistic quantum systems.
One of the novelties of relativistic theories is to describe physical phenomena on a
non-Euclidean four dimensional manifold. There is a rupture in the identification of
the Euclidean geometry with the physical space. In fact, General Relativity generalize
the spacetime structure allowing it to be any possible type of Riemannian geometry.
The geometrical properties of a Riemannian space is completely characterized
by a metric tensor. Its connection, the geometrical object that defines the covariant
derivative, is identified with the Christoffel symbol, which is completely determined
by the metric tensor. These spacetimes have two important properties, namely, all
Riemannian spaces are locally Minkowskian and the parallel transport prescription
which enables us to compare objects at different locations preserve lengths and an-
gles.
Additionally, there is still a wider class of torsion free geometrical space that was
first introduced by Weyl [1] as an attempt to include electromagnetism in the prop-
erties of spacetime. In a Weyl space, the covariant derivative is again modified to
implement a new gauge symmetry. The main idea was to incorporate electromag-
netism in the geometrical degrees of freedom and geometrize both gravitation and
electromagnetism, the only known classical long-range interactions.
From a different perspective, London proposed [2] that the geometrical space de-
veloped by Weyl could be related to quantum phenomena. London hypothesis was
that stationary states of a quantum system such as an hydrogen atom should be as-
sociated with special geometrical configurations. In particular, he obtained Bohr’s
atomic orbits for the hydrogen atom as the only possible integrable orbits on that
Weyl space. Therefore, it became admissible that Euclidean geometry could fail not
only on large scales but also on small scales. The possibility of this breakdown of
Euclidean geometry on small scales was considered by Riemann even before the de-
velopment of quantum mechanics [3]
There arises from this problem of searching out the simplest facts by which
the metric relations of space can be determined, a problem which in nature
of things is not quite definite . . . These facts are, like all facts, not necessary
but of a merely empirical certainty; they are hypothesis; one may therefore
inquire into their probability, which is truly very great within the bounds of
observation, and thereafter decide concerning the admissibility of protracting
them outside the limits of observation, not only toward the immeasurably large,
but also toward the immeasurably small.
From London’s work up to today, there has been very few but also very interesting
analysis relating non-Euclidean spaces and quantum mechanics [4–14]. In general,
all these attempts to reproduce quantum phenomena are based on Weyl spaces. After
Dirac’s work [15], the properties of a Weyl space have been considerably changed
with respect to Weyl’s original idea. In Appendix we describe in some detail what is
nowadays called a Weyl space.
Found Phys

Our aim in this work is to propose a similar but different approach to describe
quantum effects. We argue that relativistic quantum mechanics can also be under-
stood as a manifestation of what we shall call a Qwist (quantum weyl integrable
spacetime). In particular, we construct a geometrical version of Heisenberg’s uncer-
tainty principle, which is related to the variation of a vector’s length in Qwist. This
length variability also happens in Weyl spaces but with a completely different physi-
cal meaning as shall be clear in what follows.
In the next section we shall describe the formal basis of our approach. In advance,
it seems convenient to stress that to define a physical theory one has first of all to
define its kinematic properties. Given an action principle together with a set of dy-
namical fields does not completely specify the theory. In particular, each theory has
its own internal symmetries, which can be used to construct an equivalent class of
observers [16]. The point we would like to stress is that one has to define from the
beginning which are the allowed transformations of the dynamical fields.
In particular, one should not be misled by similarities of some of the equations and
confuse Qwist with any other geometrical space. Let us now define the structure of
the Qwist space that we shall identify with the physical spacetime.

1.1 Qwist Geometry

A Qwist is a geometrical manifold (gμν ; Γμν α ) endowed with a metric tensor and a

symmetric affine connection Γμν = Γ(μν) . Its symmetry group is the Manifold Map-
α α

ping Group MMG, which allow us to perform an arbitrary coordinate transforma-

tion.1 The connection of this space is defined with an extra degree of freedom given
by a scalar field Ω(x). Thus, Qwist is neither a Riemannian space nor a Weyl geom-
etry (see Appendix for more details).
We shall construct our connection such that the non-metricity condition be given
by
∇α gμν ≡ −2 (ln Ω),α gμν . (1)
One can use the above equation to solve for the connection giving
1  
α
Γμν = {μν
α
}+ Ω,ν δ αμ + Ω,μ δ αν − Ω ,α gμν . (2)
Ω
Since the connection is not equal to the Christoffel symbol, it is adequate to dis-
tinguish between two kind of covariant derivative. The Qwist covariant derivative is
constructed with the connection and we shall denote by
μ μ
∇α ξ μ ≡ ξ ;α = ξ μ,α + Γαλ ξ λ ,

and a Riemannian covariant derivative

μ μ 
ξ //α = ξ μ,α + αλ ξ λ.

1 Note that the symmetry group of a Qwist space does not include conformal transformations, which are
characteristic to Weyl spaces.
 Found Phys

The non-metricity condition (1) implies that the length of a vector is not preserved
if parallel transported
δl = −lΩ −1 Ω,μ dx μ . (3)
The above equation has an important physical meaning. It describes how the phys-
ical length of a ruler changes from point to point. Note that as long as the extra degree
of freedom is a scalar field Ω  (x  ) = Ω(x) there is no gauge freedom in (3).
Furthermore, condition (1) does not suffers from any kind of second clock effect
[17, 18] that could be present if instead of the gradient of a scalar function it were a
vector field. Considering (3), it is immediate to show that the length of a vector does
not change along a closed path

δl = 0. (4)

Thus, this property guarantees that all local measurements such as distances are
well defined and can be uniquely determined.
As usual, the curvature tensor can be written in terms of the connection as

Rαμβν = Γμβ,ν
α
− Γμν,β
α
+ Γμβ
ε α
Γνε − Γμν
ε α
Γεβ ,

which can be used to calculate its traces.

Hence, the Ricci tensor is given by
 
2∇ν (Ω,μ ) 4Ω,μ Ω,ν Ω Ω,λ Ω ,λ
Rμν = R̂μν − + − g μν + , (5)
Ω Ω2 Ω Ω2

where  ≡ ∇μ ∇ μ is the d’Alembertian operator and R̂μν is the Riemannian part of

the Ricci tensor, i.e. the Ricci tensor constructed solely with the Christoffel symbol.
For the curvature scalar R ≡ g μν Rμν we find
Ω
R = R̂ − 6  ln Ω − 6 (ln Ω),α (ln Ω),α = R̂ − 6 , (6)
Ω
where again R̂ ≡ g μν R̂μν is its corresponding Riemannian part.
It is interesting to show that it is possible to derive the geometrical properties of a
Qwist space from a palatini-like variational principle by considering the connection
as an independent field. For this purpose, consider the functional

√
I = d4 x −g Ω 2 R. (7)

If we demand this functional to be stationary with respect to variation of the connec-

tion we find

√ μν
δI = d4 x −g Ω 2 g μν δRμν = d4 x Zλ δΓμν λ
= 0, (8)

μν
where Zλ is given by

μν √ 1 √ 1 √ μ
Zλ ≡ ( −g g μν Ω 2 ); λ − ( −g g μα Ω 2 ); α δ νλ − ( −g g να Ω 2 ); α δ λ = 0. (9)
2 2
Found Phys

Taking the trace of the above expression yields

√ 
−g g μα Ω 2 = 0. (10)
;α

Substituting (10) again in (9) we finally obtain the condition that characterize a Qwist
geometry
Ω, α
gμν ; α = −2 gμν . (11)
Ω
In the following sections we will study the dynamics of a spinless charged particle
in a Qwist geometry and relate it to a relativistic quantum system. In particular, we
will show that this system yields the correct non-relativistic limit, i.e. the Schrödinger
equation for a charged particle.
We shall describe quantum mechanics using the Bohm-de Broglie causal inter-
pretation, which are amongst the well-defined formulations that reproduce the same
results of the orthodox interpretation but has the advantage of describing matter as
point-like particles.
In addition, we propose a geometrical interpretation of Heisenberg’s uncer-
tainty principle. This relativistic geometrical version combines both the position-
momentum and time-energy relations in an unique principle, which decouples into
the usual Heisenberg’s uncertainty principles in the non-relativistic limit.
Finally, we discuss how our results are related to Klein-Gordon’s equation from a
geometrical point of view and conclude in the last section with some final remarks.

2 Relativistic Quantum Mechanics

In this section, we will describe a system composed of a relativistic charged point-like

particle interacting with an external electromagnetic field and geometry. Therefore,
we are considering a charged particle wandering in a non-Euclidean spacetime that
has an independent degree of freedom.
This is perhaps the simplest relativistic system but it will be interesting enough
to present the connection between Qwist and quantum mechanics. We decided to
include an external electromagnetic field (non-dynamical) insofar as it does not veil
the main properties of Qwist. Notwithstanding, if preferable, it is possible to do the
same analysis turning the electromagnetic field off, which is equivalent to consider a
neutral particle without expense of the quantum physical content.
Our discussion will be based on a variational principle that at the same time pro-
vides the dynamical equation for the particle and naturally endows the spacetime with
an affine structure that is typical of a Qwist space. As will be clear in what follows,
the particle’s dynamics is given by integrating a relativistic Hamilton-Jacobi equation
where its momentum is described by the derivative of Hamilton’s principle function
Pμ = ∂μ S. In addition, apart from a kind of non-minimal coupling, the geometrical
sector shall be given by the Ricci scalar.
Thus, consider the following action that should be justified a posteriori

1 √  
I= d4 x −gΩ 2 Lg + Lm , (12)
κ
 Found Phys

with

Lg ≡ λ2 R, (13)
g μν
e 
e 
Lm ≡ 2 ∂μ S − Aμ ∂ν S − Aν − μ2 . (14)
 c c
In the above expressions g is the determinant of the metric gμν and R is the
Ricci scalar. As already mentioned, S is the relativistic version of the Hamilton’s

principle function of the particle which here is coupled to the gauge field Aμ = (ϕ, A)
describing an external electromagnetic field.
The constants appearing above are the speed of light c, the gravitational constant
κ ≡ 16πG/c3 , a dimensionless λ to be determined, the particle’s electric charge e
and its inverse Compton wavelength μ ≡ mc/.
We should strongly stress that there is no gravitational interaction in this system.
The gravitational constant κ appears as a global factor and it is introduce only to
adjust the dimensionality of the action [I ] = . Actually, one should not be surprised
by the introduction of κ inasmuch this is the only way to change the dimensionality
of the curvature scalar to dimension of action.
This system has three dynamical variables to be varied, namely, the dimensionless
scalar function Ω, the connection Γμν λ and S. Variation with respect to the connection

give us the geometrical structure of spacetime. Following the derivation of Sect. 1.1,
(7)–(11), we have

μν
δI = d4 x Zλ δΓμν λ
= 0 ⇒ gμν;λ = −2 (ln Ω),λ gμν . (15)

Varying the Hamilton’s function S give us a conservation-like equation,

2 √ e 
δI = − 2 d4 x −g Ω 2 ∂ μ (δS) ∂μ S − Aμ = 0,
κ c


2 √ e
= 2 d4 x −g g μν Ω 2 ∂μ S − Aμ δS = 0, (16)
κ c //ν

e 
⇒ g μν Ω 2 ∂μ S − Aμ = 0,
c //ν

where we have dropped a surface term using the four-dimensional Gauss theorem.
Note that we have written this expression using a Riemannian covariant derivative,
i.e. defined using the Christoffel symbol only.
Finally, variation with respect to Ω gives

e 
e 
∂μ S − Aμ ∂ μ S − Aμ − m2 c2 + λ2 2 R = 0. (17)
c c
Since we are supposing that the Riemannian part of the Qwist geometry is
Minkowskian, gμν = diag(1, −1, −1, −1), by choosing λ2 = 1/6, the above equa-
tion becomes

e 
e  Ω
∂μ S − Aμ ∂ μ S − Aμ − m2 c2 − 2 = 0. (18)
c c Ω
Found Phys

Equation (18) generalize the relativistic Hamilton-Jacobi equation with the inclu-
sion of the last term. This extra term which is basically the Weyl curvature scalar
respond for all relativistic quantum effects of this system.
The Weyl curvature scalar plays the same role in this relativistic scenario as the
Bohmian quantum potential for the non-relativistic quantum mechanics [19–22]. In
particular, the classical regime is attained in the limit Ω → 0.
Given a field configuration for the gauge field Aμ , (16) and (18) define a closed
system that is well defined with the specification of appropriate initial conditions.
Notwithstanding, since quantum effects are now given by modifications of the space-
time structure, it shall be convenient to use a hydrodynamical description and asso-
ciate the particle’s world-line with a time-like congruence.
Following [23–26] we shall define our canonical momentum one-form P̃ = Pμ θμ
and de Broglie’s mass respectively as
e
Pμ ≡ ∂μ S − Aμ , (19)
c
 
 2 2 Ω
M ≡ m2 + 2 R = m 1 + 2 2 . (20)
6c m c Ω

By virtue of (18)–(20) we can define a unitary time-like velocity field by

dx μ 1
Uμ = c ≡ g μν Pν =⇒ U μ Uμ = c 2 i.e. dλ2 = gμν dx μ dx ν . (21)
dλ M
However, the “Euclidean” particle velocity field is define as

dx μ 1 M2 2
Vμ =c ≡ g μν Pν =⇒ V μ Vμ = c . (22)
ds m m2
Thus, there is an intrinsic time re-parameterization along the particle’s trajectory
given by
dλ m
ds =  = dλ. (23)
1+ 2 Ω M
m2 c 2 Ω

In addition, within this geometrical interpretation, (16) can be viewed as a integra-

bility condition. One can immediately show that for any vector ξ μ we have

μ 1
4 μ 1
2 μ
ξ ;μ = Ω ξ or ξμ;ν g μν = Ω ξ ,
Ω4 //μ Ω2 //μ

which connects the covariant derivative of Qwist and Riemann geometries. Hence,
the continuity-like equation (16) can be written as
1
2 α
Ω P = Pα ; β g αβ = 0, (24)
Ω2 //α

which can be viewed as an integrability condition on Qwist for the canonical momen-
tum Pμ .
 Found Phys

Note that the original system equations (16) and (18) are equivalent to (21) and
(24). Notwithstanding the mathematical equivalence, one should bear in mind that its
physical interpretation is completely different. The Dynamical equations for the two
scalar functions, namely S and Ω is now substituted by kinematic relations for the
time-like congruence U μ .
As a matter of fact, specifying our Cauchy-surface Σc as a space-like hypersurface
where P μ is time-like and orthogonal everywhere, one can show [26–28] that the
integrability condition given by (24) guarantees that P μ will always remains time-
like. This can be proved as follows.
Consider the variation along the congruence U μ of the quantity Ω 2 MδV , where
δV is an infinitesimal 3-volume orthogonal to the congruence. It’s straightforward to
show that
d
−2 

Ω MδV = U α ∂α Ω −2 M δV + Ω −2 MU α ∂α (δV )
dλ


= U α ∂α Ω −2 M δV + Ω −2 MδV U α;α



= Ω −2 MU α δV = Ω −2 P α δV
;α ;α
−2
= Pα ; β g αβ
Ω δV = 0 . (25)

Hence, as long as δV and Ω −2 are always positive and (25) shows that
Ω −2 MδV = C te then M 2 should not change sign. Given that P μ is time-like on
the Cauchy-surface, i.e. P μ Pμ (Σc ) = M 2 (Σc ) > 0 then it also has to be time-like
everywhere Pμ P μ (x) > 0.
Our system is suitable to describe relativistic spinless charged particles. Therefore,
one might be concerned how to deal with creation-annihilation processes and if it is
possible to consistently define a time-like congruence like P μ in the presence of both
particles and anti-particles.
We should emphasize that this analyses deals only with non-interacting particle
which in a sense avoid these kind of difficulties. However, this formalism is as good
to describe particles as it is to describe anti-particles due to its invariance under time
reversal accompanied by a change of sign of the electrical charge e. A time rever-
sal is equivalent to a change of sign of the Hamilton’s function S → −S. Hence
it’s straightforward to show that the system of (16) and (18), remains unchanged if
e → −e. We shall come back later to this issues in Sect. 2.3.
Finally, consider the particle’s trajectory, which is given by integrating (21). One
should already expect that the particle should not follow a geodetic trajectory as long
as it has a quantum force acting on it. Given the particle’s velocity field we can cal-
culate this quantum force using the kinematic equations. As a matter of convenience,
we shall for the moment consider a Cartesian coordinates so that
dU μ c2 U α ∂α M μ e
= ∂ μM − U + F μα U α ,
dλ M M M
where we have define the components of the electromagnetic tensor as Fμν ≡ ∂μ Aν −
∂ν Aμ . We can readily recognize the last term as being the Lorentz’s force while the
Found Phys

two first ones are intrinsically geometrical terms that we commonly associate with
quantum effects.
The geodesic equation written in its covariant form gives

d2 x μ α
μ dx dx
β e
+ Γ = Wμ + F μ U α, (26)
dλ 2 αβ
dλ dλ Mc2 α
with
   
M 1 α M
W μ ≡ ∂ μ ln − U ∂α ln U μ. (27)
Ω c2 Ω2
Note that the quantum force W μ also depends on the velocity field W μ =
W μ (Ω, U λ ).

2.1 Non-relativistic Limit

The above system describes a charged relativistic point-like particle interacting with
an external electromagnetic field in a Q-wist geometry. This system can be under-
stood as a relativistic generalization of the Schrödinger picture of quantum mechan-
ics. As we shall now show, it is possible to recover the Schrödinger description by
redefining the Hamilton’s principal function and taking the usual non-relativistic limit
c → ∞.
One of the main difference in the description of a relativistic particle is that its
energy contains its inertial rest mass as a potential-like energy. Furthermore, inas-
much the energy is related to the Hamilton’s function by E = − ∂S ∂t , we define the
non-relativistic version of the Hamilton’s function by Snr ≡ S + mc2 t .
Substituting this ansatz in (18) we find
 
∂Snr eA0 1
 e 2 e2 A0 A0 2 ∇ 2 Ω
1− 2
+ ∇Snr − A + eA0 − 2
−
∂t mc 2m c 2mc 2m Ω
 2 
1 ∂Snr 2 ∂ 2 Ω
− − = 0.
2mc2 ∂t Ω ∂t 2
eA0
By considering the limit c → ∞ it is licit to neglect mc2
with respect to 1 and
e2 A0 A0
2mc2
to eA0 . Furthermore, the last two terms go away and we identify the spatial
part of the Weyl curvature as the non-relativistic Bohmian quantum potential Q =
2 ∇ 2 Ω
− 2m Ω [19–22].
Thus, in this limit we have
∂Snr 1
 e 2
+ ∇Snr − A + eA0 + Q = 0, (28)
∂t 2m c
which reproduce the first non-relativistic Bohmian equation, namely the Hamilton-
Jacobi like equation.
The second non-relativistic equation is derived directly from (16),
   
1 ∂ 2 ∂Snr  Ω 2 (∇S nr − e A)
 =0
Ω − mc 2
− eA 0 − ∇.
c2 ∂t ∂t c
 Found Phys

eA0
and again neglecting mc2
with respect to 1 and the terms containing 1/c2
  
∂Ω 2  
∇S nr e 
A
+ ∇. Ω 2 − = 0. (29)
∂t m mc

As it’s well known, (28) and (29) are equivalent to the Schrödinger equation.2
Hence it is in fact legitimate to view (16) and (18) as relativistic generalizations of the
Schröndinger picture of quantum mechanics for a spinless charged point-like particle.
Note that in this limit the geometrical description degenerates to

M −→ m, U μ −→ V μ , dλ −→ ds.

The de Broglie’s Mass M goes to the particle’s rest mass m while both velocity
fields coincide as well as its parameterization. In non-relativistic quantum mechanics
one can understand the quantum force as a deviation from an Euclidean geodesic
[29]. If we chose a coordinate system which has a vanishing Christoffel symbol, one
can show that
d2 x μ ∂ e
= hμν ν Q + F μν V ν ,
dt 2 ∂x m
where we have defined the projector tensor hμν along the velocity field V μ by hμν ≡
g μν − V μ V ν . Taking the non-relativistic limit (see [30–32] for more details), the
spatial part becomes

d2 x

m  + e E + v × B ,
= −∇Q
dt 2

where v is the particle’s 3-velocity. As expected, the particle feels the usual electro-
magnetic force plus the non-relativistic quantum force given by −∇Q. 

2.2 Relativistic Uncertainty Principle

The uncertainty principle is in the core of the orthodox interpretation of quantum

mechanics. In a recent paper [29], we have shown that it is possible to interpret geo-
metrically the uncertainty relation for the position and momentum by virtue of a
characteristic length scale defined by a 3-d curvature scalar.
However, as long as we have considered a non-relativistic theory, we have estab-
lished a purely spatial relation which in the relativistic context is unsatisfactory by
the requirement of covariance.
In this section, we shall generalize this “spatial” uncertainty principle to a four
dimensional relation. In a relativistic theory only four dimensional quantity acquires
physical meaning. Accordingly, the interdependence between space and time compel
us to somehow relate the uncertainty principle to the interval ds 2 = c2 dτ 2 − dl 2 . In

2 The above mentioned pair of equations are precisely the Bohm-de Broglie system of equations that one

finds when uses the polar form of the wave-function ψ = Ω exp{ i Sc } in Schrödinger’s equation.
Found Phys

addition, taking the non-relativistic limit from this relation we shall show that we
recover both uncertainty relation for position and momentum as well as for time and
energy.
As we have shown (6), the Qwist curvature scalar can be decomposed in a Rie-
mannian part plus the contribution of the extra degree of freedom Ω. If we suppose
that the Riemannian part is flat, R̂ = 0, then

6 ∂ 2Ω ∇ 2Ω
R=− + 6 .
c2 Ω ∂t 2 Ω
Apart from the speed of light c, the above equation shows that Qwist curvature
scalar which has dimension of inverse length squared is a sum of two terms one with
dimension of inverse time squared and the other with dimension of inverse length
squared. Thus, we define the Weyl length and Weyl time by
 2 −1/2  
 ∇ Ω  6 ∂ 2 Ω −1/2
Lw ≡  
6 Ω  ; Tw ≡ 
 Ω ∂t 2 
 . (30)

In a previous paper [29], we have shown that non-relativistic quantum mechanics

can be pictured also as a modification of the 3d Euclidean space which we called
Qwis. Through this analysis we were able to construct an uncertainty principle for
space and momentum based on the characteristic length above defined Lw .
If in fact the Qwist curvature scalar R is related to quantum phenomena then the
above quantities (30) have to somehow yield a measurement of departure from a
classical behavior. Recall (23)

dλ
ds =  .
2 Ω
1+ m2 c 2 Ω

In the non-relativistic limit, i.e. c → ∞, we have dλ → c.dτ . Furthermore, the

time derivative of Ω can be neglected compared to its spatial derivative. Thus, we
have
 
ds dl 2 v2 v2 p2
= 1− 2 = 1− 2 ≈1− 2 =1−
dλ dλ c 2c 2m2 c2
1 2 ∇ 2 Ω
= ≈1+ .
1+ 2 Ω 2m2 c2 Ω
m2 c 2 Ω

Comparing the above equations we find

2
p 2 . L2w = .
6
To interpret the above equation as an uncertainty-like relation we need an extra as-
sumption. Suppose that any length measurement ΔL can only measure distances big-
 Found Phys

ger than the characteristic Weyl length, i.e.

 2 −1/2
 ∇ Ω
ΔL ≥ Lw =  
6 Ω  . (31)

The reasonability of this hypothesis lies on the notion of a classical measurement.

A length measurement is made with a standard ruler which is supposed to be a stiff
object. Thus, the notion of classical standard ruler presuppose the validity of Euclid-
ean space (see [29] for a more detailed discussion). With the hypothesis (31), the
above relation becomes

Δp 2 . ΔL2 ≥ Δp 2 . L2w ⇒ Δp.ΔL ≥ √ . (32)
6
In addition, it is also possible to derive the uncertainty relation for time and energy.
For the moment we shall turn off the electromagnetic interaction (Aμ = 0). Without
the electromagnetic potential, (18) can be recast as

2 ∂ 2 Ω 2 2∇ Ω
2
E 2 = m2 c4 + p 2 c2 + −  c
Ω ∂t 2 Ω
  2 
1 p 2 ∇ 2 Ω 1 2 ∂ 2 Ω
⇒ E = mc2 1 + 2 2
− 2
+ 4 2 .
c m m Ω c m Ω ∂t 2

In Sect. 2.1, we have argued that the non-relativistic Hamilton’s principle function
should be related to the relativistic one by Snr = S + mc2 t which can be interpreted as
Enr = E − mc2 . Therefore, expanding in power of c−2 , the above equation becomes

p2 2 ∇ 2 Ω 2 ∂ 2 Ω 1
2 
Enr = − + 2 2
− 3 2
p − 2 Ω −1 ∇ 2 + O(1/c4 ).
2m 2mΩ 2mc Ω ∂t 8m c
(33)

In zeroth order,
p2 2 ∇ 2 Ω
Enr = − .
2m 2mΩ
Note that the non-relativistic energy includes a classical term p 2 /2m plus a geomet-
rical term − 2mΩ
2∇2Ω
. Using this result in the second order term we find

2 ∂ 2 Ω 2
Ecl
0= −
2mc2 Ω ∂t 2 2mc2
which gives

2
2
Ecl . Tw2 = . (34)
6
A time measurement is by definition the length of a 4-d trajectory in a fix spatial
point. However, in Qwist appears an intrinsic time re-parameterization (23) which is
Found Phys

certainly not include in the Euclidean definition of a standard clock. Generalizing the
hypothesis that a length measurement has to be greater than the weyl length Lw , we
shall suppose that a time measurement realized by a standard clock has to be greater
than the weyl time Tw , i.e. Δt ≥ Tw . Through this hypothesis, the above relation
becomes

ΔEcl .Δt ≥ ΔEcl .Tw ⇒ ΔEcl .Δt ≥ √ . (35)
6
Note that the uncertainty relation for position and momentum (32) is naturally
incorporated into the uncertainty relation for time and energy. Actually, it is the en-
tanglement between space and time that allows us to derive the above relation (35).
In orthodox quantum mechanics, the impossibility of deriving the uncertainty re-
lation for time and energy is commonly associated with the role played by time as an
external parameter, i.e. a lack of a time operator. Our geometrical approach does not
deal with operators and in fact treats space and time variables on equal footing in the
relativistic sense.

2.3 Geometrical Interpretation of Klein-Gordon’s Equation

The geometrical approach developed above describes a relativistic particle interacting

with a non-Euclidean geometry. The deviation from a pure classical behavior comes
from the non-vanishing of the Ricci scalar R. Once Ω → 0, the classical regime
is recovered. Hence, the extra degree of freedom of the Qwist spacetime, namely the
scalar function Ω is responsible for the departure from classical behavior.
Interesting enough, our geometrical variational method is formally equivalent to
a Klein-Gordon system. Suppose that instead of treating separately the particle’s and
the geometrical degrees of freedom, we have defined a wave-function by combining
both ψ = φ. exp{ i S}. The φ field has dimension of inverse length as commonly is
done in field
√ theory and is related to our dimensionless geometrical degree of freedom
by Ω = κ φ.
With respect to this hybrid object ψ , as should have been expected, (16) and (18)
can be combined into one equation, a massive Klein-Gordon equation

Dμ D μ ψ − m2 c2 ψ = 0,

where we have defined the gauge covariant derivative operator Dμ ≡ i∂μ + eAμ .
As it is well known, the above equation can be derived using a variational principle
with Lagrangian density


L = −−1 D̄μ ψ ∗ D μ ψ − m2 c2 ψ ∗ ψ , (36)

where ψ and ψ ∗ should be treated as independent dynamical variables. Modulus a

total derivative, (36) defines an action

 
1 √ Ω g μν
e 
e  m2 c2
Iψ = d4 x −g Ω 2 − 2 ∂μ S − Aμ ∂ν S − Aν + 2 .
κ Ω  c c 
(37)
 Found Phys

Note that the above action is precisely (12), i.e. I = Iψ , if we impose beforehand
a Weyl affine structure for the spacetime and identify the scalar curvature with the
first term involving derivatives of Ω. Thus, in a sense, our approach is more general
as long as the affine structure is derived as a palatini-like variational principle.
This Lagrangian density naturally defines a conserved current
1 ∗← →  
Jμ ≡ − ψ D μ ψ = Ω 2 ∂ μ S − eAμ = Ω 2 g μν Pν , (38)
2κ
and a energy-momentum tensor
2c c

Tμν ≡ D̄μ ψ ∗ D̄ν ψ − D̄λ ψ ∗ D̄ λ ψ + m2 ψ ∗ ψ gμν . (39)
 
The Klein-Gordon equation can be casted in a Schrödinger-like form by defining
a two-component wave-function (see [26, 33] for more detail). In this approach, it is
possible to identify positive and negative energy solutions. The energy is defined as
the eigenvalue of the Hamiltonian of this Schrödinger picture and it is numerically
equalto the spatial integral of the 00-component of the energy-momentum tensor, i.e.
E = d3 x T00 .
It can be shown that the negative energy solution can be mapped into the positive
energy solution by a charge conjugation operation. Therefore, as it is well known, we
can associate the negative energy solutions to anti-particle states. This charge conju-
gation is intrinsically related to the invariance of the system by a change of ψ → ψ ∗ .
Hence, charge conjugation can be imitated by a change S → −S and e → −e.
Notwithstanding, as previously emphasized, our analysis deals only with non-
interacting particle. Thus, even thought this formalism is as adequate to particles
as to anti-particles, we have not yet established how interacting process like creation-
annihilation should be described.

3 Conclusions

It had been shown in the literature that exist a very interesting connection between
non-Euclidean geometries and quantum phenomena. The predominant mechanism to
describe quantum effects by geometrical degrees of freedom has been based on Weyl
space. In the present work, we proposed a new geometrical approach based on a new
geometrical space that we called Qwist.
In Qwist, its extra scalar degree of freedom produce a length variability, which
is responsible for change in size of extended object. Furthermore, this is a physical
and in principle measurable effect. The physical interpretation of this length variabil-
ity allowed us to formulate a relativistic and geometrical version of the uncertainty
principle.
The non-Euclidean properties of Qwist provide two characteristics dimensions,
namely a Weyl length and a Weyl time defined by
 2 −1/2  
∇ Ω   1 ∂ 2 Ω −1/2
Lw ≡ 
 Ω 
 ; 
Tw ≡   .
Ω ∂t 2 
Found Phys

These quantities quantify the departure from an Euclidean geometry, which can be
used to restrict the validity of a classical measurement. Therefore, there should have
some restriction on determining the properties of the system if there is a significant
departure from Euclidean geometry, i.e. there is a significant manifestation of non-
Euclideanity in Qwist.
To support the idea that action (12) correctly describes the dynamics of a rel-
ativistic charged “quantum” particle, we have studied its non-relativistic limit and
shown that it recovers the usual Schrödinger quantum dynamics. In addition, we re-
formulated our dynamical variables to connect this relativistic system with the Klein-
Gordon equation. In particular, it was necessary to define a new complex field that
from our point of view mix geometrical degrees of freedom with the particle’s Hamil-
ton principle function.
The present formalism is adequate to describe particles as well as anti-particles.
However, it not yet clear how interactions between particles and anti-particles shall
be included in this scenario. In addition, it is still an open issue the meaning of a
many-particles system and its physical interpretation in view of the Qwist geometrical
interpretation.

Acknowledgements We would like to thank CNPq of Brazil and MN also to FAPERJ for financial sup-
port. We would also like to thank the participants of “Pequeno Seminário” of ICRA-CBPF’s Cosmology
Group for useful discussions, comments and suggestions.

Appendix: Weyl Geometry

The Weyl space is construct so as to incorporate the gauge transformation of a vec-

tor field (wμ → wμ + Λ,μ ) analogous to the electromagnetic gauge transformation
[34–38]. This new vector field is associated with the non-zero covariant derivative of
the metric tensor. The gauge transformation of the Weyl space is a combine transfor-
mation of the metric and of the weyl vector field. In a weyl gauge transformation we
have

gαβ −→ gαβ = e2Λ(x) gαβ , (40)
wμ −→ wμ = wμ + Λ,μ , (41)

while the connection is constructed so that the covariant derivative of the metric is
zero Dg = 0 (see [34, 35]), or in components
ρ
Dμ gαβ = ∂μ gαβ − gρβ Γμα
ρ
− gαρ Γμβ − 2wμ gαβ = 0. (42)

This equation can be solved to give

 
α
Γμν = {μν
α
} − δμα wν + δνα wμ − gμν w α , (43)

where {μνα } is the Christoffel symbol. Note that the covariant derivative defined in

(42) is the usual covariant derivative constructed with the connection plus an extra
term related to the weyl vector field (wμ ). To distinguish it from the conventional
 Found Phys

covariant derivative, one call it co-covariant derivative and defined it as follows. If

a tensor is transformed under a weyl gauge transformation into Aμ → Aμ = enΛ Aμ
then it is called a tensor of power n and its co-covariant derivative is given by

Dμ ξα ≡ ∇μ ξα − nξα Λμ , (44)

where the ∇ defines the covariant derivative constructed with the connection ∇ν ξμ ≡
ξμ,ν − Γνμ ξα . Hence, we have defined the metric gμν as a tensor of power 2 in the
α

weyl sense.
From (42) one can easily show that

∇μ gαβ = 2wμ gαβ . (45)


Thus, it is also straightforward to show that the length of a vector l ≡ gμν l μ l ν
parallel transported will change from point to point such that in an infinitesimal dis-
placement is given by
δl = lwμ dx μ . (46)
Note that the length of a vector is not gauge invariant. As a matter of fact, under
a weyl gauge transformation we have l → l  = eΛ(x) l. Thus, by an adequate  gauge
choice one can always make δl = 0 in a point. It suffice to choose Λ = − wμ dx μ ,
then
 
δl  = δΛ l  + eΛ δl = l  Λ,μ + wμ dx μ = 0.
However, its variation along a closed curve will not be zero since in general the
curvature Kμν ≡ wμ,ν − wν,μ = 0. There is a special case of Weyl space when the
Weyl vector wμ is the gradient of a function, i.e. wμ = ∂μ f . In this case, the length of
a vector will not change along a closed curve. Furthermore, one can choose Λ = −f
so that wμ = f,μ + Λ,μ = 0, which could motivate us to associate this subclass of
Weyl spaces with conformal transformations of the metric tensor.
However, even in this special case when the weyl vector is the gradient of a func-
tion, the Weyl space is not equivalent to a conformal transformation of the associated
Riemannian space.
A conformal transformation of the metric tensor takes for example gμν →
g̃μν = e2Λ gμν . This transformation maps a Riemannian space M characterized by
(gμν , {μν
α }) into another Riemannian space M  with (g̃μν , {
μν }). Thus, we have
α

 1 αξ   α   α 
μν ≡ g̃ g̃ξ ν,μ + g̃μξ,ν − g̃μν,ξ = μν + δ ν Λ,μ + δ αμ Λ,ν − gμν Λ,α
α
2
which can be inverted
α     α 
μν = 
μν − δ ν Λ,μ + δ μ Λ,ν − gμν Λ
α α ,α
(47)

becoming very similar to (43).

However, the new Riemannian manifold M  still satisfies the metricity condition,

namely, ∇α g̃μν = 0. The covariant derivative of the transformed metric tensor is still
Found Phys

zero. The conformal transformation does not change the metricity condition. Note
that if we blindly calculate ∇α g̃μν we find
 

∇α gμν = ∇α e2Λ g̃μν = 2Λ,α gμν (48)

which is also very similar to (45). However, contrary to (45), (48) has no physical

meaning. It is the covariant derivative in M applied to the metric tensor of M.

References

1. Weyl, H.: Sitz. Preuss. Akad. Wiss. 26, 465 (1918)

2. London, F.: Z. Phys. 42, 375 (1927)
3. Riemann: In: Smith, D.E. (ed.) A Source Book in Mathematics, vol. 2. Dover, New York (1959)
4. Israelit, M.: Found. Phys. 28, 205 (1998)
5. Israelit, M.: Found. Phys. 29, 1303 (1999)
6. Israelit, M.: Found. Phys. 32, 295 (2002)
7. Israelit, M.: Found. Phys. 32, 945 (2002)
8. Canuto, V., Adams, P.J., Hsieh, S.H., Tsiang, E.: Phys. Rev. D 16, 1643 (1977)
9. Rosen, N.: Found. Phys. 13, 363 (1983)
10. Koch, B.: arxiv:0810.2786 [hep-th] (2008)
11. Koch, B.: arxiv:0901.4106 [gr-qc] (2009)
12. Santamato, E.: Phys. Rev. D 29, 216–222 (1984)
13. Holland, P., Kyprianidis, A., Vigier, J.P.: Phys. Lett. A 107, 376 (1985)
14. Gueret, Ph., Holland, P., Kyprianidis, A., Vigier, J.P.: Phys. Lett. A 107, 379 (1985)
15. Dirac, P.M.: Proc. R. Soc. Lond. A 333, 403 (1973)
16. Anderson, J.L.: Principles of Relativity Physics. Academic Press, London (1967)
17. Novello, M., Oliveira, L.A.R., Salim, J.M., Elbaz, E.: Int. J. Mod. Phys. D 1(4), 641–677 (1992)
18. Perlick, V.: Class. Quantum Gravity 8, 1369 (1991)
19. Bohm, D.: Phys. Rev. 85, 166 (1952)
20. Bohm, D.: Phys. Rev. 85, 180 (1952)
21. Bohm, D., Hiley, B.J.: The Undivided Universe. Routledge, London (1993)
22. Holland, P.R.: The Quantum Theory of Motion. Cambridge University Press, Cambridge (1993)
23. de Broglie, L.: Non-Linear Wave Mechanics: A Causal Interpretation. Elsevier, Amsterdam (1960)
24. Halbwachs, F.: Théorie Relativiste des Fluides a Spin. Gauthier-Villars, Paris (1960)
25. Holland, P.R.: Found. Phys. 17, 345–363 (1987)
26. Cufaro-Petroni, N., Dewdney, C., Holland, P., Kyprianidis, T., Vigier, J.P.: Phys. Lett. A 106, 368–370
(1984)
27. Kyprianidis, A.: Phys. Lett. A 111, 111–116 (1985)
28. Santamato, E.: J. Math. Phys. 25, 2477–2480 (1984)
29. Novello, M., Salim, J.M., Falciano, F.T.: arxiv:0901.3741 [gr-qc] (2009)
30. Landau, L.D., Lifshitz, E.M.: The Classical Theory of Fields. Elsevier, Oxford (1975)
31. Synge, J.L.: Relativity: The Special Theory. North-Holland, Amsterdam (1958)
32. Tolman, R.C.: Relativity Thermodynamics and Cosmology. Oxford University Press, Oxford (1962)
33. Feshbach, H., Villars, F.: Rev. Mod. Phys. 30, 24 (1958)
34. Rosen, N.: Found. Phys. 12, 213 (1982)
35. Wheeler, J.T.: Phys. Rev. D 41, 431 (1990)
36. Shojai, F., Shojai, A.: Gravit. Cosmol. 9, 163 (2003)
37. Shojai, F., Shojai, A.: arxiv:gr-qc/0404102 (2004)
38. Carroll, R.: arxiv:0705.3921 [gr-qc] (2008)