1

ON A COMPREHENSIVE THEORY OF
SEMI-CLASSICAL QUANTUM GRAVOELECTRODYNAMICS

Indranu Suhendro
Department of Physics
Karlstad University
Karlstad 651 88
Sweden

Abstract

We consider a unification of gravity and electromagnetism in which electromagnetic
interaction is seen to produce a gravitational field. The field equations of gravity and
electromagnetism are therefore completely determined by the fundamental
electromagnetic laws. Insight into this unification is that although gravity and
electromagnetism have different physical characteristics (e.g., they differ in strength), it
can be shown through the algebraic properties of the curvature and the
electromagnetic field tensors that they are just different aspects of the geometry of
space-time. Another hint comes from the known speed of interaction of gravity and
electromagnetism: electromagnetic and gravitational waves both travel at the speed of
light. This means that they must somehow obey the same wave equation. This indeed
is unity. Consequently, many different gravitational and electromagnetic phenomena
may be described by a single wave equation reminiscent of the scalar Klein-Gordon
equation in quantum mechanics. Light is understood to be a gravoelectromagnetic
wave generated by a current-producing oscillating charge. The charge itself is generated
by the torsion of space-time. This electric (or more generally, electric-magnetic) charge
in turn is responsible for the creation of matter, hence also the transformation of
matter into energy and vice versa. Externally, the gravitational field manifests itself as
the final outcome of the entire process. Hence gravity and electromagnetism obey the
same set of field equations, i.e., they derive from a common origin. As a result, the
charge produces the so-called gravitational mass. Albeit the geometric non-linearity of
gravity, the linearity of electromagnetism is undisturbed: an idea which is central also in
quantum mechanics. Therefore we preserve the most basic properties of matter such
as energy, momentum, mass, charge and spin through this linearity. It is our modest
attempt to once again achieve a comprehensive unification which explains that gravity,
electromagnetism, matter and light are only different aspects of a single theory.


1. Introduction

Attempts at a consistent unified field theory of the classical fields of gravitation
and electromagnetism and perhaps also chromodynamics have been made by many great
past authors since the field concept itself was introduced by the highly original physicist,
M. Faraday in the 19
th
century. These attempts temporarily ended in the 1950s: in fact
2

Einsteins definitive version of his unified field theory as well as other parallel
constructions never comprehensively and compellingly shed new light on the relation
between gravity and electromagnetism. Indeed, they were biased by various possible
ways of constructing a unified field theory via different geometric approaches and
interpretations of the basic geometric quantities to represent the field tensors, e.g., the
electromagnetic field tensor in addition to the gravitational field tensor (for further
modern reference of such attempts, especially the last version of Einsteins
gravoelectrodynamics see, e.g., various works of S. Antoci). This is put nicely in the
words of Infeld: the problem of generalizing the theory of relativity cannot be solved
along a purely formal way. At first, one does not see how a choice can be made among
the various non-Riemannian geometries providing us with the gravitational and
Maxwells equations. The proper world-geometry which should lead to a unified theory
of gravitation and electricity can only be found by an investigation of its physical
content. In my view, one way to justify whether a unified field theory of gravitation and
electromagnetism is really true (comprehensive) or really refers to physical reality is to
see if one can derive the equation of motion of a charged particle, i.e., the (generalized)
Lorentz equation, if necessary, effortlessly or directly from the basic assumptions of the
theory. It is also important to be able to show that while gravity is in general non-linear,
electromagnetism is linear. At last, it is always our modest aim to prove that gravity and
electromagnetism derive from a common source. In view of this one must be able to
show that the electromagnetic field is the sole ingredient responsible for the creation of
matter which in turn generates a gravitational field. Hence the two fields are inseparable.
Furthermore, I find that most of the past theories were based on the Lagrangian
formulation which despite its versatility and flexibility may also cause some uneasiness
due to the often excessive freedom of choosing the field Lagrangian. This strictly formal
action-method looks like a short cut which does not lead along the direct route of true
physical progress. In the present work we shall follow a more fundamental (natural)
method and at the same time bring up again many useful classical ideas such as the
notion of a mixed geometry and Kaluza's cylinder condition and five-dimensional
3

formulation. Concerning higher-dimensional formulations of unified field theories, we
must remember that there is always a stage in physics at which direct but narrow physical
arguments can hardly impinge upon many hidden properties of Nature. In fact the use
of projective geometries also has deep physical reason and displays a certain degree of
freedom of creativity: in this sense science is an art, a creative art. But this should never
exclude the elements of mathematical simplicity so as to provide us with the very
conditions that Nature's manifest four-dimensional laws of physics seemingly take.
For instance, Kaluza's cylinder condition certainly meets such a requirement and
as far as we speak of the physical evidence (i.e., there should possibly be no intrusion of
a particular dependence upon the higher dimension(s)), such a notion must be regarded
as important if not necessary. An arbitrary affine (n + 1)-space can be represented by a
projective n-space. Such a pure higher-dimensional mathematical space should not
strictly be regarded as representing a real higher-dimensional world space. Physically
saying, in our case, the five-dimensional space only serves as a mathematical device to
represent the events of the ordinary four-dimensional space-time by a collection of
congruence curves. It in no way points to the factual, exact number of dimensions of the
Universe with respect to which the physical four-dimensional world is only a sub-space.
In this work we shall employ a five-dimensional background space simply for the sake of
convenience and simplicity.
On the microscopic scales, as we know, matter and space-time itself appear to be
discontinuous. Furthermore, matter arguably consists of molecules, atoms and smaller
elements. A physical theory based on a continuous field may well describe pieces of
matter which are so large in comparison with these elementary particles, but fail to
describe their behavior. This means that the motion of individual atoms and molecules
remains unexplained by physical theories other than quantum theory in which discrete
representations and a full concept of the so-called material wave are taken into account. I
am convinced, indeed, that if we had a sufficient knowledge of the behavior of matter in
the microcosmos, it would, and it should, be possible to calculate the way in which
matter behaves in the macrocosmos by utilizing certain appropriate statistical techniques
4

as in quantum mechanics. Unfortunately, such calculations prove to be extremely
difficult in practice and only the simplest systems can be studied this way. Whats more,
we still have to make a number of approximations to obtain some real results. Our
classical field theories alone can only deal with the behavior of elementary particles in
some average sense. Perhaps we must humbly admit that our understanding and
knowledge of the behavior of matter, as well as space-time which occupies it, is still in a
way almost entirely based on observations and experimental tests of their behavior on
the large scales. This is a matter for experimental determination but a theoretical
framework is always worth constructing. As generally accepted, at this point one must
abandon the concept of the continuous representation of physical fields which ignores
the discrete nature of both space-time and matter although it doesnt always treat matter
as uniformly distributed throughout the regions of space. Current research has centered
on quantum gravity since the departure of the 1950's but we must also acknowledge the
fact that a logically consistent unification of classical fields is still important. In fact we
do not touch upon the formal, i.e., standard construction of a quantum gravity theory
here. We derive a wave equation carrying the information of the quantum geometry of
the curved four-dimensional space-time in Section 4 by first assuming the discreteness of
the space-time manifold on the microscopic scales in order to represent the possible
inter-atomic spacings down to the order of Plancks characteristic length.
Einstein-Riemann space(-time)
4
R (a mixed, four-dimensional one) endowed with
an internal spin space
p
S is first considered. We stick to the concept of metricity and do
not depart considerably from affine-metric geometry. Later on, a five-dimensional
general background space IR
5
is introduced along with the five-dimensional and (as a
brief digression) six-dimensional sub-spaces
n
and
6
V as special coordinate systems.
Conventions: Small Latin indices run from 1 to 4. Capital Latin indices run from 1 to 5.
Round and square brackets on particular tensor indices indicate symmetric and skew-
symmetric characters, respectively. The covariant derivative is indicated either by a semi-
colon or the symbol . The ordinary partial derivative is indicated either by a comma or
the symbol . Einstein summation convention is, as usual, employed throughout this
work. Finally, by the word space we may also mean space-time.
5

2. Geometric construction of a mixed, metric-compatible four-
manifold
4
R possessing an internal spin space
p
S

Our four-dimensional manifold
4
R is endowed with a general asymmetric
connection
i
jk

and possesses a fundamental asymmetric tensor defined herein by

( )
] [ ) ( ] [ ) (
2
1
ij ij ij ij ij
g g + + =
(1)

where
) ( ) (
2
ij ij
g
will play the role of the usual geometric metric tensor with
which we raise and lower indices of tensors while
] [ ] [
2
ij ij
g will play the role of a
fundamental spin tensor (or of a skew- or anti-symmetric metric tensor). We shall also
refer both to as the fundamental tensors. They satisfy the relations

| |

=
=
=
) (
) (
) (
) (
] [
] [
) (
] [
) (
) (
c g g g g
b g g
a g g
jk
ij
jk
ij
k
i
kj
ij
k
i
kj
ij

(2)

We may construct the fundamental spin tensor as a generalization of the skew-
symmetric symplectic metric tensor in any M-dimensional space(-time), where M = 2,4,6,
, M = 2 + p, embedded in (M + n)-dimensional enveloping space(-time), where n = 0,
1,2, . In M dimensions, we can construct p = M - 2 null (possibly complex) normal
vectors (null n-legs)
p
z z z , ... , ,
2 1
with
0 =

n m
z z
( p n m , ... , 1 ,... , = and
M ..., , 1 ,... , = ). If we define the quantity

= where the skew-

symmetric, self-dual null bivector

defines a null rotation, then these null n-legs are

normal to the (hyper)plane
M n M
R
+ 2
(contained in
M
R ) defined in such a way that


( )
| | | |
( )
g det ,
1
,
0
... ! ...
... ...
... ...
| 1 | 2 1 2 1 ...
=

= =
=
+ + = =
g
g
g
z z z z p M z z z
p p

6

Here
...
is the Levi-Civita permutation symbol. Hence in four dimensions we have


0
2
1
2 1 2 1
2 1 2 1
2 1 2 1 2 1
= = =
=
= = =

z z z z tr
z z z z
z z z z z z



where

= is the generalized Kronecker delta and

= .
(The minus sign holds if the manifold is Lorentzian and vice versa.) The particular
equation
0 = tr
is of course valid in M dimensions as well. In M dimensions, the
fundamental spin tensor of our theory is defined as a bivector satisfying


| |
| |
( )

+ =
1
1
M
g g
(3a)

Hence the above relation leads to the identity


| |
| |

= g g
(3b)

In the particular case of M = 2 and n = 1, the

vanishes and the fundamental spin

tensor is none other than the two-dimensional Levi-Civita permutation tensor:


| |
| |
| |
| |
| | B
A
BC
AC
D
A
C
B
D
B
C
A
CD
AB
AB AB
g g
g g
g g

=
=
|
|

\
|

= =
0 1
1 0


where A,B = 1,2. Lets now return to our four-dimensional manifold
4
R . We now have


| |
| |
( )
| |
| | k
i
kj
ij
l
i
k
j
l
j
k
i
kl
ij
kl
ij
g g
g g

=
+ =
3
1


Hence the eigenvalue equation is arrived at:


| | | | kl
kl
ij ij
g g =
(3c)
7

We can now construct the symmetric traceless matrix
k
i
Q through


0
0
2 1 2 1
=
=
=

=
k
ik
ki ik
l
j
k
i
kl
ij
j
l s r
q p klrs
ijpq
j
l
kl
ij
k
i
u Q
tr
Q Q
u u Q
u u z z z z u u Q
Q

where
i
u is the unit velocity vector, 1 =
i
i
u u . Lets introduce the unit spin vector:


| |
| |
1
0 , 1
=
= =
=
k i
ik
i
i
i
i
k
ik i
u v g
v u v v
u g v


Multiplying both sides of (3c) by the unit spin tensor, we get


| | | |
r
kr
k
i
j
ij
u g Q u g =


In other words,


k i
k
i
v Q v = (3d)


Now we can also verify that


k
i
kj
ij
= (4)

Note that since the fundamental tensor is asymmetric it follows that


| |
( )
k
i
kj
ij
jk
ij
jk
ij
g g =
) (
(5)

The line-element of
4
R can then be given through the asymmetric fundamental tensor:


k i
ik
k i
ik
dx dx g dx dx ds
) (
2
2 = =
8

There exists in general no relation such as
k
i ji
kj
g g =
) (
] [
. However, we have the
relations

=
=
) (
) (
] [ ) ( ) (
] [
) ( ] [ ] [
) (
b g g g g
a g g g g
ij js ir
rs
ij js ir
rs
(6)

We now introduce the basis { }
l
g which spans the metric space of
4
R and its associate
{ }
l

which spans the spin space
p
S
4
R (we identify the manifold
4
R as having the
Lorentzian signature -2, i.e., it is a space-time). These bases satisfy the algebra


( ) ( )
( ) ( )
( )

=
= =
= =
=
=
) (
) (
) (
) (
) (
] [
) (
] [
] [
e g
d
c g
b g
a g
ij j i
i
j j
i
j
i
ij j i j i
k
ki i
k
ik i
g
g g
g g
g
g

(7)

We can derive all of (2) and (6) by means of (7). In a pseudo-five-dimensional space
n R
n
=
4
(a natural extension of
4
R which includes a microscopic fifth coordinate
axis normal to all the coordinate patches of
4
R ), the algebra is extended as follows:


| |
| |
| | ( )

)

= =
=
+ =
) ( ,
) ( ,
) ( ,
] [
] [
] [
c g
b g
a g C
j
ij i i
j
ij i
ij k
k
ij j i
g g n
g g n
n g g g

(8)

where the square brackets [ ] are the commutation operator,
k
ij
C stands for the
commutation functions and n is the unit normal vector to the manifold
4
R satisfying
1 ) ( = n n . Here we shall always assume that 1 ) ( + = n n anyway. In summary,
the symmetric and skew-symmetric metric tensors can be written in n R
n
=
4
as


| |
| | ) , (
) (
) (
n g g
g g
=
=
k i ik
k i ik
g
g
(8d),(8e)

9

The (intrinsic) curvature tensor of the space
4
R is given through the relations



( )
| | l i
l
jk l
l
ijk j k i k j i
i
ml
m
jk
i
mk
m
jl
i
l jk
i
k jl
i
jkl
a a R a a
R
; . ; ; ; ;
, , .
2 =
+ =


for an arbitrary vector
i
a
. The torsion tensor
| |
i
jk
is introduced through the relation


| | r
r
ik i k k i , ; ; ; ;
2 =


which holds for an arbitrary scalar field . The connection of course can be written as
| |
i
jk
i
jk
i
jk
+ =
) (
. The torsion tensor
| |
i
jk
together with the spin tensor
| | ij
g
shall play
the role associated with the internal spin of an object moving in space-time. On the
manifold
4
R , lets now turn our attention to the spin space
p
S and evaluate the tangent
component of the derivative of the spin basis
{ }
l

with the help of (7):

( ) ( )
l
k
mj
lm
ik
lk
j ik T i j
g g g g =
] [
] [
] [
, ] [
(9)

since ( )
k
k
ij
T
i j
g g = . Now with the help of (3), we have

( ) ( )
l
l
ij
l
i j
lk
j ik T i j
g g
|

\
|
=
3
1
] [
, ] [


where we have put
i
ij j
= . On the other hand one can easily show that by imposing
metricity upon the two fundamental tensors (use (7) to prove this), the following holds:


( ) ( )
k
i j
k
i j
i
jk
= = g g
(10)

Thus, solving for a tetrad-independent connection, we have


| |
| | k
i
j k jr
ir i
jk
g g =
2
1
2
3
,
(11)

10

The torsion tensor is therefore


( ) ( )
( )
| |
| | l jk
il i
j k
i
k j
j kl k jl
il i
j k
i
k j
i
jk
g g
g g g
,
, ] [ , ] [
] [
] [
4
3
4
1
4
3
4
1
+ =
+ =


(12)

where we have assumed that the fundamental spin tensor is a pure curl:


| | i j j i ij
g
, ,
= ,
| | | | | |
0
, , ,
= + +
j ki i jk k ij
g g g

This expression and the symmetric part of the connection:
( )
m
lj km
li m
lk jm
li
j kl l jk k lj
li i
jk
g g g g g g g g
] [ ) (
) (
] [ ) (
) (
, ) ( , ) ( , ) (
) (
) (
2
1
+ =


therefore determine the connection uniquely in terms of the fundamental tensors alone:

( ) ( )
| |
| | | |
( )
| |
| | | |
( )
| |
| | | |
( )
l js j ls
rs
kr
li
l ks k ls
rs
jr
li
j kr k jr
ir
r jk
ir
j
i
k j kl l jk k lj
li i
jk
g g g g g g g g g g
g g g g g g g g g
, , ) (
) (
, , ) (
) (
, , ) (
) (
, ) ( , ) ( , ) (
) (
4
3
4
3
4
3
2
1
2
1

+ + + =

(13)

There is, however, an alternative way of expressing the torsion tensor. The metric and
the fundamental spin tensors are treated as equally fundamental and satisfy the ansatz

0
; ] [ ; ) (
= =
k ij k ij
g g and 0
;
=
k ij
.

Therefore, from
0
; ] [
=
k ij
g
, we have the following:


m
jk im
m
ik mj k ij
g g g + =
] [ ] [ , ] [

11

Letting
m
ik jm jik
g W =
] [
, we have


jik ijk k ij
W W g =
, ] [
(14)

Solving for
] [ jk i
W
by making cyclic permutations of i, j and k, we get

( )
) ( ) ( , ] [ , ] [ , ] [ ] [
2
1
ij k ik j j ki i jk k ij jk i
W W g g g W + + =


Therefore

( )
) ( ) ( ) ( , ] [ , ] [ , ] [
] [ ) (
2
1
ij k ik j jk i j ki i jk k ij
jk i jk i ijk
W W W g g g
W W W
+ + + =
+ =


Now recall that
m
jk im ijk
g W =
] [
. Multiplying through by
| | il
g
, we get

( )
i
jk
m
lj km
li m
lk jm
li
j kl l jk k lj
li i
jk
g g g g g g g g
) ( ) ( ] [
] [
) ( ] [
] [
, ] [ , ] [ , ] [
] [
2
1
+ + + =
(15)

On the other hand,

( )
| |
i
jk
m
lj km
li m
lk jm
li
j kl l jk k lj
li i
jk
g g g g g g g g
] [ ] [ ) (
) (
) (
) (
, ) ( , ) ( , ) (
) (
2
1
+ + =

(16)

From (15) the torsion tensor is readily read off as

( )
m
lj km
li m
lk jm
li
j kl l jk k lj
li i
jk
g g g g g g g g
) ( ] [
] [
) ( ] [
] [
, ] [ , ] [ , ] [
] [
] [
2
1
+ + =

(17)

We denote the familiar symmetric Levi-Civita connection by


( )
( ) ( )
( )
j kl l jk k lj
li
g g g g
jk
i
, ) ( , ,
2
1
+ =
)
`

12

If we combine (13) and (17) with the help of (2), (3) and (4), after a rather lengthy but
straightforward calculation we may obtain a solution:

( )
( )
i
k
m
mj
i
j
m
mk km
li
jm
li
j kl l jk k lj
li i
jk
lj
m
g g
lk
m
g g
g g g g

] [ ] [ ] [
] [
] [
] [
, ] [ , ] [ , ] [
] [
] [
3
1
2
1
+
)
`

)
`

+
+ =
(18)

Now the spin vector
i
ik ] [
is to be determined from (12). If we contract (12) on the
indices i and j, we have

( )
k i kj k ij
ij i
ik
g g g =
4
3
4
3
, ] [ , ] [
] [
] [
(19)
But from (14):

k ij
ij
k ij
ij i
ik k
g g g g
, ] [
] [
, ) (
) (
2
1
2
1
= = =
(20)
Hence (19) becomes

| | i kj
ij
k ij
ij
k
i
ik
g g g g S
, ] [
] [
,
] [
] [
4
3
8
3
= (21)

Then with the help of (21), (18) reads

| |
( )
| | | |
( )
| |
| | | |
( )
i
k n jm
i
j n km
mn i
k j mn
i
j k mn
mn
km
li
jm
li
j kl l jk k lj
li i
jk
g g g g g g
lj
m
g g
lk
m
g g
g g g g

, , , ,
] [
] [
] [
] [
] [
, ] [ , ] [ , ] [
] [
4
1
8
1
2
1
+ +
)
`

)
`

+
+ =

(22)

So far, we have been able to express the torsion tensor, which shall generate physical
fields in our theory, in terms of the components of the fundamental tensor alone.
13

In a holonomic frame, 0
] [
=
i
jk
and, of course, we have from (16) the usual Levi-Civita
(or Christoffel) connection:


( )
j kl l jk k lj
li i
jk
g g g g
jk
i
, ) ( , ) ( , ) (
) (
2
1
+ =
)
`

=


If therefore a theory of gravity adopts this connection, one may argue that in a strict
sense, it does not admit an integral concept of internal spin in its description. Such is the
classical theory of general relativity.
In a rigid frame (constant metric) and in a pure electromagnetic gauge condition, one
may have 0
) (
=
i
jk
and in this special case we have from (15)


( )
j kl l jk k lj
li i
jk
g g g g
, ] [ , ] [ , ] [
] [
2
1
+ =


which is exactly the same in structure as
)
`

jk
i
with the fundamental spin tensor
replacing the metric tensor. We shall call this connection the pure spin connection,
denoted by

( )
j kl l jk k lj
li i
jk
g g g g L
, ] [ , ] [ , ] [
] [
2
1
+ =
(23)

Lets give an additional note to (20). Let's find the expression for
i
ki
, provided we
know that


( )
)
`

= =
=
= =
ik
i
g
g g
g g
k
k ij
ij
k ij
ij i
ik k
,
, ] [
] [
, ) (
) (
ln
2
1
2
1

14

Meanwhile, we express the following relations:


| |
| |
| |
( )
i kj
ij
k
i kj
ij
k ij
ij
i
ki
i
ik k
g g
g g g g
S
,
, ] [
] [
, ] [
] [
] [
4
3
4
3
8
3
=
=
= =


| |
| |
| | i kj
ij
k
i
ik k
m
li km
li m
lk im
li i
ik
g g
g g g g
ik
i
,
] [ ) (
) (
] [ ) (
) (
) (
4
3
4
1
+ =
=
)
`

=


Therefore


i
ik
i
ki
i
ki
i
ki
i
ki ] [ ) ( ] [ ) (
= + =

k i kj
ij
g g =
2
1
2
3
, ] [
] [


which can also be derived directly from (11). From (11) we also see that


| | | | | |
| | | |
r
jk ri
r
ik rj
k ij
r
ik rj k ij
g g
g g g
=
+ =
3
1
3
2
,
(24a),(24b)

Also, for later purposes, we derive the condition for the conservation of charges:


| | | |
k
ik ik
k
g g =
3
1
,
(24c)

Having developed the basic structural equations here, we shall see in the following
chapters that the gravitational and electromagnetic tensors are formed by means of the
fundamental tensors
) ( ) (
2
ij ij
g
and
] [ ] [
2
ij ij
g
alone (see Appendix B). In
other words, gravity and electromagnetism together arise from this single tensor. We
shall also investigate their fundamental relations and ultimately unveil their union.
15

3. Generalization of Kaluzas projective theory. Fundamental field
equations of our unified field theory

We now assume that the space-time
4
R is embedded in a general five-dimensional
Riemann space IR
5
. This is referred to as embedding of class 1. We shall later define the
space n R
n
=
4
to be a special coordinate system in IR
5
. The five-dimensional
metric tensor
) ( AB
g
of IR
5
of course satisfies the usual projective relations

=
+ =
0
) ( ) (
A
A
i
B A ij
j
B
i
A AB
n e
n n g e e g


where
A
i
A
i
x e = is the tetrad. If now { }
l
g denotes the basis of
4
R

and { }
A
e of IR
5 :

= =
= =
+ =
j
i
j
A
A
i B
A A
B
i
B
A
i
AB
B
j
A
i ij A
A
i i
A i
i
A A
e e n n e e
g e e g e
n e
,
,
) ( ) (
e g
n g e


The derivative of
i
g
4
R IR
5
is then

n g g
ij k
k
ij i j
+ =

We also have the following relations:


0 , , ,
) ( 0 , ,
; ) ( ; ) ( . ; . ; ;
; . ;
= = = = =
= = = =
C AB k ij
j
B
i
j A
i
B A
A
j
j
i
A
i
A
ij
A
j i
C
C
AB A B B A j
j
i i ij j i
g g e n e e n n e
e e e g n n g


In our work we shall, however, emphasize that the exterior curvature tensor
ij
is in
general asymmetric:
ji ij

just as the connection
i
jk

is. This is so since in general

A
j i
A
i j
e e . Within a boundary , the metric tensor
) (ij
g may possess
discontinuities in its second derivatives. Now the connection and exterior curvature
tensor satisfy

+ =
+ =
c
j
B
i
A
BC A
A
i j A ij
c
j
B
i
A
BC
k
A
A
i j
k
A
k
ij
e e n e n
e e e e e

(25a),(25b)
16

where


A
ij
C
j
B
i
A
BC
k
ij
A
k
A
i j
n e e e e + = (26)

We can also solve for
A
BC
in (25) with the help of the projective relation
n g e
A i
i
A A
n e + = . The result is, after a quite lengthy calculation,


B
j
C
A
i
i
j
A
B C
A j
C
i
B ij
j
C
i
B
k
ij
A
k
i
B C
A
i
A
BC
n e e n n
n e e e e e e e
.
) (

+
+ + =
(27)

If we now perform the calculation
( )
i k j j k
g
with the help of some of the
above relations, we have in general


( ) ( )
( )
A
A
i k j j k
A
B
i
A
BC
C
k j
C
j k A
D
k
C
j
B
i
A
BCD i k j j k
e
e e e e e e R
e
e e g
+
+ =
.
(28)

Here we have also used the fact that

( )
D
D
ABC A C B B C
R e e
.
=

On the other hand,
j
j
i i
g n
.
= , and


( )
( ) ( ) n g
n g g
lk
l
ij k ij l
l
k ij
l
mk
m
ij
l
k ij
k ij l
l
ij i j k

+ + + =
+ =
, . ,
,


Therefore we obtain another expression for ( )
i k j j k
g :

( ) ( ) ( ) n g g
il
l
jk j ik k ij l
l
k ij
l
j ik
l
ijk i k j j k
R
] [ ; ; . . .
2 + + + =

( ) ( )
A
ijk
B
i
A
BC
C
k j
C
j k
A
i k j j k
S e e e e
.
+
(29a),(29b)
17

Combining (28) and (29), we get, after some algebraic manipulations,


A
i Ajkl
D
l
C
k
B
j
A
i ABCD jk il jl ik ijkl
e S e e e e R R + =

(30a),(30b)

| |
A
Aijk il
l
jk
D
k
C
j
B
i
A
ABCD j ik k ij
n S e e e n R + = 2
; ;


We have thus established the straightforward generalizations of the equations of Gauss
and Codazzi.

Now the electromagnetic content of (30) can be seen as follows: first we split the
exterior curvature tensor
ij

into its symmetric and skew-symmetric parts:

ij ij ij ij
ij ij ij
f k
+ =
] [ ) (
] [ ) (
,

(31)

Here the symmetric exterior curvature tensor
ij
k has the explicit expression


( )
( )
A B B A
B
j
A
i
C
j
B
i
A
BC A
A
j i
A
i j A ij
n n e e
e e n e e n k
; ;
) (
2
1
2
1
+ =
+ + =
(32)

Furthermore, in our formalism, the skew-symmetric exterior curvature tensor
ij
f is
naturally equivalent to the electromagnetic field tensor
ij
F . It is convenient to set
ij ij
F f
2
1
=
. Hence the electromagnetic field tensor can be written as



( )
| |
( )
A B B A
B
j
A
i
C
j
B
i
A
BC A
A
j i
A
i j A ij
n n e e
e e n e e n F
; ;
2
=
+ =
(33)
18

The five-dimensional electromagnetic field tensor is therefore


A B B A AB
n n F =

We are now in a position to simplify (30) by invoking two conditions. The first of
these, following Kaluza, is the cylinder condition: the laws of physics in their four-
dimensional form shall not depend on the fifth coordinate
n
x
5
. We also assume that
y x
5
is a microscopic coordinate in
n
. In short, the cylinder condition is written as
(by first putting
A A
e n
5
= )


( )
( ) 0
; ; , 5 , ) (
= + = =
B C C B
C
j
B
i
A
A ij ij
n n e e n g g

where we have now assumed that in IR
5
the differential expression
i
A B
i
B A
e e
, ,

vanishes. However, from (26), we have the relation


| |
A
ij
k
ij
A
k
A
i j
A
j i
n F e e e + = 2
, ,


Furthermore, the cylinder condition implies that
0
; ;
= +
A B B A
n n
and therefore we
can nullify (32). This is often called the assumption of weakness. The second condition
is the condition of integrability imposed on arbitrary vector fields, e.g., on
i
(say) in
4
R . The necessary and sufficient condition for a vector field
i i

,
(a one-form) to be
integrable is
i j j i , ,
= . If this is applied to (28), we will then have
A
ijk
D
k
C
j
B
i
A
BCD
S e e e R
. .
= . Therefore (30) will now go into


( )
il
l
jk j ik k ij
jk il jl ik ijkl
F F F
F F F F R
] [ ; ;
2
4
1
=
=
(34a),(34b)

These are the sought unified field equations of gravity and electromagnetism. They form
the basic field equations of our unified field theory.
19

Altogether they imply that


) ( .
4
1
ik
j
k ij ik
R F F R = = (35a)

| |
0 =
ik
R
(35b)

ik
ik
F F R
4
1
=
(35c)

i
k
l
l
ik
j
ij
J F F =
. ] [
;
2
(35d)

These relations are necessary and sufficient following the two conditions we have
dealt with. These field equations seem to satisfy a definite need. They tell us a beautiful
and simple relation between gravity and electromagnetism: (34a) tells us that both inside
and outside charges, a gravitational field originates in a non-null electromagnetic field (as
in Rainich's geometry), since according to (34b), the electromagnetic current is produced
by the torsion of space-time: the torsion produces an electromagnetic source. The
electromagnetic current is generated by dynamic electric-magnetic charges. In a strict
sense, the gravitational field cannot exist without the electromagnetic field. Hence all
matter in the Universe may have an electromagnetic origin. Denoting by d a three-
dimensional infinitesimal boundary enclosing several charges, we have, from (35d)



| |

= d F u e
k
r
i r
ik .
2

We may represent a negative charge by a negative spin produced by a left-handed
twist (torsion) and a positive one by a positive spin produced by a right-handed twist.
(For the conservation of charges (currents) see Appendices C and D.) Now (35c) tells us
that when the spatial curvature, represented by the Ricci scalar, vanishes, we have a null
electromagnetic field, also it is seen that the strength of the electromagnetic field is
equivalent to the spatial curvature. Therefore gravity and electromagnetism are
inseparable. The electromagnetic source, the charge, looks like a microscopic spinning
hole in the structure of the space-time
4
R , however, the Schwarzschild singularity is
non-existent in general. Consequently, outside charges our field equations read


( )
0
4
1
;
=
=
ij
j
jk il jl ik ijkl
F
F F F F R
(36a),(36b)

which, again, give us a picture of how a gravitational field emerges (outside charges).
20

In this way, the standard action integral of our theory may take the form


( ) ( )
( ) x d g F F R R R F F R
x d g R R R I
jk il
ijkl
ijkl
ijkl
ik
ik
ik
ik
4
2 / 1
2
*
4
2 / 1
*
4
1
16
1

|
|

\
|
|

\
|
=
=

(37)

Here R
*
denotes the Ricci scalar built from the symmetric Christoffel connection alone.
From the variation of which, we would arrive at the standard Einstein-Maxwell
equations. However, we do not wish to stress heavy emphasis upon such an action-
method (which seems like a forced short cut) in order to arrive at the field equations of
our unified field theory. We must emphasize that the equations (34)-(37) tell us how the
electromagnetic field is incorporated into the gravitational field in a very natural manner,
in other words theres no need here to construct any Lagrangian density of such. We
have been led into thinking of how to couple both fields using different procedures
without realizing that these fields already encapsulate each other in Nature. But here our
space-time is already a polarized continuum in the sense that there exists an
electromagnetic field at every point of it which in turn generates a gravitational field.

__________________________________________________________________________________
Remark 1
Without the integrability condition we have, in fairly general conditions, the relation

( ) ( )
( )
| | | |
( )
B
i
C
jk
C
l
l
jk
A
BC
A
l
l
k ij
l
j ik
A
il
l
jk j ik k ij
D
k
C
j
B
i
A
BCD
A
l
l
ijk
A
i k j j k
e n e e
n e e e R e R e


+ +
+ + + =
2
2
. .
] [ ; ; . .
(a)


( ) ( )
( ) ( )
B
i
C
k j
C
j k
A
BC
A
i k j j k
B
i
C
jk
C
l
l
jk
A
BC
A
i k j j k
A
ijk
e e e e
e n e e S
+ =
+ +
] [ ] [ .
2


Hence we have


A
i Ajkl
D
l
C
k
B
j
A
i ABCD jk il jl ik ijkl
e S e e e e R R + = (b1)

| | il
l
jk
A
Aijk
D
k
C
j
B
i
A
ABCD j ik k ij
n S e e e n R + = 2
; ;
(b2)
21

These are just the equations in (30). Upon employing a suitable cylinder condition and
putting
| | ij ij
F
2
1
=
(within suitable units), we have the complete set of field equations
of gravoelectrodynamics:

( )
A
i Ajkl
D
l
C
k
B
j
A
i ABCD jk il jl ik ijkl
e S e e e e R F F F F R + =
4
1
(c1)

( )
| | il
l
jk
A
Aijk
D
k
C
j
B
i
A
ABCD j ik k ij
F n S e e e n R F F + =
; ;
2
1
(c2)


--------------------------------------------------------------------------------------------------------------
Sub-remark: Lets consider the space Y R S =
4 5
which describes a five-
dimensional thin shell where Y is the microscopic coordinate representation spanned
by the unit normal vector to the four-manifold
4
R . The coordinates of this space are
characterized by ( ) y x y
i
, =

where the Greek indices run from the 1 to 5 and where

the extra coordinate y is taken to be the Planck length:


3
c
G
y
h
=

which gives the thickness of thin shell. Here G is the gravitational constant of
Newton, h is the Planck constant divided by 2 and c is the speed of light in vacuum.
(From now on, since the Planck length is extremely tiny, we may drop any higher-order
terms in y .) Then the basis { }

of the space
5
S can in general be split into


( )
n
g
=
=
5
.


k
k
i
k
i i
y

22

It is seen that the metric tensor of the space
5
S , i.e.,

, has the following non-zero

components:


1
2
55
) ( ) (
=
=

ik ik ik
y g


The simplest sub-space of the space
5
S is given by the basis


( )
n g
g
=
=
5
0 ,
i
i i
x


where
{ }
l
g
is of course the tangent basis of the manifold
4
R . We shall denote this
pseudo-five-dimensional space as the special coordinate system n R
n
=
4
whose
metric tensor

g can be arrayed as


| |
|
|

\
|
=
1 0
0
4 4
) (
x
ik
g
g

Now the tetrad of the space
5
S is then given by ( )

=
A A
e , which can be split
into


A A
A
k
k
i
A
i
A
i
n
e y e
=
=
5
.

Then we may find the inverse to the tetrad
A
i
as follows:


A A A
k
A
i
k
i
A
i
A
y n
e y e
,
5
.
= =
+ =

23

From the above relations, we have the following:


i
k
i
k
A
k
i
A
i
k
i
k
i
A
A
k
i
k
A
k
i
A
y e
y e
.
.



+ =
=
=


The five-dimensional index of the tetrad
A
i
is raised and lowered using the metric
tensor
) ( AB
g . The inverse of the tetrad is achieved with the help of the metric tensor


k
B
i
A ik B A ik
k
B
i
A AB
e e y n n g e e G
) ( ) (
2 + =

which reduces to
) ( AB
g due to the cylinder condition. The four-dimensional metric
tensor
) (ik
g is used to raise and lower the four-dimensional index. Again, we bring in the
electromagnetic field tensor
ij
F
via the cylinder condition, which yields
ik ik
F
2
1
= .
The connection of the space
5
S is then


( )

=
. , . , . , .
2
1
2
1
2
1
2
1
, F y F y y F y F y x
i


where

is the connection of the space n R

n
=
4
with the electromagnetic field
tensor derived from it:
0 , 2 , 2
5 5 .
5
= = =

F F F
. Therefore its
only non-zero components are
ik
F . At the base of the space
5
S , the connection is

( )

, . , .
2
1
2
1
0 , y F y F x
i
= =

24

From the above expression, we see that

. 5 . 5
, .
,
2
1
2
1
F F
y F
k k k
= =
=


As can be worked out, the five-dimensional connection of the background space IR
5
is
related to that of
5
S through


( )
i
C
k
i B k
A
C B
A
C B
A
C B
A i A
BC
e F y
n F y x
.
. ,
2
1
2
1
,

+ =


Now the five-dimensional curvature tensor ( ) 0 ,
i
ABCD ABCD
x R R = is to be related
once again to the four-dimensional curvature tensor of
4
R , which can be directly derived
from the curvature tensor of the space
5
S as ( ) 0 ,
5
i
ijkl
S
ijkl
x R R = . With the help of
the above geometric objects, and after some laborious work-out, we arrive at the relation


ABCD CD AB D C B A ABCD
F F R e e e e R

+ =
2
1


where we have a new geometric object constructed from the electromagnetic field
tensor:
( ) ( )
( ) ( ) ( ) ( )
A
BCD
A
BCD
A B C C B C B A B C A
A B C C B A B C C B ABC
e
e e n e n F n x n x F
e e n e n F e n e n e F

. .
. , , .
, . . .
2
1
0 , 0 ,
2
1
2
1
2
1

+ +
+ =

25

Define another curvature tensor:


( )
D
l
C
k
B
j
A
i ABCD ABCD ijkl
e e e e R


Then we have the relation


kl ij ijkl ijkl
F F R
2
1
+ =

By the way, the curvature tensor of the space n R
n
=
4
is here given by

( ) ( )

+ = 0 , 0 ,
, , .
x x R

Expanding the connections in the above relation, we obtain

( ) ( )
( ) ( )
( ) ( )
( )

, , . . . .
, , . , . . , , . , . .
, . . , . .
, , . , , . , .
, , .
4
1
4
1
4
1
2
1
2
1
2
1
2
1
y y F F F F
y y F y F F y y F y F F
y F F y F F
y y F y F F
R
+
+ +
+ +
+ +
+ =


We therefore see that the electromagnetic field tensor is also present in the curvature
tensor of the space
n R
n
=
4

. In other words, electromagnetic and gravitational

interactions are described together on an equal footing by this single curvature tensor.
26

Direct calculation shows that some of its four-dimensional and mixed components are


| |
( )
0
2
2
1
5
; ; 5
5
.
, , .
=
+ = =
+ =
k ij
ir
r
jk j ik k ij ijk ijk
i
rl
r
jk
i
rk
r
jl
i
l jk
i
k jl
i
jkl
R
F F F R R
R


Furthermore, we obtain the following equivalent expressions:

( )
( )
ABCD CD AB
l
D
k
C
j
B A jm
m
lk l jk k jl
l
D
k
C B
i
A im
m
kl k il l ik
l
D
k
C
j
B
i
A ijkl ABCD
F F e e e n F F F
e e n e F F F e e e e R R
+ + +
+ + =
2
1
2
2
1
2
2
1
] [ ; ;
] [ ; ;

( ) ( )
l
D
k
C
j
B A jm
m
lk
l
D
k
C B
i
A im
m
kl
l
D
k
C
j
B
i
A ijkl
ABCD CD AB A D BC C BD B C AD D AC ABCD
e e e n F e e n e F e e e e R
F F n F F n F F R
] [ ] [
; ; ; ;
2 2
2
1
2
1
2
1
+ + +
+ =
where

( ) ( ) ( )
ABCD A C
M
D D
M
C MB B C
M
D D
M
C MA ABCD
n n F n F F n n F n F F =
. . . .
2
1

When the torsion tensor of the space
4
R vanishes, we have the relation

( ) ( )
l
D
k
C
j
B
i
A ijkl
ABCD CD AB A D BC C BD B C AD D AC ABCD
e e e e R
F F n F F n F F R
+
+ =
2
1
2
1
2
1
; ; ; ;

which, again, relates the curvature tensors to the electromagnetic field tensor.

27

Finally, if we define yet another five-dimensional curvature tensor:


ABCD CD AB ABCD ABCD
F F R
~
~ ~
2
1 ~
+

where
ABCD
R
~
,
AB
F
~
and
ABCD

~
are the extensions of
ABCD
R ,
AB
F and
ABCD
which
are dependent on y , we may obtain the relation


ijkr
r
l
ijrl
r
k irkl
r
j rjkl
r
i ijkl
D
l
C
k
B
j
A
i ABCD
R F y
R F y R F y R F y R e e e e
.
. . .
2
1
2
1
2
1
2
1
+
+ + + =


--------------------------------------------------------------------------------------------------------------
Lets now write the field equations of our unified field theory as


( )
| | ijk il
l
jk j ik k ij
ijkl jk il jl ik ijkl
F F F
F F F F R

+ =
+ =
2
4
1
; ;
(d)
1,2


( )
A
Aijk
D
k
C
j
B
i ABCD ijk
A
i Ajkl
D
l
C
k
B
j
A
i ABCD ijkl
n S e e e R
e S e e e e R
=
=
2

(e)
1,2

Consider the invariance of the curvature tensor under the gauge transformation


k
i
j
i
jk
i
jk ,
' + = (f)

for some function ) (x = . This is analogous to the gauge transformation of the
electromagnetic potential, i.e.,
i i i ,
' + =
, with a scaling constant ,
which leaves the electromagnetic field tensor invariant.. We define the electromagnetic
potential vector
i
and pseudo-vector
i
via
i i
k
ki
+ = where is a constant.

28

Then we see that the electromagnetic field tensor can be expressed as


( )
i k k i i k k i ik
F
, , , ,
1

= =
(g)

More specifically, the two possible electromagnetic potentials
i
and
i
transform
homogeneously and inhomogeneously, respectively, according to


C
i
B A
BC A
A
i k
k
A A
A
i i
A
A
i i
e n n e e e
e
+ =
=
,




The two potentials become equivalent in a coordinate system where g equals a
constant. Following (g), we can express the curvature tensor as


( ) ( )
) ( ) ( ) ( ) ( 2 1 jk il jl ik jk il jl ik ijkl
g g g g F F F F R + =
(h)

where
1

and
2

are invariants. (The term

kl ij
F F
0

would contribute nothing.)

Hence


) ( 2 . 1
3
ik
l
k il ik
g F F R + =
(i)
Putting
4
1
1
=
in accordance with (d)
1
and contracting (i) on the indices i and k we see
that
ik
ik
F F R
48
1
12
1
2
=
. Consequently, we have the important relations


( ) ( )
( )
rs
rs jk il jl ik
jk il jl ik jk il jl ik ijkl
F F g g g g
R g g g g F F F F R
) ( ) ( ) ( ) (
) ( ) ( ) ( ) (
48
1
12
1
4
1

+ =
(j)
29


rs
rs ik ik
l
k il ik
F F g R g F F R
) ( ) ( .
16
1
4
1
4
1
+ =
(k)

Comparing (j) and (e)
1
we find


( ) ( )
rs
rs jk il jl ik jk il jl ik
A
i Ajkl
D
l
C
k
B
j
A
i ABCD ijkl
F F g g g g R g g g g
e S e e e e R
) ( ) ( ) ( ) ( ) ( ) ( ) ( ) (
48
1
12
1
=
=

(l)
Hence also


rs
rs
rs
rs ik ik ik
F F R
F F g R g
4
1
16
1
4
1
) ( ) (
=
=

(m)
1,2

Note that our above consideration produces the following traceless field equation:


|

\
|
=
rs
rs ik
l
k il ik ik
F F g F F R g R
) ( . ) (
4
1
4
1
4
1
(n)

In a somewhat particular case, we may set ( )
ik k i ik
t u u c R g + =
2
) (
4
1
where
is a coupling constant and
ik
t
is the generalized stress-metric tensor, such that
e
R =
, where now t c
e
+ =
2
is the effective material density. We also have

( )
( )
|

\
|
+ + =
rs
rs ik
l
k il ik k i ik ik
F F g F F t u u c R g R
4
1
4
1
2
1
.
2
) (

(o)
30

The above looks slightly different from the standard field equation of general relativity:


( )
|

\
|
|

\
|
+ =
rs
rs ik
l
k il ik k i ik ik
F F g F F t u u c k R g R
4
1
2
1
.
2 *
) (
*
(p)

which is usually obtained by summing altogether the matter and electromagnetic terms.
Here
ik
R
*
and R
*
are the Ricci tensor and scalar built out of the Christoffel connection
and k is the usual coupling constant of general relativity. Lets denote by
,
0
m and c , the point-mass, material density and speed of light in vacuum. Then the
vanishing of the divergence of (p) leads to the equation of motion for a charged particle:


k i
k
k i
k
i
u F
c m
e
u u
Ds
Du
.
2
0
;
=

However, this does not provide a real hint to the supposedly missing link between
matter and electromagnetism. We hope that theres no need to add an external matter
term to the stress-energy tensor. We may interpret (n), (o) and (p) as telling us that
matter and electromagnetism are already incorporated, in other words, the
electromagnetic field produces material density out of the electromagnetic current J . In
fact these are all acceptable field equations. Now, for instance, we have
( ) t u J c R
i
i
e
+ = =
2
. From (m)
2
, the classical variation follows:


0
4
1
4
4
=
|

\
|
=
=

x d g F F R
x d g I
ik
ik

(q)

which yields the gravitational and electromagnetic equations of Einstein and Maxwell
endowed with source since the curvature scalar here contains torsion as well.
31

Finally, lets investigate the explicit relation between the Weyl tensor and the
electromagnetic field tensor in this theory. In four dimensions the Weyl tensor is


( )
( ) R g g g g
R g R g R g R g R C
jk il jl ik
il jk jk il ik jl jl ik ijkl ijkl
) ( ) ( ) ( ) (
) ( ) ( ) ( ) (
6
1
2
1
+
+ =



Comparing the above equation(s) with (j) and (k), we have

( ) ( )
( )
r
l ir jk
r
k jr il
r
k ir jl
r
l jr ik
rs
rs jk il jl ik jk il jl ik ijkl
F F g F F g F F g F F g
F F g g g g F F F F C
. ) ( . ) ( . ) ( . ) (
) ( ) ( ) ( ) (
8
1
24
1
4
1
+
+ =
(r)

We see that the Weyl tensor is composed solely of the electromagnetic field tensor in
addition to the metric tensor. Hence we come to the conclusion that the space-time
4
R
is conformally flat if and only if the electromagnetic field tensor vanishes. This agrees
with the fact that, when treating gravitation and electromagnetism separately, it is the
Weyl tensor, rather than the Riemann tensor, which is compatible with the
electromagnetic field tensor. From the structure of the Weyl tensor as revealed by (r), it
is understood that the Weyl tensor actually plays the role of an electromagnetic
polarization tensor in the space-time
4
R . In an empty region of the space-time
4
R with
a vanishing torsion tensor, when the Weyl tensor vanishes, that region possesses a
constant sectional curvature which conventionally corresponds to a constant energy
density.

__________________________________________________________________________________
Let's for a moment turn back to (30). We shall show how to get the source-torsion
relation, i.e., (35d) in a different way. For this purpose we also set a constraint
0
.
=
A
ijk
S
and assume that the background five-dimensional space is an Einstein space:


) ( AB AB
g R =


where is a cosmological constant. Whenever 0 = we say that the space is Ricci-
flat or energy-free, devoid of matter. Taking into account the cosmological constant, this
consideration therefore takes on a slightly different path than our previous one. We only
wish to see what sort of field equations it will produce.
32

We first write


( )
il
l
jk
D
k
C
j
B
i
A
ABCD j ik k ij
D
l
C
k
B
j
A
i ABCD jk il jl ik ijkl
F e e e n R F F
e e e e R F F F F R
] [ ; ;
2 2
4
1
=
+ =


Define a symmetric tensor:


ki
D C
k
B A
i ABCD ik
B n e n e R B = (38)

It is immediately seen that


0
) (
=
=
=
B A
i AB
B A
AB
ik
B
k
A
i AB
n e R
n n R
g e e R
(39a),(39b),(39c)

Therefore


ik ik
l
k il
D C
k
B A
i ABCD
B
k
A
i AB
l
k il ik
B g F F
n e n e R e e R F F R
+ + =
+ =
) ( .
.
4
1
4
1
(40)

From (38) we also have, with the help of (39b), the following:


( )
=
+ =
= =
D C B A
ABCD
B A
AB
C A AC D B
ABCD ik
ik
n n n n R n n R
n n g n n R B g B
) ( ) (

Hence we have

+ = 3
4
1
ik
ik
F F R
(41)

33

The Einstein tensor (or rather, the generalized Einstein tensor endowed with torsion)


R g R G
ik ik ik ) (
2
1



up to this point is therefore


l
k il ik
rs
rs ik ik ik
F F g F F g B G
. ) ( ) (
4
1
2
1
8
1
= (42)

From the relation


il
l
jk
D
k
C
j
B
i
A
ABCD j ik k ij
F e e e n R F F
] [ ; ;
2 2 =


we see that


j kl j
k l
kl j
k l
D j
C B
A BC
D A
A j
C
C
A
kl j
k l
D
k
j
C
k
B
A BC
D A
kj
k
J F
F n e n n R n e R
F e e e n R F
= =
=
=
] [.
] [. . . .
] [. . . ;
2
2 2 2
2 2


In other words,


k
l
l
ik i
F J
. ] [
2 =


which is just (35d). We will leave this consideration here and commit ourselves to the
field equations given by (34) and (35) for the rest of our work.
Lets obtain the (generalized) Bianchi identity with the help of (34) and (35).
Recall once again that


( )
il
l
jk j ik k ij
jk il jl ik ijkl
F F F
F F F F R
] [ ; ;
2
4
1
=
=

34

Performing the covariant derivative on
ijkl
R
, the result of the cyclic summation over the
indices k, l and m is


( ) ( )
( ) ( )
( ) ( )
im k jl l jk jk m il l im
il m jk k jm jm l ik k il
ik l jm m jl jl k im m ik l ijmk k ijlm m ijkl
F F F F F F
F F F F F F
F F F F F F R R R
; ; ; ;
; ; ; ;
; ; ; ; ; ; ;
4
1
4
1
4
1
4
1
4
1
4
1
+ +
+ +
+ = + +

(43)
Or equivalently,


| | | | | |
| | | | | | im jn
n
lk jk in
n
lm il jn
n
km
jm in
n
kl ik jn
n
ml jl in
n
mk l ijmk k ijlm m ijkl
F F F F F F
F F F F F F R R R
+ + +
+ + = + +
2
1
2
1
2
1
2
1
2
1
2
1
; ; ;

(44)

From (43) if we raise the index i and then perform a contraction with respect to the
indices i and k , we have

i jl
i
m l ji
i
m
i
m l ji
i
l m ji
m ji
i
l i jm
i
l jm l jl m l jm
i
i jml m jl
F F F F F F F F
F F F F F J F J R R R
; . ; . ; . ; .
; . ; . ; ; . ;
4
1
4
1
4
1
4
1
4
1
4
1
4
1
4
1
+ +
+ =

If we raise the index j and then contract on the indices j and l, we have the expression


( )
( )
l
m i
l
i m
i
l im
i
i
i
m
i
m
i
i m m
F F F F J
R R R R
; . ; . .
;
. ; . ;
2
1
2
1
2 2
+ =
=

35

Therefore we have


( )
m im
i
l
i m
l
m i
i
l im
i
i
i
m
i
m
F J
F F F F J R R

+ =
+ =
|

\
|

4
1
4
1
4
1
2
1
; . ; . .
;
.
(45)

where


( )
| |
l
k
k
im
i
l
l
i m
l
m i
i
l m
F F F F F
. . ; . ; . .
2
1
4
1
= =


We can also write


i ik
k
k
ik ik
F J R g R + =
|

\
|

4
1
2
1
;
) (
(46)

On the other hand, repeating the same contraction steps on (44) gives

( )
m
l
m
k
n
n
kl
k
l
l
n
n
km
l
m
k
n
n
kl
k
m
l
n
n
lk
l
k
k
n
n
lm
k
l
l
n
n
km
l
m
k
n
n
kl
m
i
i
m
i
m
i
i m m
F F
F F F F
F F F F F F F F
g R R R R

2
2
1
2
1
2
1
2
1
2 2
. . ] [
. . ] [ . . ] [
. . ] [ . . ] [ . . ] [ . . ] [
;
. ; . ;
=
+ =
+ + + =
=
Therefore we have


i
k
ik ik
g R g R
2
1
2
1
;
) (
=
|

\
|

(47a)

where
i
g , a non-linear quantity, can be seen as a complementary torsional current:

i li k
n
n
kl
i ik
k
i
F F F J g 2
2
1
. ] [
= + = (47b)
36

In the most general case, by the way, the Ricci tensor is asymmetric. If we proceed
further, the generalized Bianchi identity and its contracted form will be given by


| | | | | |
( )
ijrl
r
mk ijrk
r
lm ijrm
r
kl l ijmk k ijlm m ijkl
R R R R R R + + = + + 2
; ; ;
(48a)


| | | |
ijk
r
r
ij
j
r
r
ji
ik
i
ik ik
R R g R g R
. . . .
) (
;
) (
2
2
1
+ =
|

\
|

(48b)


Remark 2
Consider a uniform charge density. Again, our resulting field equation (45) reads


( ) ( )
s
r k
s
k r
r
s
ik k i
k
k
ik ik
F F F g J F R g R
; . ; . .
) (
.
;
) (
4
1
4
1
2
1
+ =
|

\
|



where we have set
( ) ( )
s
r k
s
k r
r
s
ik i
F F F g
; . ; . .
) (
4
1
=
. Recall the Lorentz equation of
motion:
k i
k
i
u F e
Ds
Du
c m
.
2
0
= . Setting
|

\
|
=
e

4
1
, we obtain


|
|

\
|

\
|
=
i
k
ik ik
i
R g R
Ds
Du
c m

;
) ( 2
0
2
1 1

or

| |
|
|

\
|

|

\
|
=
r
s
s
rk
ik
k
ik ik
i
R g R g R
Ds
Du
c m
.
) (
;
) ( 2
0
2
2
1 1

In the absence of charge density (when the torsion tensor is zero), i.e., in the limit
, we get the usual geodesic equation of motion of general relativity:


0
2
2
=
)
`

+
ds
dx
ds
dx
jk
i
ds
x d
k j i

37

The field equation can always be brought into the form


|

\
|
= T g T k R
ik ik ik ) (
2
1


Here the energy-momentum tensor need not always vanish outside the world-tube and
in general we can write
i ik
k
g T
2
1
;
=
. Now the Einstein tensor is


ik
rs
rs ik
l
k il ik ik ik
T F F g F F R g R G = = =
) ( . ) (
8
1
4
1
2
1

The right-hand side stands more appropriately as the field strength rather than the
classical conservative source term as
ik
ik
F F T
2
1
=
(again, is a coupling
constant). The equation of gravoelectrodynamics can immediately be written in the form


( ) ( )
( )
rs
rs ik
l
k il ik
ik
rs
rs ik
l
k il ik ik
F F g F F T
T F F g F F R g R
4
1
4
1
4
1
4
1
.
.
=
=
|

\
|
=


as expected. In this field equation, as can be seen, the material density arises directly
from electromagnetic interaction.
__________________________________________________________________________________
The (sub-)spaces n R
n
=
4
and m V
n
=
6
. The
n
covariant
derivative
We now consider the space n R
n
=
4
IR
5
, a sub-space of IR
5
with basis
{ }
A
satisfying ( ) n g ,
i A
= . Therefore this basis spans a special coordinate system in
IR
5
. We define the
n
covariant derivative to be a projective derivative which acts
upon an arbitrary vector field of the form ( ) ,
i
= or, more generally, upon an
38

arbitrary tensor field of the form
|
|

\
|
=
) (
...
...
) 1 (
...
...
) 2 (
...
...
) 1 (
...
...
...
...
, , ... , , ,
N
ij
l
N
ij
k
j
kl
i
kl
ij
kl
t t t t T T
, where
... 5
...
) 2 (
...
...
... 5
...
) 1 (
...
...
,
j
kl
j
kl
i
kl
i
kl
T t T t = =
etc., and which is projected onto the four-dimensional
physical space-time
4
R . For instance, for n g + =
i
i
then


n g
g n n g
n n g g
|

\
|
+ +
|

\
|
=
+ + =
+ + + =
i
ij j i
i
j
i
j
i
i
j j ij
i
i
i
j
j j j i
i
i
i
j j
F F
F F



2
1
2
1
2
1
2
1
; . ;
. ; ;
; ; ; ;



The projection of this onto
4
R manifests in

i
j
i
j
i
j
F
. ;
2
1
=

The stroke | represents the
n
covariant derivative which takes up the notion of
cylindricity. When this is applied, for instance, to a second rank tensor field T, we have
i
k
j j
k
i ij
k
ij
k
F t F t T T
.
) 2 (
.
) 1 (
;
2
1
2
1
=
. For a tensor of arbitrary rank, we therefore have


.. .
2
1
2
1
...
2
1
2
1
) (
...
...
) 1 (
...
...
.
) 2 (
...
... .
) 1 (
...
...
...
; ...
...
| ...

=

km
N
ij
l lm
N
ij
k
i
m
j
kl
j
m
i
kl
ij
m kl
ij
m kl
F t F t
F t F t T T
(49)

So the electromagnetic field extends the covariant derivative. To pass on from general
relativity to our unified field theory we merely need to replace ordinary (horizontal)
covariant derivatives with
n
covariant derivatives. Recall that the equation of geodesic
motion in general relativity is
0
;
=
j i
j
u u
where ds dx u
i i
/ = is of course the unit
velocity vector. Generalizing by letting ) , ( =
i
u u and setting
0
| ;
=
j i
j
j i
j
u u u u
, we have
0
2
1
. ; |
=
|

\
|
=
j i
j
i
j
j i
j
u F u u u
.
39

Setting
|
|

\
|
=
2
0
2
c m
e
where, again,
e
is the electric charge,
0
m is the point-mass
and c is the speed of light in vacuum, we have the Lorentz equation of motion:


j i
j
i
j i
j
i
u u
Ds
Du
u F
c m
e
Ds
Du
; .
2
0
, =

The fifth component of the momentum is consequently given by


|
|

\
|
=
2
5
2
c
e
p
(50)

We will now show that the quantity is indeed constant along the world-line. First we
write
n g u + =
i
i
u
, then, as before, we have



n g u
|

\
|
+ +
|

\
|
=
i
ij j i
i
j
i
j j
u F F u
2
1
2
1
; . ;


Applying the law of parallel transport in
n R
n
=
4

, i.e., 0
4
= u
u
(in the
direction of u
4
) where u
4
represents the ordinary tangent four-velocity field, we get
two equations of motion:


0
2
1
0
2
1
,
. ;
=
|

\
|
+
=
|

\
|

j i
ij j
j i
j
i
j
u u F
u F u
(51a),(51b)

The first of these is just the usual Lorentz equation in a general coordinate system, the
one we've just obtained before using the straightforward notion of vertical-horizontal

n
covariant derivative. Meanwhile, the second reads, due to the vanishing of its
second term:
0 =

ds
d
, which establishes the constancy of with respect to the world-
40

line. Therefore it is justified that forms a fundamental constant of Nature in the sense
of a correct parameterization. We also have
0
0
= =
ds
m d
ds
e d
. Again, theres a certain
possibility for the electric charge and the mass, to vary with time, perhaps slowly in
reality. We shall now consider the unit spin vector field in the spin space
p
S :


k i
ik i
i
u g u g v
] [
4
= (52)

which has been defined in Section 2. This spin (rotation) vector is analogous to the
ordinary velocity vector in the spin space representation. For the moment, let
( )
k
ik i i
u g v v
] [
; , = = v
where n v + =
i
i
v . Therefore we see that


n
n n
n n g g v
|

\
|
+ +
|

\
|
=
+ + + =
+ + + =
k l
j kl j i kj
ik i
j
j j j i
i
i
i
j
j j j i
i
i
i
j j
v F g F g v
v v
u u
. ] [ ,
] [
;
; ; ; ;
; ; ; ;
2
1
2
1






(53)

with the help of (7). If the law of parallel transport
0
4
= u
u
applies for the velocity
field u , it is intuitive that in the same manner it must also apply to the spin field v :

0
4
= v
u
(54)

This states that spin is geometrically conserved. We then get


0
2
1
0
2
1
. ] [ ,
] [
;
=
|

\
|
+
=
|

\
|

j k l
j kl j
j
kj
ik i
j
u v F g
u F g v

(55a),(55b)

which are completely equivalent to the equations of motion in (51a) and (51b).

41

Lets also observe that


ik
k
ik
k
G G
; |
=
(56a)


j ki i jk k ij j ki i jk k ij
F F F F F F
; ; ; | | |
+ + = + +
(56b)

i.e., the vertical-horizontal
n
covariant derivative operator when applied to the
Einstein tensor and the electromagnetic field tensor equals the ordinary covariant
derivative operator. We shall be able to prove this statement. First
i
j
k k
j
i ik
j
ik
j
ik
j
F Y F X G G G
. . ; | ;
2
1
2
1
=
where
i i i
Y G X = =
5
(due to the
symmetry of the tensor
ik
G ,
5 5 i i
R G =
), so that
i
k
k ik
k
ik
k
F Y G G
. ; |
2
1
=
. Now the
five-dimensional curvature tensors (the Riemann and Ricci tensors) in
n
are


C
EB
E
AC
C
EC
E
AB
C
B AC
C
C AB AB
A
ED
E
BC
A
EC
E
BD
A
D BC
A
C BD
A
BCD
R
R
+ =
+ =
, ,
, , .


In this special coordinate system we have

( )
( )
( )
k
k
j
j j
j
k
k
k
k k
k
k
i k
k
i i i
k
i j
j
i
k
i
k k
i
F F
F
F F
5 . . 5
55
5
5 .
5
5
. . 5
0
2
1
2
1
0
2
1
2
1
2
1
= = = = =
= = = = =
= = =
g g n g
g n n n
g g n g


( )
( )
( ) 0
2
1
2
2
1
2
1
. ] [ ; .
. . , .
5 5 , 5 , 5 5
= + =
+ =
+ =
+ =
i i
k
l
l
ik
k
k i
l
k
k
li
l
i
k
lk
k
k i
k
li
l
k
l
kl
k
i
k
i k
k
k i i
J J
F F
F F F
R

with the help of (35d). Hence
ik
k
ik
k
G G
; |
=
. (56b) can also be easily proven this way.
42

As a brief digression, consider a six-dimensional manifold
m V
n
=
6
where m is the
second normal coordinate with respect to
4
R . Let
| | i i
g
5
,
5
a ,
6
b ,
)
`

k
i
i
k
5
.
,
)
`

k
i
i
k
6
.
,
| | 56
g and
| | i i
g
6

. Casting (12) and (22) into six
dimensions, the electromagnetic field tensor can be written in terms of the fundamental
spin tensor as the following equivalent expressions:

| | | |
( ) ( ) ( )
| |
( ) ( )
i k k i r ik
r
i k k i i kr k ir
r
ik
g
g g F
, , ,
, , , ,
4
5
4
5


=
=
(57a)

( ) ( ) ( )
| | | |
( )
| | | |
( )
m
i km
m
k im
m
i km
m
k im km im
l
i k k i i k k i i kl l ik k li
l
ik
g g b
g g a
li
m
g
lk
m
g
b a g g g F
. .
. . ] [ ] [
, , , , , ] [ , ] [ , ] [
2
2 2


|
|

\
|
)
`

)
`

+ =

(57b)
Since the basis in this space is given by ( ) m , n g g ,
i
=

, the fundamental tensors are

| |
| |
| |
|
|
|

\
|

=
|
|
|

\
|
=
0
0 ,
1 0 0
0 1 0
0 0
4 4
] [
4 4
) (
) (
i
i i
i
x
ik
x
ik
g
g
g
g




(58a),(58b)
__________________________________________________________________________________
Remark 3 (on the modified Maxwell's equations)

Using (49) we can now generalize Maxwell's field equations through the new extended
electromagnetic tensor
i k k i ik
F
| |
= where

| | | | l
l
ik ik l
l
ik i k k i i k k i
old
ik
F F + = + = = 2 2
, , ; ;
) (

43

where
ik k i k i
F
2
1
; |
=
. Now =
5
is taken to be an extra scalar potential.
Therefore

ik i k k i ik
F F =
, ,
or


( ) ( )
i k k i ik
F
, ,
1
1 + =

( )
i k k i , ,
(a)

For instance, the first pair of Maxwell's equations can therefore be generalized into

( )
|
|

\
|

+ =


r
r
r
t
A
c
E
1
1
1
(b1)
( ) A x B
r r r
+ =
1
1 (b2)
( ) ( ) ( ) A x A x B div
r r r r r r r
+ + + =
1 2
1 1 (b3)

( ) ( ) ( ) ( )
1 1 1
1
1
1
1
1

+

= +
t
B
c
B
t c
E curl
r
r r
(b4)
where A
r
is the three-dimensional electromagnetic vector potential:
( )
a
A A =
r
,

is the
electromagnetic scalar potential, E
r
is the electric field, B
r
is the magnetic field and
r
is
the three-dimensional (curvilinear) gradient operator and 4
2
= . Here we
define the electric and magnetic fields in such a way that

( ) ( )
3
1
12
1
4
1 , 1 B F E F
a a

+ = + =
( ) ( )
1
1
23
2
1
31
1 , 1 B F B F

+ = + =

We also note that in (b3) the divergence A x
r r r
is in general non-vanishing when
torsion is present in the three-dimensional curved sub-space. Direct calculation gives


| |
( )
( )
| | d
d
ab
ab
c c
d c
d
ab d
d
cab
abc
grad curl
A A R A curl div
, . .
; .
2
2
1
=
=
r


So the magnetic charge with density
m
in the infinitesimal volume d is given by


| |
( )

= d A A R
d c
d
ab d
d
cab
abc
; .
2
2
1

44

4. Dynamics in the microscopic limit. Spin-curvature tensor of
p
S .
Wave equation describing the geometry of
4
R

We now investigate the microscopic dynamics of our theory. Let's introduce an
infinitesimal coordinate transformation into
n
through the diffeomorphism


i i i
x x + =
'


with an

external Killing-like vector ( ) ) ( ), ( : , x x
i i i
= = = (not to be
confused with the internal Killing vector which describes the internal symmetry of a
particular configuration of space-time or which maps a particular space-time onto itself).
The function

here shall play the role of the amplitude of the quantum mechanical
state vector . Recall that
4
R represents the four-dimensional physical world and n is
a microscopic dimension. In its standard form ( )
)) , , ( ( ) / 2 ( z y x t E h i
e C x
r p

is the
quantum mechanical scalar wave function; h is the Planck constant, E is energy and p is
the three-momentum. Define the extension of the space-time
n R
n
=
4

by


) (
2
1
ij ij
g D

=
(59)

We would like to express the most general symmetry, first, of the structure of
n
and
then find out what sort of symmetry (expressed in terms of the Killing-like vector) is
required to describe the non-local statics or non-deformability of the structure of
the metric tensor g (the lattice arrangement). Our exterior derivative is defined as the
variation of an arbitrary quantity with respect to the external field

. Unlike the ordinary
Killing vector which maps a space-time onto itself, the external field and hence also the
derivative
) ( ij
g D

map
4
R onto, say,
4
' R which possesses a deformed metrical
structure of g
4
R , g' . We calculate the change in the metric tensor with respect to the
external field, according to the scheme ( ) ) (
' ' '
) ( ) ( j i ij j i ij
g g g g g g = = , as follows:



i i i i j i j i ij
D D D g D = =
|
|

\
|
+
|
|

\
|
= g g g g g g g
'
) (
;

The Lie derivative

D denotes the exterior change with respect to the infinitesimal
exterior field, i.e., it dynamically measures the deformation of the geometry of the space-
time
4
R .
45

Thus


( ) ( )
j i j i ij
g D g g + =
) (


where


n g
|

\
|
+ +
|

\
|
=
k
ki i k
k
i
k
i i
F F
2
1
2
1
, . ;



By direct calculation, we thus obtain



i j j i ij
g D
; ; ) (
+ =

(60)
As an interesting feature, we point out that the change
) (ij
g D

in the structure of the
four-dimensional metric tensor does not involve the wave function

. The space-time
4
R will be called static if g does not change with respect to . The four-dimensional
(but not the five-dimensional) metric is therefore static whenever
0
; ;
= +
i j j i

.
Now let
4
R be an infinitesimal copy of
4
R . To arrive at the lattice picture, let also
... , , ,
" "
4
' "
4
"
4
R R R be n such copies of
4
R . Imagine the space
n
consisting of these
copies. This space is therefore populated by
4
R and its copies. If we assume that each
of the copies of
4
R has the same metric tensor as
4
R , then we may have


( ) , 0 =


We shall call this particular fundamental symmetry normal symmetry or spherical world-
symmetry. Then the n copies of
4
R exist simultaneously and each history is independent
of the four-dimensional external field
4
and is dependent on the wave function

only. In other words, the special lattice arrangement ( ) , 0 = gives us a condition for
4
R and its copies to co-exist simultaneously independently of how the four-dimensional
external field deforms their interior metrical structure. Thus the many sub-manifolds
n
4
R represent many simultaneous realities which we call world-pictures. More
specifically, at one point in
4
R , there may at least exist two world-pictures. In other
46

words, a point seen by an observer confined to lie in
4
R may actually be a line or curve
whose two end points represent the two solutions to the wave function

. The splitting
of
4
R into its copies occurs and can only be perceived on the microscopic scales with
the wave function

describing the entire process. Conversely, on the macroscopic

scales the inhabitants of n R
n
=
4
may perceive the collection of the n
4
R (sub-
spaces of
4
R ) representing a continuous four manifold ...
"
4
'
4 4 4
= R R R R .
There may be an infinite number of n
4
R composing the space-time
4
R , possessing the
same fluctuating metric tensor g , i.e., that of
4
R (the total space). On the microscopic
scales, fluctuations in the metric do occur. In special cases, the cylinder condition may
prevent the topology of the spaces n
4
R to change which cannot be perceived directly by
an external observer in
4
R for it takes place along the microscopic coordinate y . The
fluctuations induce the many different world-pictures. Our task now is to find the wave
equation describing the entire process. Now, with the help of (7) and (8) we can write


( ) ( ) n n g g + =
i
j
j i i


( ) ( ) ( )
j
i j j i
g g g + =
2
1


( ) ( ) ( ) ( ) n n g g g + +
i
j
i j j i
2
1


( ) | |
i
j
j
j
ij
g D g n g ,
2
1
) (


=

( ) ( )
j
ij l k
kl
l
l
ki j
jk
g g C g g g g n
] [
] [ ] [
2
3
+ +

( )
j
l k
kl
ij
g g
2
3


| |
( ) ( )
(

=
i
j
j l k
kl j
ij
g g D g n n g g ,
2
3
2
1
) (




| |
( ) ( )
j
l k
kl
ij l
l
ki j
jk
C g g g g n +
2
3

(61)

47

Setting

( ) ( )
j
j l k
kl
g n n g S =
] [
2
3


| |
( ) ( )
j
l k
kl
ij l
l
ki j
jk
i
C g g g g n X =
2
3


we have

| |
i i
j
ij i
g D X g S g + + = ,
2
1
) (

(62)

To see that the vector S represents a spin (rotation) vector is not difficult as we know
that
( ) 0 = n g
i
and hence
( ) 0 = n g
i
D
where the change in n can be represented
by an internal rotation:
| | n S n , = D
. With the help of (8), we now have


( ) | | ( ) | | ( ) ( )
i i i i i
S D = = = = S g n S n S g n g , ,


Therefore

( ) ( ) ( )
i i i i
S = = = n n S

Now

| |
( )
n
n g S
+ =
+ =
i
i
j i
ij i
i
S
g S


2
3
(63)

( )
k
k
i i i i
F S
. ,
2
1
= = n


| |
( )
j i
ij
g g =
2
3
(64a),(64b)

|

\
|
=
ji i j
ij
F g
2
1
2
3
;
] [

48

We can also calculate the exterior variation of the electromagnetic field tensor
ik
F :


( )
( ) ( )
| | ( ) ( )
( ) ( )



k j
j
i k j
j
i k i
k i k i
k i k i
k i ik
F F
D
D D
D F D
|

\
|

(

=
+ =
+ =
=
g g g S S
g n g n S
g n g n
g n
. .
2
1
, 2 2
2 , 2
2 2
2

(65a)
Hence we can write


( ) | | ( ) ( )
( ) ( )
|

\
|
+ + =

|

\
|
=
=
jk
j
i jk
j
i ik
k j
j
i
l
kl j
j
i k i
k k j
j
i k i ik
X F g D F H
F g D F
F F D
. ) ( .
. ) ( .
.
2
1
2
2
1
2
, 2





X g g g S
g S g S

(65b)
where we have just defined the spin-curvature tensor
ik
H :

( )
k i ik
H = S (66)

which measures the internal change of the spin in the direction of the spin basis. We
further posit that the spin-curvature tensor satisfies the supplementary identities (which
are deduced from the conditions 0
4
= u
u
and 0
4
= S
u
)

0 =
i
ik
u H (67a)

0 = H tr
(67b)

The transverse condition (67a) reproduces the Lorentz equation of motion while (67b)
describes the internal properties of the structure of physical fields which corresponds to
the quantum limit on our manifold (for details see Appendix B).
49

The explicit expression of
ik
H can be found to be

( )
k i ik
H = S

| | i k
l
i kl
S F g
; .
2
1
= (68)

from (64a) and (64b). Furthermore, still with the help of (64a) and (64b), we obtain, after
some simplifications,


( )
l k k l
l
i i k ik
F H
; ; . ; ;
4
1
+ =
( )
i
l
l
k lk
l
i
F F F
;
. .
2
1
4
1
+ (69)

However, we recall that


( )
jk il jl ik ijkl
F F F F R =
4
1
,
l
k il ik
F F R
.
4
1
=

and therefore we obtain the spin-curvature relation in the form

( )
l k k l
l
i ik i k ik
F R H
; ; . ; ;
4
1
+ =
( )
i
l
l
k
F
;
.
2
1
+
(70)

If we contract (70) with respect to the indices i and k, we have

H = ( R )
i
i
k i
ik
J F
2
1
;
+ +
(71)

where is the covariant four-dimensional Laplacian, again, R is the curvature scalar
and
i
J is the current density vector. However, using (67a) and (67b) and associating
with the space n R
n
=
4
the fundamental world-symmetry ( ) , 0 = , then we
obtain, from (70), the equation of motion


ik ik i k
R H + =
; ;
(72)

50

From (67a), (67b) and (72), we obtain the wave equation

( R ) 0 = (73a)


This resembles the scalar Klein-Gordon wave equation except that we have the
curvature scalar R in place of ( )
2
0
2
/ h c m M = (we normally expect this in generalizing
the scalar Klein-Gordon equation). Note also that the ordinary Klein-Gordon and Dirac
equations do not explicitly contain any electromagnetic terms. This means that the
electromagnetic field must somehow already be incorporated into gravity in terms of
M . Since

is just the amplitude of the state vector , we can also write

( R ) 0 = (73b)

If the curvature scalar vanishes, there is no source (or actually, no electromagnetic
field strength) and we have 0 = which is the wave equation of massless
particles.

__________________________________________________________________________________
Remark 4
Recall (65):


|

\
|
+ + =
jk
j
i jk
j
i ik ik
X F g D F H F D
. ) ( .
2
1
2

(a)

where


( )
k i ik
ik i k k i ik
H
g D


=
= + =
S
2
; ; ) (



Meanwhile, for an arbitrary tensor field T , we have in general

51


...
...
;
...
... ;
...
...
;
...
... ;
...
...
...
; ...
...
...

+ + + =
j
m
im
kl
i
m
mj
kl
m
l
ij
km
m
k
ij
ml
ij
m kl
m ij
kl
T T
T T T T D




Therefore


l
k il
l
i lk l ik
l
ik
F F F F D
; ; ;
+ + =

(b)

Comparing this with (a), we have, for the spin-curvature tensor
ik
H ,


( ) ( )
lk
l
i l k k l
l
i
l
k il
l
i lk l ik
l
ik
X F F F F F H
. ; ; . ; ; ;
2
1
4
1
2
1
+ + + = (c)


in terms of the electromagnetic field tensor.
__________________________________________________________________________________

Finally, lets define the following tensor:


i k
k
j j i ij
F S A
| . ;
2
1

(74)

where, as before,


ki i k i k
F
2
1
;
=


is the
n
covariant derivative of
k

, the notion of which we have developed in Section

3 of this work, and


k
k
i i i
F S
. ,
2
1
=


is the spin vector (64a)
.
The meaning of the tensor (74) will become clear soon. It has no
classical analogue. We are now in a position to decompose (74) into its symmetric and
alternating parts. The symmetric part of (74):

52


|

\
|
+ =
i k
k
j j k
k
i i j j i ij
F F S S A
| . | . ; ; ) (
2
1
2
1
2
1

(75a)


may be interpreted as the tension of the spin field.
Now its alternating part:


| |
|

\
|
+ =
i k
k
j j k
k
i i j j i ij
F F S S A
| . | . ; ;
2
1
2
1
2
1

(75b)

represents a non-linear spin field. (However, this becomes linear when we invoke the
fundamental world-symmetry ( ) , 0 = .) If we employ this fundamental world-
symmetry, (75a) and (75b) become


( )
ij i j j i ij
R S S A + + =
; ; ) (
2
1
(76a)

| |
( )
i j j i ij
S S A
; ;
2
1
=
(76b)

With the help of (64a) and the relation


| | r
r
ik i k k i , ; ; ; ;
2 =

(76b) can also be written


| | | | k
k
ij ij
A
,
=
(77)

Now


( )
ki
k
j i k
k
j
j
k
k
i j i ij
F F F F A
. ; .
;
. ; ;
4
1
2
1
2
1
+ =


( )
|

\
|
+ + =
i k
k
j
j
k
k
i ij j i
F F R
; .
;
. ; ;
2
1
2
1



53

Therefore the Ricci tensor can be expressed as

( )
|

\
|
+ + =

i r
r
k
k
r
r
i k i k i i r
r
k ik
F F S F R
; .
;
. ; ; ;
1
| .
1
2
1
2
1
2
1
(78)

We can still obtain another form of the wave equation of our quantum gravity theory.
Taking the world-symmetry
( ) , 0 =
, we have, from
n r r = '
,


r
r
i i i i
F g n h g
. ,
2
1
+ =
(79)

where
i i ,
' r h
is the basis of the space-time
4
' R . Now the metric tensor of the space-
time
4
R is


( )
s
k
r
i rs kr
r
i
k i k i ir
r
k i k ik
k i ik
F F g F
F h
g
. . ) (
2
.
, , , . ,
) (
4
1
2
1
2
1
.


+ +
+ + =
= g g


where ( )
k i ik
h h h is the metric tensor of
4
' R , ) n h
i i
( and ( )
k i ik
g h .
Direct calculation shows that


ik ik ik
i i
F g

2
1
) (
,
+ =
=


Then we arrive at the relation


s
k
r
i rs k i ik ik
F F g h g
. . ) (
2
, , ) (
4
1
=
(80)


Now from (60) we find that this is subject to the condition


0
) (
=
ij
g D

(81)
54

Hence we obtain the wave equation


s
k
r
i rs k i
F F g
. . ) (
2
, ,
4
1
= (82a)

Expressed in terms of the Ricci tensor, the equivalent form of (82a) is


ik k i
R
2
, ,
= (82b)

Expressed in terms of the Einstein tensor R g R G
ik ik ik ) (
2
1
= , (82b) becomes


ik s r
rs
ik
s
k
r
i
G g g
2
, ,
) (
) (
2
1
=
|

\
|

(83)

If in particular the space-time
4
R has a constant sectional curvature, then
( )
) ( ) ( ) ( ) (
12
1
jk il jl ik ijkl
g g g g R = and
) (ik ik
g R = , where
R
4
1

is constant,
so (82b) reduces to


) (
2
, , ik k i
g = (84)


Any axisymmetric solution of (84) would then yield equations that could readily be
integrated, giving the wave function in a relatively simple form. Multiplying now (82b) by
the contravariant metric tensor
) (ik
g
, we have the wave equation in terms of the
curvature scalar as follows:

R g
k i
ik 2
, ,
) (
= (85)

Finally, lets consider a special case. In the absence of the scalar source, i.e., in void,
the wave equation becomes


0
, ,
) (
=
k i
ik
g
(86)

This wave equation therefore describes a massless, null electromagnetic field where


0 =
ik
ik
F F

55

In this case the electromagnetic field tensor is a null bivector. Therefore, according to
our theory, there are indeed seemingly void regions in the Universe that are governed
by null electromagnetic fields only.



































56

5. Conclusion

We have shown that gravity and electromagnetism are intertwined in a very natural
manner, both ensuing from the melting of the same underlying space-time geometry.
They obey the same set of field equations. However, there are actually no objectively
existing elementary particles in this theory. Based on the wave equation (73), we may
suggest that what we perceive as particles are only singularities which may be interpreted
as wave centers. In the microcosmos everything is essentially a wave function that also
contains particle properties. Individual wave function is a fragment of the universal wave
function represented by the wave function of the Universe in (73). Therefore all objects
are essentially interconnected. We have seen that the electric(-magnetic) charge is none
other than the torsion of space-time. This charge can also be described by the wave
function alone. This doesnt seem to be contradictory evidence if we realize that nothing
exists in the quantum realm save the quantum mechanical wave function (unfortunately,
we have not made it possible here to carry a detailed elaboration on this statement).
Although we have not approached and constructed a quantum theory of gravity in the
strictly formal way (through the canonical quantization procedure), internal consistency
of our theory awaits further justification. For a few more details of the underlying
unifying features of our theory, see the Appendices.




















57

Appendix A: Embedding of generalized Riemannian manifolds (with
torsion) in N = n + p dimensions

In connection to Section 3 of this work where we considered an embedding of class 1,
we outline the most general formulation of embedding theory of class p in N = n + p
dimensions where n now is the number of dimensions of the embedded Riemannian
manifold. First, let the embedding space
N
R be an N-dimensional Riemannian manifold
spanned by the basis { }
A
e . For the sake of generality we take
N
R to be an N-
dimensional space-time. Let also
n
R be an n-dimensional Riemannian sub-manifold
(possessing torsion) in
N
R spanned by the basis { }
l
g where now the capital Latin
indices A,B, run from 1 to N and the ordinary ones i,j, from 1 to n. If now
( )
B A AB
g e e =
and ( )
j i ij
g g g = denote the metric tensors of
N
R and
n
R ,
respectively, and if we introduce the p-unit normal vectors (also called n-legs)
) (
n
(where the Greek indices run from 1 to p and summation over any repeated Greek
indices is explicitly indicated otherwise there is no summation ), then

+ =
=
+ =
+
+ + =
+ =
=
= + =
= =
=
+ =
=

A
ij
C
j
B
i
A
BC
k
ij
A
k
A
j i
A
ij
A
j i
C
j
B
i
A
BC A
A
j i A ij
B
j
C
A
i
ki
kj C B
A
A j
C
i
B ij
j
C
i
B
k
ij
A
k
i
C B
A
i
A
BC
C
j
B
i
A
BC
k
A
A
j i
k
A
k
ij
C
C
AB B A
ij j i ij k
k
ij j i
A
A
i
B A ij
j
B
i
A AB
AB
B
j
A
i ij
n e e e e
n e
e e n e n
n e e g n n
n e e e e e e e
e e e e e
n e
n n g e e g
g e e g
) ( ) ( ) (
,
) ( ) ( ) (
;
) (
,
) ( ) (
) ( ) ( ) ( ) (
,
) (
) ( ) ( ) (
,
,
,
) ( ) ( ) (
;
) ( ) ( ) (
,
) ( ) ( ) ( ) (
) (
) ( ) (
,
1 , ) (
0

e e
n g n g g
n n


58

Now since

+ =

) ( ) ( ) (
n g e
A i
i
A A
n e
and
) (
;
) (

j A
A
i ij
n e =
for the
asymmetric p-extrinsic curvatures, we see that

+ =
|

\
|
=

) ( ) ( ) (
;
) ( ) (
;
) (
;
) ( ) ( ) ( ) (
B
A
j A j B
j A B
A A
B ij
i
B
n n n n
n n n e


Lets define the p-torsion vectors by

i i
A
i A i
n n
=
=
) ( ) (
;


Hence


( )

A i
k
A ki i A
n e n

+ =
) ( ) ( ) (
;

or

+ =

) ( ) ( ) ( ) (
;
n g n
i
k
ki i


Now

+
+ =

,
) ( ) (
) ( ) (
;
) ( ) ( ) ( ) ( ) ( ) (
; ;
A k i
r
A rk i A k i A sk ri
rs i
A ik k i A
n
e n n g e n


Hence we obtain the expression

k i k i sk ri
rs A
k i A
g n n

+ + =
) (
;
) ( ) ( ) ( ) (
; ;





59

Meanwhile,


( )
B
ik B A
C
k
B
i C B A
A
k i B A
C
k
B
i C B A
C
k
C
B
i B A k i A
B
i B A i A
n n e e n
e n e e n
e e n n
e n n
) ( ) ( ) ( ) (
;
) (
; ;
;
) (
;
) (
; ;
;
) (
;
) (
; ;
) (
;
) (
;

+ =
+ =
=
=


but


i
B
k
A ki
i
B i A B A
e e e n n
) ( ) (
;
) (
;

= =


Hence


0
0
) ( ) (
;
) ( ) (
;
=
=
A
B A
B
B A
n n
n n




We also see that


C
k
B
i
A
C B A
A
k i A
e e n n n n
) ( ) (
; ;
) ( ) (
; ;

=


Consequently, we have


( ) ( )
D
k
C
i
B A
ABCD
C
k
B
i
A
B C A C B A
A
i k A k i A
e e n n R
e e n n n n n n
) ( ) (
) ( ) (
; ;
) (
; ;
) ( ) (
; ;
) (
; ;


=
=


On the other hand, we see that


( ) ( )
( )
D
k
C
i
B A
ABCD i k k i
sk ri si rk
rs
i k k i
A
i k A k i A
e e n n R
g n n n
) ( ) ( ) (
) ( ) ( ) ( ) (
; ;
) ( ) (
; ;
) (
; ;

+ +
+ =

60

Combining the last two equations we get the Ricci equations:


( ) ( )

i k k i si rk sk ri
rs
i k k i
g
) ( ) ( ) ( ) ( ) (
; ;


Now from the relation

+ =

) ( ) ( ) (
,
n g g
ij k
k
ij j i


we obtain the expression


) ( ) ( ) ( ) (
,
) ( ) ( ) (
) ( ) ( ) (
, ,

n
g g

\
|
+ + +
|
|

\
|
+ =
k ij k ij
r
ij rk
r sk ij
rs r
sk
s
ij
r
k ij jk i
g


Hence consequently,


( )
| |
( )
( )

+
+ +
|

\
|
+ =

,
) ( ) ( ) ( ) (
) ( ) ( ) (
;
) (
;
) (
) ( ) ( ) ( ) ( ) (
. , ,
2
n
n
g g g
j ik k ij
ir
r
jk j ik k ij
r ij sk ik sj
rs r
ijk kj i jk i
g R


On the other hand, we have


( )
( )
A
C
k j
B
i
A
BC
C
j
B
k i
A
BC
C
k
B
j i
A
BC
A
jk i
A
D
k
C
j
B
i
A
ED
E
BC
A
D BC jk i
e e e e e e e
e e e
e
e g
, , , ,
, ,
+ + + +
+ =



( )
A
A
ijk
D
k
C
j
B
i
A
BCD kj i jk i
S e e e R e g g
. . , ,
+ =


where, just as in Section 3,


( )
B
i
A
BC
C
j k
C
k j
A
kj i
A
jk i
A
ijk
e e e e e S + =
, , , , .


61

Combining the above, we generalize the Gauss-Codazzi equations into


( )
| |
( )

+
+ =
+ =

j ik k ij
ir
r
jk
A
Aijk
D
k
C
j
B
i ABCD j ik k ij
A
i Ajkl
D
l
C
k
B
j
A
i ABCD jk il jl ik ijkl
n S e e e n R
e S e e e e R R
) ( ) ( ) (
) ( ) ( ) ( ) (
;
) (
;
) ( ) ( ) ( ) ( ) (
2



Finally, when
n
R is embedded isometrically in
N
R , i.e., when the embedding manifold
N
R is an Euclidean or pseudo-Euclidean N-dimensional space(-time) or if we impose a
particular integrability condition on the n-vectors in
n
R the way we derived (34) in
Section 3, we have the system of equations


( )
| |
( )
( ) ( )

=
+ =
=

i k k i si rk sk ri
rs
i k k i
j ik k ij ir
r
jk j ik k ij
jk il jl ik ijkl
g
R
) ( ) ( ) ( ) ( ) (
; ;
) ( ) ( ) ( ) ( ) (
;
) (
;
) ( ) ( ) ( ) ( ) (
2


Turning to physics, a unified field field theory of gravity and electromagnetism may be
framed by the geometric quantities above either in 5 = 4 + 1 background dimensions or,
if we wish to extend it, in N = 4 + p background dimensions. As an alternative, in N =
4 + p dimensions we may also observe that the following assumptions can be made
possible:

1) The first alternating p-exterior curvature
| |
) 1 (
ik

is equivalent to the ordinary

electromagnetic field tensor
ik
F .
2) The remaining p-exetrior curvatures
| | | |
) (
) (
) 1 (
) (
) ( ) 2 (
,..., , ..., ,
p
ik ik
p
ik ik
represent fields
beyond the known electromagnetic and gravitational fields.
3) As in Section 3, if desired, the four-dimensional metric tensor
ij
g
may not
depend on the extra p-coordinates { }
N N N
x x x ,..., ,
3 4
. Thus the cylinder
condition represented by the equations


0
) (
;
) (
;
= +

A B B A
n n
or
0
) (
) (
=

ij


is, again, arrived at.
62

Appendix B: Formulation of our gravoelectrodynamics in the sense of
the theory of distributions. Massive quantum electromagnetic field
tensor

In Section 3 we have assumed a type of parallel transport applied to the pseudo-five-
dimensional velocity field ( ) n g u + =
i
i A
u u in n R
n
=
4
, i.e.,
0
4
= u
u

where
ds
dx
u u
i
i
i
i
= = ;
4
g u
and
0 ;
2
2
0
=

= =
ds
d
n u
c m
e
A
A
. To describe
non-diverging point-like objects which may experience no change in energy even when
accelerated, like electrons, we now introduce a special kind of autoparallelism through
the relation
0 = u
i
. This says that the pseudo-five-dimensional velocity field
) , ( =
i
u u is an autoparallel vector field in the sense of the theory of distributions,
whose magnitude is independent of the four coordinates of
4
R . In other words,


k
k
i i
i
k
i
k
u F
F u
. ,
. ;
2
1
2
1
=
=


From these we have


0
0
;
.
2
0
=

=
=
ds
d
u
u F
c m
e
Ds
Du
i
i
k i
k
i


Were now in a position to derive the field equations of gravoelectrodynamics with the
help of these relations. Afterwards, we shall show that these relations, indeed, lead to
acceptable equations of motion in both four and five dimensions. We may also
emphasize that the five-dimensional space

IR
5
is a Riemann space. First we note that


r ik
r
l l ik
ik l l ik l k i
u F F F
F F u
. ;
, ; ; ;
4
1
2
1
2
1
2
1
=
+ =


63

Now


| |
( ) ( )
r ik
r
l il
r
k k il l ik
r
kl r i r
r
ikl k l i l k i
u F F F F F F
u u R u u
. . ; ;
; . ; ; ; ;
4
1
2
1
2
+ =
=


Therefore we have


( )
| |
( ) + + =
jr
r
kl k jl l jk i jk
i
l jl
i
k i
i
jkl
F F F u F F F F u R 2
2
1
4
1
; ; . . .


But
A
A
i i
u u = and
A
A
n u = , so by means of symmetry, we can lop off the
A
u :


( )
| |
( )
A
jr
r
kl k jl l jk
A
i jk
i
l jl
i
k
A
i
i
jkl
n F F F F F F F R + + = 2
2
1
4
1
; ; . . .



From this relation, we derive the unified field equations


( )
| | ir
r
jk j ik k ij
jk il jl ik ijkl
F F F
F F F F R
=
=
2
4
1
; ;


as expected. We have thus no need to assume that the curvature of the background five-
dimensional space IR
5
vanishes. To strengthen our proof, recall that
n g
ij j i
=
;
.
Hence


r
r
k ij k ij
k ij k ij k j i
g n
n n g
. ;
; ; ; ;


=
+ =



( ) ( )
r
r
k ij
r
j ik j ik k ij j k i k j i
g n g g
. . ; ; ; ; ; ;
+ =


However,


| |
n g g g
ir
r
jk r
r
ijk j k i k j i
R = 2
. ; ; ; ;

64

Combining the above, we have


| | ir
r
jk j ik k ij
jk il jl ik ijkl
R


=
=
2
; ;


Invoking the cylinder condition, again we get


( )
| | ir
r
jk j ik k ij
jk il jl ik ijkl
F F F
F F F F R
=
=
2
4
1
; ;


We now assume that the five-dimensional equation of motion in IR
5
is in general not a
geodesic equation of motion. Instead, we expect an equation of motion of the form


B A
B
C B
A
BC
A
u F
ds
dx
ds
dx
ds
x d
.
2
2
~
= +


where
i A
i
A
u e u = , and


( )
2 4
) (
) ( ) (
2 5 2
ds dx dx g
dx dx n n g e e dx dx g ds ds
k i
ik
B A
B A ik
k
B
i
A
B A
AB
= =
+ = = =


( )
B CD D BC C DB
DA A
BC
g g g g
, ) ( , ) ( , ) (
) (
2
1
+ =


( )
0
0
~
; ;
=
=
=
=
+
B
k
A
i AB
B
AB
AB
B
k
A
i ik
A B B A AB
AB AB AB
e e
n F
F e e F
n n F
F F


Note also that the connection transforms as

B
k
C
A
i
i
k C B
A A k
C
i
B ik
k
C
j
B
i
jk
A
i
i
C B
A
i
A
BC
n e e F n n n e e F e e e e e
. , ,
2
1
2
1
+ + + =

65

Now with the help of the relation


A
ik
C
k
B
i
A
BC
r
ik
A
r
A
k i
n F e e e e
2
1
,
+ =


we have


ds
dx
ds
dx
ds
dx
ds
dx
e
ds
x d
e
ds
x d
e
ds
dx
ds
dx
e
ds
dx
e
ds
d
ds
x d
C B
A
BC
k j
i
jk
A
i
i
A
i
i
A
i
k i
A
k i
i
A
i
A
+ =
+ =
|
|

\
|
=
2
2
2
2
,
2
2


Hence


B A
B
k j
i
jk
i
A
i
C B
A
BC
A
u F
ds
dx
ds
dx
ds
x d
e
ds
dx
ds
dx
ds
x d
.
2
2
2
2
~
= |

\
|
+ = +

Setting 0 =
AB
and
2
0
c m
e
= , since
k i
k
k B
k
i
A
A
B
B i
A
A
B
u F u e e F u e F
. . .
= =
,
we then obtain the equation of motion:


k i
k
i
u F
c m
e
Ds
Du
.
2
0
=

A straightforward way to obtain the five-dimensional equation of motion is as follows:
from our assumptions we have


( )
ik k A
A
i
F u e =
2
1
;


ik
B
k
A
i B A
A
A ik
F e e u n u F = +
2
1
;


Therefore recalling that 0 =
A
A
n u and multiplying through by
k
u , we get


k i
k
B i
A
A
B
u F u e u
. ;
2
1
=

66

Again, noting that
Ds
Du
u u u u
A
k A
k
B A
B
= =
; ;
and multiplying through by
C
i
e
, we get


A C
A
C
A
A
C
u F
c m
e
n
Ds
Du
n
Ds
Du
.
2
0
=

But


0
2
1
2
1
; ;
= =
=
= = =
k i
ik
k i
A ik
A
k A
k A
B A
B A
A
A
A
A
u u F
u e F u
u u n u u n
Ds
Du
n
Ds
Dn
u


Therefore


B A
B
A
u F
c m
e
Ds
Du
.
2
0
=


Lets now recall the spin-curvature tensor (66) as well as (63) and (64):



( )
k i ik
H = S


| | i k
l
i kl
S F g
; .
2
1
=

| |
( )
n
n g S
+ =
+ =
i
i
j i
ij i
i
S
g S


2
3


( )
k
k
i i i i
F S
. ,
2
1
= = n


| |
( )
j i
ij
g g =
2
3


|

\
|
=
ji i j
ij
F g
2
1
2
3
;
] [

67

Recall (67a):

0 =
i
ik
u H

Multiplying (68) by
i
u , we get the equation of motion:



| |
i r
i kr
k
u F g
Ds
DS
.
2
1
=

Equivalently,



| | r
kr
ik
i
u F g
Ds
DS

2
1
=


We can now compare this with (55a). Now this leads us to consider a case in our theory
in which the spin vector
i
S is normalized (in the quantum limit). Then it is given by


| |
k
ik i i
u g v S = =

Multiplying once again by
| | ij
g
, we get


| |
r
jr
i
ij
u F
Ds
DS
g
2
1
=


k i
k
i
u F
Ds
Du
.
2
1
=



which is the Lorentz equation of motion. In this case then it automatically follows that


0
2
2
0
=
=
|
|

\
|
=
ds
d
c m
e

Corresponding to our ongoing analysis, lets also recall (67b):

0 = H tr
68

This gives the divergence equation


| |
ik
ik
i
i
F g S
2
1
;
=

or


| |
0
2
1
;
=
|

\
|

k i ik
ik
u F g

On the other hand, from (64a), and by employing our fundamental symmetry, we have

=
i
i
S
;


Therefore, with the help of (73a), we see that

| |
R F g
ik
ik
= =
2
1


| |

ik
ik
ik
ik
F F F g
4
1
2
1
=

The simplest solution for massive, electrically charged particles of this would then be


| | ik ik
g F 2 =

which expresses the proportionality of the already quantized electromagnetic field
tensor to the fundamental spin tensor of our unified field theory. In other words,


| | ik ik
g
c m
e
F
1
2
0
4

|
|

\
|
=

Finally, with the help of (85), we get the fundamental quantum relations
(
0 R
, 0 ):

| |
2
2
0
1
4

=
|
|

\
|
=
R
g R
c m
e
F
ik ik

69

Now, following (1), we can write our asymmetric fundamental tensor
ik
as


|
|

\
|
|
|

\
|
+ =
ik ik ik
F
e
c m
g
2
0
) (
4
1
2
1


which satisfies (4):
k
i
kj
ij
= . On the other hand, now we see that


|
|

\
|
|
|

\
|

|
|

\
|
+ =
|
|

\
|
|
|

\
|
+ =
k
i
k
i
k
i
rk
ir
k
i
k
i
kr
ir
R
e
c m
F
e
c m
R
e
c m
.
2
2
2
0
.
2
0
.
2
2
2
0
4
1
2
1
2
1
4
1
2
1




Hence we get the following expression for the Ricci tensor:


) (
2
2
2
0
4
ik ik
g
c m
e
R

|
|

\
|
=

We now define an inverse wave function:


1
2
0
2

|
|

\
|
=
c m
e


We can then express the Ricci and the electromagnetic field tensors as


| | ik ik
ik ik
g F
g R
=
=
2
) (
2


Then it follows from (2b) and (2c) that


| |
| | 2
.
1
4
2
1
2
= =
=

ik
ik
r
i kr ik
F g R
R g F


We have now therefore fulfilled our promise in the beginning (for instance, at the end of
Section 2) to express gravity and electromagnetism in terms of the components of the
fundamental tensor, i.e.,
) ( ) (
2
ik ik
g and
] [ ] [
2
ik ik
g , alone.
70

Appendix C: On the conservation of currents

We need to recall the basic field equations:


( )
| | ir
r
jk j ik k ij
jk il jl ik ijkl
F F F
F F F F R
=
=
2
4
1
; ;



| |
| |
s
r
r
ks
ik i
rs k
rs
i
k
k i
k
rk
k
i
r
rk i
k r
ik
k
rk i
r
ik
F g J
F F J F
F F F F R
F F R
.
) (
. .
; . ; . ;
.
2
4
1
4
1
4
1
4
1
4
1
=
+ =
=
=


To guarantee conservation of currents, we now introduce the one-form
i

n
:

n g
i
k
ik i
R + =

where
i
represents another current. Requiring, in the sense of the theory of
distributions, that the covariant derivative of
i

vanishes at all points of
4
R also means
that its covariant divergence also vanishes:


0
;
=
i
i



Hence


0
2
1
2
1
, 0
. ; ;
; ; ; ;
= +
= + + +
k
k
i
i i
i ik
ik
k
ik
i
i
i i
i k i
ik
k
ik
i
F F R R
R R
g n n g
n n g g




Then we have

0
2
1
;
. ;
=
=
i
i
k i
k
ik
k
F R

71

Comparing the last two equations above with the fourth equation, one must find


| |
( )
kl i
kl
i i
F J =
2
1


Meanwhile, in the presence of torsion the Bianchi identity for the electromagnetic field
tensor and the covariant divergence of the four-current J are


| | | | | |
( )
| |
( )
i
kl i
kl
i
i
jr
r
ki ir
r
jk kr
r
ij j ki i jk k ij
F J
F F F F F F
;
;
; ; ;
2
=
+ + = + +


Now since
0
;
=
i
i

, we see that


| |
( ) 0
;
=
i
kl i
kl
F

and (in a general setting)


| |
i kl i
kl
i
F J 2 =

for some constant , are necessary and sufficient conditions for the current density
vector J to be conserved. Otherwise both
i
and
i
J are directly equivalent to each
other. Of course one may also define another conserved current through


0
|| ||
= =
i
i
ik
k
i
j F j


where the double stroke represents a covariant derivative with respect to the symmetric
Levi-Civita connection. In general the current J will automatically be conserved if the
orthogonality condition imposed on the twist vector, derived from the torsion tensor,
and the velocity vector:

0 =
i
i
u
where


| |
k
ki i
=

holds. For more details of the conservation law for charges, see Appendix D below.

72

Appendix D: On the wave equations of our unified field theory

We start again with the basic field equations of our unified field theory:


( )
| |
| |
r
s
s
kr
ik i
ir
r
jk j ik k ij
jk il jl ik ijkl
F g J
F F F
F F F F R
.
) (
; ;
2
2
4
1
=
=
=


We remind ourselves that these field equations give us a set of complete relations
between the curvature tensor, the torsion tensor, the electromagnetic field tensor and the
current density vector. From these field equations, we are then able to derive the
following insightful algebraic relations:


( )
| |
( )
| |
( )
r
q
k
r
k
sq r
rs q
ik
i
r
q
p
r
p
sq r
rs q
kp
ik i
jk il
ijkl
ijkl
ijkl
ik
ik
jl ik
ijkl
ik
ik
r
k ir risk
rs
ik
rk
r
l rl
r
k ijkl
ij
R F R F J R
R F R F g R J
F F R R R R R
F F R R R R
R F R F R F
F R F R R F
. . .
2
. . .
) ( 1
2
.
1
. .
2
2
4
1
4
1
+ =
+ =
+ + =
+ =
+ =
=

where the density corresponds to possible electric-magnetic charge distribution. As
has been shown previously, we also have the traceless field equation:


|

\
|
=
rs
rs ik
l
k il ik ik
F F g F F R g R
) ( . ) (
4
1
4
1
4
1


Now recall that the tetrad and the unit normal vector satisfy


A
ik
A
k i
n F e
2
1
;
=


A
k
k
i
A
i
e F n
. ;
2
1
=

73

Then we see that


A
r
rk
ik
A
i
A
k j i
jk
A
r
r
k ij
A
k ij
A
k j i
e F F n J e g
e F F n F e
4
1
2
1
4
1
2
1
; ;
) (
. ; ; ;
=
=


With the help of the basic field equations, we obtain the tetrad wave equation of our
unified field theory:


A
k
k
i
A
i
A
i
e R n J e
.
2
1
=

This expression gives a wave-type equation of the tetrad endowed with two sources: the
electromagnetic source, i.e., the electromagnetic current density vector and the Ricci
curvature tensor which represents the gravitational source in standard general relativity.

In addition, we can also obtain the following wave equation:

A A
i
i A
n R e J n =
2
1


In other words,

(
R +
)
A
i
i A
e J n
2
1
=


Furthermore, it follows that


k
A
e
k
i
A
i
R e
.
=

A
n
i
A
i
J e
2
1
=

A
n
R n
A
=


i
A
e
i A
J n
2
1
=




74

We may now express the current density vector as


A i
n J 2 =
i
A
e

Again, we obtain the conservation law of electromagnetic currents as follows:


A
i
i
i
n J
; ;
2 =

i
A
e
A
n 2 +
i
i A
e
;

=
A
k
k
i
e F
.

i
A
e
=
i
k
k
i
R F
. .

= 0

Here we have also used the relation
2
1
;
=
i
i A
e
( ) 0
.
=
A
i
i
n F .

Meanwhile, since the equation 0
) (
=
ik
g must be satisfied unconditionally by the
metric tensor
A
k Ai ik
e e g =
) (
and also since 0
; ) ( ; ) (
= =
C
i C AB i AB
e g g , we then get


Ak
e
A
i
A
i
e e =
ik Ak
R e 2

Now the curvature tensor can be expressed in terms of the tetrad as

( )
B
j k
A
i l
B
j l
A
i k AB ijkl
e e e e g R =
) (


Then the Ricci tensor is


B
k
r A
i r AB ik
e e g R =
) (


We can also express this as


A
i AB ik
e g R
) (
=
B
k
e

which can be written equivalently as

( )
B
i
r
r
A
k
B
k
r
r
A
i AB ik
e e e e g R + =
) (
2
1

75

Consider now a source-free region in the space-time
4
R . As weve seen, the absence of
source is characterized by the vanishing of the torsion tensor. In such a case, if the
space-time has a constant sectional curvature K , we obtain the tetrad wave equation for
the empty region:

( R
4
1
+ ) 0 =
A
i
e

In other words,

( K + ) 0 =
A
i
e

Combining our tetrad wave equation
A
k
k
i
A
i
A
i
e R n J e
.
2
1
= with the equation
for the Ricci tensor we have derived in the quantum limit (in Appendix B), which is


) (
2
2
2
0
4
ik ik
g
c m
e
R

|
|

\
|
=

we obtain the following wave equation in the presence of the electromagnetic current
density:

( R
4
1
+ )
A
i
A
i
n J e
2
1
=


| | | |
( )
s r
i
i
rs
ik
rs
r
ik
dx R R C R
. ..
2 4 + =

where K C 4 = is constant.

Finally, lets have a look back at the wave equation given by (73b):

( R ) 0 =

If we fully assume that the space-time
4
R is embedded isometrically in IR
5
spanned by
the time coordinate ct = and the four space coordinates y w v u , , , which together
76

form the line-element
2 2 2 2 2 2 2 2 2
dy d d dy dw dv du d ds = ,
and if the wave function represented by the state vector does not depend on the
microscopic fifth coordinate y , we can write the wave equation in the simple form:


= |

\
|

2
2
2
2
R






























77

Appendix E: A more compact form of the generalized Gauss-Codazzi
equations in IR
5


We write the generalized Gauss-Codazzi equations (see (30)

for instance) in IR
5
once
again:


A
i Ajkl
D
l
C
k
B
j
A
i ABCD jk il jl ik ijkl
e S e e e e R R + =


| |
A
Aijk ir
r
jk
D
k
C
j
B
i
A
ABCD j ik k ij
n S e e e n R + = 2
; ;


in terms of the general asymmetric extrinsic curvature tensor. Here, as usual,


( )
B
i
A
BC
C
j k
C
k j
A
kj i
A
jk i
A
ijk
e e e e e S + =
, , , , .


From the fundamental relations


A
k
k
i
A
i
A
ij
A
j i
e n
n e
. ;
;

=
=


we have


A
r
r
k ij
A
k ij
A
k j i
e n e
. ; ; ;
=


Hence

( ) ( )
A
r ij
r
k ik
r
j
A
j ik k ij
A
j k i
A
k j i
e n e e
. . ; ; ; ; ; ;
+ =

With the help of the generalized Gauss-Codazzi equations above and the identity


B
A
B
A
B
i
i
A
n n e e =

we see that that the generalized Gauss-Codazzi equations in IR
5
can be written somewhat
more compactly as a single equation:


| |
A
ijk
A
ir
r
jk
D
k
C
j
B
i
A
BCD
A
r
r
ijk
A
j k i
A
k j i
S n e e e R e R e e
. . . ; ; ; ;
2 + =




78

Appendix F: The fundamental geometric properties of a curved
manifold

Let us present the fundamental geometric objects of an n dimensional curved
manifold. Let
i
i
a i
a
i
a
E X E
x
X
=

= (the Einstein summation convention is

assumed throughout this work) be the covariant (frame) basis spanning the
n dimensional base manifold

C with local coordinates ( )

k a a
X x x = . The
contravariant (coframe) basis
b
is then given via the orthogonal projection
b
a a
b
= , , where
b
a
are the components of the Kronecker delta (whose value is
unity if the indices coincide or null otherwise). The set of linearly independent local
directional derivatives
i
i
i
X
E =

= gives the coordinate basis of the locally flat

tangent space ) (M T
x
at a point

C x . Here M denotes the topological space of the

so-called n tuples ( ) ( )
n
x x h x h ..., ,
1
= such that relative to a given chart ( ) ( ) x h U , on
a neighborhood U of a local coordinate point x , our

C differentiable manifold itself

is a topological space. The dual basis to
i
E spanning the locally flat cotangent space
) (
*
M T
x
will then be given by the differential elements
k
dX via the relation
k
i i
k
dX = , . In fact and in general, the one-forms
k
dX indeed act as a linear map
IR M T
x
) ( when applied to an arbitrary vector field ) (M T F
x
of the explicit form
a
a
i
i
x
f
X
F F

= . Then it is easy to see that

i i
X F F = and
a a
x F f = , from
which we obtain the usual transformation laws for the contravariant components of a
vector field, i.e.,
a i
a
i
f X F = and
i a
i
i
F x f = , relating the localized components
of F to the general ones and vice versa. In addition, we also see that
k k k
F X F F dX = = , .
The components of the symmetric metric tensor
b a
ab
g g = of the base manifold

C are readily given by

b a ab
g , =

satisfying

b
a
bc
ac
g g =

79

where
b a ab
g , = . It is to be understood that the covariant and contravariant
components of the metric tensor will be used to raise and the (component) indices of
vectors and tensors.

The components of the metric tensor ( )
k i
ik N
dX dX x g = describing the locally
flat tangent space ) (M T
x
of rigid frames at a point ( )
a
N N
x x x = are given by

( ) 1 ,..., 1 , 1 , = = diag E E
k i ik


In four dimensions, the above may be taken to be the components of the Minkowski
metric tensor, i.e., ( ) 1 , 1 , 1 , 1 , = = diag E E
k i ik
.

Then we have the expression

k
b
i
a ik ab
X X g =

The line-element of

C is then given by

( )
k i b
k
a
i ab
dX dX x x g g ds = =
2


where
i a
i
a
dX x = .

Given the existence of a local coordinate transformation via ( )

x x x
i i
= in

C , the
components of an arbitrary tensor field

C T of rank ) , ( q p transform according to

x x x x x x T T
h d c
g b a g ab
h cd
= ... ...
...
...
...
...


Let
p
p
i i i
j j j
...
...
2 1
2 1
be the components of the generalized Kronecker delta. They are given by

|
|
|
|
|

\
|
= =
p
p p p
p
p
p
p
p
p
i
j
i
j
i
j
i
j
i
j
i
j
i
j
i
j
i
j
i i
j j j
i i i
j j j

...
... ... ... ...
...
...
det
2 1
2
2
2
1
2
1
2
1
1
1
1
2 1
2 1
2 1
...
...
...
...


80

where ( )
p p
j j j j j j
g
... ...
2 1 2 1
det = and
( )
p p
i i i i i i
g
... ...
2 1 2 1
det
1
= are the covariant and
contravariant components of the completely anti-symmetric Levi-Civita permutation
tensor, respectively, with the ordinary permutation symbols being given as usual by
q
j j j ...
2 1
and
p
i i i ...
2 1
. Again, if is an arbitrary tensor, then the object represented by

p
p
p p
i i i
i i i
j j j j j j
p
...
...
... ...
2 1
2 1
2 1 2 1
!
1
=

is completely anti-symmetric.

Introducing a generally asymmetric connection via the covariant derivative

c
c
ab a b
=

i.e.,

( ) | |
c
ab
c
ab a b
c c
ab
+ = = ,

where the round index brackets indicate symmetrization and the square ones indicate
anti-symmetrization, we have, by means of the local coordinate transformation given by
( )

x x x
a a
= in

b a c
c
ab a b
e e e e =

where the tetrads of the moving frames are given by

x e
a a
= and
a a
x e

= . They
satisfy
a
b b
a
e e

= and

=
a
a
e e . In addition, it can also be verified that

c a
cb b
a a
b
c b a
bc
a a
e e e e
e e e e

=
=


We know that is a non-tensorial object, since its components transform as

b a
c
a b
c c
ab
e e e e e + =

However, it can be described as a kind of displacement field since it is what makes
possible a comparison of vectors from point to point in

C . In fact the relation

81

c
c
ab a b
= defines the so-called metricity condition, i.e., the change (during a
displacement) in the basis can be measured by the basis itself. This immediately
translates into

0 =
ab c
g

where we have just applied the notion of a covariant derivative to an arbitrary tensor
field T :

g ab
p cd
p
hm
g ab
h cp
p
dm
g ab
h pd
p
cm
p ab
h cd
g
pm
g ap
h cd
b
pm
g pb
h cd
a
pm
g ab
h cd m
g ab
h cd m
T T T
T T T T T
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...

+ + + + =


such that ( )
g ab
h cd m
g ab
h cd m
T T
...
...
...
...
= .

The condition 0 =
ab c
g can be solved to give

( )
| | | | | |
( )
e
da be
e
db ae
cd c
ab bd a ab d da b
cd c
ab
g g g g g g g + + + =
2
1


from which it is customary to define

( )
bd a ab d da b
cd c
ab
g g g g + =
2
1


as the Christoffel symbols (symmetric in their two lower indices) and

| | | | | |
( )
e
da be
e
db ae
cd c
ab
c
ab
g g g K + =

as the components of the so-called contorsion tensor (anti-symmetric in the first two
mixed indices).

Note that the components of the torsion tensor are given by

| |
( )

b c c b c b b c
a a
bc
e e e e e + =
2
1


where we have set

c c
e = , such that for an arbitrary scalar field we have

82

( )
| |
=
c
c
ab a b b a
2

The components of the curvature tensor R of

C are then given via the relation

( )
| |
s ab
r cd w
w
pq
s
wpq
w ab
r cd
b
wpq
s aw
r cd
a
wpq
s wb
r cd
w
rpq
s ab
w cd
w
dpq
s ab
r cw
w
cpq
s ab
r wd
s ab
r cd q p p q
T
R T R T R T
R T R T R T T
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
2
...
...


+ + + =


where

( )
d
ec
e
ab
d
eb
e
ac
d
ab c
d
ac b
d
abc
d
ec
e
ab
d
eb
e
ac
d
ab c
d
ac b
d
abc
K K K K K K B
R
+ + =
+ =



where

denotes covariant differentiation with respect to the Christoffel symbols alone,

and where

( )
d
ec
e
ab
d
eb
e
ac
d
ab c
d
ac b
d
abc
B + =

are the components of the Riemann-Christoffel curvature tensor of

C .

From the components of the curvature tensor, namely,
d
abc
R , we have (using the metric
tensor to raise and lower indices)

( )
| | | |
( )
| | | | | |
acb
abc
d
cd
b
ab
ac c
bc a
ab a
a
d
cd
c
ab
c
ac b
d
cb
c
ad
c
ab c ab
c
acb ab
K K g g B R R
K K K K B R R
=
+ + =
2

4
2

where ( ) ( )
c
acb ab
B B are the components of the symmetric Ricci tensor and
( ) ( )
a
a
B B is the Ricci scalar. Note that
d
bc ad abc
K g K and
a
de
be cd acb
K g g K .

Now since

( ) ( )
( ) ( )
| |
b
ab a
b
ab
a
b
ab
b
ba
b
ba
g
g
+ =
= = =
2 det ln
det ln


83

we see that for a continuous metric determinant, the so-called homothetic curvature
vanishes:

0 = =
c
ca b
c
cb a
c
cab ab
R H

Introducing the traceless Weyl tensor W , we have the following decomposition
theorem:

( )
( ) ( )
( ) R g g
n n
R g R R g R
n
W R
ac
d
b ab
d
c
d
c ab ab
d
c
d
b ac ac
d
b
d
abc
d
abc

+
+

+ =
2 1
1
2
1


which is valid for 2 > n . For 2 = n , we have

( )
ab
d
c ac
d
b G
d
abc
g g K R =

where

R K
G
2
1
=

is the Gaussian curvature of the surface. Note that (in this case) the Weyl tensor
vanishes.

Any n dimensional manifold (for which 1 > n ) with constant sectional curvature R
and vanishing torsion is called an Einstein space. It is described by the following simple
relations:

( )
R g
n
R
R g g
n n
R
ab ab
ab
d
c ac
d
b
d
abc
1
) 1 (
1
=

=


In the above, we note especially that

( )
( )
( ) =
=
=
B R
B R
B R
ab ab
d
abc
d
abc

84


Furthermore, after some lengthy algebra, we obtain, in general, the following generalized
Bianchi identities:

| | | | | | | | | | | |
( )
| | | | | |
( )
| | | |
cdb
a
a
cd
d
c
c
da
ab ab ab
a
a
bfd
f
ec
a
bfc
f
de
a
bfe
f
cd
a
bec d
a
bde c
a
bcd e
e
bc
a
ed
e
db
a
ec
e
cd
a
eb
a
db c
a
cd b
a
bc d
a
dbc
a
cdb
a
bcd
R R g R g R
R R R R R R
R R R
+ =
|

\
|

+ + = + +
+ + + + + = + +
2
2
1
2
2

for any metric-compatible manifold endowed with both curvature and torsion.

In the last of the above set of equations, we have introduced the generalized Einstein
tensor, i.e.,

R g R G
ab ab ab
2
1


In particular, we also have the following specialized identities, i.e., the regular Bianchi
identities:

0
2
1

0

0
= |

\
|

= + +
= + +
B g B
B B B
B B B
ab ab
a
a
bec d
a
bde c
a
bcd e
a
dbc
a
cdb
a
bcd


In general, these hold in the case of a symmetric, metric-compatible connection. Non-
metric differential geometry is beyond the scope of our present consideration.

We now define the so-called Lie derivative which can be used to define a
diffeomorphism invariant in

C . for a vector field U and a tensor field T , both

arbitrary, the invariant derivative represented (in component notation) by

g
m
m ab
h cd
b
m
g am
h cd
a
m
g mb
h cd
m
h
g ab
m cd
m
d
g ab
h cm
m
c
g ab
h md
m g ab
h cd m
g ab
h cd U
U T U T U T
U T U T U T U T T L

+ + + + =
...
...
...
....
...
...
...
...
...
...
...
...
...
...
...
...
...
...


defines the Lie derivative of T with respect to U . With the help of the torsion tensor
and the relation

85

| |
( )
c a
bc
a
bc
a
b
c a
cb
a
b
a
b
U U U U U = = 2

we can write

| | | | | |
| | | | | |
p g ab
m cd
m
hp
p g ab
h cm
m
dp
p g ab
h md
m
cp
p m ab
h cd
g
mp
p g am
h cd
b
mp
p g mb
h cd
a
mp
g
m
m ab
h cd
b
m
g am
h cd
a
m
g mb
h cd
m
h
g ab
m cd
m
d
g ab
h cm
m
c
g ab
h md
m g ab
h cd m
g ab
h cd U
U T U T U T
U T U T U T
U T U T U T
U T U T U T U T T L
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
....
...
...
...
...
...
...
...
...
...
...
...
...
2 ... 2 2
2 ... 2 2
...
...
+
+ + + +

+ + + + =


Hence, noting that the components of the torsion tensor, namely,
| |
i
kl
, indeed
transform as components of a tensor field, it is seen that the
s ij
r kl U
T L
...
...
do transform as
components of a tensor field. Apparently, the beautiful property of the Lie derivative
(applied to an arbitrary tensor field) is that it is connection-independent even in a curved
manifold.