Volume 4 PROGRESS IN PHYSICS October, 2007

SPECIAL REPORT

A Four-Dimensional Continuum Theory of Space-Time and

the Classical Physical Fields
Indranu Suhendro
Department of Physics, Karlstad University, Karlstad 651 88, Sweden
E-mail: spherical [email protected]

In this work, we attempt to describe the classical physical fields of gravity, electromag-
netism, and the so-called intrinsic spin (chirality) in terms of a set of fully geometrized
constitutive equations. In our formalism, we treat the four-dimensional space-time con-
tinuum as a deformable medium and the classical fields as intrinsic stress and spin fields
generated by infinitesimal displacements and rotations in the space-time continuum it-
self. In itself, the unifying continuum approach employed herein may suggest a possible
unified field theory of the known classical physical fields.

1 Introduction dimensional curvilinear coordinates of Riemannian geome-

try, nothing more. As such, any possible interaction between
Many attempts have been made to incorporate the so-called the physical fields (e.g., the interaction between gravity and
standard (Hookean) linear elasticity theory into general rela- electromagnetism) has not been investigated in detail.
tivity in the hope to describe the dynamics of material bodies In the present work, we develop a fully geometrized con-
in a fully covariant four-dimensional manner. As we know, tinuum theory of space-time and the classical physical fields
many of these attempts have concentrated solely on the treat- in which the actions of these physical fields contribute di-
ment of material bodies as linearly elastic continua and not rectly to the dynamics of the space-time geometry itself. In
quite generally on the treatment of space-time itself as a lin- this model, we therefore assume that a physical field is di-
early elastic, deformable continuum. In the former case, tak- rectly associated with each and every point in the region of
ing into account the gravitational field as the only intrinsic space-time occupied by the field (or, a material body in the
field in the space-time continuum, it is therefore true that case of gravity). This allows us to describe the dynamics of
the linearity attributed to the material bodies means that the the space-time geometry solely in terms of the translational
general consideration is limited to weakly gravitating objects and rotational behavior of points within the occupied region.
only. This is because the curvature tensor is in general quad- Consequently, the geometric quantities (objects) of the space-
ratic in the the so-called connection which can be said to time continuum (e.g., curvature) are directly describable in
represent the displacement field in the space-time manifold. terms of purely kinematic variables such as displacement,
However, in most cases, it is enough to consider an infinitesi- spin, velocity, acceleration, and the particle symmetries them-
mal displacement field only such that the linear theory works selves.
perfectly well. However, for the sake of generality, we need As we have said above, at present, for the sake of sim-
not assume only the linear behavior of the properly-stressed plicity, we shall assume the inherently elastic behavior of the
space-time continuum (and material bodies) such that the pos- space-time continuum. This, I believe, is adequate especially
sible limiting consequences of the linear theory can be readily in most cosmological cases. Such an assumption is nothing
overcome whenever it becomes necessary. Therefore, in the but intuitive, especially when considering the fact that we
present work, we shall both consider both the linear and non- do not fully know the reality of the constituents of the fab-
linear formulations in terms of the response of the space-time ric of the Universe yet. As such, the possible limitations of
geometry to infinitesimal deformations and rotations with in- the present theory, if any, can be neglected considerably until
trinsic generators. we fully understand how the fabric of the space-time contin-
A few past attempts at the full description of the elas- uum is actually formed and how the properties of individual
tic behavior of the space-time geometry in the presence of elementary particles might contribute to this formation.
physical fields in the language of general relativity have been
quite significant. However, as standard general relativity de- 2 The fundamental geometric properties of a curved
scribes only the field of gravity in a purely geometric fash- manifold
ion, these past attempts have generally never gone beyond
the simple reformulation of the classical laws of elasticity in Let us present the fundamental geometric objects of an n-
i
the presence of gravity which means that these classical laws dimensional curved manifold. Let !a = @X i
@xa Ei = @a X Ei
of elasticity have merely been referred to the general four- (the Einstein summation convention is assumed throughout

34 I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields
October, 2007 PROGRESS IN PHYSICS Volume 4

this work) be the covariant (frame) basis spanning the n- Given the existence of a local coordinate transformation
dimensional 1 xa = via xi = xi (x ) in C1 , the components of an arbitrary ten-

base manifold C with local coordinates
a k
= x X . The contravariant
(coframe) 
b
basis  is then given sor field T 2 C1 of rank (p; q ) transform according to
via the orthogonal projection b ; !a = ab , where ab are the
ab:::g ::: @ xa @ xb : : : @ xg @ x   : : : @ x
components of the Kronecker delta (whose value is unity if Tcd:::h = T:::    c  @d x h :
the indices coincide or null otherwise). The set of linearly in-
@ = @ gives i i :::i
dependent local directional derivatives Ei = @X i i Let j11 j22 :::jpp be the components of the generalized Kro-
the coordinate basis of the locally flat tangent space Tx (M) necker delta. They are given by
at a point x 2 C1 . Here M denotes the topological
space of 0 1
the so-called n-tuples h (x) = h x1 ; : : : ; xn such that rel- ji11 ji12 ::: ji1p
ative to a given chart (U; h (x)) on a neighborhood U of a B
ji21 ji22 ::: ji2p C
:::jp =2j1 j2 :::jp 2
ji11ji22:::i p i1 :::ip = det B C
local coordinate point, our C1 -differentiable manifold itself B
@ ::: ::: ::: :::
C
A
is a topological space. The dual basis to Ei spanning the lo- jip1 jip2 ::: jipp
cally flat cotangent space Tx (M) will then
be given by the
differential elements dX k via the relation dX k ; @i = ik . p
In fact and in general, the one-forms dX k indeed act as a
where 2j1 j2 :::jp = det (g ) j1 j2 :::jp and 2i1 i2 :::ip = p
i1 i2 :::ip
det(g)
linear map Tx (M) ! IR when applied to an arbitrary vector are the covariant and contravariant components of the com-
field F 2 Tx (M) of the explicit form F = F i @X @ = fa @ .
i @xa
pletely anti-symmetric Levi-Civita permutation tensor, re-
Then it is easy to see that F i = F X i and f a = F xa , from spectively, with the ordinary permutation symbols being
which we obtain the usual transformation laws for the con- given as usual by j1 j2 :::jq and i1 i2 :::ip . Again, if ! is an
travariant components of a vector field, i.e., F i = @a X i f a arbitrary tensor, then the object represented by
and f i = @i xa F i , relating the localized components of F to 1 i1 i2 :::ip
the  !j =  !

general  ones and vice versa. In addition, we also see that 1 j2 :::jp
p! j1 j2 :::jp i1 i2 :::ip
dX k ; F = F X k = F k .
The components of the symmetric metric tensor g = is completely anti-symmetric.
= gab a
b of the base manifold C1 are readily given by Introducing a generally asymmetric connection via the
covariant derivative
gab = h!a ; !b i
@b !a = c !
ab c
satisfying
gac g bc = ab i.e.,


c
ab = hc ; @b !a i = c + c
(ab) [ab]
where g ab = a ; b . It is to be understood that the covari-
ant and contravariant components of the metric tensor will where the round index brackets indicate symmetrization and
be used to raise and the (component) indices of vectors and the square ones indicate anti-symmetrization, we have, by
tensors. means of the local coordinate transformation given by xa =
The components of the metric tensor = xa (x ) in C1
g (xN ) = ik dX i
dX k @b ea = c e
ab c
 
 e e
a b
describing the locally flat tangent space Tx (M) of rigid where the tetrads of the moving frames are given by e a=
frames at a point xN = xN (xa ) are given by = @a x and ea = @ xa . They satisfy ea eb = ba and ea ea =
=  . In addition, it can also be verified that
ik = hEi ; Ek i = diag (1; 1; : : : ; 1) :
In four dimensions, the above may be taken to be the com- @ ea =   ea a eb ec ;
bc  
ponents of the Minkowski metric tensor, i.e., ik =hEi ; Ek i= @b ea = ea   eb a ec :
= diag (1; 1; 1; 1). cb 
Then we have the expression We know that is a non-tensorial object, since its com-
ponents transform as
gab = ik @a X i @b X k :
The line-element of C1 is then given by
c
ab = ec @b ea + ec  
 e e :
a b


ds2 = g = gab @i xa @k xb dX i
dX k However, it can be described as a kind of displacement
field since it is what makes possible a comparison of vectors
where a = @i xa dX i . from point to point in C1 . In fact the relation @b !a = cab !c

I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields 35
Volume 4 PROGRESS IN PHYSICS October, 2007

defines the so-called metricity condition, i.e., the change (dur- where r
^ denotes covariant differentiation with respect to the
ing a displacement) in the basis can be measured by the basis Christoffel symbols alone, and where
itself. This immediately translates into
B d (
) = @
d @
d +
e
d
e
d
abc b ac c ab ac eb ab ec
rc gab = 0 are the components of the Riemann-Christoffel curvature ten-
1
where we have just applied the notion of a covariant derivative sor of C .
to an arbitrary tensor field T : From the components of the curvature tensor, namely,
Rd abc , we have (using the metric tensor to raise and lower
rm T ab:::g = @m T ab:::g + a T pb:::g + b T ap:::g + : : : indices)
cd:::h cd:::h pm cd:::h pm cd:::h
g ab:::p
: : : + pm Tcd:::h p ab:::g
cm Tpd:::h
p ab:::g Rab  Rc acb = Bab (
) + r^c Kab
c c Kd
Kad
dm Tcp:::h : : : cb

::: p ab:::g 2r^b [ac] + 2Kab d[cd]

c c
hm Tcd:::p

such that (@m T )cd:::h = rm Tcd:::h .

ab:::g ab:::g R  Raa = B (
) 4 g ab r
^a
c
[bc]
The condition rc gab = 0 can be solved to give 2 g [ab] d[cd]
ac b Kabc K acb
c 1 where Bab (
)  Bacb c (
) are the components of the sym-
ab = g cd (@b gda @d gab + @a gbd ) +
2   metric Ricci tensor and B (
)  Baa (
) is the Ricci scalar.
+ c[ab] g cd gae e[db] + gbe e[da] Note that Kabc  gad Kbcd and K acb  g cd g be K a .
de
Now since
from which it is customary to define  p 
b =
bba =
bab = @a ln det (g )
ba
1

cab = g cd (@b gda @d gab + @a gbd ) b
 p 
b
2 ab = @a ln det (g ) + 2 [ab]
as the Christoffel symbols (symmetric in their two lower in-
we see that for a continuous metric determinant, the so-called
dices) and
homothetic curvature vanishes:
 
c = c
Kab [ab] g cd gae e[db] + gbe e[da] Hab  Rc cab = @a c
cb @b c
ca = 0:
as the components of the so-called cotwist tensor (anti- Introducing the traceless Weyl tensor W , we have the fol-
symmetric in the first two mixed indices). lowing decomposition theorem:
Note that the components of the twist tensor are given by 1 d

Rdabc = Wabc
d +  R + g Rd  d R g Rd +
1   n 2 b ac ac b c ab ab c
a = ea @c eb @b ec + eb  c ec  b
[bc] 1

2 + d g d g R
(n 1) (n 2) c ab b ac
where we have set    
c =  ec , such that for an arbitrary which is valid for n > 2. For n = 2, we have
scalar field  we have


Rdabc = KG bd gac cd gab
(ra rb rb ra )  = 2 [ab] rc  :
c
where
The components of the curvature tensor R of C1 are then KG = R
1
given via the relation 2
is the Gaussian curvature of the surface. Note that (in this
(rq rp rp rq ) Tcd:::r
ab:::s = T ab:::s Rw + T ab:::s Rw +
wd:::r cpq cw:::r dpq case) the Weyl tensor vanishes.
ab:::s Rw
+ : : : + Tcd:::w wb:::s Ra
Tcd:::r aw:::s Rb
Tcd:::r Any n-dimensional manifold (for which n > 1) with con-
rpq wpq wpq
stant sectional curvature R and vanishing twist is called an
::: ab:::w Rs
Tcd:::r wpq 2 w r ab:::s
[pq] w Tcd:::r Einstein space. It is described by the following simple rela-
tions:
where 1

Rdabc = d g cd gab R ;
n(n 1) b ac
Rdabc = @b d
ac @c d + e d
ab ac eb
e d
ab ec
1
= B dabc (
) + r d r
^b Kac ^c Kab + Kac
d e Kd e Kd ;
Kab Rab = g R:
eb ec n ab

36 I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields
October, 2007 PROGRESS IN PHYSICS Volume 4

In the above, we note especially that we can write

Rdabc = B dabc (
) ;

ab:::g
LU Tcd:::h = rm Tcd:::h
ab:::g m ab:::g
U + Tmd:::h rc U m +
Rab = Bab (
) ;
ab:::g
+ Tcm:::h rd U m + : : : + Tcd:::m
ab:::g
rh U m mb:::g
Tcd:::h rm U a
R = B (
) :
am:::g
Tcd::::h rm U b ::: ab:::m r U g +
Tcd:::h m

+2 a T mb:::g U p + 2 b T am:::g U p +
Furthermore, after some lengthy algebra, we obtain, in [mp] cd:::h [mp] cd:::h
g ab:::m U p 2 m T ab:::g U p +
general, the following generalized Bianchi identities: : : : + 2 [mp] Tcd:::h [cp] md:::h
Rabcd + Racdb + Radbc = 2 @d a[bc] + @ b a[cd] + ab:::g p : : : 2 m T ab:::g U p :
+2 m [dp] Tcm:::h U [hp] cd:::m


+ @c a[db] + aeb e[cd] + aec e[db] + aed e[bc] ; Hence, noting that the components of the twist tensor,
namely, i[kl] , indeed transform as components of a tensor
re Rabcd + rc Rabde + rd Rabec = ij:::s
field, it is seen that the LU Tkl:::r do transform as components
=2 f a f a f a
of a tensor field. Apparently, the beautiful property of the
[cd] R bfe + [de] R bfc + [ec] R bfd ; Lie derivative (applied to an arbitrary tensor field) is that it is
  connection-independent even in a curved manifold.
1 ab
ra Rab
2
g R = 2 g ab c Rd + a Rcdb
[da] c [cd] a We will need the identities derived in this Section later on.

for any metric-compatible manifold endowed with both cur- 3 The generalized four-dimensional linear constitutive
vature and twist. field equations
In the last of the above set of equations, we have intro-
duced the generalized Einstein tensor, i.e., We shall now present a four-dimensional linear continuum
theory of the classical physical fields capable of describing
1
Gab  Rab g R microspin phenomena in addition to the gravitational and
2 ab electromagnetic fields. By microspin phenomena, we mean
those phenomena generated by rotation of points in the four-
In particular, we also have the following specialized iden-
dimensional space-time manifold (continuum) S4 with local
tities, i.e., the regular Bianchi identities:
coordinates x in the manner described by the so-called
a a a
B bcd + B cdb + B dbc = 0 ; Cosserat continuum theory.
We start with the following constitutive equation in four
r B + r B + r B = 0;
^
e
a
bcd
^ a
c ^ bde
a
d bec
dimensions:  
1  1 
  T  = C  D = R g R
1  2
r^a B ab
2
g ab B = 0 :
where now the Greek indices run from 0 to 3. In the above
equation, T  are the contravariant components of the gener-
In general, these hold in the case of a symmetric, metric- ally asymmetric energy-momentum tensor, C 
are the
compatible connection. Non-metric differential geometry is mixed components of the generalized four-dimensional elas-
beyond the scope of our present consideration. ticity tensor, D are the contravariant components of the
We now define the so-called Lie derivative which can be four-dimensional displacement gradient tensor, R are the
used to define a diffeomorphism invariant in C1 . for a vec- contravariant components of the generalized (asymmetric)
tor field U and a tensor field T , both arbitrary, the invariant four-dimensional Ricci curvature tensor,  = 8  is the Ein-
derivative represented (in component notation) by 
stein coupling constant (in geometrized units), and R = R is
ab:::g ab:::g m ab:::g the generalized Ricci four-dimensional curvature scalar.
LU Tcd:::h = @m Tcd:::h U + Tmd:::h @c U m + Furthermore, we can decompose our four-dimensional
ab:::g ab:::g
+ Tcm:::h @d U m + : : : + Tcd:::m @h U m elasticity tensor into its holonomic and anholonomic parts as
follows:
mb:::g am:::g
Tcd:::h @m U a Tcd::::h @m U b ::: ab:::m @ U g
Tcd:::h m C  = A + B 
defines the Lie derivative of T with respect to U . With the where
help of the twist tensor and the relation A = A(()) = A

@b U a = rb U a a Uc = rb U a a 2 a
U c B  = B [ []] = B 

cb bc [bc]

I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields 37
Volume 4 PROGRESS IN PHYSICS October, 2007

such that dimensional macroscopic displacement field vector,  are

C  = C  : the components of the four-dimensional microscopic (micro-
polar) intrinsic spin vector, e is a constant proportional to the
Therefore, we can express the fully covariant components electric charge, and ' are the components of the electromag-
of the generalized four-dimensional elasticity tensor in terms netic four-potential vector. We assume that in general   ,  ,
of the covariant components of the symmetric metric tensor and ' are linearly independent of each other.
g (satisfying, as before, g g  =  ) as The intrinsic four-dimensional macroscopic spin (“angu-
lar momentum”) tensor is then given by
C = g g + g g + g g =
1
= g g +  (g g + g g ) + ! (g g g g )
 = (r  r  ) :
2
where ,  , , , and ! are constitutive invariants that are not Likewise, the intrinsic four-dimensional microscopic (mi-
necessarily constant. It is therefore seen that cropolar) spin tensor is given by
A = g g +  (g g + g g ) 1
S = (r  r  ) :
2
B = ! (g g g g )
Note that this tensor vanishes when the points are not al-
An infinitesimal displacement (diffeomorphism) in the lowed to rotate such as in conventional (standard) cases.
space-time manifold S4 from an initial point P to a neigh- Meanwhile, the electromagnetic field tensor is given by
boring point Q is given as usual by
F = r ' r ' :
x (Q) = x (P ) +  
In this case, we especially note that, by means of the con-
where   are the components of the four-dimensional infinite- dition  
[ ]  = 0, the above expression reduces to the usual
simal displacement field vector. The generally asymmetric Maxwellian relation
four-dimensional displacement gradient tensor is then given
by F = @ ' @ ' :
D = r  :
We can now write the intrinsic spin tensor as
The decomposition D = D( ) + D[ ] and the sup-
plementary infinitesimal point-rotation condition  
[ ]  = 0 ! =
 + S + eF :
allow us to define the symmetric four-dimensional displace-
ment (“dilation”) tensor by Hence the full electromagnetic content of the theory be-
comes visible. We also see that our space-time continuum can
1 1
 = D( ) = (r  + r  ) = L g be considered as a dynamically polarizable medium possess-
2 2 ing chirality. As such, the gravitational and electromagnetic
from which the “dilation” scalar is given by fields, i.e., the familiar classical fields, are intrinsic geometric
objects in the theory.
1
 =  = D = g  L g = r   Furthermore, from the cotwist tensor, let us define a geo-
2 metric spin vector via
as well as the anti-symmetric four-dimensional intrinsic spin
(vorticity) tensor by A  K
 =2  :
[]
1
! = D[ ] = (r  r  ) : Now, in a somewhat restrictive case, in connection with
2 the spin fields represented by   ;  ; and ' , the selection
Let us now decompose the four-dimensional infinitesimal
displacement field vector as follows: A = c1  + c2  + 2 ec3 ' = 2 
i.e.,
 = @F + :
2 = c1  ++c2++22ee'c3 '
Here the continuous scalar function F represents the in-   
tegrable part of the four-dimensional macroscopic displace- will directly attribute the cotwist tensor to the intrinsic spin
ment field vector while the remaining parts are given by fields of the theory. However, we would in general expect the
 via
    intrinsic spin fields to remain in the case of a semi-symmetric
=  +  + 2 e' connection, for which A = 0 and so we cannot carry this
where   are the components of the non-integrable four- proposition any further.

38 I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields
October, 2007 PROGRESS IN PHYSICS Volume 4

At this point, we see that the holonomic part of the gen- divergence
eralized four-dimensional elasticity tensor given by A is
responsible for (centrally symmetric) gravitational phenom- r T  = g r () +  r D +
ena while the anholonomic part given by B owes its ex- + r D + D r  + D r
istence to the (con)twist tensor which is responsible for the
existence of the intrinsic spin fields in our consideration. need not vanish in general since
Furthermore, we see that the components of the energy-  
1 1 
momentum tensor can now be expressed as r T = r R


 g R =
2
T = g  +  D + D : 1      R

= 2g [] R + []  :
In other words,

In an isotropic, homogeneous Universe, for which the
T( ) = g  + ( + )  ; constitutive invariants ; ; ; ; and ! are constant, the
above expression reduces to
T[ ] = ( ) ! :
r T  = g r  +  r D + r D :
Alternatively,
If we require the above divergence to vanish, however, we
1 1
T( ) = g g  L g + ( + ) L g ; see that the motion described by this condition is still more
2 2 general than the pure geodesic motion for point-particles.
Still in the case of an isotropic, homogeneous Universe,
T[ ] = ( ) (
 + S + eF ) :
possibly on large cosmological scales, then our expression
We may note that, in a sense analogous to that of the or- for the energy-momentum tensor relates the generalized Ricci
dinary mechanics of continuous media, the generally asym- curvature scalar directly to the “dilation” scalar. In general,
metric character of the energy-momentum tensor means that we have
a material object in motion is generally subject to distributed 1
R =  (4  +  + )  =   =   g  L g :
body couples. 2
We also have
Now, for the generalized Ricci curvature tensor, we obtain
1
T = T  = (4  +  + )  =
 R: the following asymmetric constitutive field equation:
  
1
R =  T g T =  ( g +  D + )
Let us briefly relate our description to the standard mate-
2 
rial description given by general relativity. For this purpose,
let us assume that the intrinsic spin fields other than the elec- where
1
tromagnetic field are negligible. If we denote the material = (2 +  + )  :
density and the pressure by  and p, respectively, then it can 2
be directly verified that In other words,


 4p R( ) =  g + ( + )  ;
=
4 +  +
is a solution to the ordinary expression R[ ] =  ( ) ! :
T( ) =  u u pg Inserting the value of , we can alternatively write
   
1 1 1
F F g F F  R( ) = 8 g + ( + ) L g
4   4   2
where u are the covariant components of the unit veloc- R[ ] = 8 ( ) (
 + S + eF ) :
ity vector. This is true whether the electromagnetic field is
present or not since the (symmetric) energy-momentum ten- Hence, the correspondence between the generalized Ricci
sor of the electromagnetic field given by curvature tensor and the physical fields in our theory becomes
  complete. The present theory shows that in a curved space-
1  1  time with a particular spherical symmetry and in a flat Min-
J = F F g F F
4   4   kowski space-time (both space-times are solutions to the
is traceless. equation  = 0, i.e., L g = 0) it is in general still pos-
At this point, however, we may note that the covariant sible for the spin fields to exist. One possible geometry that

I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields 39
Volume 4 PROGRESS IN PHYSICS October, 2007

complies with such a space-time symmetry is the geome- formulation presented in the preceding section. The result-
try of distant parallelism with vanishing space-time curvature ing non-linear constitutive field equations will therefore not
(but non-vanishing Riemann-Christoffel curvature) and non- be limited to weak fields only. In general, it can be shown
vanishing twist. that the full curvature tensor contains terms quadratic in the
Now let us recall that in four dimensions, with the help of displacement gradient tensor and this gives us the reason to
the Weyl tensor W , we have the decomposition express the energy-momentum tensor which is quadratic in
the displacement gradient tensor.
R = W + We start with the non-linear constitutive field equation
1  
+ (g R + g R g R g R ) + T  = C  D + K  D D = 1 R 1 g  R
2    2
1
+ (g g g g ) R : where
6
We obtain, upon setting  = 12 ,  = 12  ,  = 12  , K = a1 g g g + a2 g g g +
 = 16  
and  + a3 g g g + a4 g g g + a5 g g g +
R = W + 2 (g g g g ) + + a6 g g g + a7 g g g + a8 g g g +
+  (g D + g D g D g D ) + + a9 g g g + a10 g g g + a11 g g g +
+ a12 g g g + a13 g g g + a14 g g g +
+  (g D + g D g D g D ) +
+ a15 g g g
+  (g g g g )  :
where the fifteen constitutive invariants a1 , a2 , . . . , a15 are
Therefore, in terms of the anholonomic part of the gener- not necessarily constant.
alized elasticity tensor, we have We shall set

R = W + 2 B + K = K = K = K :
!
+  (g D + g D g D g D ) + Letting
+  (g D + g D g D g D ) + K = P + Q ;
P = P( )()() ;
+  (g g g g )  :
In the special case of a pure gravitational field, the twist
Q = Q[ ][][] ;
we have
of the space-time continuum vanishes. In this situation our
intrinsic spin fields vanish and consequently, we are left P = P = P = P ;
simply with
Q = Q = Q = Q :
R = W + Introducing the eleven constitutive invariants b1 , b2 , . . . ,
1
g D ) + b11 , we can write


+  +  (g D + g D g D
2 K = b1 g g g + b2 g (g g + g + g ) +
+  (g g g g )  :
+ b3 g (g g g g ) + b4 g (g g + g g ) +
In standard general relativity, this gives the explicit form
of the Riemann-Christoffel curvature tensor in terms of the + b5 g (g g g g ) + b6 g (g g + g g ) +
Lie derivative L g = 2 . For a space-time satisfying the + b7 g (g g g g ) + b8 g (g g + g g ) +
symmetry L g = 0, we simply have R = W , i.e.,
the space-time is devoid of material sources or “empty”. This
+ b9 g (g g g g ) + b10 g (g g + g g ) +
condition is relatively weaker than the case of a space-time + b11 g (g g g g ) :
with constant sectional curvature, R = const. for which the
Weyl tensor vanishes. The energy-momentum tensor is therefore given by


T =  + b1 2 +2 b2   +2 b3 ! !  g +
4 The generalized four-dimensional non-linear constitu-
tive field equations +  D + D + 2 (b4 + b6 )  +
+ 2 (b5 + b7 )  ! + 2 b8 D  + 2 b9 D ! +
In reference to the preceding section, let us now present, in
a somewhat concise manner, a non-linear extension of the + 2 b10 D  + 2 b11 D ! :

40 I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields
October, 2007 PROGRESS IN PHYSICS Volume 4

In other words, The generalized Ricci curvature scalar is then

T( ) =   + b1 2 + 2b2   + 2b3 ! ! 
g +

R =  h1  + h2 2 + h3   + h4 ! ! 
+ ( + )  2 (b4 + b6 )  + (b8 + b10 )  where


h1 = 4 c1 + c5 + c6 ;
 D  + D  + (b9 + b11 ) D ! + D ! ;
h2 = 4 c2 + c5 ;
T[ ] = ( ) ! + 2 (b4 + b6 )  ! + (b8 + b10 ) 


h3 = 4 c3 + c9 + c11 ;
 D  D  + (b9 + b11 ) D ! D ! :
h4 = 4 c4 + c10 + c12 :
We also have
Finally, we obtain, for the curvature tensor, the following
T = 1  + 2 2 + 3   + 4 ! !  expression:
where we have set
R = W +
1 = 4  +  + ;

+ f1  + f2 2 + f3   + f4 ! !  
 2 = 4 b1 + 2 ( b4 + b6 ) ;

3 = 8 b2 + 2 (b8 + b10 ) ;
 (g g g g ) +  + f5  g  + g 



g  g  +  + f6  g ! + g !
 4 = 8 b3 + 2 ( b9 b11 ) ;

for the sake of simplicity. g ! g ! +  g D + g D
For the generalized Ricci curvature tensor, we obtain

n
g D g D + f7 D  g +


R =  c1  + c2 2 + c3   + c4 ! !  g + + D  g D  g D  g +

+ c5 D + c6 D + c7  + c8 ! + c9 D  + + f8 D ! g + D ! g D ! g
o


+ c10 D ! + c11 D  + c12 D ! D ! g + f9 D  g + D  g


where D  g D  g + f10 D ! g +
1
c1 = (2 +  + ) ; c7 = 2 (b4 + b6 ) ;

2 + D ! g D ! g D ! g
c2 = (b1 + b4 + b6 ) ; c8 = 2 (b5 + b7 ) ;
where
c3 = (2b2 + b8 + b10 ) ; c9 = 2b8 ;
f1 = c1 =  +  ; f6 = c8 ;
c4 = (2b3 + b9 b11 ) ; c10 = 2b9 ;  
2 1
c5 =  ; c11 = 2b10 ; f2 = 1  c + c ; f7 = c9 ;

3  2 6 7
c6 = ; c12 = 2b11 ; 2 1
f3 = 1  c +  (c9 + c11 ) ; f8 = c10 ;
i.e.,

3  3 6
n 2 1
R( ) =  c1  + c2 2 + c3   + f4 = 1  c4 +  (c10 c12 ) ; f9 = c11 ;
3 6


+ c4 ! !  g + (c5 + c6 )  + c7  + f5 = c7 ; f10 = c12 :
1

+ (c9 + c11 ) D  + D  + At this point, the apparent main difficulty lies in the fact
2
1
o that there are too many constitutive invariants that need to be
+ (c10 + c12 ) D ! + D ! ;
2 exactly determined. As such, the linear theory is compara-
n tively preferable since it only contains three constitutive in-
R[ ] =  (c5 c6 ) ! + c8  ! + variants. However, by presenting the most general structure
1
of the non-linear continuum theory in this section, we have
+ (c9 + c11 ) D  D  + acquired a quite general picture of the most general behavior
2
1
o of the space-time continuum in the presence of the classical
+ (c10 +c12 ) D ! D ! : fields.
2

I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields 41
Volume 4 PROGRESS IN PHYSICS October, 2007

5 The equations of motion where the operator Ds D = u r represents the absolute co-

variant derivative. In the above equations we have introduced
Let us now investigate the local translational-rotational mo- the following invariants:
tion of points in the space-time continuum S4 . Consider an    1=2
Du Du
infinitesimal displacement in the manner described in the pre-  = g ;
ceding section. Keeping the initial position fixed, the unit ve- Ds Ds
locity vector is given by Dv  
 = 2 u v  z ;
d  dx Ds
u = = ;  
ds ds Dz  Dz  1=2
' = g :
1 = g u u ; Ds Ds
In particular, we note that, the twist scalar  measures the
such that, at any proper time given by the world-line s, the twist of the curve ` in S4 due to microspin.
parametric representation At this point, we see that our equations of motion describe
d  = u (x ; s) ds a “minimal” geodesic motion (with intrinsic spin) when  =0.
In other words, if
describes space-time curves whose tangents are everywhere Du
directed along the direction of a particle’s motion. As usual, = 0;
Ds
the world-line can be parametrized by a scalar & via s =
Dv 
= a& + b, where a and b are constants of motion. =  w ;
The local equations of motion along arbitrary curves in Ds
the space-time continuum S4 can be described by the quadru- Dw
=  v  + 'z  ;
plet of unit space-time vectors (u; v; w; z ) orthogonal to each Ds
other where the first three unit vectors, or the triplet (u; v; w), Dz 
= 'w :
may be defined as (a set of) local tangent vectors in the (three- Ds
dimensional) hypersurface  (t) such that the unit vector z is
However, in general, any material motion in S4 will not
normal to it. More explicitly, the hypersurface  (t) is given
follow the condition  = 0. This is true especially for the
as the time section t = x0 = const of S4 . This way, the equa-
motion of a physical object with structure. In general, any
tions of motion will be derived by generalizing the ordinary
physical object can be regarded as a collection of points (with
Frenet equations of orientable points along an arbitrary curve
different orientations) obeying our general equations of mo-
in three-dimensional Euclidean space, i.e., by recasting them
tion. It is therefore clear that  , 0 for a moving finite phys-
in a four-dimensional manner. Of course, we will also include
ical object (with structure) whose material points cannot be
effects of microspin generated by the twist of space-time.
homogeneously oriented.
With respect
to the anholonomic space-time basis ! = Furthermore, it can be shown that the gradient of the unit
= ! x (X k ) = ei @X@ i , we can write velocity vector can be decomposed according to
u = u ! ; r u =  +  + 16 h  + u a
v = v  ! ; where
w = w  ! ;
h = g u u ;
z = z  ! ;
1
we obtain, in general, the following set of equations of motion  = h h (r u + r u ) =
of points, i.e., point-like particles, along an arbitrary curve ` 4
1   ^  1
in the space-time continuum S4 : = h h r u + r ^ u h h K  u ;
4 2   () 
Du 1
=  v ;  = h h (r u r u ) =
Ds 4
Dv  1   ^  1
=  w  u ; = h h r u r ^ u h h K  u ;
Ds 4 2   [] 
Dw  = r u ;
=  v + ' z ;
Ds
Dz  Du
= 'w ; a = :
Ds Ds

42 I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields
October, 2007 PROGRESS IN PHYSICS Volume 4

Note that We then arrive at the following invariant integral:

h u =  u =  u = 0; Z 



 
I= T  r(  )  + T  r(  ) ! +
K( ) = g  g [ ] + g [] ; S4
1 1
K[ ] = [ ] : + A   + B  ! !  +
2 2
1 1
Meanwhile, with the help of the identities + P     + Q ! !  ! 


3 3
u r r u = r u r u (r u ) r u = 

D
T + u (r  ) (f   u ) d 
= r a (r u ) r u ;
p
u (r r r r ) u = R u u 2  u
[ ] r
 u ; where d = det (g ) dx0 dx1 dx2 dx3 is the proper four-
dimensional differential
p volume.
we obtain
Writing L = det (g ) L and employing the variational
D
= r a (r u )(r u ) R u u +2 [ ] u r u principle, we then have
Ds Z 
@L @L @L
for the “rate of shear” of a moving material object with re- I = T  +    +  !  +
@T  @ @!
spect to the world-line. S4 
@L
+  (r  ) d  = 0 :
6 The variational principle for the theory @ (r  )
Now
Let us now derive the field equations of the present theory Z Z  
@L @L
by means of the variational principle. Considering thermody-
 (r  ) d  = r  d 
namic effects, in general, our theory can best be described by @ (r  ) @ (r  ) 
the following Lagrangian density: S4 S4
Z   Z  
@L @L
L = L 1 + L 2 + L 3 r  d  =
@ (r  ) 
r  d 
@ (r  ) 
where S4 S4
1p
L 1 = det (g )  since the first term on the right-hand-side of the first line is
  an absolute differential that can be transformed away on the
1

 R (r  D )
2
 D R ; boundary of integration by means of the divergence theorem.
Hence we have
 Z 
p 1  @L @L @L
L 2 = D D +
det (g )
C I = T  +    +  ! 
2   
@T  @ @!
S4
1       
+ K  D D D D 
T ; @L
3 r @ (r  )
 d  = 0
p
L 3 = det (g ) u (r  ) (f  u ) ; where each term in the integrand is independent of the others.
where  is a thermal coefficient,
T is (the change in) the We may also note that the variations T  ,   , !  , and
temperature, and f is a generally varying scalar entity. Note  are arbitrary.
that here we have only explicitly assumed that  = r   . From @T@L = 0, we obtain
Alternatively, we can express L as follows:
 
 = r(  ) ;
1p 1 
L 1 = det (g ) R g R (r  D ) : ! = r[  ] ;
 2
Hence we have i.e., the covariant components of the “dilation” and intrinsic
 spin tensors, respectively.
p
L = det (g ) T  (r  D ) + From @@L = 0, we obtain
 
1 1 1 ( ) 1 
+ C  D D + K  D D D T ( ) = R g R =
2 3   2
D 
T + u (r  ) (f  u ) :
  = A  + P    g 
T

I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields 43
Volume 4 PROGRESS IN PHYSICS October, 2007

i.e., the symmetric contravariant components of the energy- If we restrict our attention to point-like particles, the body
momentum tensor. force vanishes since it cannot act on a structureless (zero-
In other words, dimensional) object. And since the motion is geodesic, i.e.,
  a = 0, we have the conservation law
1  1 
T  = R g R = r T  = 0 :
 2
In this case, this conservation law is true regardless of
= C  D + K  D D  g 
T :
whether the energy-momentum tensor is symmetric or not.
Finally, we now show in detail that the fourth variation Let us now discuss the so-called couple stress, i.e., the
yields an important equation of motion. We first see that couple per unit area which is also known as the distributed
moment. We denote the couple stress tensor by the second-
@L   
rank tensor field M . In analogy to the linear constitutive re-
= T  + u f u :
@ (r  ) lations relating the energy-momentum tensor to the displace-
ment gradient tensor, we write
Hence      
  M =J  L +H  L L
@L
r @ (r  )
= r T  + r (f u )   + where
J = E + F ;
+ f u r  r (u ) u u r u :
H = U + V :
Let us define the “extended” shear scalar and the mass These are assumed to possess the same symmetry proper-
current density vector, respectively, via ties as C and K , respectively, i.e., E have the
same symmetry properties as A , F have the same
l = r (fu ) ;
symmetry properties as B , U have the same sym-
metry properties as P , and V have the same
J  =  u :
symmetry properties as Q .
We can now readily identify the acceleration vector and Likewise, the asymmetric tensor given by
the body force per unit mass, respectively, by L = L( ) + L[ ]
Du
a = u r a = ; is comparable to the displacement gradient tensor.
Introducing a new infinitesimal spin potential via  , let
Ds
the covariant dual form of the intrinsic spin tensor be
1 

b =

l  + f (1 r J  ) u : given by
1
!  =
2
2 ! = 12 (r  r  ) :
In the conservative case, the condition r J = 0 gives
Let us now introduce a completely anti-symmetric third-
D
=  r u : rank spin tensor via
Ds 1

S  = ( ) 2  :
In the weak-field limit for which u = 1; uA (where 2
A = 1; 2; 3) we obtain the ordinary continuity equation, As a direct consequence, we see that
@
r S  = ( ) ! 
+ rA  uA = 0 :
@t In other words,
1  [ ] 
Finally, we have
Z
r S  = T [ ] N  =

R 
(r T  +  b  a )  d  = 0 where
S4 N  = 2 (b4 + b6 )  !  + (b8 + b10 ) 



i.e., the equation of motion  D  D  + (b9 + b9 ) D !  D !  ;
r T  =  (a b ) 1

 = c8  !  + (c9 + c11 ) D  D  +
or   2
1  1

r R
2
g R =  (a b ) : + (c10 + c11 ) D !  D !  :
2

44 I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields
October, 2007 PROGRESS IN PHYSICS Volume 4

We can now form the second Lagrangian density of our in necessary detail.
theory as Firstly,

p 1 @H
H = det (g ) M  (r  L ) + J  L L + = M  2 S  u + u (h Is ) :
2 @ (r  )
1 
+ H  L L L 2  (r  ) S  u +

3 
Then we see that
 
+ u (r  ) (h Is )
  
r @H
= r M 
@ (r  )
where h is a scalar function, I is the moment of inertia, and 2 T [] u 2 S  r u + r (hu )  +
s are the components of the angular velocity vector.
Letting L( ) = X and L[ ] = Z , the corresponding + hu r  I r (u ) s Iu r s :
action integral is
Z  We now define the angular acceleration by



J= M  r(  ) X + M  r[  ] Z +
 = u r s =
Ds
S4 Ds
1 1
+ E  X X  + F  Z Z  + and the angular body force per unit mass by
2 2  
1  1 1   D 
+ U  X X  X  + V  Z Z  Z   = l + h I (r J  ) s
3 3  Ds

2 (r  ) S  u + u (r  ) (h Is ) d  : where l = r (hu ).
We have
p Z 
As before, writing H = det (g )H and performing the  
variation J = 0, we have r M  2 T [] u + S  r u +
Z  S4 
@H @H
J = M  +  X  + +   I  d  = 0 :
@M  @X
S4   
@H @H Hence we obtain the equation of motion
+
@Z 
Z  r @ (r  )
 d  = 0
r M  = 2 
with arbitrary variations M , X , Z , and  .   
From @M @H = 0, we obtain
  T [] N  u + S  r u +  (I  )
X = r(  ) ; i.e.,

Z = r[  ] : r M  = 2 

 
@H 1  [] 
From @X  = 0, we obtain  
R  u + S  r u +  (I  ) :
M ( ) = E  X  + U  X  X  :
@H
From @Z = 0, we obtain 7 Final remarks


We have seen that the classical fields of physics can be uni-

M [ ] = F  Z  + V  Z  Z  :
fied in a simple manner by treating space-time itself as a
We therefore have the constitutive relation four-dimensional finite (but unbounded) elastic medium ca-
pable of undergoing extensions (dilations) and internal point-
M  = J  L + H  L L : rotations in the presence of material-energy fields. In the
present framework, the classical physical fields indeed appear
Let us investigate the last variation on an equal footing as they are of purely geometric character.
Z   In addition, we must note that this apparent simplicity still
@H
r @ (r  )
 d = 0 leaves the constitutive invariants undetermined. At the mo-
S4
ment, we leave this aspect of the theory to more specialized

I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields 45
Volume 4 PROGRESS IN PHYSICS October, 2007

attempts. However, it can be said, in general, that we ex- possible effects arising from this consideration, it is often not
pect the constitutive invariants of the theory to be functions of sufficient to model the space-time continuum itself as contin-
the known physical properties of matter such as material den- uous, isotropic, and homogeneous. Furthermore, the rather
sity, energy density, compressibility, material symmetry, etc. predominant presence of twisting paths may give rise to par-
This way, we have successfully built a significant theoretical ticles exhibiting micropolar structure.
framework that holds in all classical physical situations. In geometrizing microspin phenomena, we emphasize
We would also like to remark that once the constitutive that the initial microspin variables are not to be freely chosen
invariants are determined and incorporated into the possible to be included in the so-called elasticity scalar functional of
equations of state, the fully non-linear formulation of the pre- the space-time continuum which is equivalent to a Lagrangian
sent theory should be very satisfactory for describing the dy- density. Rather, one must first identify them with the internal
namics of astrophysical objects especially various fluids geometric properties of the space-time continuum. In other
which exhibit the characteristics of non-degenerate relativis- words, one must primarily unfold their underlying geometric
tic and non-Newtonian fluids. existence in the space-time continuum itself. This is precisely
We have seen that the general dynamical behavior of a what we have done in this work.
material body as determined by the equations of motion given Finally, we note that geometric discontinuities can also
in Section 5, is intrinsically related to the underlying geome- be incorporated into the present theory. Such discontinuities
try of the space-time continuum which in turn is largely deter- can be seen as topological defects in the space-time contin-
mined by the constitutive relations given in Sections 3 and 4. uum. Holographic four-dimensional continua with cellular,
In Section 6, we have also constructed a framework in which fibrous, or foamy structure may indeed represent admissible
the motion of a point-like particle is always subject to the con- semi-classical models of the Universe which can be realized
servation law of matter and energy regardless of the particle’s in the framework of the present theory. In such a case, the
intrinsic spin. metric must therefore be quantized. It remains to be seen how
We also note that a material body in our continuum this might correspond to any conventional quantum descrip-
representation of space-time can be regarded as the three- tion of the space-time continuum.
dimensional boundary of a so-called world-tube such that out-
side the world-tube the region is said to be free or empty. This Acknowledgements
three-dimensional boundary can be represented by a time-like
hypersurface. Such hypersurfaces can be seen as disturbances I would like to sincerely thank D. Rabounski and S. J. Cro-
in the space-time continuum. Furthermore, such disturbances thers for their kind assistance and the numerous discussions
are equivalent to three-dimensional representations of mate- devoted to making this work appear in a somewhat more read-
rial waves (not necessarily gravitational waves). In this con- able form.
text, one may formulate the dynamic discontinuity conditions Submitted on June 22, 2007
as purely geometric and kinematic compatibility conditions Accepted on July 04, 2007
over the hypersurfaces.
In common with standard general relativity, a region of References
the space-time continuum is said to be statical if it can be
covered by a space-time coordinate system relative to which 1. Landau L. D. and Lifshitz E. M. Theory of elasticity. Pergamon,
the components of the metric tensor are independent of time. 1975.
It may be that such a region can be covered by one or more 2. Forest S. Mechanics of Cosserat media — an introduction.
such coordinate systems. As such, material waves are propa- Ecole des Mines de Paris, Paris, 2005.
gated into a fixed (three-dimensional) curved space along tra- 3. Sakharov A. D. Dokl. Akad. Nauk USSR, 1967, v. 177, 70.
jectories normal to the family of hypersurfaces given by the 4. Sokolov S. N. Gen. Rel. Grav., 1995, v. 27, 1167.
successive positions of a material body in the fixed space. In
various cases, such trajectories can be represented as curves
of zero length in the space-time continuum.
The microscopic substructure of the space-time contin-
uum provides us room for additional degrees of freedom. In
other words, there exist intrinsic length scales associated with
these additional degrees of freedom. Correspondingly, one
may define the so-called microrotational inertial field. In fact,
the internal rotation of the points in the space-time contin-
uum is seen as representing the intrinsic spin of elementary
particles. On microscopic scales, the structure of the space-
time continuum can indeed appear to be granular. Due to

46 I. Suhendro. A Four-Dimensional Continuum Theory of Space-Time and the Classical Physical Fields