Professional Documents
Culture Documents
net/publication/225549761
CITATIONS READS
233 308
1 author:
Friedrich W. Hehl
University of Cologne
272 PUBLICATIONS 10,675 CITATIONS
SEE PROFILE
All content following this page was uploaded by Friedrich W. Hehl on 29 March 2019.
REVIEW ARTICLE
F.W. HEHL
Institut f~r Theoretische Physik der Technischen Universit~v
Clausthal, D-3392 Clausthal-Zellerfeld, W. Germany
Received 5 June 1972
ABSTRACT
From physical arguments space-time is assumed to possess
Das Wirkliche zieht in den Raum nicht ein wie in eine recht-
winklig-gleichf~rmige Mietskaserne, an welcher all sein wechselvol-
les Kr~ftespiel spurlos vor~bergeht, sondern wie die Schnecke baut
und gestaltet die Materie selbst sich dies ihr Haus.
Hermann Weyl
if
w~ = -~ d ~ ( ~ , ~ , g , ~ g ) . (i.i)
momentum tensor:
d l
(e def (det(_gij))2; for the exact definition of the variational
derivative see (3.7) of Part If). (1.2) should be regarded as the
proper definition of oij, because in the aforementioned kinematical
definition homogeneity of space-time is assumed --furthermore, one
is not led to a unique localisation of energy-momentum. Such a
localisation of energy-momentum is given by (1.2), however. Hence
by (1.2), from the infinite number of kinematically defined energy-
momentum tensors, one is selected as the proper one.
Also from (1.2) it is clear that dij has to do with translations,
because the variation of the metric means a small change of the
mutual distance between the points of space-time. Hence gij is of
the type of a strain tensor.
Z = Z(~,~,g,~g,K,~K). (1.3)r
Then the dynamical definition of spin angular momentum would read
eTkj i def 6Z
BKij k . (1.4)
It also would select the proper spin angular momentum tensor from
the variety of kinematically defined spin tensors.
Just as the variation of the metric in (1.2) represents a small
deformation of space-time by displacements, the variation of Kij k
in (1.4) should cause a small deformation of space-time by contor-
tions independent of the displacements. Hence the hypothesis (1.4)
demands a non-Riemannian geometry for space-time characterized,
apart from the metric gij, by a contortion tensor Kij k. Due to
(1.4) Kij k should be antisymmetric in j and k and should have the
dimension (length) -I .
1.3 Contortion
The affine connection of the geometry looked for in this context
was given explicitly and discussed extensively for the first time
by KrGner and the author [34,28]. One is led to it, without diffi-
culties, by the following heuristic line of reasoning:
First let us introduce the postulate about parallel displace-
ment of vectors. Let C3 by a contravariant vector. Assuming lin-
earity in the infinitesimal we get for its change under parallel
displacement by the coordinate differential dx I ('affinity postu-
late')
dC~ = - r ~ j c J ~ i, (1.s)
sijk d------
e f 89 k
- ~ji) s F~ij], (1.6)
Vkgij = O, (1.7)
where the indices, if necessary, have been raised and lowered with
the help of gij or its reciprocal. {k} is Christoffel's symbol
of the second kind with respect to the metric gij" A 4-dimensional
space with the connection (1.8) will be called a U4.
One arrives at conventional GR, if additionally to postulates
41.5,7) one requires the symmetry of the connection Sij k = 0 in ac-
cordance with Euclidean geometry. Then 41.9) vanishes and the con-
nection has a Riemannian form. We will drop this additional assump-
tion, however, since it does net seem plausible when spin is present
(see section 1.4). Thus we get a Riemann-Cartan geometry with the
connection 41.8). In comparison with Riemannian geometry it is
distinguished, say, by a new tensor Kij k with 24 independent com-
ponents and the dimension (length) -1.
In order to understand the geometrical meaning of Kij k, one can
introduce in a continuous way a tetrad at each world polnt. Due
338 F.W. HEHL
to the existence of a metric we will choose the tetrads to be ortho-
normal.
In (3.56,61) of Part II we will show the following: If one trans-
fers parallelly an orthonormal tetrad with the help of the connec-
tion (1.8) from one world point to an infinitesimally neighboring
point, the tetrad will rotate. The rotation consists of two parts.
One part, the Riemannian part, depends only on the choice of the
tetrad field --in special cases it can vanish --whereas the other
part of the rotation is proportional to Kij k. A variation of Xij k,
which is at the same time a variation of the connection (1.8) of
space-time, leads to a small rotation of the parallelly transferred
tetrad. Consequently one can claim that a space-time with the con-
nection (1.8) possesses rotational degrees of freedom which are
described
9
by Ki~k.
3
Hence Kij k is called the contorsion tensor
Thls tensor has the properties required by (1.4), which we expres-
sed previously by the same notation9 Thereby our problem has been
solved.
Taking mainly (1.2,4,8) as ingredients, one is able to develop
a formalism with reference to GR with the help of a variational
principle, which demonstrates the feasibility of the program stated
up to this point.
1.4 A P r i o r i Evidence
The extension of GR treated here is not founded on a principle
as obvious as the equivalence principle used to derive conventional
GR. Rather one starts from the somewhat formal requirement that
space-time should carry an image of spin, i.e. the 'aether' should
possess independent rotational degrees of freedom. In GR one recog-
nizes gravitation a p ~ o ~ as the interaction which leads to a geo-
metrization of energy-momentum, whereas for spin angular momentum
regarded here the relevant (very weak spin-spin contact) inter-
action only emerges after the completion of the theory. Because
of (1.4) one merely knows that the Kijk-field has spin angular mom-
entum as its source.
There are phenomenological hints, nevertheless, which make it
rather plausible to describe space-time by a U4. Take for instance
a piece of matter, the spins of which are parallelly oriented within
a macroscopic domain, say a ferromagnet. In the rest system of the
magnet there is no macroscopic motion. Still, at least from a clas-
sical point of view --and we are always using classical field theory
here-- one has to imagine an internal ordered motion of the matter
as sketched in figure i. The Einstein-deHaas experiment can be
referred to as confirmation of this conception.
In the case of a ferromagnet this spin motion originates from
fermions ('spinor matter'). Hence it is not possible to eliminate
this motion through transition to a suitable rotating frame of ref-
erence. Also in the case of bosons ('tensor matter') not much is
changed. Imagine a Proca field the spins of which are ordered ac-
cording to figure I, and change over to a suitable rotating frame
of references as outlined in figure i. Depending on the microscopic
SPIN AND TORSION IN GENERAL RELATIVITY 339
I
Figure i - View on the cross-section of a ferromagnetic bar.
The small vortices represent the imagined spin motion.
7 ($,~$)da. (2.1)
~i § Vi. (2.2)
(For fields with vanishing rest mass see the discussion in section
4.7 of Part If). In the Riemannian space V 4 of GR the substitution
(2.2) can be looked upon as if one were going over to curvilinear
coordinates. Not so in a U4: In this case in general there enter
explicitly torsion dependent terms, as expected according to the
discussion of section 1.4.
In conventional GR the minimal coupling principle of gravitationf
is an expression of the equivalence principle; like this, it is of
a heuristic nature, as statements about flat space like (2.1), can
never lead unequivocally to statements about curved space. Experi-
ments could force us to introduce terms of the Pauli-type without
altering the theory too much--a situation which is similar both for
a V 4 and a U4 theory. Thus we obtain for the action function
% This principle had already been used implicitly for some time,
compare e.g. Rosenfeld [$5]. It was explicitly formulated and dis-
cussed by Sciama [58]. The criticism in [65] seems unjustified to
us, because the prescription (2.2), according to the minimal coupl-
ing principle, is applied only in the Lagrangian density.
SPIN AND TORSION IN GENERAL RELATIVITY 343
minimally coupled to the U4, with the help of (1.8,9), (2.],2)
Wm c c
Wf = ~ ~(g,~g,S,SS)d~. (2.4)
In o r d e r t o come b a c k to GR f o r v a n i s h i n g s p i n , k must be e x p r e s -
sed t h r o u g h N e w t o n ' s g r a v i t a t i o n a l c o n s t a n t y i n t h e f o l l o w i n g way:
6w = o . (2.73
~g
6~ l ~V 6Z 1 6~
(2.9,10)
69i j 2k 69ij ' 6Sij k 2k 6sijk "
1 @s ~V = kT ijk. (2.15)
2e 6Sj]k s
2.5 Tetd~ads
As long as one treats material tensor fields, both in the con-
text of the theory described here and in conventional GR, a general
covariant representation in usual (holonomic) coordinates is appro-
priate. Introducing material spinor fields, however, it is appro-
priate, for reasons of clarity, to express everything in (anholo-
nomic) Cartesian coordinates, due to the preference of Cartesian
coordinates in connection with spinors. Therefore a tetrad formal-
ism has the same position in a U4-theory as in conventional GR (in
contrast to what is often stated in the literature).
and
+l 0 0 0
0 -i 0 0
gij = eaie6jg~B with g~B =
(2.17)
o 0 -i o
o o o -i
6s i ~V ~ ~V 6gk~
- - - - (2.19)
~e~ i 2k 6ea i 2k 6gk~ 6ea i
~gks = 2 i (2.21)
~e~ i e~(k6Z)'
ACKNOWLEDGMENTS
For support and discussions the author would like to thank Dr.
B,K. Datta Gairola, Prof. Dr. H.D. Doebner, Dr. Th. Engel~ Prof.
Dr. E. KrOner, Dr. W. L~cke, and Dr. P. Rosenthal.
REFERENCES
1. Anderson, J.L. (1967). Principles of Relativity Physics,
(Academic Press, New York).
2. Belinfante, F.J. (1939). Physica, 6, 887.
3. Bi~&k, J. (1966). Czech. J. Phys., B]6, 95.
4. BilLy, B.A. (1960). Progr. Solid Mech. ~mst.), 1, 331.
5. Bilby, B.A., Bullough, R. and Smith, E. (1955). Proc. Roy.
SOC., A231, 263.
6. B~rner, G. and D~rr, H.P. (1969). Nuovo Cimento, A64, 669.
7. Cap, F., Majerotto, W., Raab, W. and Unteregger, P. (1966).
Fortschr. Phys., 14, 205.
8. Caftan, E. (1922). C.R. Acad. Sci. Paris, 174, 593.
9. Caftan, E. (1923). Ann. Ec. Norm. Sup. (3), 40, 325; (1924).
Ann. Ec. Norm. Sup. (3), 41, i; (1925). Ann. Ec. Norm. Sup. (3),
42, 17.
10. Clerc, R.-L. (1971). C~ Acad. Sci. Paris, A272, 1145.
11. Clere, R.-L. (1971). C~ Acad. Sci. Paris, A272, 1760.
12. Clerc, R.-L. (1972). C.R. Acad. Sci. Paris, A274, 525.
13. Corson, E.M. (1953). Introduction to Tensors, Spinors, and
Relativistic Wave-Equations, (Blackie, London).
14. Cosserat, E. et s (1909). Th~orie des corps d#formables,
(Hermann, Paris).
15. Costa de Beauregard, O. (1942). C.R. Acad. Sci. Paris, 214, 904.
16. Costa de Beauregard, O. (1943). J. Math. Pures Appl., 22, 85.
17. Costa de Beauregard, O. (1963). Phys. Rev., 129, 466.
18. Datta, B.K. (1971). NuoVo Cime~to, 33, 171.
348 F.W. HEHL
19. Datta, B.K. (1971). Nuovo Cimento, B6, i.
20. Datta, B.K. (1971). Nuovo Cimento, B6, 16.
21. D~rr, H.P. (1971). Nuovo Cimento, A4, 187.
22. Einstein, A. (1960). Grundz~ge der Relativit~tstheorie, (2nd
edn.), (Vieweg, Braunschweig).
23. Finkelstein, R. (1960).J. Math. Phys., 1, 440.
24. Finkelstein, R. (1961). Ann. Phys. (N.Y.), 12, 200.
25. Finkelstein, R. (1966). J. Math. Phys., 7, 1632.
26. GUnther, W. (1958). Abh. Braunschweig. Wiss. Ges., 10, 195.
27. Hayashi, K. (1973). Gen. Rel. Gray., 4, i.
28. Hehl, F. (1966). Abh. Brauschweig. Wiss. Ges., 18, 98.
29. Hehl, F. (1968). In Mechanics of Generalized Continua (IUTAM-
Symposium), (ed. KrOner, E.), (Springer, Berlin), see p. 347.
30. Hehl, F.W. (1970). Spin und Torsion in der allgemeinen Relativ-
it~tstheorie oder die Riemann-Cartansche Geometrie der Welt,
(Habilitation thesis, Technische Universit~t Clausthal).
31. Hehl, F.W. (1971). Phys. Lett., A36, 225.
32. Hehl, F. and KrOner, E. (1965). Z. Naturforsch., 20a, 236.
33. Hehl, F. and KrOner, E. (1965). Z. Phys., 187, 478.
34. Hehl, F.W. and KrOner, E. (1965). Z. Phys., 187, 478.
35. Hilbert, D. (1915). Nachr. Ges. Wiss. C~ttingen, math.-phys. K1.,
p. 395.
36. Kibble, T.W.B. (1961). J. Math. Phys., 2, 212.
37. Kondo, K. (1952). Proceedings of the 2nd Japan National Congress
for Applied Mechanics, p. 41.
38. Kondo, K. (ed.) (1955,58,62,68). RAAG Memoirs of The Unifying
Study of Basic Problems in Engineering and Physical Sciences
by Means of Geometry, Vols 1,2,3,4, (Gakujutsu Bunken Fukyu-
Kai, Tokyo).
39. KrOner, E. (1960). Arch. Rat. Mech. Anal., 4, 273.
40. Kr@ner, E. (1963). Int. J. Eng. Sci., ], 261.
41. Kr6ner, E. (1963). Plastizit[t und Versetzungen, in Sommerfeld,
A. (1963). Vorlesungen ~ber theoretische Physik, Vol. 2, (6th
edn.), (Akademische Verlagsgesellschaft, Leipzig), Chapter 9.
42. KrOner, E. (ed.) (1968). Mechanics of Generalized Continua
(IUTAM-Symposium), (Springer, Berlin).
43. Kucha9, K. (1965). Acta Phys. Polon., ZS, 695.
44. Landau, L.D. und Lifsehitz, E.M. (1963). Feldtheorie, (ikademie-
Verlag, Berlin), (translated from the Russian).
45. Lur~at, F. (1964). Physics, ], 95.
46. Minkowski, H. (1958). Raum und Zeit, in Lorentz, H.A. (1958).
das Relativit~tsprinzip, (6th edn.), (Wissenschaftliche Buch-
gesellschaft, Darmstadt), p. 54.
47. Moen, I.O. and Moffat, J.W. (1969). Phys. Rev., 179, 1233.
48. Moffat, J.W. (1968). Lectures in Theoretical High Energy
Physics, (ed. Aly, H.H.), (Wiley, London and New York), p. 423.
49. Morgan, T.A. and Peres, A. (1962). Phys. Rev. Lett., 9, 79.
50. Muraskin, M. (1968). J. Math. Phys., 9, 2056.
51. Palatini, A. (1919). Rend. Circ. Matem. Palermo, 43, 203.
52. Papapetrou, A. (1949). Phil. Mag. (7), 40, 937.
53. Ponomariev, V.N. (1971). Bull. Acad. Polon. Sci., S~r. sci.
math~ astron, et phys., 19, 545.
SPIN AND TORSION IN GENERAL RELATIVITY 349
54. Rodichev, V.I. (1961). SoY. Phys. JETP, ]3, 1029.
55. Rosenfeld, L. (1940). M~m. icad. Roy. Belg., C1. Sc.~ ]8,
fasc. 6.
56. Schouten, J.i. (1954). Ricci-Calculus, (2nd edn.), (Springer~
Berlin).
57. Schr~dinger~, E. (1960). Space-Time Structure, (Cambridge Univer-
sity Press), (reprinted with corrections).
58. Sciama, D.W. (1962). In Recent Developments in General Rela-
tivity, (Pergamon Press, Oxford), p. 415.
59. Sciama, D.W. (1961). J. Math. Phys., Z, 472.
60. Sciama, D.W. (1964). Rev. Mod. Phys., 36, 463.
61. Smrz, P. (1970). Nuovo Cimento, B65, 147.
62. Thorne, K.S. and Will, C.M. (1971). Astrophys. J., 163, 595.
83. Tonnelat, M.-A. (1955). La th~orie du champ unifid drEinstein
et quelques-uns de ses dCveloppements, (Gauthier-Villars, Paris).
64. Tonnelat, M.-A. (1965). Les thdories unitaires de l'dlectro-
magnetisme et de la gravitation, (Gauthier-Villars, Paris).
65. Trautman, A. (1965). Lectures on General Relativity, in Bro~d-
eis Sumner Institute in Theoretical Physics 1964, Vol. 2,
(Prentice-Hall, Englewood Cliffs, New Jersey), p. 1.
66. Trautman, A. (1972). Gen. Rel. Gray., 3, 167.
67. Uhlenbeck, G.E. and Goudsmit, S. (1925). Naturwiss., 13, 953.
68. Uhlenbeck, G.E. and Goudsmit, S. (1926). Nature, 117, 264.
69. Utiyama, R. (1956). Phys. Rev., 101, 1597.
70. Weyl, H. (1929). Z. Phys., ~6, 330.
71. Weyl, H. (1950). Phys. Rev., 77, 699.
72. Duff, H.P. (1973). Gen. Rel. Gray., 4, 29.
73. Hehl, P.W. and yon der Heyde, P. (1973/4). Ann. Inst. Henri
Poincar@, 19, No.2, (in press).
74. Kopczy~ski, W. (1972). Phys. Lett., A$9, 219. See also [75].
75. Trautman, A. (1972). On The Structure of The Einstein-Cartan
Equations, (Preprint, Warsaw University), (based on a Report
for the 'Convegno di Relativita T, held in Rome).