You are on page 1of 185

Lecture Notes

in Physics
Edited by H. Araki, Kyoto, J. Ehlers, Mijnchen, K. Hepp, Zijrich
R. Kippenhahn, Mijnchen, H. A. Weidenmiiller, Heidelberg
and J. Zittartz, KGln
Managing Editor: W. Beiglbijck

246

Field Theory, Quantum Gravity


and Strings
Proceedings of a Seminar Series Held at DAPHE,
Observatoire de Meudon,
and LPTHE, Universit6 Pierre et Marie Curie, Paris,
Between October 1984 and October 1985

Edited by H. J. de Vega and N. S6nchez

Springer-Verlag
Berlin Heidelberg New York Tokyo
Editors

H. J. de Vega
Universite Pierre et Marie Curie, L.P.T.H.E.
Tour 16, ler Stage, 4, place Jussieu, F-75230 Paris Cedex, France

N. Sanchez
Observatoire de Paris, Section dAstrophysique de Meudon
5, place Jules Janssen, F-92195 Meudon Principal Cedex, France

ISBN 3-540-16452-g Springer-Verlag Berlin Heidelberg NewYork Tokyo


ISBN O-387-16452-9 Springer-Verlag NewYork Heidelberg Berlin Tokyo

This work is subject to copyright. All rights are reserved, whether the whole or part of the material
is concerned, specifically those of translation, reprinting, re-use of illustrations, broadcasting,
reproduction by photocopying machine or similar means, and storage in data banks. Under
5 54 of the German Copyright Law where copies are made for other than private use, a fee is
payable to Verwertungsgesellschafi Wart, Munich.
0 by Springer-Verlag Berlin Heidelberg 1986
Printed in Germany
Printing and binding: Beltz Offsetdruck, HemsbachIBergstr.
2153/3140-543210
PREFACE

Perhaps the main challenge in t h e o r e t i c a l physics today is the quantum u n i f i c a t i o n of


a l l i n t e r a c t i o n s , including g r a v i t y . Such a u n i f i c a t i o n is strongly suggested by the
b e a u t i f u l non-Abelian gauge theory of strong, electromagnetic and weak i n t e r a c t i o n s ,
and, in addition, is required for a conceptual u n i f i c a t i o n of general r e l a t i v i t y and
quantum theory.
The r e v i v a l of i n t e r e s t in s t r i n g theory since 1984 has arisen in t h i s context.
Superstring models appear to be candidates f o r the achievement of such u n i f i c a t i o n . A
consistent description of primordial cosmology ( t ~ t Planck) r e q u i r e s a quantum the-
ory of g r a v i t y . Since a f u l l quantum theory of g r a v i t y is not yet available, d i f f e r e n t
types of approximations and models are used, in p a r t i c u l a r , the wave function of the
Universe approach and semiclassical treatments of g r a v i t y . A nice p o s s i b i l i t y for a
geometrical u n i f i c a t i o n of g r a v i t y and gauge theories arises from higher-dimensional
theories through dimensional reduction f o l l o w i n g Kaluza and K1ein's proposal. Perturba-
t i v e schemes are not s u f f i c i e n t to elucidate the physical content of d i f f e r e n t f i e l d
theories of i n t e r e s t in d i f f e r e n t contexts. Exactly solvable theories can be helpful
for understanding more r e a l i s t i c models; they can be important in four (or more) dimen-
sions or else as models in the two-dimensional sheet of a s t r i n g . In addition, the
development of powerful methods f o r solving non-linear problems is of conceptual and
p r a c t i c a l importance.
A seminar series "Seminaires sur les ~quations non-lin~aires en th~orie des champs"
intended to f o l l o w current developments in mathematical physics, p a r t i c u l a r l y in the
above-mentioned areas, was started in the Paris region in October 1983. The seminars
take place a l t e r n a t e l y at DAPHE-Observatoire de Meudon and LPTHE-Universit~ Pierre et
Marie Curie (Paris Vl),and they encourage regular meetings between t h e o r e t i c a l physi-
c i s t s of d i f f e r e n t d i s c i p l i n e s and a number of mathematicians. Participants come from
Paris VI and VII, IHP, ENS, Coll~ge de France, CPT-Marseille, DAPHE-Meudon, IHES and
LPTHE-Orsay. The f i r s t volume "Non-Linear Equations in Classical and Quantum Field
Theory", comprising the twenty-two lectures delivered in t h i s series up to October
1984, has already been published by Springer-Verlag as Lecture Notes in Physics,
Voi.226. The present volume "Field Theory, Quantum Gravity and Strings" accounts flor
the next twenty-two lectures delivered up to October 1985.
I t is a pleasure to thank a l l the speakers f o r accepting our i n v i t a t i o n s and f o r
their i n t e r e s t i n g c o n t r i b u t i o n s . We thank a l l the p a r t i c i p a n t s f o r t h e i r i n t e r e s t and
f o r t h e i r s t i m u l a t i n g discussions. We also thank M. Dubois-Violette at Orsay and J.L.
Richard at Marseille, and B. Carter and B. Whiting at Meudon for t h e i r cooperation and
encouragement. We acknowledge Mrs. C. Rosolen and Mrs. D. Lopes for t h e i r typing of
part of these proceedings.
JV

We p a r t i c u l a r l y thank the S c i e n t i f i c Direction "Math6matiques-Physique de Base" of


C.N.R.S. and the "Observatoire de Paris-Meudon" f o r the f i n a n c i a l support which has
made t h i s series possible. We extend our appreciation to Springer-Verlag f o r t h e i r co-
operation and e f f i c i e n c y in publishing these proceedings and hope that the p o s s i b i l i t y
of making our seminars more widely available in t h i s way w i l l continue in the f u t u r e .

Paris-Meudon H.J. de Vega


December 1985 N. S~nchez
TABLE OF CONTENTS

LECTURES ON QUANTUM COSMOLOGY


S.W. Hawking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

SOLITONS AND BLACK HOLES IN 4, 5 DIMENSIONS


G.W. Gibbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

TRUNCATIONS IN KALUZA-KLEIN THEORIES


C.N. Pope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

CANONICAL QUANTIZATION AND COSMIC CENSORSHIP


P. H a j i c e k . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

QUANTUM EFFECTS IN NON-INERTIAL FRAMES AND QUANTUM COVARIANCE


D. Bernard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

STOCHASTIC DE SITTER (INFLATIONARY) STAGE IN THE EARLY UNIVERSE


A.A. S t a r o b i n s k y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

SOME MATHEMATICAL ASPECTS OF STOCHASTIC QUANTIZATION


G. J o n a - L a s i n i o ~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . : ...... 127

SUPERSTRINGS AND THE UNIFICATION OF FORCES AND PARTICLES


M.B. Green . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

CONFORMALLY INVARIANT FIELD THEORIES IN TWO DIMENSIONS


CRITICAL SYSTEMS AND STRINGS
J.-L. Gervais . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

LIOUVILLE MODEL ON THE LATTICE


L.D. Faddeev ( * ) and L.A. T a k h t a j a n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

EXACT SOLVABILITY OF SEMICLASSICAL QUANTUM GRAVITY IN


TWO DIMENSIONS AND LIOUVILLE THEORY
N. S~nchez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~80

SOME FEATURES OF COMPLETE INTEGRABILITY IN SUPERSYMMETRIC GAUGE THEORIES


D. Devchand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190

MONOPOLES AND RECIPROCITY


E. C o r r i g a n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Vl

NON-LOCAL CONSERVATION LAWS FOR NON-LINEAR SIGMA MODELS WITH FERMIONS


M. Forger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

INVERSE SCATTERING TRANSFORM IN ANGULAR MOMENTUMAND APPLICATIONS TO


NON-LOCAL EFFECTIVE ACTIONS
J. Avan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

GENERAL STRUCTUREAND PROPERTIES OF THE INTEGRABLE NON-LINEAR


EVOLUTION EQUATIONS IN I+I AND 2+I DIMENSIONS
B.G. Konopelchenko . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

HIERARCHIES OF POISSON BRACKETS FOR ELEMENTS OF THE SCATTERING'MATRICES


B.G. Konopelchenko . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

MULTIDIMENSIONAL INVERSE SCATTERING AND NON-LINEAR EQUATIONS


A . I . Nachman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298

AN SL(3)-SYMMETRICAL F-GORDON EQUATION Z B = ~ ( e Z - e -2Z)


B. Gaffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

THE SOLUTION OF THE CARTAN EQUIVALENCE PROBLEM FOR d2y = F(x,y, dy)
UNDER THE PSEUDO-GROUP~ = ~(X), y = ~ ( x , y ) ~ dx
N. Kamran(*) and W.F. Shadwick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320

QUANTUM R MATRIX RELATED TO THE GENERALIZED TODA SYSTEM: AN ALGEBRAIC


APPROACH
M. Jimbo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

SOLUTION OF THE MULTICHANNEL KONDO-PROBLEM


N. Andrei . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362

THE DIRECTED ANIMALS AND RELATED PROBLEMS


Deepak Dhar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368

INCOMMENSURATE STRUCTURESAND BREAKING OF ANALYTICITY


S. Aubry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373

L i s t of Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

(*) Lecture given by t h i s author


Lectures on Quantum Cosmology

S, W. Hawking

Department of Applied Mathematics & Theoretical Physics.

Silver Street,

Cambridge CB3 9EW.

1. IntroduoUon,

The aim of cosmology Is to describe the Universe and to explain why it

should be the way it is. For this purpose one constructs a mathematical model of

the universe and a set of rules which relate elements of the model to observable

quantities, This model normally consists of two parts:

[11 Local Laws which govern the physical fields in the model, tn classical

physics, these Laws are normally expressed as differential equations

which can be derived from an action I. In quantum physics the Laws

can be obtained from a path integral over all field configurations

weighted with e x p ( i I ),

[2] Boundary Conditions which pick out one particular state from among the

set of those allowed by the Local Laws. The classical state can be

specified by the boundary conditions for the differential equations at

some initial time and the quantum state can be determined by the

asymptotic conditions on the class C of field configurations that are

summed over in the path integral,

Many people would say that the boundary conditions for the universe

were not a question for science but for metaphysics or religion. However. in
classical general relativity one cannot avoid the problem of boundary conditions

b e c a u s e there are a number of t h e o r e m s ['1] which show that the universe must

have started out with a s p a c e t l m e singularity of infinite density and s p a c e t i m e c u r v a -

ture. At this singularity all the Laws of physics would break down. Thus one

could not predict how the universe would e m e r g e from the Big Bang singularity but

would have to impose it as a b o u n d a r y condition. One can, however, Interpret

the singularity t h e o r e m s in a different way: namely, that they indicate that the g r a v i -

tational field was so strong in the very early universe that classical g e n e r a l relativity

breaks down and that quantum gravitational effects have to be taken into account.

There does not seem to be any necessity for singularities in quantum gravity and.

as I shall show, one can avoid the problem of b o u n d a r y conditions.

I shall a d o p t what Is called the Euclidean a p p r o a c h to quantum gravity.

In this one performs a path integral over Euclidean i.e. positive definite metrics

rather than over metrics with Lorentzian signature (- + + +) and then analytically

continues the result to the Lorentzian regime. The basic assumption of the

Euclidean a p p r o a c h is that the "probability" of a positive definite 4 - m e t r i c g/zv and

matter field configuration is proportional to

exp(- I[g~v,O]) (1.1)

where I = - iI is the Euclidean action.

i[g/Lv,~] = .--P-
z6rr
/2

_ ~ 2KhZ/2d3x
aM (1.2)

- f(R - 2A - m(g~v,~'))gl/2d4x
M P

where h i j Is the 3 - m e t r i c on the b o u n d a r y aM and K Is the trace of the second

fundamental form of the boundary. The surface term in the action is n e c e s s a r y


because the curvature scalar R contains second derivatives of the metric. The

physics of the universe is g o v e r n e d by probabilities of the form above for all 4 -

metrics g/zv and matter field configurations b e l o n g i n g to a certain class C. The

specification of this class d e t e r m i n e s the quantum state of the universe.

There seem to be two and only two natural c h o i c e s of the class C:

a) C o m p a c t Metrics

b) Non-compact metrics which are asymptotic to metrics of maximal s y m -

metry, i.e. flat Euclidean s p a c e or Euclidean a n t i - d e Sitter s p a c e

Boundary conditions of type b) define the usual vacuum state. In this state the

expectation values of most quantities a r e defined to be zero so the vacuum state is

not o as much Interest as the quantum state of the universe. In particle scattering

calculations one starts with the vacuum state and one c h a n g e s the state by creating

particles by the action of field o p e r a t o r s at infinity in the infinite past. One lets the

particles interact and then annihilates the resultant p a r t i c l e s b y t h e action of other

field o p e r a t o r s at future infinity. This gets one back to the vacuum state. If one

s u p p o s e d that the quantum state of the universe was s o m e such particle scattering

state, one one would loose all ability to p r e d i c t the state of the universe b e c a u s e

one would have no idea what was coming in. One would also e x p e c t that the

matter In the universe would become concentrated in a certain region and that it

would decrease to zero at large distances instead of the roughly homogeneous

universe that we observe,

In particle scattering p r o b l e m s , one is interested in o b s e r v a b l e s at infin-

ity. One is therefore c o n c e r n e d only with metrics which are c o n n e c t e d to Infinity:

any d i s c o n n e c t e d c o m p a c t parts of the metric would not contribute to the scattering

of particles from infinity. In c o s m o l o g y , on the other hand, one is c o n c e r n e d with

o b s e r v a b l e s In a finite region In the middle of the s p a c e and It does not matter

whether this region is c o n n e c t e d to an infinite asymptotic region. Suppose that

the class C which defines the quantum state of the universe consists of metrics of
4

type b ) , The expectation value of an o b s e r v a b l e In a finite region will be given by

a path integral which contains contributions from two kinds of metric.

I) Connected asymptotically Euclidean or a n t i - d e Sitter metrics

ii) Disconnected metrics which consist of a c o m p a c t part which contains the

region of observation and an asymptotically Euclidean or a n t i - d e Sitter

part

One cannot exclude d i s c o n n e c t e d metrics from the path integral b e c a u s e they can

be a p p r o x i m a t e d by c o n n e c t e d metrics in which the different parts were joined by

thin tubes. The tubes could be chosen to have n e g l i g i b l e action. Similarly, topo-

logically non-trivial metrics cannot be excluded because they can be a p p r o x i m a t e d

by t o p o l o g i c a l l y trivial metrics. It turns out that the d o m i n a n t contribution to the

path integral c o m e s from d i s c o n n e c t e d metrics of the second kind. Thus, as far

as observations in a finite region are c o n c e r n e d , the result of c h o o s i n g the class C

that defines the quantum state to be non-compact metrics of type b) would be

almost the same as c h o o s i n g it to be c o m p a c t metrics of type a ) . It would t h e r e -

fore seem more natural to c h o o s e C to be the class of all c o m p a c t non-singular

metrics. This would mean that the universe would be c o m p l e t e l y self-contained

without any singularities at which the laws of physics break clown and without any

edges at which boundary conditions would have to be set. It should be

emphasised, however, that this is only a orocosal for the quantum state of the

universe. One cannot derive It from some other principle but merely show that It

Is a natural choice, The ultimate test is not whether It Is aesthetically a p p e a l i n g

but whether it e n a b l e s one to make predictions that a g r e e with observations. I

shall e n d e a v o u r to do this for a simple model.

2. The Wavefunctlon of the Universe

In practice, one is normally interested in the probability, not of the

entire 4-metric, but of a more restricted set of o b s e r v a b l e s . Such a probability

can be derived from the basic probability ( ] . ] ) by Integrating over the unobserved
quantities. A particularly Important case Is the probability P [ h i j , ~ o ] of finding a

closed c o m p a c t 3 - s u b m a n l f o l d S which divides the 4 - m a n i f o l d M Into two parts M

and on which the induced 3 - m e t r i c is h i j and the matter field configuration is o

is

P[hij,d>0] = fd[g#v]d[d>]exp(-~[g/zv,~] ) (2. 1)

where the Integral Is taken over all 4 - m e t r i c s and matter field configurations b e l o n g -

Ing to the class C which contain the submanlfold S on which the Induced 3 - m e t r i c

is h i j and the matter field configuration Is Do, This probability can be factorized

into the product of two amplitudes or wave functions ~' [ h i j ,~0].

P [ h i j , ~ O] = ~ ' + [ h i j , C ~ o ] ~ _ [ h i j , ~ o ] where

@[hij,~o] = fd[g#v]d[~]exp(-~[g#v,~]) (2.2)


C

The path integral Is over the classes C+ of metrics on the compact manifolds M+

with boundary S. With the choice of c o m p a c t metrics for C, ~z+ = ~_ and both

are real. I shall therefore drop the subscripts + and - and refer to ~z as the

"Wavefunction of the Universe'.

In a neighbourhood of S in M, one can introduce a time coordinate t ,


i
which is zero on S, and three space coordinates x and one can write the metric

in the 3 + 1 form

ds 2 = _ (N 2 - NiNi)dt2 + 2Nidxidt + hijdxidxJ (2.3)

A Lorentzian metric corresponds to the lapse N being real and a Euclidean metric

corresponds to N negative imaginary. The shift vector N i is real in both cases.

In the Lorentzian case the classical action is


6

I = I(Lg + Lm)d3xdt (2, 4)

where

mD ijklw h%/2 (2.5)


Lg = 16nN(G ~.ijKkl + 3R)

I
Kij = ~N - -at + 2N( ilJ ) J (2.6)

is the second fundamental form of S and

G ijkl = -1/2 h~%(hikh jl + hilh jk - 2h ijhkz) (2.7)

In the case of a massive scalar field

f
I -2 a~ _ 2N ia~a~ (2.8)
N2

-[hiJ- NiN___3_'
_a_~_N
2 J]axiax
jS~ m2~21

In the Hamiltonian treatment of General Relativity one r e g a r d s the c o m -

ponents h i j of the 3 - m e t r l c and the field , as the c a n o n i c a l coordinates The

c a n o n i c a l l y c o n j u g a t e momenta are

, a~ hh ~"-m
m 2
7r13 = ---- = - 16
167/P ( K i j - hiJK) (2.9)
at%ij

an
TP4~ = _ _ m -_ N-lhlh I~ - t"i--a+-]
axZJ (2.10)
8+
The H a m l l t o n l a n Is

H = ~(~iJF*ij + ?r~ ~ - Lg - Lm)d3x (2.11)

= I(NH 0 + NiHi)d3x

where

HO " 16Xrmp2Gijkl TrijTrkl - 167r


mph%/~ 3R (2.12)

k
+ ~,2h]/z 7r + hiJ a~. a. + m2~2
axZax 3 1
H i = _ 2 ijl j + hiJ a~. (2.18)
ax 3

and

Gijkl = ~/2h-~/~(hikhjl + hilhjk - hijhkl) (2.14)

From its path integral definition, the w a v e f u n c t i o n ~, is a function only of

the 3 - m e t r i c hij and the m a t t e r field c o n f i g u r a t i o n D 0 on S but it is not a function

of t , which is m e r e l y a c o o r d i n a t e that can be given any value. It t h e r e f o r e fol-

lows that ~I, will be u n c h a n g e d if the surface S is displaced a d i s t a n c e N a l o n g the

n o r m a l s and shifted an a m o u n t N i a l o n g itself. The c h a n g e in u n d e r that d i s -

p l a c e m e n t will be the q u a n t u m H a m t l t o n i a n o p e r a t o r acting on "#'. Thus ~ will o b e y

the zero e n e r g y S c h r o e d l n g e r e q u a t i o n .

H~ = 0 (2.15)
8

where the Hamlltonian operator is obtained from the classical Hamiltonian by the

replacements

TriJ(x) ~ -- i ~ i O(j x ) ' 7r#(x) --', -- i---~


5 ~ 0( x ) (2, ] 6 )

3 Quantlzatlon

The wavefunction ~" can be r e g a r d e d as a function on the infinite d i m e n -

sional manifold W of all 3 - m e t r i c s h i j and matter fields on S. A t a n g e n t vector

to W is a pair of fields (Tij,P,) on S where ~ / i j can be r e g a r d e d as a small

change of the metric h i j and /~ can be r e g a r d e d as a small c h a n g e of ~. For


2
each c h o i c e of N on S there is a natural metric F ( N ) on W

ds2 = J [321; ~ij~kl + I/2hlh/~2 (3. "1)

The zero e n e r g y S c h r o d i n g e r equation

H~' = 0 (3.2)

can be d e c o m p o s e d into the m o m e n t u m constraint

H ~' -= fNiHid3x ~' (3.3)

= $hv'~ i 2 ~-; =j axJ 8~(x)j

This Implies that ~" is the same on 3 - m e t r i c s and matter field configurations that

are related by c o o r d i n a t e transformations In S. The other part of the S c h r o e d l n g e r

equation, c o r r e s p o n d i n g to
HI~ = o (3.4)

where H I = "j N H o d 3 X is called the Wheeler-DeWitt equation. There is one

Wheeler-DeWitt equation for each choice of N on S. One can regard them as a

system of s e c o n d o r d e r partial differential equations for ~I, on W. There is some

ambiguity in the choice of operator ordering in these equations but this will not

affect the results of this paper. We shall assume that II I has the form 2

( - z/zv 2 + ~RE + v)~" = o (3.5)

where v 2 is the Laplacian in the metric F(N). RE is the curvature scalar of this

metric and the potential V Is

2 /
V = j.hl/ZN _ mp 3R + E + U d3x (3.6)
167r

T OO 2
where U ~'2n,. The c o n s t a n t E can be r e g a r d e d as a renormalization of

the c o s m o l o g i c a l c o n s t a n t A. We shall assume that the r e n o r m a l i z e d A is zero. We

shall also assume that the coefficient ~ of the s c a l a r curvature RE of W is zero.

Any wavefunctton ~I, which satisfies the momentum constraint and the

Wheeler-DeWitt equation for each choice of N and N i on S d e s c r i b e s a possible

quantum state of the Universe. We shall be c o n c e r n e d with the p a r t i c u l a r solution

which represents the quantum state defined by a path integral over compact 4-

metrics without boundary. In this case

= Id[g~v]d[~]exp( - I(g~v,~)) (3.7)

where I is the Euclidean action obtained by setting N negative imaginary. One can

regard (3.7) as a b o u n d a r y condition on the Wheeler-DeWItt equations. It implies

that tI, tends to a constant, which can be normalized to one. as h i 3 goes to zero.
10

4 Unperturbed Friedman Model

References 3,4,5 considered the Minisuperspace m o d e l which consisted

of a F r i e d m a n m o d e l with m e t r i c

ds 2 = 02( - N2dt 2 + a2dN~) (4.])

w h e r e dn~ is the metric of the unit 3 - s p h e r e . The n o r m a l i z a t i o n factor 0 2 = 2 2


3Trmp

has been included for convenience, The model contains a scalar field (21/2T/O)-I~

with mass u-lm which is constant on surfaces of constant t. One can easily gen-

eralize this to the case of a s c a l a r field with a potential V ( ~ ) . Such g e n e r a l i z a t i o n s

include m o d e l s with h i g h e r derivative q u a n t u m c o r r e c t i o n s 6. The action is

a2 N 2 tdtJ + m2~2 (4.2)

The classical H a m l l t o n l a n Is

H = ~2N( - a-l~ a
2 + a - 3 n~2 - a + a3m2~ 2) (4.3)

where

ada 7T# a3d# (4.4)


7Ta Ndt = N dt

The classical H a m i l t o n i a n c o n s t r a i n t is H = o. The classical field e q u a t i o n s are

+ = o (4.5)
a t tN + a d-- d t

Nd f! da N2am2~2
11

The W h e e l e r - D e W i t t e q u a t i o n is

]/zNe-3a[ (35
a2
a,2
+ 2Vl~(a,#) = 0 (4.7)

where

V = zAz(eeam22 - e 4=) (4.8)

and ~x = t,n a. One can r e g a r d e q u a t i o n (4.7) as a h y p e r b o l i c e q u a t i o n for ~' In

the flat s p a c e with c o o r d i n a t e s (~z,~) with a as the time c o o r d i n a t e . The b o u n -

dary c o n d i t i o n that gives the q u a n t u m state defined by a path Integral over c o m p a c t

4-metrics is ~ -* 1 as o~ -. - ~o If o n e i n t e g r a t e s e q u a t i o n (4.7) with this b o u n -

dary condition, one finds that the wavefunction starts oscillating in the region

V >0, I#1 > 1 (this has been c o n f i r m e d numerically 5). One can i n t e r p r e t the

o s c i l l a t o r y c o m p o n e n t of the w a v e f u n c t i o n by the WKB a p p r o x i m a t i o n :

= Re ( C e iS ) (4.9)

where C is a slowly varying amplitude and S Is a raplclly varying phase. One

c h o o s e s S to satisfy the classical H a m i t t o n - J a c o b i e q u a t i o n :

H(Yra,rr#,a,#) = o (4. lO)

where

~'a = aa-~'
s
~~ = as
a-~ (4,11)

One can write (4. "10) in the form

I/zfab as as + e-3~'v = o (4.12)


aqaaq b
12

where fab is the inverse to the metric F(1):

fab = e-3~diag(-i,i) (4. ]3)

The wavefunetlon (4, 9) will then satisfy the Wheeler-DeWltt equation If

v2c + 2ifab aC a S + iCV2S = 0 (4. ]4)


aga~q b

where V 2 is the Laplacian in the metric l a b ' One can ignore the first term in

equation ( 4 . 1 4 ) and can integrate the equation along the trajectories of the vector

field X a = d r~- = l a b a.__S and so determine the amplitude C. These trajectories


aq b

correspond to classical solutions of the field equations. They are parameterized by

the coordinate time t of the classical solutions.

The solutions that correspond to the oscillating part of the wavefunction

of the Minisuperspaee model start out at V = O, I~J > 1 with ~da = d_~
dt = o.
They expand exponentially with

S = - ~el 3=m ~1(1 - m - 2 e - 2 = ~ - 2 ) " - ~e3=ml~l (4, 15)

dl~l 1
~ = ml~l = - z-m (4.16)
dt ' dt

After a time of order 3 m - ] ' ( l # . l l - 1), where ~1 is the initial value of ~. the field

starts to oscillate with frequency m. The solution then becomes matter dominated

and expands with e a proportional to t 2/3. If there were other fields present, the

massive scalar particles would decay Into light particles and then the solution would

expand with e ~z proportional to t z/z, Eventually the solution would reach a maximum

9~ 2
radius of order e x 9 ( - ' ~ - ) or e x p ( 9 ~ l ) depending on whether it is radiation or matter

dominated for most of the expansion. The solution would then recollapse in a

similar manner.
13

5 The Perturbed Friedman Model

We assume that the metric is of the form ( 2 . 3 ) except the right hand
2
side has been multiplied by a normalization factor o The 3-metric h i 3 has the

form

2
hij = a (nij + Eij) (5.])

where Nij Is the metric on the unit 3-sphere and Eij Is a perturbation on this

metric and may be expanded in harmonics:

z3 E [61/2 ~ _n 6%/2 n 2]/2 0 S " n


' = n,l,m an~m 3 ij~Jim + bn~m (Pij )~m + CnEm ( z 3 )~m

+ 2%/2 c e e n d 0 n de Ge n ] (5.2)
nm (Sij)Im + 2 nm (Gij)Im + 2 nero ( ij)~m I

The coefficients a . m , b . ,c .
d de ,c e
n n~m' nm' n~m are functions of the time c o o r -
nLm nm
i
dlnate t but not the three spatial coordinates x .

The Q ( x z) are the standard scalar harmonics on the 3-sphere. The

P i j ( x 1) are given by (suppressing all but the i , j indices)

1
Pij (n 2 1- I) Qlij + 3-~ijQ (5.3)

i
They are traoeless, P i = 0. The S i j are defined by

Sij = Sil j + Sjl i (5.4)

where Si are the transverse vector harmonics, sill-o. The Gij are the

transverse traceless tensor harmonics. Gi i = Gij I j = 0. Further details about

the harmonics and their normalization can be found in appendix A.


14

The lapse, shift and the scalar field ~(xi,t) can be expanded in terms

of harmonics:

{
N = N O i + 6- ~ F. gnm Qm
n, ,m
n) (5.5)

n + 2 ~ Jnm (Si);m ]
N i . e (= Y. {6-]/2 knm (Pi)m (5.6)
n, ,m

= o-1 1 #(t) + F. fnm Qm


n, ,m
nl (5, 7)

1
where P i Qli" Hereafter. the labels n , 9 . , m , o and e will be denoted
(n 2 - l)
simply by n. One can then expand the action to all orders in terms of the "back-

ground" quantities a,#,N 0 but only to second order in the "perturbations"

a n , b n , On, t i n , f n , g n , k n , j n :

I = I o ( a , # , N O) + F.I (5.8)
n n

where I O is the action of the unperturbed model ( 4 . 2 ) and In is quadratic in the

perturbations and is given in appendix B.

One can define conjugate momenta in the usual manner. They are:

~a = - NLle3a& + quadratic terms (5, 9)

~ = NLIe3~ ~ + quadratic terms (5.10)

77 = - NLle3a[~ n + &(a n - gn ) + !e-a k ] (5.11)


3 nJ
an

2
= NLIe3U iD__=_~I [~n + 4&bn - ~l e - a k n,] (5.12)
~bn (n 2 - 1)
15

/;c = N; le3'~ (n2 - 4) [~n + 4&c n - e-aJn ] (5. ]3)


n

(5. ] 4 )

I
~rf = Nole3(~ fn + ~(3an - gn )
n l (5. 15)

The q u a d r a t i c terms In e q u a t i o n s (5.9) and (5.]0) are given in a p p e n d i x B. The

H a m i l t o n l a n can then be expressed in terms of t h e s e m o m e n t a and the o t h e r q u a n t i -

ties:

.-.o I .,o ~.?~+ ~n Hn,~I nE{knSH~I + Jn VHn_lj} (5. ]6)

The subscripts 0,1,2 on the "1 and H_ d e n o t e the o r d e r s of the quantities in the

perturbations and S and V d e n o t e the s c a l a r and v e c t o r parts of the shift part of

the H a m i l t o n i a n . HIO is the H a m i l t o n i a n of the u n p e r t u r b e d m o d e l with N = 1 :

"g0 a ~ + - (5. 17)

S n Vn Tn
The s e c o n d order Hamlitonian is given by H i 2 = E H / 2 = F.( HI2 + Hi2 + HI2)
n - 13
where

+ n,.

_ ~2 + L_~_:!/ 2 + ~f2 ~ + 8bn~ b n ~ - 6an~ f n ~#


an (n2-4) ~bn n + 2an~an

- 2)an + (n-~--i) n
16

+ e'=m
[ n + 6anfn~)l + [2 n -
(n2_l) nJj (5. ]8)

VHI2n = Z/2e-3aI(n2_4)c2[lOTr2+ 6~] + i_.__(n2_4)


Tr2Cn+ 8Cn/TCnTr +

(n2-4)C2n[2e 4(z - 6e6am2~2]] (5. "19)

TH,2n -~'2e" -3~, lan


f.2 [1OTr2 + 6/T~] + TrC~n + 8dnTrdnTr +

d2n[(n2+l)e 4(z - 6e6(Zm2#2]] (5.20)

The first order Hamlltonlans are

H[1 1/'ze- 3a
= an n

+ m2e6(Z[2fn. + 3an.21 - 2e~a[(n2-4)bn + (n2+~'~)anll (5.21)

The shift parts of the Hamlltonlan are

(5.22)
_1 = "~e - nan n (n2_1) nj ~z

VnH_/ = e -(z{n,cn + 4(n2-4) Cn~ a] (5.23)

The classical field equations are given in appendix B.

Because the Lagrange multipliers No,gn,k n , j n are Independent. the


zero energy Sohroedlnger equation

H~!" = 0 (5.24)
17

can be d e c o m p o s e d as before into m o m e n t u m constraints and Wheeler-DeWitt equa-

tions. As the m o m e n t u m constraints are linear in the m o m e n t a , there is no a m b i -

guity in the o p e r a t o r o r d e r i n g . One therefore has

S n i -3~ a _ + 4(n 2 4) bn Ba
H_.I~, = - ~e a~n I I an (n 2 - 1) 1
a 3fn ~l~!, = 0 ( 5, 25)
ab n

H..,].~" = e
-I ~ < + 4(n 2 - 4) c n ~" = 0 (5.26)

n
The first order Hamiltonians Hll give a series of finite dimensional

second o r d e r differential e q u a t i o n s , one for each n. In the o r d e r of a p p r o x i m a t i o n

that we are using, the ambiguity in the o p e r a t o r o r d e r i n g will consist of the possible
a
addition o f terms linear in ~-~. The effect of such terms can be c o m p e n s a t e d for

by multiplying the wavefunction by powers of e a. This will not affect the relative

probabilities of different observations at a given value of a. We shall therefore

ignore such ambiguities and terms.

2 4e 2
(5.27)

Finally, one has an Infinite d i m e n s i o n a l s e c o n d o r d e r differential equation

I.,o + ~.,s.?2
+. v"?2+ T.~2,1. = o (5.28)

where HIo Is the operator in the Wheeler-DeWitt equatlon of the unperturbed


18

Friedman Mlnlsuperspace model:

= ]/,ze_3(z[ (92 82 + e6(Zm2,2 _ e4~Z] (5.29)


,o l~ a~2

and

HI2 = ]/2e -3(z !0_/._n_2-4) b21 (92 [15a2 + 6--(n2-4) b21 a 2


I-I =a n + (n2_l) nJaa2 --- [2 n (n2_l) nJ a,2
--

a2 (n.2-1__). a 2 a2 a a 8b -=(9 a__ a a


aan 2 (n2-4) (gbn2 afn 2 2anSana~ nabna~ + 6an~fna~

- e4al'l[~(n 2..~)5a 2n + (n2-7_.1 (n2-!). b 2 + 2(n2 4 ) anbn _ (n2_l)f2n]j


3 (n 2 -1 ) 17.

+ e6amZ[fZn + 6anfn# ] + e6(Xm2"2[3a2,


(~ n - 6(n2-4)b2]](n2_l)
njj (5.30)

aa
VHI2n = ]/2e-3(~
t t (ga
a2 1 j__2_2_
(n2-4) (9cn
8Cna ~-(gu *
n

(n2-4)CZn[2e 4a - 6e6am2~2]] (5.31)

T l-If2
n = ]he-3a - d 2 [10 a2 + 6
n( aa2 ad 2 8d na--~-~na~
a a
+
n

d2[(n2+l)e 4u - 6e6Um2#2]] (5.82)

We shall call equation (5.28) the master equation. It is not hyperbolic


a2
because, as well as the positive second derivatives -a~
- 2 in HIO, there are the posi-
19

a2 S n
tlve second derlvatlves----
aa 2 in e a c h Hi2. However. one can use the m o m e n t u m
n
c o n s t r a i n t (5, 25) to substitute for the partial derivatives with r e s p e c t to a n and then

solve the resultant differential equation on a n = o. Similarly, one can use the

momentum constraint (5.26) to substitute for the partial derivatives with r e s p e c t to

cn and then solve on c n = O. One thus obtains a modified equation which is

hyper bolic for small fn" If one knows the w a v e f u n c t i o n on a n = 0 = c n , one ca n

use the momentum constraints to c a l c u l a t e the wav e f u n ct io n at o t h e r values of a


n

and c n .

6 The WavefuncUon

Because the perturbation modes are not co u p le d to each o t h e r, the

wavefunctlon can be expressed as a sum of t e r m s of the f o rm

= Re (~I,0(z,@)I-[tI,(n)(cz,#,an,bn,Cn,dn,fn)) (6. l )
n

= Re ( Ce is )

w h e r e S Is a rapidly varying function of a and @ and C Is a s l o w l y ' v a r y i n g function

of all the variables, If one substitutes (6. ] ) Into the m a s t e r equation and divides

by ~, one obtains

(6.2)
n~m 2~!'(n)~!
"(m)

+ E t, + e _ 3 a V(~,~) = o
n
20

where v 2 is the Laplaclan in the Minisuperspace metric f a b = e3=diag(-l'l) and


2

the dot product Is with respect to this metric.

An individual perturbation mode does not contribute a significant fraction

of the sums in the third and fourth terms in equation ( 6 . 2 ) . Thus these terms can

be replaced by

(V2~I') _(v2~I'(n)) I v2~I'(n)}2


~" "~n ~(n) + VZlEn--~((n'l~'- (6.3)

(v2~I,(n)) I v "#'(n)]2
" - n In1,"
,fE " tJ

In order that the ansatz (6.]) be valid, the terms in (6.2) that depend on

an,bn, Cn,dn,fn have to cancel out. Thts implies

(v2~I').(v2~F(n)) + V2v22~I,(n)= ~2-~,(n) (6.4)


II, ~I, -

( - I/2V2 + e -3e V + ~2J,J)~Z 0 = 0 (6.5)

V ~,(n)
2
where J = ~..--:-:--:-
~ ~,(n)

In regions In which the phase S is a rapidly varying function of ~z and

~, one can neglect the second term in ( 6 . 4 ) In comparison with the first term.

One can also replace the T/'CZ and Tt# which appear in H~2 by _~S and ~ respec-

tively. The vector X a = f a b 8_S_ obtained by raising the covector v2S by the
aq b
a
inverse minisuperspace metric f a b can be regarded as ~ where t is the time
21

parameter of the classical Friedman metric that corresponds to 1/, by the WKB

approximation. One then obtains a time dependent Schroedtnger equation for each

mode along a trajectory of the vector field x a :

a~.( n )
i- H:L~(n)n (6.6)
at iz

Equation ( 6 . 5 ) can be Interpreted as the Wheeter-DeWitt equation for a

two dimensional minlsuperspace model with an extra term ] / z J . J arising from the

perturbations, in o r d e r to make J finite, one will have to make subtractions. Sub-


n
tracting out the ground state energies of the HI2 corresponds to a renormalization

of the cosmological constant A. There Is a second subtraction which corresponds

to a renormallzatlon of the Planck mass mp and a third one which corresponds to a

curvature squared counterterm, The effect of such higher derivative terms in the
6
action has been considered elsewhere

One can write ~,n as

~n = S~(n)(a'#'an'bn'fn) V~(n)(a'#'Cn) T~z(n)(a'#'dn) (6.7)

where S~,(n),V~I,(n) and T~,(n) obey independent Schroedinger equations with


S.n V.n Tn
HI2, HI2 and HI2 respectively.
22

7 The B o u n d a r y Conditions

We want to find the solution of the master equation that c o r r e s p o n d s to

[hij,~] -- fd[g~v]d[~ ] exp(-I) (7.1)

where the integral is taken over all c o m p a c t 4 - m e t r i c s and matter fields which are

bounded by the 8 - s u r f a c e S. If one takes the scale p a r a m e t e r u to be very n e g a -

tive but keeps the other p a r a m e t e r s fixed, the Euclidean action I tends to zero like

e 2a. Thus one would expect ~ to tend to one as ~z tends to minus infinity.

One can estimate the form of the scalar, vector and tensor parts

S~(n),V~'(n),T~ t(n) of the perturbation ~(n) from the path integral (7. "1). One

takes the 4 - m e t r i c g/.Lv and the scalar field 4) to be of the b a c k g r o u n d form

ds 2 = 02( - N2dt 2 + ~_2~z( t )=~2


u*=3) (7.2)

and #() respectively plus a small perturbation described by the variables

(an,bn, fn) , c n and d n as functions of t , In o r d e r for the b a c k g r o u n d 4 - m e t r i c

to be c o m p a c t , it has to be Euclidean when o~ = - ,= ie N has to be purely n e g a -

tive imaginary at u = - =o, which we shall take to be = 0. In regions in which

the metric is Lorentzian. N will be real and positive. In o r d e r to allow a smooth

transition from Euclidean to Lorentzlan, we shall take N to be of the form - i e i/~

where /z = 0 at t = 0. In o r d e r that the 4 - m e t r i c and the scalar field be regular

at t = 0, an,bn, Cn,dn,fn have to vanish there,

The tensor perturbations d n have the Euclidean action

T~n = ~/zfdt d n TD dn + boundary term (7.3)

where
23

(7.4)

.2
+ 4iN0 e3a + Vze-2a 3 2,2 _ _..!__#. 3&2 1 df }l
- ~-m ~ 2(iN O)2 2(iN O)2

The last term In (7.4) vanishes If the background metric s a t i s f i e s the background

field e q u a t i o n s . The a c t i o n Is e x t r e m l z e d w h e n d n s a t i s f i e s the e q u a t i o n

TD d = o (7,5)
n

F o r a d n that satisfies ( 7 . 5 ) , the action is just the b o u n d a r y term

]
T~el l e3a[ d (~ + 4&d21 (7.6)
n = 2iN 0 [ n n nj

The path Integral o v e r d n will be

Id[dn] exp(- Tin) (detTD) -~/2 exp(- T~cl (7.7)


= n )

One now has to integrate (7.7) over different background metrics to obtain the

wavefunction T~(n) One expects the dominant contribution to come from back-

ground m e t r i c s that a r e n e a r a s o l u t i o n of the c l a s s i c a l b a o k g r o u n d field e q u a t i o n s .

F o r such metrics o n e can e m p l o y the a d i a b a t i c a p p r o x i m a t i o n in which one regards

a to be a slowly v a r y i n g f u n c t i o n of t . Then the s o l u t i o n of ( 7 . 5 ) which o b e y s the

b o u n d a r y c o n d i t i o n d n = 0 at t = o is

d = A(e vr - e -vT) (7.8)


n

where v = e-U(n2-1) ~'z and ~- = f i N 0 d t . This approximation will be valid for

background f i e l d s which are near a solution of the background field e q u a t i o n s a n d


24

for which

I~001 ~ ne -a
(7.9)

For a regular Euclidean metric, I~ol = e -a near t = o. If the metric is a

Euclidean solution of the background field equations, then I~ool<e-a. Thus the

adiabatic a p p r o x l m a t l o n should hold for l a r g e values of n Into the r e g i o n in which

the solution of the background field equations becomes Lorentzian and the WKB

a p p r o x i m a t i o n can be used. The w a v e f u n c t i o n T,~,(n) will then be

(7. I0)

in the Euclidean region, ~" will be real and positive. For large values of n,

co'ch(vT) " 1. In the Lorentzian region w h e r e the WKB a p p r o x i m a t i o n applies, r

will be c o m p l e x but It will still have a positive real part and c o t h ( v T ) will still be

a p p r o x i m a t e l y 1 for l a r g e n. Thus

T~(n) = B exp[ - 2i ~aSd2n _]/zne2ad2n] (7.11)

The n o r m a l i z a t i o n constant B can be c h o s e n to be 1. Thus, a p a r t from a phase

factor, the g r a v i t a t i o n a l wave m o d e s e n t e r the WKB r e g i o n in t h e i r g r o u n d state.

We now c o n s i d e r the v e c t o r part V~,(n) of the w a v e f u n c t i o n . This is

pure g a u g e as the quantities c n can be given any values by g a u g e t r a n s f o r m a t i o n s

p a r a m e t e r l z e d by the J n ' The f r e e d o m to m a k e g a u g e t r a n s f o r m a t i o n s Is reflected

q u a n t u m m e c h a n i c a l l y in the c o n s t r a i n t

e +4( ~'= o (7. 12)


25

One can integrate ( 7 . 1 2 ) to give

(a,{Cn}) = ~(a - 2 E(nZ-4)C2n,O) (7. 13)


n

where the d e p e n d e n c e on the other variables has been suppressed, One can also
aS
replace ~ by i~F. One can then solve for V~,(n).

V~(n) = exp[2i (r12-4.)C2n~]@S (7. 14)

The scalar perturbation modes a n , b n and f n involve a combination of

the behavlour of the tensor and vector perturbations. The scalar part of the action

is given in appendix B. The action is extremized by solutions of the classical e q u a -

tions

NO d( e 3aan.
No )
+ l(n2 - 4)No2ea(an+ bn) + 3e3a(~n -
No2m2@fn) =

d le2(zknl
N2{3e3am2@2 - ~(n2+2)ealgn + e3a~gn - 3/N ~[ N0J (7. 15)

d e3a~n 1 2
N O dr( ~0 ) - ~n _
1 )No2e~(a n + bn) =
~n 2 _
1 )N2e~gn

+ 3!NO d [ 2akn] (7.18)

NO ~t (e 3afn
NO ) +
3e3~a
n
+
tWo[m2 e3 a
.2[
+
(n2-l)e~Z]fn

e3~/- 2N2m2~gn + ~gn- e-a~kn I (7."17}


26

There is a three parameter family of solutions to ( 7 . ] 5 ) to ( 7 . ] 7 ) which

obey the boundary condition a n = b n = f n = 0 at t = o. There are however,

two constraint equations:

an + Ln2--416 + 3fn~ " e-= k (7.18)


(n2_1) n = agn (n2_l) n

.2
3an ( _ &2 + ) + 2(~ n _ &~n)

2 2 -20~ [(n 2 -4)b n + (n 2 + ]/2)an ]


+ N O2 m2(2fn @ + 3an~2 ) _ ~N0e

.2
=-2&e-Uk3 n + 2gn( _ &2 + @ ) ( 7 19)

These correspond to the two gauge degrees of freedom parameterized by k n and g n

respectively. The Euclidean action for a solution to equations (7. "15) to (7. ]9) is

S~cl 1 30~I .(_n_2-4lb ]~ + fnfn


n = ~00 e - anan + (n2_l) n n

+ &I -a2 + 4(n2-4)b2] * 3~anfn


n (n 2-z) nj

(7.20)
(nZ_z) njj

where the background field equations have been used.

In many ways the simplest gauge to work in is that with g n = k n = o.

However, this gauge does not allow one to find a compact 4 - m e t r i c which Is

bounded by a 3-surface with arbitrary values of a n , b n and f n and which is a solu-

tion of the equations (7. ]5) to (7. ]7) and the constraint equations. Instead, we

shall use the gauge a n = b n = 0 and shall solve the constraint equations ( 7 . 1 8 )
27

a n d (7. 19) to find g n and k n :

(7,21)
(n2-4)& ~ + s ~2
" I I -

k = 3 ( n 2 - 1 ) e a~. . . . . . (7.22)

With t h e s e s u b s t i t u t e d . ( 7 . 17) becomes a second order equation for f


n

NO dt~ No + N.2[
0[m 2e 3a + (n2_l)e=]fnj =

(7.23)

For large n we can again use the adiabatic approximation to e s t i m a t e

the s o l u t i o n of ( 7 . 2 3 ) when I#1 > 1:

fn = Asinh(vr) (7, 24)

where v 2 = e-2a(n2-1). Thus f o r t h e s e m o d e s

. 2o~.2 . aS ]
S~(n)(u,~, 0 , 0 ,fn) *, e x 9 - ~2ne rn - ]/2Z~gnf n (7.25)

This Is of the ground state f o r m apart from a smal l phase factor. The val ue of

S~.(n) at n o n - z e r o v a l u e s of a n and b n c a n be f o u n d by i n t e g r a t i n g the c o n s t r a i n t

equations (5.25) and ( 5 . 2 7 ) .


28

The tensor and scalar modes start off In thetr ground states, apart p o s -

sibly from the modes at low n. The vector modes are pure g a u g e and can be
H(n)t,(n)
neglected. Thus the total e n e r g y E = n ~ - - ~ ( - ~ - of the perturbations will be small

when the ground state energies are subtracted. But E = i(vzS).J where

V2~( n )
J = F. t,(n" ~ - Thus J is small. This means that the wavefunction ~0 will obey
n
the Wheeler-DeWitt equation of the unperturbed m i n i s u p e r s p a c e model and the phase

factor S will be approximately - i l n ~ O. However the homogeneous scalar field

mode # will not start out in its ground state. There are two reasons for this:

first, regularity at t = o requires an = b n = cn = d n = fn =o. but does not

require # = o. Second, the classical field equation for ~ i s o f the form f o r a damped

h a r m o n i c oscillator with a c o n s t a n t f r e q u e n c y m rather than a d e c r e a s i n g frequency


-5
e n. This means that the a d i a b a t i c approximation is not valid at small t and that

the solution of the classical field equation is # a p p r o x i m a t e l y constant. The action

of such solutions is small, so large values of I$1 are not d a m p e d as they are for

the other variables. Thus the WKB trajectories which start out from large values of

I$1 have high probability. They will c o r r e s p o n d to classical solutions which have a

long inflationary period and then go over to a matter d o m i n a t e d expansion. In a

realistic model which included other fields of low rest mass, the matter e n e r g y in

the oscillations of the massive scalar field would d e c a y into light particles with a

thermal spectrum. The model would then expand as a radiation dominated

universe.

8 Growth of Perturbations

The tensor modes will obey the S c h r o e d l n g e r equation

iaT~ (n)
at THI2T~(n ) (8. I)

8dn i ~ a
n [ [~J adn 2 ad n
29

d2n{(n2+l)e4a - 6e6Xm2#2}} (8, 2)

One can write

T~(n) = exp(-2a) exp{-2i @S d2} T,(on) (8.3)

then

.aT$1n)
I at ~ze Sa I _ ___
a2
ad2 + d2(n2_l)e4~
j
T~(n)
v (8.4)

The WKB a p p r o x i m a t i o n to the background Wheeler-DeWitt equation has been used

in deriving (8.4). Then (8.4) has the form of the S c h r o e d i n g e r e q u a t i o n for an

oscillator with a time dependent frequency v = (n2-1)VZe -a, Initially the

w a v e f u n c t i o n T,~'(on) will be in the g r o u n d state ( a p a r t from a n o r m a l i z a t i o n factor )

and the f r e q u e n c y v will be l a r g e c o m p a r e d to (~. In this case o n e can use the

a d i a b a t i c a p p r o x i m a t i o n to show that T~(on) r e m a i n s in the g r o u n d state

T~(n), exp[ - ~ne2=d2n} (8.5)

The adiabatic approximation will break down when v - <z ie the

w a v e l e n g t h of the g r a v i t a t i o n a l m o d e b e c o m e s e q u a l to the horizon scale in the inf-

l a t i o n a r y period. The w a v e f u n c t i o n T~/,(on) will then " freeze " :

T~o(n) , exp
{ - ~ne
1 (8.6)

where a, is the value of <z at which the mode goes outside the horizon. The

w a v e f u n c t i o n T~'(on) will r e m a i n of the form (8.6) until the m o d e r e - e n t e r s the h o r -

izon in the matter or radiation d o m i n a t e d era at the much g r e a t e r v a l u e ~ze of a.


30

One can then apply the adiabatic a p p r o x i m a t i o n again to (8.4) but T~n)r will no

l o n g e r be in the ground state; it will be a superposition of a number of highly

excited states, This is the phenomenon of the amplification of the ground state

fluctuations in the gravitational wave modes that was discussed in r e f e r e n c e s


7,8,9

The behaviour of the scalar modes is rather similar but their description

is more complicated because of the gauge degrees of freedom. In the previous

section we evaluated the wavefunction S ~ ( n ) on a n = ]on = o by the path integral

prescription. The ground state form ( in f n ) that we found will be valid until the

adiabatic approximation breaks down ie until the wavelength of the m o d e e x c e d e s the

horizon distance during the inflationary period. In o r d e r to discuss the s u b s e q u e n t

behaviour of the wavefunction, it is c o n v e n i e n t to use the first order Hamlltonlan

constraint ( 5 . 2 7 ) to evaluate S~.(n) on a n ~ o , b n = f n = o. One finds that

S~(n)(u,#,an, o, o )= B exp{iCa2] S,(n)(=,#,an) (8, 7)

The normalization and phase factors B and C d e p e n d on ~z and ~ but not a n .

C = l[aS]
2 [ a a J -if [8S}
Lao~J 2 - l(n2-4)e4(Z] (8.8)

At the time the wavelength of the mode equals the horizon d i s t a n c e during the infla-

tionary period, the wavefunction S~'(on) has the form

S~(n) = exp - I/2 n I y;2 e 2=, a (8.9)

where Y, is the value of y = ~r~l.~.~a[a.~j-z when the mode leaves the horizon,

y, = 3#,. More g e n e r a l l y , in the case of a scalar field with a potential V(~),


- r _~_Y3- z
y = 6V[a~j .
31

One can obtain a Schroedinger equation for S~,(n)


-0 by putting

bn = fn = o in the scalar Hamlltonlan


S HI2
n a a
and substituting for ~--~-- and ~ - - from
n n
the momentum constraint (5.25) and the first o r d e r Hamlltonian constraint (5.27)

respectively. This gives

aS (n)
tO
z at ]/ze 3a _ y2 6a 2

e4a(n2_4)[12- 1 4a[asI-2] a2 I St(on) (8. 10)


~y

where terms of o r d e r -12 have been neglected, The term e4a[aS]


[8oe.I -2 will be small
n
1
compared to ~ except near the time of maximum radius of the background solution.
Y
The S c h r o e d i n g e r equation for S t ( o n ) ( a n ) is very similar to the equation for

Tt(on)(dn),- (8.4), except that the kinetic term is multiplied by a factor y 2 and the
2
potential term is divided by a factor y . One would therefore expect that for

wavelengths within the horizon, St(on) would have the ground state form

exp(-
.. -2 2aa2n )
%,2ny e and this is bourne out by ( 8 . 9 ) . On the other hand. when

the wavelength becomes larger than the horizon, t h e S c h r o e d i n g e r equation (8. ]0)

indicates that Tt(on) will freeze in the form ( 8 . 9 ) until the mode r e - e n t e r s the h o r -

izon In the matter dominated era. Even if the equation of state of the Universe

changes to radiation dominated during the period that the wavelength of the mode is

greater than the horizon size. it will still be true that St(on) is frozen in the form

(8.9). The ground state fluctuations in the scalar modes will therefore be amplified

in a similar m a n n e r to the tensor modes. At the time of r e - e n t r y of the horizon

the rms fluctuation in the scalar modes, in the gauge in which b = f = o. will
n n
be greater by the factor y , than the rms fluctuation in the tensor modes of the

same wavelength.
32

9 Comparison with Observation

From a knowledge of T~(on) and S~(on) one can calculate the relative

probabilttes of observing different values of d n and a n at a given point on a t r a j e c -

tory of the v e c t o r field X i ie at a given value of <x and ~ in a b a c k g r o u n d metric

which is a solution of the classical field e q u a t i o n s . In fact, the d e p e n d e n c e on

will be u n i m p o r t a n t and we shall n e g l e c t it. One can then c a l c u l a t e the p r o b a b i l i -

ties of o b s e r v i n g different amounts of a n l s o t r o p y In the m i c r o w a v e b a c k g r o u n d and

can c o m p a r e these p r e d i c t i o n s with the u p p e r limits set by o b s e r v a t i o n .

The t e n s o r and s c a l a r p e r t u r b a t i o n m o d e s will be in highly excited states

at large values of ,v. This means that we can treat their development as an

e n s e m b l e evolving a c c o r d i n g to the classical equations of m o t i o n with initial d i s t r i b u -

tions in d n and a n p r o p o r t i o n a l to IT~(on)12 and IS~(on)12 respectively. The initial

distributions in (in and =4,n will be proportional to T ~rO


, , , ( n ) ," d TPno I and
n
IS~("n)n'a,, S~'(n)lv respectively. In fact. at the time that the modes re-enter the
n

horizon, the distributions will be c o n c e n t r a t e d at d = a. = O.


n n

The surfaces with b n = fn = 0 will be surfaces of constant energy

density in the classical solution during the Inflationary period. By local c o n s e r v a t i o n

of e n e r g y , they will r e m a i n surfaces of c o n s t a n t e n e r g y density in the e r a after the

inflationary period when the e n e r g y is d o m i n a t e d by the c o h e r e n t oscillations of the

homogeneous background scalar field . If the scalar particles decay Into light

particles and heat up the universe, the surfaces with b n = f n = 0 will be surfaces

of c o n s t a n t temperature. The surface of last scattering of the microwave back-

ground will be such a surface with t e m p e r a t u r e T s. The m i c r o w a v e radiation can

be considered to have propagated freely to us from this surface. Thus the

o b s e r v e d t e m p e r a t u r e will be

T
T s (9. "I)
o I + z
33

where z ts the redshlft of the surface of last scattering, Variations In the

o b s e r v e d t e m p e r a t u r e will arise from v a r i a t i o n s In z in different d i r e c t i o n s of o b s e r -

vation. These a r e given by

1 + z = .t/Zn /z (9.2)

e v a l u a t e d at the surface of last s c a t t e r i n g w h e r e n/L ts the unit n o r m a l to the s u r -

faces of c o n s t a n t t in the gauge gn = kn = J n = o and b n = fn = o on t h e

surface of last s~atterlng and p/L Is the p a r a l l e l l y p r o p a g a t e d t a n g e n t v e c t o r to the

null g e o d e s i c from the o b s e r v e r n o r m a l i z e d by /~n/z= 1 at the p r e s e n t time, One

can c a l c u l a t e the evolution of t./Zn/z down the past light c o n e of the o b s e r v e r :

d fl;Ln/zJ]
d-i[ = n ~ ; v t'p't'v (9.3)

where k is the afflna parameter on the null geodesic, The only n o n - z e r o com-

p o n e n t s of n # . ; v are

nis j = e2=I&nlj + F.(A


n n + &an~ljQ + F.(5
n n + (xbn)Pij
"

+ E(dnn + &dn)Gijl (9.4)

In the g a u g e that we are using, the d o m i n a n t a n l s o t r o p l a terms In ( 9 , 4 )

on the scale of the horlzon, will be t h o s e involving & a n and &dn. These will give

t e m p e r a t u r e a n i s o t r o p i e s of the form

((AT/T)2> '= <a2n > or - <d2n > (9.5)

The n u m b e r of m o d e s that c o n t r i b u t e to a n i s o t r o p i e s on the scale of the horizon Is

of the o r d e r of n 3. From the results of the last section


34

<a2n> = y2n-le -20~* (9.6)

<d2n> = n-Ze-2~* (9, 7)

The dominant contribution comes from the scalar modes which give

<(AT/T)2> ~, y,n
2 2 e -2~, (9.8)

-a.
But ne *, & , , the value of the Hubble constant at the time that the present

horizon size left the horizon during the inflationary period. The observational upper

limit of about 10- 8 on < ( A T / T ) 2> restricts this Hubble constant to be less than

about 5.10-5m (Ref. ]0) which in turn restricts the mass of the scalar field to be
P
less than 1014 GeV .

10 Conclusion and Summary

We started from the proposal that the quantum state of the Universe is

defined by a path integral over compact 4-metrics. This can be regarded as a

boundary condition for the Wheeler-DeWltt equation for the wavefunctlon of the

Universe on the infinite dimensional manifold, superspace, the space of all 3-

metrics and matter field configurations on a 3-surface S. Previous papers had

considered finite dimensional approximations to superspace and had shown that the

boundary condition led to a wavefunctlon which could be interpreted as c o r r e s p o n d -

Ing to a family of classical solutions which were homogeneous and isotropic and

which had a period of exponential or inflationary expansion. In the present paper

we extended this work to the full superspace without restrictions, We treated the

two basic homogeneous and Isotroplc degrees of freedom exactly and the other

degrees of freedom to second order. We justified this approximation by showing

that the inhomogeneous or anisotropic modes started out in their ground states.
35

We derived time d e p e n d e n t S c h r o e d l n g e r equations for each mode. We

showed that they r e m a i n e d In the ground state until their wavelength e x c e d e d the

horizon size during the Inflationary period. In the s u b s e q u e n t expansion the ground

state fluctuations got frozen until the wavelength r e - e n t e r e d the horizon during the

radiation or matter d o m i n a t e d era, This part of the calculation is similar to e a r l i e r


7 ll, 12 .
work on the d e v e l o p m e n t of gravitational waves and density perturbations m

the inflationary universe but It has the a d v a n t a g e that the assumptions of a period of

exponential expansion and of an initial ground state for the perturbations are justi-

fied. The perturbations would be c o m p a t i b l e with the upper limits set by o b s e r v a -

tions of the microwave b a c k g r o u n d if the s c a l a r field that drives the Inflation has a

mass of ].014 GeV or less.

In section 8 we c a l c u l a t e d the s c a l a r perturbations in a g a u g e In which

the surfaces of c o n s t a n t time are surfaces of c o n s t a n t density. There are thus no

density fluctuations in this gauge. However. one can make a transformation to a

g a u g e in which a n = b n = o. In this g a u g e the density fluctuation at the time that

the wavelength c o m e s within the horizon is

.2
<(Ap/p)2> ==
},2 Pe .2
(2, (lO.l)
.2 2
aeO e

Because y and & , d e p e n d only logarithmically on the wavelength of the p e r t u r b a -

tions, this gives an almost scale free s p e c t r u m of density fluctuations. These f l u c -

tuations can evolve a c c o r d i n g to the classical field equations to give rise to the f o r -

mation of galaxies and all the other structure that we observe in the Universe.

Thus all the c o m p l e x i t i e s of the present state of the Universe have their origin In the

ground state fluctuations In the tnhomogeneous modes and so arise from the

H e l s e n b e r g Uncertainty Principle.
36

Referenoes

1 S.W. Hawking & G . F . R . Ellis. "The L a r g e - S c a l e Structure of S p a c e -

Time". (Cambridge University Press, ,1973).

2 S.W, Hawking and D, N, Page. "Operator Ordering and the Flatness of

the Universe', DAMTP preprlnt (,1985)

3 S,W. Hawklng In: "Relativity. Groups and Topology I1", Los Houches

]983, Session XL, edited by B.S. DeWitt & R. Stora (North Holland

Amsterdam. ]984)

4 S.W. Hawking. Nucl. Phys. B239 257 ( 1 9 8 4 ) ,

5 S.W, Hawktng and Z.C, Wu, Phys, Lett ,151B ,15 ( ] 9 8 5 )

6 S,W, H a w k t n g & J.C, Luttrelh NueI, Phys, B247 250 (1984)

7 V.A. Rubakov. M.V. Sazhin & A, V. Veryaskin. Phys. Lett. 1,15B ,189

(]982)

8 L.P. Grischuk, Zh, Eksp. Teor, Fiz 6_7 825 ( ] 9 7 4 ) [Sov, Phys, JETP

40 409 (1975)]" Ann. N.Y. Sci 30__? 439 (1977)

9 A . A , Staroblnsky. PIs'ma Zh. Eksp. Teor. FIz 30 719 ( ] 9 7 9 )

.10 S.W. Hawktng, Phys. Lett. 1508 339 (1985)

11 S.W. Hawking, Phys. Lett. "115B 295 (,1982)

12 A . H . Guth & S.Y. Pi, Phys. Rev. Lett. 49 1110 (,1982) [JETP Lett. 30

682 (,1979)]

13 E,M. Llfshitz and I.M. Khalatnlkov, Adv. Phys. ]2 ,185 (.11963)

,14 U . H , G e r l a c h and U.K, Sengupta. Phys. Rev D,18 .11773 (,1978)


37

Appendix A: Harmonics on the 3 - s p h e r e

In this a p p e n d i x we d e s c r i b e the p r o p e r t i e s of the s c a l a r , v e c t o r and

tensor harmonics on the 8 - s p h e r e S 3. The m e t r i c on S 3 is G i j and so

the line e l e m e n t is

d 2 = f~..dxZdx 3 = dx 2 + sin2x(d82 + sin28d# 2) (A])


z3

A vertical stroke will denote covariant differentiation with respect to the

metric R i j ' Indices i,j,k are raised and lowered using R i j "

(1) Scalar Harmonics

The s c a l a r spherical harmonics O/_m


n ( x,O,#) are scalar eigenfunctions of

the Laplacian o p e r a t o r on S 3 ' Thus, they satisfy the e i g e n v a l u e e q u a t i o n

Q(n) Ik (n 2 I) Q(n)
Ik . . . . n = 1,2,3... (A2)

The most g e n e r a l solution to ( A 2 ) . for given n, Is a sum of solutions

n-i J
Q(n)(x,O,~ ) ~ ~ An n (A3)
= =o m = - Im Q ~ m (x''~)

w h e r e A nl m are a set of a r b i t r a r y constants. n


The QP.m are given explicitly

by

n
Qjm(X,e,) = n;(x)Ytm(e,#) ( A4 )

where Ylrll ( 8 , # ) are the usual harmonics on the 2-sphere, S 2, and


13, )4 n
[I ( X ) are the Fock harmonics The spherical harmonics Qim

constitute a complete orthogonal set for the e x p a n s i o n of any s c a l a r field

on S 3.
38

(2) Vector Harmonics

The t r a n s v e r s e v e c t o r harmonics (Si)jm(x,e,#)


n
are vector eigenfunctions

of the Laplaclan operator on S3 which are transverse. That is, they

satisfy the e i g e n v a l u e e q u a t i o n

s(n) Ik = _ (n 2 . 2 ) S~ n ) n = 2,3,4... (AS)


l Ik l

and the t r a n s v e r s e c o n d i t i o n

Shn ' '"i " = o (A6)


1

The most g e n e r a l s o l u t i o n to ( A 5 ) and ( A 6 ) is a sum of s o l u t i o n s

n-i
s(n)(x,8,#) = E E Bn n
~=I m=- ~m (Si)~m(X'e'#) (A7)

w h e r e B nm a r e a set of a r b i t r a r y constants. Explicit e x p r e s s i o n s f o r the


n
(Si)m a r e given in r e f e r e n c e 14 w h e r e it is a l s o e x p l a i n e d how they a r e

c l a s s i f i e d as odd (o) or e v e n (e) using a parity t r a n s f o r m a t i o n . We thus

have two linearly independent transverse vector harmonics Si and sei

( n , , m s u p p r e s s e d )

n
Using the scalar harmonics Qm we may construct a third vector
n
harmonic (Pi)m. defined by ( n , . m suppressed)

1
Pi - Qli n = 2,3,4... (AS)
(n 2 - 1)

It may be shown to satisfy

Ik ( n 2 - 3) P. and p. I i = - Q (A9)
Pilk = - z x
39

The three vector h a r m o n i c s S i ' S ei and P i constitute a c o m p l e t e o r t h o g o n a l

set for the expansion of any vector field on S 3,

(3) Tensor Harmonics

n
The transverse traceless tensor harmonics (Gi3)9.m (x,e,~) are tensor

elgenfunctions of the Laplacian o p e r a t o r on S 3 which are transverse and

traceless. That is. they satisfy the e i g e n v a l u e equation

G(n) Ik G(.n) (AI0)


ij Ik = - (n2 - 3) 13 n = 3,4,5...

and the transverse and traceless conditions

c~n )li = o G~ n)i = 0 (All)


l3 ' 1

The most g e n e r a l solution to ( A l l ) and (A12) is a sum of solutions

n-i
G(.n)(x,8,#) = )'1 ~" C n (Gij); (A12)
13 9=2 m=- P. Em m( x ' e ' # )

where Cnm are a set of arbitrary constants. As in the vector case they

may be classified as odd or even. Explicit expressions for ( GO 13 and


i j )m

( Gei j )9,m
n are given In r e f e r e n c e 14

Using the transverse vector harmonics ( S io )n E m and (s ei)m,


n we m a y
o n e n
construct traceless tensor harmonics (Sij) m and (Sij) m defined, both

for odd and even, by ( n , , m s u p p r e s s e d )

Sij = Sil j ~ Sjl i (A'I3)

i
and thus S , = 0 since S . is transverse. In addition, the S i j may be
1 1
shown to satisfy
40

S.. Ij = ._ ( n 2 _ 4 ) S, (A]4)
13 1

S,,.i~l i = o (A]5)
13

Ik (n 2 6) (AI6)
Sijlk = - _ Si3

n
Using the scalar harmonics Qm" we may construct two tensors
n n
(Qij)Jtm and ( P i j )~m defined by ( n , . m suppressed)

1
Qij = 3-flijQ n = 1,2,3 (A17)

1
and Pij (n 2 - i) Qlij 3~ij Q n ~ 2,3,4 (A]8)

i
The Pij are traceless "Pi = 0, and in addition, may be shown to

satisfy

Ij 2 2
Pij = - ~n - 4) Pi (A19)

Pijlk Ik = -- ( n 2 - 7) Pij (A20)

.
]ij = ~2( n 2 - 4) Q (A2I)
P13

The six t e n s o r h a r m o n i c s Qij,Pij, S i j ' S ei j ' Gi j and G ei j constitute

a c o m p l e t e o r t h o g o n a l set for the e x p a n s i o n of any s y m m e t r i c s e c o n d rank

tensor field on S 3,

(4) Orthogonality and Normalization

The n o r m a l i z a t i o n of the scalar, v e c t o r and t e n s o r harmonics Is fixed by

the o r t h o g o n a l i t y relations, We d e n o t e the i n t e g r a t i o n m e a s u r e on S 3 by

dp.. Thus
41

d/z = d3x (dehflij)]/2 = sin2X sine dxded~ (A22)

n
The _Otto are normalized so that

fd# Qim
n Q~'m'
n' = onn' 8~, 8mm, (A23)

This implies

n i n' = _ 1 6nn' 8mm, (A24)


fd/A (Pi)~m (P)~'m' (n 2 - I) ~i'

and

n ij n' = 2(n 2 - 4) 6nn' 8mm, (A25)


fd~ (Pij)im (P )'m' 3(n 2 i) 6,

The ( S i ) nm . both odd and even, are normalized so that

n (S z )'m'
;d# (Si)m n' = 8nn' 8, 8mm, (A26)

This implies

~d/L (Sij)Imn (siJ)n',m, = 2(n2 -. 4) 8 nn' 8i , 6ram ' (A27)

Finally, the ( G i 3 )m'


n both odd and even, are normalized so that

fd/~ (Oij n (dijon ' = 8nn' (A28)


)m " "'m' 8, mm'

The information given in this a p p e n d i x about the spherical harmonics

is all that is n e e d e d to perform the derivations presented in the main text,

Further details may be found in r e f e r e n c e s 13 & 14


42

Appendix B: Action and Field Equations of the Infinite Dimensional Model

The action

I = lo((Z,~,N0) + EI (Bl)
n rl

where I 0 Is the action of the unperturbed model:

Io = - %fdt
.
SOe
3a{&
/~-
2 e
-2a
-
.2
_~_ + m z c z
2
j (B 2 )
[N o NO

{n is quadratic in the p e r t u r b a t i o n s and may be written

In = Idt(L3 + L mn ) (B3)

where

Ln = //ae~o ~i n 2-25_)an2 + (n2-7_/~ 4 / b 2 _ 2(n2_4)C2n _ (n2l)d2n ~(n2-4)anbn


2
g 3 (n2_l) n

gn['32-(n2-4)bn 4 32-(n2I/2)anl+ ~ I 3(n2-i!)k2


I + (n2
n 4)j2nl

2 .2
3(z _ ~2 LD/-4)62 ~ (n2 4)62 + d
+ I~ 0 n (n2_l) n - n

& I - 2an&n + 81D~-n~/b


(n2 i) n 6n + e(n2- 4)CnCn + Sdndnl

+ 21 - ~32
an + 6 02+
(n21) n
6(n2 - 4)C2n + 60:I

+ gn[2 o+ 23an Onl


43

and

!
L n = VaNo e3a 1 f n + j _ m 2 .2 + - ( - )n
m

-I- 3 ~ - m 2 , ~ 2 2 4(n2-4-)b 2 - 4(n2-4)C2n - 4,d + - . ~ g 2n


2 -~-2-No an- (n2_].) n 0

I
- gn 2m2fn@ + 3m2an 2 + 2--~-2- + 3-~D'--~-
NO
fn~ "21
e--a
N 2 J - ~knfn
" (B5)

The full expressions for n a and nl~ are

7/'a ++3+,
-~-0 [ - & + F..
n
I_ anon 2
+ 4(n -4)b 6 + 4(n2-4)Cn6 n + 4rindn
(n2 I) n n
1

n 2 n (n2_l) n

F. gn{~n t &(3a n - gn) + 3 e kn] (B7)


n

+~=.- ( ~-(~ _ +__..: _ +<o.c: _ ~:ii +

n n + 3an~- e-+kn~olJ (B7)

The classical field equations may be obtained from the action (B1) by

varying with respect to each of the fields in turn. Variation with respect
44

to = and # gives two field equations, similar to those obtained In section

II. 4. but modified by terms quadratic in the perturbations:

N~m2~ = quadratic terms (88)


I o"l " :'

NO d E &l(j + 3 # .2 - N2e -2a

_ _ (~2 + # - INOe + N m2#2

Variation with respect to the perturbations an,bn,Cn,d n and fn leads to

five field equations:

d. 3dan.
NO ~E(e
1 2
~ ; ) + ~ ( n - 4)N20en'(an -f b n ) + 3 e 3 e ( ~nf , - No2m2#fn ) -

N213e3am2#2 - ~(n2+2)ea}gn + e3a&gn - ~N ~dl f e ~ J d ~oJ (8,o,

d e3Ul~n 1 2 2 a 1 n2
~(
N O dt( N;) - ~n - l)Noe (an + bn ) = - 1 )N2eagn

(8ll)
+ ?o dt! "ol

d. 3c~Cn. d [ 2aJn}
~(e Eo,= ~{e EoI (812)

d e3adn (B'I3)
NO dt( ~00 ) + (n2 - l)N2e~dn = 0

" n [~o[m
N O d~tt(e3afN-~) + 3e3a~a . 2[ 2 e 3a + (n2-1 ) e a l f n =
0
45

e3={ - 2N~m2#g n + ~C3n - e-O~#knJ (814)

In obtaining (Bl0) - (B14). the field equations (B8) and (89)have been

used and terms cubic in the perturbations have been droppped.

Variation with respect to the Lagrange multipliers kn, J n , g n and NO

leads to a set of constraints. Variation with respect to k n and 3n leads


to the momentum constraints;

&n 4. U~J-415 4- 3~n~ = " e-= k (815)


(n2_l) n ~gn (n2_l) n

6 n = e-aJn (816)

Variation with respect to gn gives the linear Hamiltonian constraint:

.2

+ NoZ ma(2fn ~ + San@2) - 3-~0 2e-Zaf[(n 2-4)bn 4. (n2 + 1/Z)anJ

.2

"-~e-~Zk3 n + Zgn( _ &2 4- @ ) (817)

Finally. variation with respect to NO yields the Hamiltonian constraint.


which we write as

~e3a - - - ~~2 .+ - - - e -2~ 4. ~ $ z = quadratic terms (818)


NO NO
SOLITONS AND B L A C K HOLES IN 4,5 D I M E N S I O N S

G.W. Gibbons
Department of A p p l i e d M a t h e m a t i c s and T h e o r e t i c a l Physics,
University of Cambridge, Silver Street, C a m b r i d g e CB3 9EW
U.K.

Contents

I) Introduction
2) Topology and Initial Data
3) The Black Hole as Soliton
4) Solitons in 5-dimensions
5) Pyrgon-Monopole duality

I. Introduction

This is the w r i t t e n version of two lectures given in Paris in


February 1985. Since the m a t e r i a l as given has now a p p e a r e d elsewhere
[1,2] I have d e c i d e d not to r e p e a t the lectures verbatim but rather
to comment on the general problem of solitons in gravity, in p a r t i c u l a r
on the importance or otherwise of spatial and spacetime topology
contrasting the situation in 4 and in 5 spacetime dimensions. My m a i n
p o i n t will be that while there are many similarities w i t h the s i t u a t i o n
in Y a n g - M i l l s - H i g g s theory there are s i g n if i c a n t differences. In
particular the a p p a r e n t l y inevitable occurrence of spacetime singularities
and their c o n j e c t u r e d shielding by event h o r i z o n s (Cosmic Censorship)
means that one cannot assume that the time e v o l u t i o n of initial data
is continuous. This substantially alters ones views of the importance
of topology in the c l a s s i c a l theory. It is highly likely that the
quantum theory - should it make m a t h e m a t i c a l sense - will be similarly
affected. The plan of the article is as follows: in section 2 I will
discuss some t o p o l o g i c a l aspects of the initial value problem. In
section 3 I will d e s c r i b e why I don't feel one can regard black holes
as solitons except in the extreme Reissner-Nordstrom case, and the
relation of this to supergravity. In section 4 I will contrast the
situation with that in 5 - d i m e n s i o n s and I will argue that the true
analogue of m a g n e t i c monopoles in Y a n g - M i l l s theory are the m u l t i - T a u b
NUT solutions whose importance for K a l u z a - K l e i n theory was first
stresse d by Gross, Perry and Sorkin. Their relation to black holes
will also be described. In section 5 I will d e s c r i b e a duality
47

conjecture analogous to that of O l i v e & Montonen in the Y a n g - M i l l s


case.

2. Topology and the Initial Data

It is an a t t r a c t i v e idea that the way to study solitons and o t h e r


topological features in General Relativity is to start with an initial
data set {E,gij,Kij} where E is a 3 - d i m e n s i o n a l manifold, gij a
Riemannian metric and K.. the second f u n d a m e n t a l form. The m e t r i c
13
and second fundamental form just satisfy c e r t a i n c o n s t r a i n t s and be
asymptotically flat. Indeed one could imagine more than one a s y m p t o t i c
region, just as there is in the S c h w a r z s c h i l d vacuum solution. The k
asymptotic regions may be i m a g i n e d to be c o m p a c t i f i e d to give a compact
manifold ~ , Z being diffeomorphic to E with k points removed.
There is no complete topological classification of 3 m a n i f o l d s but it
is k n o w n [3] that for o r i e n t a b l e manifolds and factors E may be
expressed uniquely as the c o n n e c t e d sum of a number of "prime m a n i f o l d s "
E
l

~ ZI # Z2 "'" # Zn (I)

A complete list of prime manifolds is not k n o w n but it is k n o w n that


for instance S2 x S I and e l l i p t i c spaces S3/F where F is a suitable
d iscret e subgroup of SO(4) with free action on S3 are prime. Initial
data satisfying the c o n s t r a i n s which are o r i e n t a b l e are, according to
Schoen and Y a u [4] p r o b a b l y limited to a sum of S 2 S1's and e l l i p t i c
spaces.

The existence of a u n i q u e factorization has led W i t t e n [5] to


argue that there are no solitons in 4 - d i m e n s i o n a l gravity because if
there were one w o u l d expect an a n t i s o l i t o n 3-metric such that one
could write:

S3 : Z # ~ (2)
s s

where Zs is the soliton 3-space topology and ~s that of the anti-


soliton. If E is prime this is ruled out by the u n i q u e n e s s One
s
now seems to have a p r o b l e m with CPT since (2) implies that the
soliton antisoliton pair cannot have the q u a n t u m numbers of the vacuum.
The way out of this particular difficulty would seem to be that topology
is not a "good q u a n t u m number". This seems reasonable because it
appears that any t o p o l o g i c a l l y non-trivial initial data set m u s t evolve
to give spacetime singularities in its future [6].
48

A c c o r d i n g to the widely believed but still as yet u n p r o v e d Cosmic


C e n s o r s h i p Hypothesis [7] these singularities will be shielded inside
event horizons. Furthermore it is also widely believed that the final
state (in the classical theory) will consist of one or more time
independent black holes. These black holes will have the m e t r i c of
the Kerr solution.

The consequences of this are rather d i s a p p o i n t i n g as far as


spatial topology is concerned. Suppose one started with for instance
one of Sorkin's n o n - o r i e n t a b l e wormholes [8]. That is E = P the
non-orientable S2 bundle over SI It is not d i f f i c u l t to construct
initial d a t a w i t h this t o p o l o g y [9]. This has a number of fascinating
topological properies [3]. For instance, topologically:

P # (S 2 S I) ~ P # P (3)

w h i c h one m i g h t i n t e r p r e t as saying that two n o n - o r i e n t a b l e w o r m h o l e s


could turn into a n o n - o r i e n t a b l e w o r m h o l e and a c o n v e n t i o n a l o r i e n t a b l e
wormhole. All of this however will be invisible from infinity since
p r e s u m a b l y each or maybe both will be s u r r o u n d e d by event horizons and
the fact that they are t o p o l o g i c a l l y n o n - t r i v i a l will play no role in
the e x t e r i o r dynamics. The final black hole s o l u t i o n will be a
S c h w a r z s c h i l d or Kerr metric and no hint of the interior t o p o l o g y will
show up in that.

Very much the same applies to the s i g n i f i c a n c e of the 8-vacuum


structure of the initial data. One might v i e w the c o n f i g u r a t i o n space
Q for gravity as the space of R i e m a n n i a n metrics on E factored by
the set of d i f f e o m o r p h i s m s Diff,(E) having a point on E (the point
at infinity) and its tangent space invariant. If Diff,(~) is not
c o n n e c t e d the c o n f i g u r a t i o n space Q will not be simply connected and
% - v a c u u m analogous to those in Yang-Mills theory are possible [10].
A p a r t i c u l a r instance of this is the beautiful work of Sorkin and
Friedman [11] on spin from gravity. Because Q is not simply connected
a r o t a t i o n of the spacetime relative to infinity may result in one
moving around a closed loop in Q which is not h o m o t o p i c to the constant
path. The q u a n t u m wave function could in principle change sign under
such a rotation. As an example consider as they do ~ to be S3/F
where F is the 8 element group c o n s i s t i n g of the q u a t e r n i o n s and their
negatives t o g e t h e r w i t h I . It is quite easy to construct time
symmetric initial data c o r r e s p o n d i n g o this space. The r e s u l t i n g space
can be thought of as c o n t a i n i n g 7 black holes s u i t a b l y i d e n t i f i e d [9].
Despite the exotic t o p o l o g y it seems rather likely that the end result
49

will be just one large b l a c k hole. Again there will be no sign in the
external metric of the initial exotic topology.

Finally as a final argument against the s i g n i f i c a n c e of 3-space


t opolog y let me remind the r e a d e r of the well k n o w n t h e o r e m of Serini,
Einstein, Pauli and L i c h n e r o w i c z which I like to p a r a p h r a s e as "No
solitons without horizonts" The t h e o r e m states that there are no
regular globally static solutions of the v a c u u m E i n s t e i n equations
other than the flat one. The a r g u m e n t depends on the fact that if
goo = v2' with V ~ 0 on ~ and V I at i n f i n i t y the field e q u a t i o n s
imply that

V.VIv = 0 (4)
1

w here Vi is c o v a r i a n t differentiation with respect to the spatial


metric gij" The m a x i m u m principle immediately shows that V = I.
The r e m a i n i n g field e q u a t i o n now reads

R = 0 (5)
l]

w here Rij is the Ricci tensor of gij which in ~ - d i m e n s i o n s shows


that gij and hence the 4 - d i m e n s i o n a l m e t r i c must be flat.

3. The B l a c k Hole as S o l i t o n

The remarks in s e c t i o n 2 have been i n t e n d e d to c o n v i n c e the reader


of the importance of the 4-dimensional d y na m i c s of the t h e o r y as opposed
to that of 3 - d i m e n s i o n a l initial data. This does not m e a n that one can
necessarily regard black holes as solitons. Far from it. They have no
fixed mass or angular m o m e n t u m even in the c l a s s i c a l theory. Indeed
the n o n - d e c r e a s i n g property of the e v e n t h o r i z o n area is a n y t h i n g but
solitonic. The s i t u a t i o n is even w o r s e in the q u a n t u m theory since we
know from the work of H a w k i n g [12] that black holes are u n s t a b l e against
thermal evaporation. We are still ignorant of the final outcome of this
process w h i c h may not be c a l c u l a b l e in E i n s t e i n t h e o r y but m a y require
a consistent quantum theory of gravity. A plausible guess is that the
hole simply d i s a p p e a r s in a puff of radiation. If this is true the
b lack hole should be r e g a r d e d in the q u a n t u m theory as an u n s t a b l e
"intermediate state", rather than a stable p a r t i c l e - l i k e state.

The e x c e p t i o n to this w o u l d be if the hole carried a "central"


charge. By central I mean completely conserved and not c a r r i e d by any
of the f u n d a m e n t a l fields of the theory. For e x a m p l e in N=2 ungauged
50

extended s u p e r g r a ~ i t y [13] there is a Maxwell field. The fields of the


N=2 supergravity m u l t i p l e t s are the graviton, the photon and the
gravitino. These are all e l e c t r i c a l l y neutral with respect to the
Maxwell field - that is why the theory is "ungauged". It is quite
possible for black holes to carry this charge - e s s e n t i a l l y because
the lines of flux are "trapped in the topology" as people used to say
in the days of "Geometrodynamics". The m e t r i c of such holes (if non-
rotating) is that of Reissner and Nordstrom. It is p a r a m e t e r i z e d by
the mass M and charge Q . Because of the duality, invariance of the
theory of any m a g n e t i c charge may be rotated to zero by a suitable
d u a l i t y rotation. The s i n g u l a r i t y is clothed by an event horizon if

M ~ IQI/K (6)

where K 2 = 4~G and G is N e w t o n ' s constant. I have d e s c r i b e d in more


detail e l s e w h e r e [I,2] how one m a y v i e w (6) as a B o g o m o l n y type i n e q u a l i t y
[see also 14,15,16,17]. The electric charge Q is truly central in
the sense of the s u p e r s y m m e t r y algebra and the inequality in (6) is
saturated by extreme black holes which are "supersymmetric" in that
they possess "Killing spinors". There exist a whole family of multi-
black hole m e t r i c s [17] satisfying (6). These are the Papapetrou-
M a j u m d a r metrics [18] which are included in the general class of
I s r a e l - W i l s o n metrics [19]. Tod [20] has shown that the I s r a e l - W i l s o n
metrics e x h a u s t all the metrics with Killing spinors in N=2 supergravity.
It has been known for some time that the throat of the extreme Reissner-
N o r d s t r o m metric has the g e o m e t r y of the R o b i n s o n - B e r t o t t i solution,
i.e. the product metric on S 2 x (ADS)2 when (ADS) 2 is 2 - d i m e n s i o n a l
anti-de Sitter space. The R o b i n s o n - B e r t o t t i metric shares with flat
space the p r o p e r t y of being m a x i m a l l y s u p e r s y m m e t r i c - i.e. of having
the largest possible number of K i l l i n g spinors. Thus the extreme
R e i s s n e r - N o r d s t r o m metrics spatially interpolate b e t w e e n the 2 possible
"vacua" of N=2 u n g a u g e d supergravitY. The p o s s i b l e relevance of this
remark for spontaneous c o m p a c t i f i c a t i o n is intriguing. For the present
let me remark that this is typically s o l i t o n - l i k e behaviour.

Since the charge is central it cannot be lost during H a w k i n g


e v a p o r a t i o n and so a hole with an initial charge must settle down to
the lowest mass state with that charge. This is the extreme (zero
temperature) state. This extreme R e i s s n e r - N o r d s t r o m holes seem to
behave just like solitons. The hole with the opposite charge is clearly
the a n t i s o l i t o n and it seems e x t r e m e l y p l a u s i b l e that a s o l i t o n - a n t i -
soliton might completely annihilate one another. They cannot do this
51

classically if C o s m i c Censorship holds since by H a w k i n g ' s area theorem


the final event h o r i z o n m u s t have n o n - v a n i s h i n g area but the r e s u l t a n t
Schwarzschild black hole can then e v a p o r a t e thermally.

The m a i n way in w h i c h the extreme holes d i f f e r from solitons is


that there seems to be no w a y of fixing their mass or charge - i.e.
no q u a n t i z a t i o n rule.

Since the e x t r e m e holes (which need not all have the same mass),
can remain in e q u i l i b r i u m it is r e a s o n a b l e to c o n s i d e r departures
from e q u i l i b r i u m perturbatively. To lowest order they should move
on g e o d e s i c s on a suitable "moduli space", that is to lowest order the
parameters specifying the solution should change slowly. This is the
same approximation as has been used successfully in Y a n g - M i l l s theory
[21,22]. In the p r e s e n t case the P a p a p e t r o u - M a j u m d a r solution
(representing N b l a c k holes) is s p e c i f i e d by giving the positions of
N points in ~3 . In p r i n c i p a l the points could coincide though I will
argue in a short w h i l e that this doesn't happen. If the holes, having
equal masses, were identical one w o u l d factor by the a c t i o n of the
permutation group SN on the N positions. Thus we k n o w the m o d u l i
space. The m e t r i c is not known. However if one makes the a p p r o x i m a t i o n
that one hole is very m u c h smaller than all the others one can anticipate
that the m o t i o n of the small hole in the field of the others should be
g i v e n by the s t a n d a r d equation for a charged g e o d e s i c (with charge =
mass x K). In the slow m o t i o n limit this does indeed give non-relativistic
geodesic motion in the m e t r i c

ds 2 = U3dx 2 (7)

w here
i=N-1 GM.
l
U = I + Z (8)
i=I Ix -

This m e t r i c is complete on ~ 3 _ {xi} " In this approximation the holes


would take an infinite time to merge or coalesce.

The q u a n t u m scattering of e x t r e m e holes could be studied in the


non-relativistic limit by looking at the S c h r 6 d i n g e r equation on the
moduli space. This w o u l d presumably correspond to the scalar L a p l a c i a n
with respect to the m e t r i c on the m o d u l i space, though it is also
possible that p o t e n t i a l terms m i g h t appear due to one loop effects. In
the case that the holes all had equal mass one should divide out by
the p e r m u t a t i o n group. The moduli space w o u l d have fundamental group
S N. The wave function could in p r i n c i p l e then be even or odd u n d e r
52

permutation. Thus one could imagine "fermionic" black holes! This is


the analogue of the e f f e c t of S o r k i n and F r i e d m a n I d e s c r i b e d above.

It is p o s s i b l e to find e x t r e m e b l a c k holes in the N=4 ungauged


e x t e n d e d s u p e r g r a v i t y t h e o r y as w e l l [23]. T h e y s h o u l d also p r o b a b l y
be t h o u g h t of as solitons. Like the e x t r e m e holes in N=2 they also
have no natural mass q u a n t i z a t i o n - at least as far as c l a s s i c a l or
semi-classical considerations are concerned. To get a s a t i s f a c t o r y
q u a n t i z a t i o n rule one seems f o r c e d to turn to K a l u z a - K l e i n theory.

4. Solitons in 5 - d i m e n s i o n s

M u c h of the d i s c u s s i o n about the r e l e v a n c e of t o p o l o g y in s e c t i o n


2 could be r e p e a t e d here w i t h 4 replacing 3. The d e t a i l s of the
t o p o l o g i c a l d i s c u s s i o n w o u l d d i f f e r and we c e r t a i n l y d o n ' t have d e t a i l e d
singularity theorems and black h o l e u n i q u e n e s s theorems in h i g h e r
d i m e n s i o n s - i n d e e d we k n o w v e r y little about b l a c k holes in h i g h e r
dimensions. However in h i g h e r d i m e n s i o n s gravity is e v e n m o r e a t t r a c t i v e
(having a force i n v e r s e l y as d i s t a n c e to the p o w e r of the d i m e n s i o n of
spacetime minus 2) than in 4-dimensions. In 5 - d i m e n s i o n s it d e p e n d s on
distance in the same way as the r e p u l s i v e centrifugal force (which is
i n v e r s e l y as d i s t a n c e cubed in all d i m e n s i o n s ) . In h i g h e r d i m e n s i o n s
it rises e v e n m o r e r a p i d l y t h a n the c e n t r i f u g a l repulsion. This would
seem to m a k e g r a v i t a t i o n a l c o l l a p s e and s p a c e t i m e singularities even
more l i k e l y in h i g h e r d i m e n s i o n s .

H o w e v e r there is an i m p o r t a n t d i f f e r e n c e . We are no l o n g e r o b l i g e d
nor w o u l d we w i s h to c o n f i n e o u r s e l v e s to i n i t i a l d a t a w h i c h are
asymptotically Euclidean. If we do so the a r g u m e n t that the v a n i s h i n g
of the Ricci t e n s o r implies that the 4 - s p a c e is flat still goes t h r o u g h
a c c o r d i n g to S c h o e n and Yau's P o s i t i v e A c t i o n T h e o r e m [25]. If we
d o n ' t r e q u i r e that the 4 - m e t r i c be a s y m p t o t i c a l l y E u c l i d e a n there are
m a n y c o m p l e t e Ricci flat 4-metrics, i n c l u d i n g one - t h a t on the K3
s u r f a c e - w h i c h is compact. Any gravitational i n s t a n t o n will give a
static 5 - m e t r i c w i t h no horizons. Note that if we have no h o r i z o n we
are still f o r c e d to have V = I, that is the m e t r i c m u s t be a p r o d u c t
on ~ x M , where M is the 4-manifold. In the o l d e r l a n g u a g e the
s p a c e t i m e w o u l d be said to be "ultrastatic".

N o t all of these o b j e c t s w i l l be c l a s s i c a l l y stable. The s t a b i l i t y


will be g o v e r n e d by s p e c t r u m of the L i c h n e r o w i c z L a p l a c i a n a c t i n g on
s y m m e t r i c t e n s o r s on M . If M has a s e l f - d u a l m e t r i c this is k n o w n to
be p o s i t i v e and h e n c e the c o r r e s p o n d i n g static lump w i l l be c l a s s i c a l l y
stable. If M has a m e t r i c w h i c h is not s e l f - d u a l the s p e c t r u m is not
58

likely to have a positive spectrum and the c o r r e s p o n d i n g lump will be


unstable. Examples of this are the E u c l i d e a n Schwarzschild solution [26]
and the "Taub-Bolt solution" [27].

The e v o l u t i o n of these objects in the full n o n - l i n e a r theory is


unclear. The E u c l i d e a n Schwarzschild solution has the same a s y m p t o t i c s
as the flat m e t r i c on ~3 S I so p r e s u m a b l y it loses energy to gravi-
tational radiation and attempts to settle down to the flat m e t r i c but
it can't do this w i t h o u t forming some sort of s i n g u l a r i t y since this
would involve a spatial topology change. It seems likely that a black
hole will be formed but this is not known. The same remarks apply to
the T a u b - B o l t metric which presumably tries to settle down to the Taub-
NUT metric. Again black hole formation seems likely. It is p o s s i b l e
that these black holes appear regular w h e n ~ i e w e d from a 4 - d i m e n s i o n a l
stand-point in w h i c h case they should be i n c l u d e d w i t h those d e s c r i b e d
in [24] and [28]. The H a w k i n g effect m a y then cause these black holes
to evolve to the flat or the T a u b - N U T solution.

The b o u n d a r y conditions of i n t e r e s t for K a l u z a - K l e i n theory is that


the m e t r i c be w h a t has been called in this context asymptotically flat -
i.e. that it a p p r o a c h the flat p r o d u c t metric on R3 x SI at infinity
or that it be a s y m p t o t i c a l l y locally flat. The typical example of the
latter is the self-dual Taub-NUT metric, or m u l t i - T a u b - N U T with N
centres. The t o p o l o g y at infinity in this case is ~ BN when BN is
the S1 bundle over S2 w i t h Hopf invariant N - i.e. the lens space
L(N,I) .

Gross, Perry and S o r k i n [29] have p o i n t e d out that the T a u b - N U T


solution plays the role of a m a g n e t i c monopole in K a l u z a - K l e i n theory.
Perry and m y s e l f [30] have shown that the m o n o p o l e moment P of any
asymptotically locally flat solution should satisfy the B o g o m o l n y type
inequality

IPl < M [9)


2< --

with equality in the s u p e r s y m m e t r i c self-dual case. It is i n t e r e s t i n g


to note that the g r a v i t a t i o n a l instanton solution of A t i y a h & Hitchin
[22] is s e l f - d u a l but has a negative mass. This is p r e s u m a b l y because
it has the topology at i n f i n i t y of ~ x (S3/F) where F is the b i n a r y
dihedral group. The crucial point here is w h e t h e r or not suitable
solutions of the W i t t e n equation exist.

The m u l t i Taub-NUT solutions are s p e c i f i e d by giving N non-


coincident points in ~3. Permutating the points gives the same m e t r i c
54

so the moduli space is the well k n o w n c o n f i g u r a t i o n space { (jR3) N - A } / S N


where A is the points in (jR3)N where two or more points coincide
and SN is as before the p e r m u t a t i o n group on N symbols. The m e t r i c
on the m o d u l i space is under study. Again the q u a n t u m m e c h a n i c s offers
the p o s s i b i l i t y of m u l t i v a l u e d wave functions though whether these
monopoles can r e a l l y be thought of as fermions remains at p r e s e n t unclear.

An i m p o r t a n t property of the T a u b - N U T solutions is that the m a g n e t i c


charge P satisfies the Dirac q u a n t i z a t i o n condition:

eP = 2~ (10)

where e is the basic unit of charge in K a l u z a - K l e i n theory.

This in turn implies that (using the e q u a l i t y in (9)) the mass M


is quantized:

M - 1 (11)
4~<e

Given their stability and the q u a n t i z a t i o n of the mass and m a g n e t i c


charge it seems reasonable to regard the T a u b - N U T solutions as
representing solitons though this does require, as in section 2, that
some of the t o p o l o g i c a l numbers associated with the o b j e c t are not
conserved. In the p r e s e n t case two such numbers are of interest. The
Hirzebruch signature and the Euler number. The m u l t i p l e monopole
has n o n - v a n i s h i n g Hirzebruch signature. Roughly it c o r r e s p o n d s to
magnetic charge. Since this can be read off from the asymptotic
boundary conditions one might expect this to be conserved. The Euler
number is a d i f f e r e n t matter however. This cannot be d e t e r m i n e d from
infinity and g i v e n the likely occurrence of s i n g u l a r i t i e s there seems
to be no good r e a s o n for it to be conserved. Another argument, due to
Hawking, is that the E u c l i d e a n action in General Relativity is not
scale invariant. This means that it may cost arbitrarily little action
to pass from one t o p o l o g i c a l configuration to another. This is unlike
the case in Y a n g - M i l l s theory in 4-dimensions where the a c t i o n is scale
invariant and typically topologically different configurations
differ by an amount 8~2/g 2 where g is the coupling constant. If
one does accept them as solitons one sees a number of s t r i k i n g
resemblances with the m a s s i v e modes of the K a l u z a - K l e i n theory. This
is the subject of the next section.
55

5. Pyrgon-Monopole duality

The p h y s i c a l content of the 5-dimensional Kaluza-Klein theory when


viewed from the p o i n t of v i e w of 4 - d i m e n s i o n s

I) A set of m a s s l e s s states, the graviton, graviphoton and


dilaton
2) A tower of m a s s i v e states of spin 0, 1 and 2 each with mass
m and charge e given by

m = n lel
2K (12)

w here n = 1,2,3,...

At the linearized level all the m a s s i v e states are t r i v i a l l y stable.


W h e n one takes into account interactions one m i g h t expect the h i g h e r
mass states to d e c a y into lower mass states but a charged state cannot
decay into a neutral state. Thus the lowest mass states, n=1 , should
be a b s o l u t e l y stable except against annihilation w i t h their antiparticle
state. These stable lowest mass states have been called Pyrgons [31].
Thus the p e r t u r b a t i v e physical Hilbert space consists of m a s s l e s s states,
pyrgons and antipyrgons. In a s u p e r s y m m e t r i c theory the pyrgons fit
into m a s s i v e supermultiplets with central charge. In N=8 for example
the r e l a t i o n (12) c o r r e s p o n d s to the m a x i m a l central charge allowed.
This is n e c e s s a r y to avoid states with spin greater than 2.

N o w the G-P-S m o n o p o l e s possess in the N=8 supergravity model of


Cremmer [32] the m a x i m u m permitted number of K i l l i n g spinors and hence
supersymmetries. As shown in [30] they fit into s u p e r m u l t i p l e t s when
the zero modes are taken into account. There is a rather close
analogy, indeed one is tempted to say a duality, between the m o n o p o l e s
of K a l u z a - K l e i n theory and the p y r g o n s . This suggested duality is
analogous to that w h i c h has been suggested in Y a n g - M i l l s theory [33].
In the p r e s e n t case we suggest that there m i g h t exist in the full
quantum theory operators which create and a n n i h i l a t e monopole states.
In a d d i t i o n there will be o p e r a t o r s which create the m a s s l e s s states.
If these satisfy an e f f e c t i v e field theory it is e s s e n t i a l l y unique -
it m u s t be the o r i g i n a l field t h e o r y of the pyrgons. This is e s s e n t i a l l y
because of the s u p e r m u l t i p l e t s structure. Thus we have the c o n j e c t u r e d
dualities:

monopole ++ pyrgon

massless fields ++ massless fields

antimonopole + anti-pyrgon.
56

It is d i f f i c u l t to see with present day techniques how such a


conjecture could be verified. In the Yang-Mills case some partial
e v i d e n c e has come from a study of m a g n e t i c and e l e c t r i c dipole moments.
It has been v e r i f i e d that the g y r o m a g n e t i c ratio of the o r d i n a r y Yang-
Mills particles equals the g y r o e l e c t r i c ratio of the m o n o p o l e s plus
fermionic zero-modes [34]. It is known that the g y r o m a g n e t i c ratios
of the pyrgons are anomalous and equal unity, rather than the Dirac
value of 2 [35]. It w o u l d be i n t e r e s t i n g to calculate the electric
dipole moments of G-P-S m o n o p o l e s with their fermionic zero-modes.

Further insight into this c o n j e c t u r e d d u a l i t y m i g h t come from a


study of m o n o p o l e - p y r g o n interactions. A number of authors [36] have
pointed out that there is no "Callan-Rubakov" effect [37] which would
catalyze the decay of pyrgons. This is m o s t easily seen from the fact
that scalar modes on Taub-NUT are well defined and u s i n g the c o v a r i a n t l y
constant spinor fields on T a u b - N U T one can obtain all solutions of the
Dirac equation.

Thus if g is a c o v a r i a n t l y constant spinor or T a u b - N U T and


a solution of the wave e q u a t i o n w i t h energy ~ then:

~ = [~ ~i (~)]~

are solutions of the Dirac e q u a t i o n with the same energy.

A s t r i k i n g fact about the scalar modes on the T a u b - N U T b a c k g r o u n d


is that the m a s s i v e scalar Pyrgon wave e q u a t i o n separates in 2 d i f f e r e n t
c o o r d i n a t e systems. One system is the standard radial v a r i a b l e s in
w h i c h the m e t r i c is

ds 2 : (I + ~ ) - 1 4 N 2 ( d ~ + c o s @ d * ) 2 + 1 1 +~) (dp2+p2(d@2+sin2@d*2)) (13)

where 0 ! ~ ! 47 . Thus 8 ~ N = 2~R K where RK is the radius of the


K a l u z a - K l e i n circle. The scalar field has the form
,n
~ = e -i~t elY~n Yim(0) eim~f (r) (14)
n

where ~Yzm(8)eim~ is a spin w e i g h t e d spherical harmonic and where


z
f (r) ls a n o n - r e l a t i v i s t i c Coulomb wave f u n c t i o n w i t h angular m o m e n t u m
n
but where the Coulomb potential is energy dependent, i.e. depends
u p o n ~ , that is f satisfies

12 d (p2 df) i(i+1)f


2 + (2N 2 - n 2 )f + ( 2 _ n 2 ) f = 0 (15)
p dp dp p N P 4N 2
57

There are no bound states, just scattering states. Since the radial
equation (15) is a Coulomb one one might anticipate that scattering is
better described using parabolic coordinates, defined by

= (I + cose)

q = (I - cose) (16)

This is in fact true. The wave equation also separates in the t,~,~,q
coordinates. Using them one can give a simple description of the
scattering. The classical orbits are especially simple being conic
sections. They are, when projected into the 3-space spanned by
p, 8 and ~, the intersection of a cone centred at p=0 with a plane,
the intersection being a hyperbola in general.

The existence of 2 different coordinate systems in which the wave


equation separates is often taken as the indication of hidden symmetries
and indeed of a "spectrum generating algebra". The precise nature of
this algebra in the present case has not been worked out. It is tempting
to speculate that it may be related to the known existence of Kac-Moody
algebras in K a l u z a - K l e i n theory [38].

Ano%her tempting speculation is that these ideas will find their


full expression in string theory. Mike Green [39] has remarked that
if one considers 10-dimensional string theory on where 10-D of the
spacelike dimensions form a torus, each of whose radii equals R one
obtains string states with masses satisfying

M~ R2N~
(mass)2 = i=0[ (--!lR
2 + ___~i)~,2 + _~2~, (NO + No ) (17)

No and No are occupation numbers for higher string states. The


integers {M i} are K a l u z a - K l e i n charges resulting from the periodicity
in the 10-D compact dimensions. The integers {N i } are topological
charges associated with the number of times a closed string winds
round the i'th compact dimension. Consider the limit

R 0 and ~'
-R- constant = I

The resulting D - d i m e n s i o n a l field theory has an infinite number of


massive spin 2 supermultiplets whose masses are determined by 1
This theory is apparently identical to the theory obtained by starting
with 10-dimensional N=2 supergravity and c o m p a c t i f y i n g on a hypertorus
with (10-D) dimensions having finite radii R =I
58

Now set D=5 . The reduction of N=2 d=10 to 5 dimensions gives


Cremmer's N=8 D=5 model, with fts pyrgon states. On the other hand
from (17) we see that the states corresponding to zero K a l u z a - K l e i n
charge but non-vanishing topological winding numbers will survive in
this limit. These presumably correspond to the magnetic monopole states.

References

[I] G.W. Gibbons in "Supersymmetry, Supergravity and Related Topics"


ed. F. del Aguila, J.A. de Azcarraga and L.E. Ibanez, World
Scientific 1985.
[2] G.W. Gibbons in "Non Linear Phenomena in Physics" ed. F. Claro,
Springer Proceedings in Physics #3. Springer Verlag 1985.
[3] J. Hempel "Topology of 3-Manifolds", Princeton University Press
(1976).
[4] R. Schoen and S.T. Yau, Phys. Rev. Lett. 43 1457 (1979).
[5] E. Witten, Commun. Math. Phys. 100 197 (1985).
[6] D. Gannon, J. Math. Phys. 2364 (1975).
D. Gannon, G.R.G. 7 219 (1976).
C.W. Lee, Commun. Math. Phys. 51 157 (1976).
[7] R. Penrose, Ann. N.Y. Acad. Sci. 224 125 (1973).
[8] R. Sorkin, J. Phys. At0 717 (1977).
[9] G.W. Gibbons, unpublished.
[10] C.J. Isham, Phys. Lett. I06B 188 (1981).
C.J. Isham, in "Quantum Structure of Space and Time", eds. C.J.
Isham & M.J. Duff. Cambridge University Press (1982).
[11] J. Friedman & R. Sorkin, Phys. Rev. Lett. 44 1100 (1980);
see also B. Witt: Milwaukee preprint.
[12] S.W. Hawking, Nature (Lond.) 248 30 (1974).
S.W. Hawking, Commun. Math. Phys. 43 199 (1975).
[13] S. Ferrara & P. van Nieuwenhuizen, Phys. Rev. Lett. 37 1669
(1976).
[14] G.W. Gibbons, in Heisenberg Memorial Symposium, ed. P. B r e i t e n l o h n e r
and H.P. Durr, Springer Lecture Notes in Physics #160.
[15] G.W. Gibbons & C.M. Hull, Phys. Letts. I09B 190 (1982).
[16] G.W. Gibbons, in Proc. 4th Silarg Symposium, ed. C. Aragone,
World Scientific.
[17] J.B. Hartle and S.W. Hawking, Commun. Math. Phys. 26 87 (1982).
[18] A. Papapetrou, Proc. Roy. Irish. Acad. A51 191 (1947).
S.D. Majumdar, Phys. Rev. 72 390 (1947).
[19] W. Israel and G.A. Wilson, J. Math. Phys. 1 3 865 (1972).
[20] P. Tod, Phys. Lett. B121 241 (1983).
[21] N. Manton, Phys. Rev.
N. M a n t o n in "Monopoles in Quantum Field Theory" ed.
N. Craigie, P. Goddard and W. Nahm, World Scientific (1982).
[22] M.F. Atiyah and N. Hitchin, Phys. Lett. I07A 21 (1985).
[23] G.W. Gibbons, Nucl. Phys. B207 337 (1982).
[24] W. Simon, G.R.G. 17 761 (1985).
[25] R. Schoen and S.T__Yan, Phys. Rev. Lett. 42 547 (1979).
[26] D. Page, Phys. Rev.
B. Allen, Phys. Rev. D30 1153 (1984).
[27] R.E. Young, Phys. Rev. D28 2420 (1983).
[28] G.~. Gibbons and D. Wiltshire, Annals of Phys., in press.
[29] D. Gross and M.J. Perry, Nucl. Phys. B226 29 (1983).
R. Sorkin, Phys. Rev. Lett. 51 87 (1983).
[30] G.W. Gibbons and M.J. Perry, Nucl. Phys. B248 629 (1984).
[31] E.W. Kolb and R. Slansky, Phys. Lett. 135B 378 (1984).
[32] E. Cremmer in "Superspace and Supergravity", ed. S.W. Hawking and
S.W. Hawking & M. Rocek, Cambridge University Press 1981.
59

[33] C. Montonen and D.I. Olive, Phys. Lett. 72B 117 (1977).
[34] H. Osborn, Phys. Lett. 115B 226 (1982).
Bo-Yu. Hou, Phys. Lett. 125B 389 (1983).
[35] A. Hoysoyer et al., Phys. Lett. 134B (1984).
[36] P.C. Nelson, Nucl. Phys. 238B 638 (1984).
H. Ezawa and A. Iwasaki, Phys. Lett. 138B 81 (1984).
M. Kobayashi and A. Sugamoto, Prog. Theor. Phys. 72 122 (1984).
F.A. Bais and P. Batenburg, Nucl. Phys. B245 469 (1984).
[37] V. Rubakov., Pisma Zh. Eksp. Teor. Fiz. 33 658 (1981); Nucl.
Phys. 203B 311 (1982).
C.G. Callan, Phys. Rev. D25 2141 (1981).
[38] A. Salam and J. Strathdee, Annals of Phys. 141 316 (1982).
L. Dolan and M.J. Duff, Phys. Rev. Lett. 52 14 (1984).
[39] M. Green "The Status of Superstrings" undated Queen Mary College
preprint.
TRUNCATIONS IN KALUZA-KLEIN THEORIES

C.N. Pope

Blackett Laboratory, Imperial College, Prince Consort Road, London


SW7 2BZ, UK.

Certain mathematical aspects of Kaluza-Klein theories are discussed,


concerned with the ability to truncate the four-dimensional spectrum of
states to a finite subsector, including the graviton and Yang-Mills
gauge bosons. This yields a criterion by means of which certain
exceptional theories are singled out from the generic case.

i. INTRODUCTION
Kaluza-Klein theories provide a natural and geometrical unification
of gravity and gauge fields, in which general coordinate invariance and
local gauge invariance both arise as subsectors of general coordinate
invariance in a higher dimension. However, in a successful unification
of all the fundamental forces in nature, one would like the unifying
theory to be unique, and in this respect Kaluza-Klein theories seem at
first sight to fare rather badly. Not only does one have the usual
freedom , as with four-dimensional theories, to pick and choose what
fields are to be included in the Lagrangian, but one also has the
additional freedom to choose one's favorite dimension!
In order to try to single out the 'right' theory from all the
candidate theories, one requires some rather powerful criteria which
can be used to restrict the possibilities. Broadly speaking, such
criteria tend to divide into two categories; on the one hand there are
those based on physical principles derived from phenomenological
considerations, while on the other hand there are mathematical
principles based on the requirement of self-consistency of the theory.
Examples of physical principles would be the requirement that the
theory admit a realistic gauge group with chiral fermions, and a
Minkowski space ground state. Superficially such requirements seem
very reasonable, but it should be borne in mind that the natural
unification sca~ of any quantum theory of gravity is the Planck scale,
19
I0 GeV or i0 cm, whilst the physical principles mentioned above are
based on observations in particle accelerators at energy scales lower
by about 17 orders of magnitude, and cosmological comparisons involve
a ~urther extrapolation of about 40 orders of magnitude. Seen in this
light, it is perhaps rather premature to be imposing these physical
requirements from the outset. Of course the 'correct' theory should
ultimately be able to explain the observed phenomena, but it may well
61

be that they emerge in a highly non-trivial way as low-energy


collective phenomena which are by no means manifest in the fundamental
theory. The phenomenon of superconductivity is possibly a good analogy;
it can be understood very satisfactorily within the framework of the
non-relativistic SchrSdinger equation applied to the theory of
electrons in metals, but it is a consequence of highly non-trivial
collective effects which were only understood long after the
development of quantum mechanics. It would not have been very fruitful
at the time when quantum mechanics was being developed to have demanded
as a prerequisite of a successful microscopic theory that it be seen at
the outset to be able to explain superconductivity.
An illustration of the way in which a quantum theory of gravity
might look very different on microscopic and macroscopic scales is
provided by the idea of 'spacetime foam', originally suggested by
I 2
Wheeler and developed by Hawking. In order that spacetime be
macroscopically flat, these studies suggested at the Planck scale it
should be topologically complex and 'foamlike', with an effective
cosmological constant of order -i in Planck units, compared with the
observed A at large scales which is smaller by at least 120 orders of
magnitude. Thus a huge negative cosmological constant might even be a
desirable feature of the fundamental theory! This simple model should
probably not be taken too seriously at present, but it does serve to
illustrate the point that one should perhaps be wary of imposing our
everyday low-energy prejudices at the Planck scale.
We will therefore adopt the view that for now we should be guided
more by mathematical principles of self-consistency than by physical
considerations. Ultimately, an obvious requirement would be that the
theory should be finite, and it is possible that this may be such a
powerful criterion that it would lead to a unique unified theory. The
recent surge of interest in superstring theories is based in part on
such a hope. The issues involved here are far too complicated to be
answerable at present, but the same general principle of mathematical
consistency is one which can be applied at many more elementary levels.
This paper is primarily concerned with certain consistency questions in
classical Kaluza-Klein theories, and it is to these that we now turn.

2. INCONSISTENCIES IN GENERIC KALUZA-KLEIN THEORIES


Any (4+k)-dimensional theory which admits spontaneous
compactifications to ground state solutions of the form M4x Mk, where
M4 is four-dimensional spacetime and M k is a compact internal space,
admits a Kaluza-Klein interpretation in which fluctuations of the
82

(4+k)-dimensional fields around their ground-state values have a

4-dimensional interpretation as excitations of infinite towers of


massless and massive states (see, for example, ref.3). Retaining all
these states is necessarily consistent, since it is just a rewriting of
the original higher dimensional theory. However, if one tries to
truncate to just a finite number of states, then in a generic Kaluza-
Klein theory one is liable to run into inconsistencies. These arise
because in general the states which are retained act as inhomogeneous
source terms for the states which are discarded, and hence setting
these states to zero is inconsistent with their equations of motion.
To see this in detail, let us consider the case of pure gravity with
a cosmological constant in (4+k) dimensions, with field equation

- -- R + A = 0, (i)
~N 2 %N %N

We use the notation that the (4+k)-dimensional fields are 'hatted', and
M,N,... are world indices running over 4+k values. These will be
decomposed as M=(~,m), etc., where ~ runs over M4and m runs over M k.
Equation (I) admits ground state solutions of the form M4 x Mk,
where Mk is a compact Einstein space satisfying

2A
Rmn =-- gmn' (2)
k+2

and M4 is a 4-dimensional spacetime satisfying the Einstein equation


with cosmological constant. We will assume A is positive, so Mk can be
a space with continuous symmetries, such as the k-sphere with SO(k+l)
isometry group.
The standard Kaluza-Klein ansatz, which is designed to truncate to
just the gravity and gauge boson degrees of freedom in 4 dimensions,
corresponds to setting the (4+k)-dimensional metric to be

^ A ^ B
gMN = eM eN ~AB' (3)

where

e (x,y) = e ~(x) dx , (4)

~a(x,y) = e a m(y) d y m - K1a(y)


" A i ~(x) dx , (5)

where A,B .... are (4+k)-dimensional local Lorentz indices decomposed


m
into ~, ~,... running over M4 and a,b,.., in Mk, x and y are the
63

coordinates on M4and M k respectively, K i are the Killing vectors on Mk


generating its isometry group G (i=l , ...dim G) , and A i ~(x) are the
Yang-Mills gauge potentials. The k-bein eam(y ) is related to the metric
on Mk, which is unchanged from the ground state and still satisfies
(2), by ea eb
m n 6ab= gmn"
A straightforward calculation of the Riemannian curvature for (4)
and (5) shows that the Ricci tensor is given by

I Ki a Kj Fi Fj Y (6)
R~ = R~ 2 a ~y ~ '

^ i Ki K3 b Fi Fje~
Rab = Rab + --~ a e8 , (7)

^ 1 F i ~ (8)
R~b - 2 K1b DI3 ~ '

where

F i a~ = 2 V[ a Ai~]+ Cijk A j Ak~, (9)

the structure constants c.. k are defined by


i3

[K i , K j ] = cij k K k , (io)

where K i = K im 5m, and D~ is the Yang-Mills covariant derivative. Re~


and Rab are the Rieci tensors for the vielbeins e ~ (x) and eam(y ) on M4
and Mkrespectively.
In the local Lorentz basis eAM, the (4+k)-dimensional field equation
(I) reads

-- R + A = O. (ii)
RAB 2 "lAB 'lAB

Substituting (6) , (7) and (8) into II) now reveals the
inconsistencies. There are three cases to consider, corresponding to AB
= ~, ~b and ab in (ii). The first of these yields

1 = ~1 ( Fi ~y Fj ~ y ---~I Fi y8 Fjy8 K ia K j (12)


R 8 - - -2R ~8 + A ~8 ~ ) a'

where R = R~ ~ is the 4-dimensional Ricci scalar. All the terms in


(12) depend only on the spacetime coordinates x ~, except for the
Killing vectors K i, which depend upon the coordinates ym on M k. In fact
in general one has
64

K ia K j = 8ij + yij(y) (13)


a

where y13(y) is symmetric and tracefree in i and j. Thus defining T 13


by

TiJ~ = F(i~Y Fj ) ~y - ~I Fi 78 Fj 78 ~ 8 ' (14)

' 4

(12) implies that the tracefree part of T13 ~ must vanish

Tij~ - (dim G) -I 6ij T kke~ = 0, (15)

while the y-independent part of (12) gives the standard Einstein


equation with Yang-Mills source term. However, the algebraic constraint
(15) on the Yang-Mills fields means that the ansatz (4), (5) is
inconsistent with (e~ (x), A i~(x) ) being an arbitrary solution of the
4-dimensional Einstein-Yang-Mills equations.
The only way of overcoming this inconsistency of the truncation in
(4) and (5) is to restrict the Killin~ vectors to just those which
generate some subgroup G' of G for which

K i'a KJ' a = 8 i'j' (16)

Since (16) implies that each Killing vector has constant (unit) length,
the subgroup G' can be non-trivial only if the Euler number X of Mk
vanishes (since if X ~ 6 then all vector fields must vanish somewhere).
Even if X = O, then G' is usually a lot smaller than G. For example if
Mk is the SO(k+l)-invariant k-sphere, then G' is SU(2) when k=4n+3, and
G' can only be U(1) when k=4n+l. On a group manifold H, with G = HL x
HR, G' can be either HLOr HR .
If one were prepared to include more 4-dimensional fields in the
ansatz (4), (5), consistency could be restored for the entire isometry
group G by including massive spin 2 fields, which would introduce a
balancing y-dependence on the left-hand-side of (12). But we know on
general grounds that massive spin 2 can only beconsistently coupled to
5
gravity by coupling infinitely many such fields , which would defeat
the object of the truncation. If the entire isometry group G is
retained, then setting, these massive ~pin 2 fields to zero is
inconsistent with their equations of motion
The ~b components of (ii) present no difficulties, yielding the
Yang-Mills equation
65

D F = 0. (17)

However the ab components of (ii) yield

K i a KJb" F i a8 Fj~ = 0, (18)

another unacceptable constraint. This time there is no subgroup of G


for which this inconsistency is avoided. The resolution here is that
one must include scalar degrees of freedom
i'j' (x) in (4) and (5),
corresponding to allowing the metric components gmn to fluctuate as
~i'j'~ K i' K j' which means that now (18) is of the general form
Q,i'j'~ Fi~ n.'
a~ F3'e~ Of course we must still also restrict to the
subgroup G' for which (16) holds, since the introduction of the scalar
fields does not resolve the previous inconsistency problem in (12).
The situation described above in the case of the (4+k)-dimensional
pure Einstein equation is reasonably representative of any generic
Kaluza-Klein theory. In order to write an ansatz which extracts a
finite number of 4-dimensional fields including the graviton and gauge
bosons, one must restrict the gauge group from the isometry group G to
the subgroup G' whose Killing vectors satisfy (16), and also include
i'j'
the Kaluza-Klein scalars ~ , which are in the symmetric product of
the adjoint of G' with itself. The need to restrict from G to G' can be
understood from a group theoretical point of view: In general the
ansatz (4), (5) must be invariant under a transitively acting subgroup
K of the isometry group of ~k, and G' is a subgroup of G which is
centralized by K in G ' Of course, in order for G' to be
non-trivial, this means that Mk must certainly be an homogeneous
space.
The need to restrict from G to G', and include the scalars, arose
from insisting that the ansatz should satisfy the higher dimensional
field equations. In an alternative approach, the view is sometimes
taken that one should substitute the ansatz into the higher dimensional
action, and integrate out over y, to obtain an effective 4-dimensional
action. In our dicussion, this would be equivalent to averaging (12)
over Mk, in which case the y-dependent term in (13) would disappear,
and omitting the the scalars
oi'j' from the ansatz ( so the
possibility of varying them in the 4-dimensional effective action would
of course never arise). Thus one could obtain a 4-dimensional
Einstein-Yang-Mills action with gauge group G. However this procedure
is a prescription for satisfying certain components of the higher-
dimensional field equations and violating others, and such an approach,
66

which would have to be justified on physical grounds, runs counter to


the philosophy that it is premature to be imposing low- energy physical
prejudices at the Planck scale. One would also lose the correspondence
between extrema of the 4-dimensional action and extrema of the
higher-dimensional action.
However, perhaps the most compelling reason for adopting the
'consistent' approach is that in certain very exceptional theories,
remarkable consistent truncations can be made which cannot be
understood within the framework of the general discussion of this
Section. Thus we have a mathematical criterion which singles out
certain Kaluza-Klein theories as being very special. It is to one such
theory that we now turn.

3. ELEVEN DIMENSIONAL SUPERGRAVITY.


8
We will examine the bosonic sector of d=ll supergravity , which
^
comprises the metric tensor ^ZMN and a 3-form potential AMNP, satisfying
the field equations

I ~ 1 ^ ~BCD E I ^~
RAB . 2 . ~AB . ~ .( FACDE
. 8 F ~ ~AB )
' (19)

^VA 9ABCD _ i
576
BCDEI...E8 ?E ?E5
E I...E4 ...E8 , (20)

A
where FABCD = 4 ~[A ABCD]' and we are using local Lorentz indices.
9
These admit ground state Freund-Rubin solutions on M 4 x M7, in which
A
one sets all components of FABCD t o zero except in spacetime, where
~e~y6 = 3m~e~y6; M7 is an Einstein space satisfying Rab = 6m 2 6ab , and
in spacetime M4, R ~ = - 1 2 m 2 ~ .
One can easily show that already at the linearized level it is
necessary to augment the elfbein ansatz (4) and (5) by an ansatz on
~ABCD including the gauge bosons, in order to extract the massless spin
A
1 degrees of freedom in M 4. The correct ansatz for FABCD turns out to
be '

^ = !_ F i~6 Ki (21)
~y6 = 3m g ~ y 6 ' F~cd 2m E ~ Y 6 Vc d"

Substituting into (19) and choosing AB = ~8, this yields

1 1
- -- R -12m 2 = -- T i3 K ia Kj + -1 - 2 YaK i b yaK jb ) (22)
R8 2 ~ ~ 2 e~ a
2m

where TiJ~ is defined by (14). This equation should be compared with


87

(12). In this case, the equation is consistent for a subgroup G' of G


whose Killing vectors satisfy

Ki'a KJ'a + 1 2 Va Ki'b yaK j'b = 6i'j' (23)


2m

'a i Ki, a
It is easy to see that if Ki satisfies (16) then -- satisfies
2
(23), so the question arises as to whether more Killing vectors can
satisfy (23) than (16). For a generic Einstein space M7 the answer
seems to be no, but in just one case the situation is different. On the
SO(8~-invariant round 7-sphere, all 28 Killing vectors of S0(8) satisfy
(23) .
Of course the above calculation has only been concerned with the ~
components of the Einstein equation (19). Full consistency of (19) and
(20) would certainly require the inclusion of Kaluza-Klein scalars in
the ansatze for ~MN and AMN P. However, the calculation has already
exhibited a property of the 7-sphere compactification of d=ll
supergravity that seems to be unique; no other known Kaluza-Klein
theory could possibly yield all the gauge bosons of a non-abelian
isometry group G in 4-dimensions, since all other known theories yield
an inconsistency in the ~ components of the Einstein equation.
In fact there are strong indications that the full d=ll supergravity
7
theory, compactified on S , can be consistently truncated to just the
massless N=8 supergravity multiplet, whose bosonic sector comprises
the graviton, 28 gauge bosons, and 35 each of scalars and pseudo-
scalars. The complexities of the spin 0 sector have so far defeated all
attempts to perform a complete check of the consistency, but partial
ii 12
results in this direction are encouraging ' . Note that the fact that
only a 35 of scalars is needed is another remarkable property of this
truncation. Generically, one would have expected to need all the
r e p r e s e n t a t i o n s occuring in the symmetric product of 28 with itself.
The full N=8 truncation has been intractable to date because of the
complications due to the spin 0 fields. A simpler problem, which still
e x h i b i t s some remarkable properties of the theory, is to first truncate
to N=3 supergravity (for which the supermultiplet contains no spin 0
fields) and then discard the fermions. One is then left with just the
Einstein-Yang-Mills system, with SU(2) gauge group. The remarkable
point here is that consistency should be achievable with no scalar
fields atall. The calculations, which is still quite involved, is
described in detail in ref. 13. Restricting to the appropriate SU(2)
subgroup of S O ( 8 ) , it t u r n s out t h a t the a n s a t z (21) for ~ A B C D is
correct to all orders, and one finds that (4), (5) and (21) yield an
68

i
exact solution of (19) and (20), where ( g~v(x), A ~(x)) is an
arbitrary solution of the 4-dimensional Einstein-Yang-Mills equations
with SU(2) gauge group. There is no other theory known to admit this
kind of non-trivial embedding of solutions of the Einstein-Yang-Mills
equations.
A remarkable feature of this SU(2) truncation is that consistency of
the purely bosonic subsector depends crucially upon the presence in the
M.. S ^ A A
eleven dimensional theory of the ~ " FM...N Fp...Q AR... S term in
the Lagrangian, with precisely the coefficient demanded by super-
symmetry. Thus consistency and supersymmetry seem to be intimately
related, although the precise way in which this works remains unclear.

4. CONCLUSION
We have seen in section 2 that in a generic Kaluza-Klein theory it
is often not possible to make an ansatz which extracts just the
massless four-dimensional fields and which satisfies the higher-
dimensional equations of motion. In such cases the only way to restore
consistency is to reinstate some of the previously truncated fields. In
the example of section 2, this would include infinitely many massive
spin 2.
Suppose, however, that one were prepared to take the point of view
that one should simply take the massless ansatz (even though it is
inconsistent) and substitute it into the higher-dimensional action,
thereby obtaining an effective four-dimensional action describing just
the massless fields. What would go wrong?
As discussed in ref.3, the problem is one of non-uniqueness.
Specifying that one should make a massless ansatz is merely a statement
that at the linearized level the fluctuations around the ground state
are to be expanded as spacetime fields times zero-mode harmonics of the
relevant mass-operators on M k. Provided that one respects the
symmetries of the system, one is free to make any non-linear
modification of this prescription that one wishes, since it will not
affect the criterion of masslessness. For example, one could add a term
2 y(1)
(x) (y) to the right-hand-side of the massless ansatz ~(x,y)=(x)
y(0)(y)," where yt0)" " is a zero-mode and ykI)~" is a non-zero mode of the
relevant mass operator. Although such a modification will leave the
quadratic terms in the effective four-dimensional Lagrangian unchanged,
it will of course drastically alter the interaction terms. Which is the
correct choice?
In the case of a Kaluza-Klein theory admitting a consistent
truncation, the answer is unambiguous: the correct choice is the one
69

that ensures that the ansatz satisfies the equations of motion. (Of
course one still has the freedom to make field redefinitions amongst
the massless fields.) However if there is no consistent truncation,
then there simply is no unique choice of massless ansatz, and so most
of the interaction terms in the effective four-dimensional Lagrangian
are arbitrary, i.e. their coefficients depend upon which particular
massless ansatz one chooses. Thus only for a consistent truncation does
it make sense to study the non-linear structure of a Kaluza-Klein
theory.

ACKNOWLEDGEMENT
I would like to thank M.J. Duff, G.W. Gibbons, B.E.W. Nilsson and
K.S. Stelle for many helpful discussions.

REFERENCES

I) J.A. Wheeler, in: Relativity, groups and topology, eds B.S. deWitt
and C.M. deWitt (Gordon and Breach, New York, 1964).

2) S.W. Hawking, Nucl. Phys. B144 (1978) 349.

3) M.J. Duff, B.E.W. Nilsson and C.N. Pope, Phys. Report, in print.

4) M.J. Duff, B.E.W. Nilsson, C.N. Pope and N.P. Warner, Phys. Lett.
149B (1984) 90.

5) D. Boulware and S. Deser, Ann. Phys. 81 (1975) 193.

6) N. Manton, UCSB preprint NSF-ITP-83-04.

7) M.J. Duff and C.N. Pope, Nucl. Phys. B255 (1985) 355.

8) E. Cremmer, B. Julia and J. Scherk, Phys. Lett. 76B (1978) 409.

9) P.G.O. Freund and M.A. Rubin, Phys. Lett. 97B (1980) 233.

i0) M.J. Duff and C.N. Pope, in: Supersymmetry and supergravity 82, eds
S. Ferrara, J.G. Taylor and P. van Nieuwenhuizen (World Scientific,
Singapore, 1983).

ii) B.E.W. Nilsson, Phys. Lett. 155B (1985) 54.

12) B. de Wit, H. Nieolai and N.P. Warner, Nucl. Phys. B255 (1985) 29.

13) C.N. Pope, Class. Quantum Gray. 2 (1985) L77.


CANONICAL QUANTIZATION AND COSMIC CENSORSHIP

P. Hajicek
I n s t i t u t e for Theoretical Physics
University of Bern
Sidlerstrasse 5, CH-3012 Bern, Switzerland

I . Introduction

One of the most d i f f i c u l t problems of quantum g r a v i t y originates from the w e l l -


known feature of the theory that the causal structure of spacetime - the system of
l i g h t cones - is a function of the dynamical f i e l d i t s e l f . Moreover, r e a l i s t i c models
of matter f i e l d s and g r a v i t y w i l l be unstable with respect to the gravitational col-
lapse and formation of black holes. This means, in this context, that the deviations
of causal structures of possible dynamical developments from each other, or from some
standard structure l i k e e.g. that of the Minkowski spacetime, can be d r a s t i c a l l y large.
(We have assumed here that black-hole-like objects e x i s t in nature, but this is very
plausible due to cumulating observations - see, e.g. [I] .) By the way, such an as-
sumption could lead to some l i m i t s on r e a l i s t i c models for g r a v i t y : i t seems that
the classical l i m i t of such models had to y i e l d more than j u s t the perturbation series
for general r e l a t i v i t y around the f l a t background. Then, a l l such models, even with
high d e r i v a t i v e s , with a compound graviton, supergravity, Kaluza-Klein, string theory,
etc. w i l l suffer from the above problem.

Most of the today investigations are based on the expansion in the number of
loops. This means that the true l i g h t cones are approximated by the l i g h t cones of a
given, fixed classical solution (corresponding, say, to the ground state). However,
such an approximation is dangerous even in the purely classical theory - t h e series
leading to divergent integrals in higher order contributions 12] .

Another way to avoid the problem seems to be offered by the Euclidean regime: the
dynamical equations become e l l i p t i c and there is no e x p l i c i t causal structure. How-
ever, the d i f f i c u l t y seems to reappear at a d i f f e r e n t l e v e l : the corresponding quantum
theory becomes acausal [4 , non-unitary [4] , not asymptotically complete [5] ,
and leads to the loss of quantum coherence [4] .

Within the canonical quantization, the above problem takes on the following form.
For any canonical formalism to work, we have to f o l i a t e the spacetime by Cauchy hyper-
71

surfaces. However, the existence of such a f o l i a t i o n , the so-called global hyperbol-


i c i t y is a very special property of the causal structure. As the causal structure i t -
self is a function of the dynamical f i e l d , we do not know whether or not we are in
conflict with the quantum dynamics, i f we require the global hyperbolicity from the
outset.

Let us explain by an example what is meant by a "conflict with the dynamics".


First, i t is possible to weaken the global hyperbolicity somewhat. Let us consider
only asymptotically f l a t spacetimes. Then, for the scattering problems to be well-
defined, i t is sufficient that the part [ I - ( ~ +) n I + ( ~ ' - ) ] of each element M' from
a "large" class of such spacetimes is globally hyperbolic. This means, roughly speak-
ing, that in each spacetime M' satisfying the requirement, there are no singularities
which are visible from 3 + and at the same time influenceable from ~ - . Large class
means that the exceptions form a set of measure zero. I t is convenient to put the
measure on the space of regular Cauchy data for asymptotically f l a t spacetimes. Then,
the requirement is clearly closely related to the so-called Weak Cosmic Censorship [6].
In the classical form of this hypothesis, one assumes that the whole system of the
classical f i e l d equations is satisfied by all spacetimes M'. Thus, a violation of the
Weak Cosmic Censorship could be considered as a sort of conflict between the classical
dynamics and the weakened global hyperbolicity. Now, i t is well known that this form
of Cosmic Censorship is very l i k e l y to be violated [7] .

However, the dynamics we consider is the quantum dynamics. The corresponding


Quantum Cosmic Censorship Hypothesis (Q.C.C.H.) is not equivalent to the classical one.
In fact, some people believe that the singularities of the classical general r e l a t i v i t y
w i l l be avoided in the corresponding quantum theory. I f this is true, then the Q.C.C.H.
w i l l be more l i k e l y to be satisfied than the classical one. We shall touch this problem
in more detail later on.

Another, more subjective d i f f i c u l t y with the existence of Cauchy hypersurfaces is


that the globally hyperbolic spacetimes are "topologically dull" (as Hawking puts i t ) :
no change of the space topology is possible (this i s , roughly, the content of a clas-
sical Geroch theorem [ 6 ] ) . However, as f a r as I know, there is no proof that the
corresponding quantum dynamics w i l l also p r o h i b i t any change of space topology and
s p l i t , in t h i s way, into the corresponding superselection sectors.

Sometimes, one compares the general r e l a t i v i t y with the string theory. Any clas-
sical ( i . e . non-quantized) string i s , on one hand, a two-dimensional spacetime with
well-defined dynamics. On the other hand, changes in topology of the time-constant
folii of the classical string manifold are possible (and even necessary in order that
there is any interaction between s t r i n g s ) . Why, so one asks, are these two facts com-
patible in the s t r i n g theory, and not compatible in the general r e l a t i v i t y ? The
72

answer is simple: the strings are able to join t h e i r ends to form a regular internal
string point or to be torn into pieces with regular end points. No such discontinu-
i t i e s are allowed for classical spacetimes.

2. The method of canonical reduction

Suppose we have some self-consistent f i e l d - t h e o r e t i c a l model containing gravity


in the form of spacetime metric. Then, we can t r y to quantize i t by the so-called
canonical reduction method (see, e.g. [ 8 ] ). This method consists roughly in the
following steps. F i r s t , one adds some gauge conditions to the dynamical equations.
The gauge conditions have the form of equations (mostly d i f f e r e n t i a l equations) con-
taining the dynamical variables of the model. Then, the variables are divided into
the following classes: the true dynamical variables, the dependent variables, the
gauge variables and the Lagrange m u l t i p l i e r s . The dynamical equations are divided into
gauge conditions, gauge propagating equations, constraints and the true dynamical
equations, The gauge conditions, gauge propagating equations and constraints are solved
for the dependent variables, gauge variables and Lagrange m u l t i p l i e r s , and the true
dynamical equations are expressed through the true dynamical variables only. In this
way, the constraints and gauge freedom disappear and we obtain only mutually in-
dependent dynamical variables. Such a reduction can be performed within a Lagrangian
(second order) or Hamiltonian ( f i r s t order) formulation and the reduced theory can be
quantized in the standard way.

In most cases, i t is impossible to perform this program e x p l i c i t l y (one has to


solve a system of d i f f e r e n t i a l equations with arbitrary c o e f f i c i e n t functions). How-
ever, for our purposes, the abstract existence of solutions to these equations is
sufficient. Indeed, one can transform the resulting quantum theory to the form, which
is independent of a p a r t i c u l a r gauge condition, and which enables calculation of rele-
vant physical quantities ( l i k e the S-matrix) without an e x p l i c i t reduction [9, I0] .

For some theories, there is no gauge condition which works for the whole space-
time manifold and for the t o t a l i t y of possible f i e l d s (Gribov ambiguity, see, e.g.
[II I ). However, what we r e a l l y want to do is the deparametrization of the system,
that i s , only a partial reduction so that the gauge condition fixes j u s t the space-
like foliation. This should always be possible, or else no reasonable dynamics would
exist.

Within the reduction method, the problems with causal structure become even more
numerous and involved. F i r s t , the family of hypersurfaces defined by the gauge condi-
tion can become degenerate (containing, e.g. intersection of the hypersurfaces)
73

or non-spacelike for some values of the dynamical f i e l d . Even i f the f o l i a t i o n is


l o c a l l y regular and spacelike, i t need not represent, globally, a f o l i a t i o n by Cauchy
hypersurfaces, i r r e s p e c t i v e l y whether or not the spacetime to be f o l i a t e d is globally
hyperbolic. And, f i n a l l y , the spacetime need not be globally hyperbolic. However, we
have also more conditions on kinematically possible metrics: the gauge conditions,
gauge propagating equations and constraints must hold before any dynamics is set up.
We call these "predynamical equations". Let us, now, formulate all such predynamical
assumptions more carefully.

a) Let the spacetime mandifold (M~g) be I) smooth (C2), and 2) asymptotically


Minkowskian, 3) causal, orientable and time-orientable.

In many proofs, the requirement al) can be weakened. However, in this i n i t i a l state
of investigation of these problems, i t is very comfortable. The assumption a2) means
that (M',g) has a complete ~ [12] ; scattering problems can be formulated.

b) Let ~ = 0 be a gauge condition in the form of a ( d i f f e r e n t i a l ) equation for the


metric which (possibly supplied with a boundary condition) defines a regular space-
l i k e f o l i a t i o n of some neighbourhood N of i in M' by asymptotically f l a t hyper-
surfaces. Let the corresponding time parameter t coincide asymptQtically with a
proper time and has the bounds t ~ (- ~, ~).

A f o l i a t i o n t ( x ) = const is regular and spacelike in N, i f the vector f i e l d t , i (normal


to the hypersurfaces) is well-defined, continuous and timelike everywhere in N.

c) Let there be some C> 0 such that the hypersurfaces t = t o for all t o < - C can be
extended, as solutions of ~ = 0 in M', to form a regular spacelike f o l i a t i o n of
some part Nl of M' by hypersurfaces, each of which is complete with respect to
the positive definite metric induced on i t by the metric of (M',g).

Thus, in the "remote past", the f o l i i represent a regular, i n f i n i t e , asymptotically


f l a t space. I t can have a n o n - t r i v i a l topology and contain incoming extremal black
holes (Hawking temperatureT = 0). This is the "assumption of regular i n i t i a l data".

Let us define M to be the maximal connected neighbourhood of i in M' to which


the hypersurfaces t = const can be extended, as solutions of ~ = O, to form a regular
spacelike f o l i a t i o n . Thus, M contains N and NI .

d) We assume that each hypersurface t = const in M is either complete or has a boundary


in M' which coincides with an apparent horizon (AH).

The content of assumption d) is twofold: l ) I t states that the kinematically possible


fields are regular. Hence, the maximal extension of the f o l i a t i o n ~ = 0 leads either
to complete hypersurfaces, or ends at points, which are regular points of M'. Such
points can only be singular with respect to the f o l i a t i o n : points of intersection of
74

different t : const hypersurfaces, points, where the hypersurfaces cease to be space-


l i k e , etc. 2) I t is a requirement on the gauge condition ~ = O: the singular points
of the corresponding f o l i a t i o n must coincide with the AH of the spacetime (M',g). We
shall see in the next section that this can, at least for some models, easily be done.
This condition developed during the investigation from the attempt to f o l i a t e only
the region which was outside of black holes. The AH is chosen because i t is locally
well-defined.
The assumption of regularity of kinematically possible f i e l d s , as contained in
d) i s , on one hand, a very weak analog of assumptions which are usually done, i f one
constructs the quantum dynamics of some f i e l d . For example, for the linear fields
~131 , one considers f i r s t some space of Cauchy data which can serve as test fields
- they are C~ and, say, of compact support. Then, one finds some norm, or scalar
product with respect to which the spaces can be completed to Banach, or Hilbert
spaces. The n:orm, or at least the corresponding topology should be preserved by the
dynamics in order that the dynamics of the more singular elements be well-defined.
Thus, for a construction of quantum dynamics, i t seems necessary that the kinematical-
l y possible fields are dense in some suitable functional space. On the other hand,
the assumptions c) and d) together remind us strongly on the weak cosmic censorship
hypothesis. Of course, there are differences: for example, we require that the singu-
l a r i t i e s are hidden beyond AH instead of beyond event horizons. The r e a l l y important
difference i s , however, that we do not require the v a l i d i t y of the f u l l system of the
classical dynamical equations. We require, f i r s t , just the predynamical equations
and, second, the p o s s i b i l i t y to construct a reasonable quantum dynamics.

Let us call the points c) and d) together with the assumption that one can con-
struct a reasonable quantum dynamics with them, a Quantum Weak Cosmic Censorship Hypo-
thesis. I t is clear that the quantum censorship could be true even i f the classical
one is invalid.

e) We assume, f i n a l l y , that the whole system of the predynamical equations holds in M.

We shall call our f i e l d theoretical model for gravity to be completely foliable, i f


the assumptions a) - e) imply the following properties of M:

(i) M is asymptotically Minkowskian,


( i i ) each hypersurface ~ = 0 in M is a Cauchy hypersurface for M.

The gauge condition ~ can, then, be called a "complete f o l i a t i o n " .

Thus, at the end, the singularities beyond the apparent horizon are not v i s i b l e from
75

3. BCMNmodel

The most simple known f i e l d theoretical model in which the dynamics can lead to
a t r u l y nontrivial causal structure (namely the ~ormation of black hole horizons)
is the Berger-Chitre-Moncrief-Nutku model [14~. We show in this section that the
model is completely foliable.

Let us f i r s t b r i e f l y introduce the model. I t results from the Einstein-Maxwell


system to which an uncharged scalar f i e l d is minimally coupled. All dynamical degrees
of freedom are frozen except for the spherically symmetrical ones. We have the follow-
ing variables: a metric gab on a two-dimensional manifold M' (t-r-surface of the
original four-dimensional spacetime), a real scalar f i e l d @on M' (r-coordinate), and
the real scalar f i e l d ~ on M' (the original scalar). The action has the form:

(see [151 ). Here, G is the Newton constant, Q and P are the electric and the magnetic
charges of the possible (incoming) black hole, g is the determinant of gab and R is the
curvature scalar corresponding to gab"

Let us choose the gauge condition

where na is a normal vector to the t = const surfaces, and supplement i t by the bound-
ary condition at i n f i n i t y

(3) }oo :

The condition (3) specifies the time parameter t up to an additive constant. In


an asymptotically Minko~skian spacetime M',the f o l i a t i o n (2), (3) w i l l be spacelike
near i and the assumption b) w i l l be satisfied. The t = const hypersurfaces in the
maximal extension M with a regular spacelike f o l i a t i o n can only have AH as boundaries
in M'. This is clear from the following considerations.

The f o l i a t i o n is regular at a l l points, where the direction of na is uniquely


determined by the condition (2), that i s , where @a@is a non-zero vector. Any c r i t i c a l
point p of @satisfies
78

where ~a and ka are the two independent null directions at p. Thus, p is a future and
past AH simultaneoulsy, or "double" AH (DAH).

The f o l i a t i o n ceases to be spacelike, i f ~a~ becomes null, that is either at a


future AH (FAH) or at a past AH (PAH).

In [161 , the following theorem has been shown.

Theorem l : I f the conditions a) - e) are satisfied, then M is asymptotically Minkowskian

Let us consider the predynamical equations. I f we reduce the theory in i t s Lagran-


gian form, then the equations read as follows: the Hamiltonian constraint:

f~

the momentum constraint:

and the gauge propagating equations:

~x
C6) w'X )_ (~,) = ~ I ( ~

(~- 0 ,

Here, we have chosen the x coordinate to be

The following relations hold:

- ~ , ~ = -9 ~~

- C
77

so ~ is the lapse, B the s h i f t function and ma i s the unit tangential vector to


t = const hypersurfaces.

I t has been shown in [17! that the three equations (4), (5) and (6) are equi-
valent to the following tensorial equation:

SI

Thus, i t is e a s i l y transformable to any coordinate system in M; i f written out in a


double null coordinates, eq. (7) implies the following

Theorem 2: Let H be the future (past) h a l f of an outgoing (incoming) null hypersurface


though a future (past) AH p. Let p l i e on the boundary of M and H inside of M. Then,
the divergence of the null geodetic generators of H is non-positive (non-negative).

Using the theorems 1 and 2, one can show that the FAH cannot be v i s i b l e from ~+
the the PAH cannot be influenced from ~ - , as well as that the world tube of an ap-
parent horizon at the boundary of M is not timelike.

These properties and the assumptions a) - e) imply the following

Theorem 3: All hypersurfaces t = const in M are complete with respect to the induced
n~tri c.

From theorem 3 and the r e g u l a r i t y of the f o l i a t i o n in M, i t e a s i l y follows that


the t = const are Cauchy hypersurfaces for M. Hence, the BCMNmodel is completely
f o l i a b l e and the conditions (2), (3) represent a complete f o l i a t i o n .

4. Hawking e f f e c t , p o s i t i v i t y problem and u n i t a r i t y

The picture which the complete f o l i a b i l i t y of the BCMNmodel gives concerning


the kinematically possible t r a j e c t o r i e s is quite d i f f e r e n t from the current ideas
about the expectation value of the n~tric in a spherically symmetric spacetime with
a collapse. According to these ideas, a black hole horizon w i l l appear and the most
of the information about the collapsing object w i l l be l o s t inside of i t . There w i l l
also be a radiation going from the collapse region out to the i n f i n i t y - the so-called
Hawking radiation. The o r i g i n of t h i s radiation w i l l be localized to a neighbourhood
of the black hole horizon. The energy necessary for the Hawking radiation w i l l be
taken d i r e c t l y from the black hole: a l o c a l l y defined current of negative energy w i l ]
be pouring through the horizon from outside. Due to t h i s negative current, an apparent
horizon can form outside the black hole horizon. Such an AH w i l l be v i s i b l e from ~ +,
and i t s world tube w i l l be t i m e l i k e .
78

The crucial question in this respect is which properties of the kinematically


possible trajectories can survive the quantization and can, in this way, appear as
properties of the expectation value of the metric. Nali;ve]~ i t could seem that all
such properties must survive, because the expectation value of the metric can be calf
culated as a path integral average over the kinematically possible trajectories. How-
ever, this is not true in general. For example, the energy density of, say, Klein-
Gordon field is everywhere non-negative for any kinematically possible trajectory but
its expectation value can be negative at some points.

In [16] , this question has been discussed at some length. One possibility, due
to Ashtekar and Horowitz [18] , is the following: i f a given general property of all
kinematically possible trajectories can be considered as a property of the configura-
tion space of the system, then i t will survive the quantization. Such properties are,
for example, the absence of AH in M (this is the absence of critical points of
along t = const surfaces), or the fact that M is asymptotically Minkowskian.

Another, more obvious possibility, is to look at the operators which represent


the components of the metric in the quantum theory to see which sort of spacetime
they are likely to yield. In our gauge, the component gll = I/y of the metric gives
information whether the t = const surfaces are spacelike (gll > O) or null (gll = 0).
The l a t t e r case would mean an AH. After the reduction, gll is a dependent variable,
given by [19] :

i
b
where

= 4- G

is a positive function, and

&
with ~(y) being the canonical momentum and ~(y) the canonical coordinate (true dynam-
ical variables), gll(x) is clearly positive classically, but can i t be made to an
operator with positive spectrum ? Some subtraction procedure is necessary to define
g l l ( x ) ; for example, i f this procedure can be applied directly to T(x) - T(y) (so
that the exponential of i t is already well-defined), then gll(x) will be positive even
i f T(x) - T(y) i t s e l f is not.

I t is interesting to notice that the problem with positivity of gll(x) is closely


related to the well-known "positivity problem" in the canonical quantization [ 8 ] :
79

A canonical coordinate Q which describes gravity (in two-dimensional theories, t h i s


can be g l l ( x ) indeed) has a l i m i t e d range of i t s values. I f the corresponding operator
Q s a t i s f i e s the canonical commutation rules (CCR) together with i t s conjugated momentum
~Q, then the spectrum of Q cannot be l i m i t e d . Two solutions to the problem have been
proposed in the l i t e r a t u r e :

I ) Choose other variable ( l i k e log Q) and i t s conjugate [20] , or 2) use the pair
{Q~(Q~Q)} as the pair of variables whose CCR determine the operator algebra [21] . A
t h i r d solution, possfble only within the reduction method, is to choose such a quantity
as a dependent one. (This is analogous to what one does in ordinary quantum mechanics
with the canonically conjugated pair {time, energy}).

Another important point to discuss is the u n i t a r i t y of the resulting quantum


theory. We have shown that the canonical qunatization is applicable to the BCMNmodel,
because the relevant part of each kinematically possible spacetime can be f o l i a t e d by
Cauchy hypersurfaces. The dynamical development of the quantum states from one such
surface to another w i l l be unitary. However, t h i s relevant part M is not the whole
spacetime in general. Thus, the f i n a l , t = ~, Cauchy hypersurface can contain a part
of the boundary of M in M', e.g. an event horizon, say. I t i s , however, impossible to
perform measurements along the event horizon, and one w i l l be interested to know the
state only along that part of the t = ~ hypersurface which does not contain the hor-
izons. Such a state w i l l , in general, be mixed, and we seem to lose the quantum co-
herence, and u n i t a r i t y , again.

Here, the Hawking e f f e c t could, in f a c t , help. I f a l l energy of the collapsing


object is radiated away again, then we can end up with a "clean" horizon, that i s , no
horizon at a l l , or that one which has been present before the collapse (incoming hole).
For t h i s to work, i t is necessary that the Hawking radiation carries away a l l inform-
ation about the collapsing object. Such a transfer of information seems to be impos-
sible according to the current ideas which l o c a l i z e the o r i g i n of Hawking radiation
to a neighbourhood of the event horizon.

There was another school of thinking about the Hawking e f f e c t [22] , l e t us call
i t Boulware school, which localized the o r i g i n of Hawking radiation to the inside of
the collapsing object. The energy of the radiation was taken d i r e c t l y from the object
so that the horizon could never form. The calculations of the Boulware school, however,
did not reveal any better information transfer than Hawking's. This was due to the as-
sumption that a fixed classical background gravitational f i e l d was well-defined every-
where and that i t was the only source of the radiation; such a f i e l d had certain proper-
t i e s which did not depend on the d e t a i l s of the collapse. The struggle about where the
o r i g i n of Hawking radiation is to be localized has been won by the Hawking school a f t e r
the consens has been achieved about the regularization and renormalization of the
80

stress-energy tensor: the expectation value, <T~>, of this tensor gives a l o c a l l y


well-defined c-number energy current (see, e.g. [23] and the references given
there).

Let me very b r i e f l y c r i t i z e these theories. F i r s t , such fine l o c a l i z a t i o n of the


origin of the Hawking radiation that i t can distinguish between the inside of the
collapsing object and a neighbourhood of the corresponding event horizon need not make
any sense at a l l , in spite of the results about <T >. Indeed, the a r b i t r a r i l y detailed
l o c a l l y defined c-number energy current given by <T > could have some physical r e a l i t y
U~
only i f the corresponding mean squared deviation, <~T2 >, would be negligible with
respect to <T >. As far as I know, this was never shown.

Second, i t is a n o n - t r i v i a l problem, whether the whole perturbation scheme


used to calculate the Hawking e f f e c t is applicable. That is the semiclassical aprox-
imation: f i r s t , one calculates the classical solution for the collapsing object and
the surrounding gravitational f i e l d ; second, one investigates the behaviour of small
quantum disturbances around t h i s fixed classical background; f i n a l l y , one couples the
classical metric to the expectation value of the stress-energy tensor of these disturb-
ances. In our case, the classical background is unstable and the e f f e c t i t produces,
is large.

In t h i s way, my speculations about the information transfer in the Hawking effect


need not be completely wrong and u n i t a r i t y could be saved.

F i n a l l y , I would l i k e to stress that my c r i t i c s concern only the so-called dyn-


amical Hawking e f f e c t and not the existence and properties of the thermal quantum
states on the s t a t i c black hole background.

Acknowledgement: I am very indepted to R. Penrose and J. Hartle for important


c r i t i c a l remarks.
81

References

1 D.R. Whitehouse, A.M. Cruise: Nature 315 (1985) 554.


2 D. Christodoulou, B.G. Schmidt: Convergent and Aysmptotic Iteration Method in
General Relativity. Preprint MPI-PAE/Astro 177, 1979.
3 S.W. Hawking: in "Qantum Gravity. 2nd Oxford Symposium". Ed. by C.J. Isham,
R. Penrose, D.W. Sciama. Oxford, Clarendon Press, 1981.
4 S.W. Hawking: Phys.Rev. Dl4 (1976) 2460.
5 S.W. Hawking: Commun.Math.Phys. 87 (1982) 395.
6 S.W. Hawking, G.F.R. Ellis: The Large Scale Structure of Spacetime. Cambridge,
Cambridge University Press, 1973.
7 D. Christodoulou: Commun.Math.Phys. 93 (1984) 171.
8 C.J. Isham: in"Quantum Gravity. An Oxford Symposium". Ed. by C.J. Isham,
R. Penrose, D.W. Sciama, Clarendon Press, Oxford, 1975.
9 J.B. Hartle, K. Kuchar: J.Math.Phys. 25 (1983) 57.
I0 E.S. Fradkin, G.A. Vilkovisky: Preprint TH-2332 CERN, 1977;
E.S. Fradkin, I.V. Tyutin: Phys.Rev.D2 (1970) 2841.
II T.P. Killingback: Commun.Math.Phys. I00 (1985) 267.
12 R. Geroch, G.T. Horowith: Phys.Rev.Lett. 4__00(1978) 203.
13 B . S . Kay: Commun.Math.Physo 62 (1978) 55.
14 B . K . Berger, D.M. Chitre, V.E. Moncrief, Y. Nutku: Phys.Rev. D5 (1972) 2467.
15 P. Thomi, B. Isaak, P. Hajicek: Phys.Rev. D30 (1984) 1168.
16 P. Hajicek: Phys.Rev. D31 (1985) 787.
17 P. Hajicek: Phys.Rev. D31 (1985) 2452.
18 A. Ashtekar, G.T. Horowitz: Phys.Rev. D26 (1982) 3342.
19 P. Hajicek: Phys.Rev. D30 (1984) 1178.
20 C.W. Misner: in "Magic without Magic. John Archibald Wheeler. A Collection of
Essays in Honor of His Sixtieth Birthday". Ed. by J.R. Klauder. Freeman, San
Francsico, 1972.
21 J.R. Klauder: Phys.Rev. D2 (1980) 272;
J.R. Klauder: in " R e l a t i ~ t y " . Ed by M.S. Carmeli, S.I. Flicker, L. Witten.
New York, Plenum, 1970;
C.J. Isham, A.C. Kakas: Classical Quantum Gravity 1 (1984) 621.
22 D.G. Boulware: Phys.Rev. DI3 (1976) 2169.
23 N.D. B i r r e l l , P.C.W. Davies: "Quantum Fields in Curved Space". Cambridge,
Cambridge University Pres, 1982.
QUANTUM EFFECTS IN NON INERTIAL FRAMES

AND QUANTUM COVARIANCE

Denis BERNARD
Groupe d'Astrophysique Relativiste
C.N.R.S. - Observatoire de Paris-Meudon
92195 Meudon Principal Cedex - France.

Abstract :

We review recent results in non-inertial quantum field theory. By formulating Q.F.T.


in a large class of accelerated frames, the classical and the quantum aspects of the
theory are unified. We describe the thermal effects, their asymptotic character and
the role of the P.C.T. symmetry. A discussion of quantum covariance and detection
processes is also given.

0. INTRODUCTION

Quantum field theory in accelerated frames is a possible approach to the understan-


ding of gravitational effects. Most of the known results about quantum field theory
in curved space-time, such as the Hawking radiation, can be described by analogies
with non-inertial effects in flat space-time. Therefore, non-inertial quantum field
theory makes possible the setting up of a "laboratory" for studying field quantiza-
tion in curved space-time. However~ it is also a possible way to discuss non-iner-
tial detection process or a possible way to search for a quantum covariance princi-
ple. This paper reviews recent results about non-inertial quantum field theory and
presents some new ones, too. In particular, we analyse thermal effects and their
asymptotic character and we relate them to proposed quantum covariance laws. A criti-
cal discussion of detection processes and their link with quantum field theory in
non-inertial frames is presented.
The content of this paper is :
I. Q.F.T. in Rindler frame : the role of the P.C.T. symmetry
II. Q.F.T. in analytic accelerated frames
III. The asymptotic character of thermal effects from a local principle
IV. A hamiltonian formulation
V. Vacuum fluctuation in accelerated frames
88

VI. Discussion.

I. QUANTUM FIELD THEORY IN RINDLER FRAME : THE ROLE OF THE P.C.T. SYMMETRY.

We shall begin with the Davies-Unruh's III result about the quantification of a sca-
lar field ~ in a uniformly accelerated frame (Rindler frame). There are many ways
based on Bogoliubov transformations, Green's functions, and others 121, to obtain
this famous result . But here, we want to present a global description where the
role of the P.C.T. symmetry is illustrated 13l. In particular this symmetry becomes
crucial for the analysis of the state identification proposed by t'Hooft. First in
flat space-time, accelerated trajectories are completely describe by the transport
law of the tetrad carrying by the observers. This equation is the Fermi-Walker
transport equation 141 which reads :
--+ ._._7 A n
Jet,r) + e. w re(r) - 0
(1.1) ~.f_

where + ~'n(r) is the tetrad and O m~r) the generator of this transport. @~m
is a generator of Lorentz transformation and can be written as

(1.2) @ = 4__ e L~# w~h E ~# + = o

where L ~ are the generators of Lorentz transformations. It is useful to intro-


duce the acceleration--~ and the rotation ~ as the "electric" and "magnetic"
part of the antisymmetric tensor

(1.3) -~ and _.4 * E ~ # E ~# = -d._Cl


"-"
4
We are looking for accelerating trajectories for which ~'E -~- ~ > 0 so that
the operator ~ is a "good" generator of temporel evolution"

for any time-like vector . y ~ < ~ are the translation generators). ~t is interes-
ting to introduce the non-inertial coordinates ( -~p ~ ) defined with respect to
the inertial ones ~ by

where J =
dr' ~'
In terms of these coordinates, the metric take the form
84

Then t h e h a m i l t o n i a n

.A

(1.~) Ht - "~ ~,__~_ ~ -~-


2

becomes simply the ~-time evolution generator

(1.~) ILl _- _ i

It is generally supposed that these generators represent the observer's hamiltonian.


For the sake of simplicity (and because we can always take it as an approximation at
least during a small duration ~ ) , we choose ~ ~-independant. For ~ and
parallel, the hamiltonian becomes

(1.9)

L.
where ~ and ~ are respectively the boost-generator and the angular-momentum
in the direction i. Moreover, the previous non-inertial coordinates become the rota-
ting Rindler coordinates

(I.i0)

p=fJ
in cylindrical coordinate
The accelerated coordinates (-~ ~ ~I ~) cover only a s u b m a n i f o l d ~ 7 of the

~
Minkowski space. (see figure) ~+

)X

E~ d
The region ~ = ~ , >0 , is the field of communi-
cation of "Rindler accelerated observers". R-and ~*J are the past and futur event-
horizons of these regions.
The quantum particle states for this "observers" are chosen to be eigen-functions of
85

the hamiltonian

and we shall require that these wave functions vanish on ~ E or on ~ .


Because m is the generator of the Lorentz transformation (1.5), the wave functions
1"" satisfy the following transformation law

where /~(~)is the Lorentz transformation (1.5).


!

That property characterizes t h e f u n c t i o n --~Ep--' b u t i t is simplest to i n t r o d u c e a

plane-wave
decomposition(l.13)
16, /(I~ ~{~" ~) 6E,,
~.l~,~) = ~ ,~(~) ; Ek= f d ~
- C,2~)~ o ~
Use of (1.12), yields a differential equation for

(i.14) [0 E~ ~ + i~- ~ ~ ] G, (~) = -; c 6~o(~)


(where ~ is the cylindrical angle of t ),
whose solutions are

(i.15) ( ~ k + ~ -+ )m ' o -"~


M 0 ei~

where m is the angular momentum ; ~ = q +m


and
Therefore,
+-
~1 c a n be c h o s e n a s
we c a n b u i l t , ~fter
~ = va~+ ~{ ,
normalization, a wave f u n c t i o n basis, 6,1, m
and 6~qcm ,. which can be used to construct the Fock-space of the quantum field:

(1.16)

[in the discrete notation] +


The operators of creation-annihilation,,__ OG~,rm and ~l~q Ir~ , define the va-
cuum state I0> : Q~,mlO~'- O
i

Because the ~ , q , m have positive minkowskian-energy, this vacuum is the Minkowski


one. Now, the region R~ is outside the field of communication of the accelerator
"observers" inside R I. Therefore we would like to diagonalize the hamiltonian
separatly inside the region R I and R~. Thanks to the P.C.T. symmetry, we can link
86

the value of the wave function inside R I to that inside R]I. From (1.13) and (1.15)
we get :

(1.17) ~c(~,m
: e -[o,-
(In this region, the logarithm in the equation (1.15) has been defined on the half-
upper complex plane). Since,

(1.18)

we have

,I
and a similar relation for ~ ~O .
Therefore, the states
~aemD],"
I ~'

'~,~,~
. -7~ &ml..~-----
-x 1

J-6,'l,-,~ J
vanish in the region R ~ and are eigenfunctions of H.
Similarly, we define

(1.21) ~,9,~ zl 6%-


Z
which is the P.C.T. symmetric image of T ~j~ . The [ ~ vanish in the re-
gion R I and are eigenfunctions of H, too. The normalized wave functions i~ and ~
and their complex conjugates make up a wave function basis which defines the Rindler
mode. The quantum field ~ reads

(1.23) "-- ~
(,q,m
I3~C6j~119 16,q,m 4- ~[ ,q,m 6,~,m ]

and from (1 32), "(J)~ = (~)-I(~1--/@ where @ is the antiunitary P,C.T.operator.


The creation-annihilation operators ~C6~c]jm and -~hC = ~-~z C ~ define the
Rindler vacuum: IO~ ~ IC I O ~ = ~CIO~> = O
Because, the definition (1.20) mixes positive and negative frequencies, the Rindler
vacuum is not equivalent to the minkowski-one. The different creation-annihilation
operators are related by the Bogoliubov transformation
87

and similarly for ~C .


Therefore, the Minkowski vacuum I05 contains Rindler modes. The density of Rindler
modes :

d e s c r i b e s a P l a n e k i a n spectrum. The a c c e l e r a t i o n plays the r o l e o f the t e m p e r a t u r e


T=o/~'~ and the rotation velocity appears as a chemical potential.
The unitary transformation linking the Rindler mode to the Minkowski-one can be
written as :
Io5 -- 1110 >

(1.26)

The pure Minkowski vacuum state contains pairs of Rindler modes. (like the B.C.S.
state). Each pair contains one "particle" created in the region R I and another crea-
ted outside the horizons ~ o But, if we restrict ourselves to observable, ~ say
whose support is restricted to the region RI, it is better to introduce a density
matrix ~ by

(1.27) : IO><O1

that is, by taking the trace over the states built from ~+ Then, the expecta-
tion of the observable ~ , in the Minkowski vacuum takes the form :

A
And the density matrix ~ , describing a thermal mixed state, is
I

(1.29) = I

n
where I~)~,~,~> = (~!)~ (~q~m)I0~ are the n-Rindler mode states.
This thermal character persists in the presence of interactions. By using a path
integral approach, W. Unruh and N. Weiss i51 have shown that a thermal quantum field
theory in a Rindler frame coincides, for the Hawking-Unruh temperature, with the eu-
clidean Q.F.T. in an inertial frame.
88

Remark on electromagnetic analogies.

The description of the accelerated trajectories in terms of Lorentz generators like


(4_~) illustrates, once more, the analogie between classical electromagnetic and
gravitationals effectsJ I. The tensor E ~ becomes the analog of (~) times the
electromagnetic tensor. In particular all stationary trajectories (such that 6 w ~
is ~ - i n d e p e n d a n t ] can be found directly from the study of trajectories in constant
electromagnetic fields. (see ref.(6 bis) and ref. (25) for another derivation of
these trajectories).
These analogies persist at the quantum level. Indeed, the Schwinger Lagrangian in
presence of an electric field E (B = 0)

8-ir '~ ~ <,.~E - 6


can be written as

where =

It is quite natural to interpret 2ms as the excitation energy du to virtual particle


production and ~'(S) as the spectral function describing the density of virtual
excitations. Then, we may define the temperature describing the average excitation of
the ground state in presence of the field E by

T - .lm _ eft

~]Crn

This value is the analog of the Unruh-Hawking temperature

2. Q.F.T. IN ANALYTIC ACCELERATED FRAMES.

The thermal effects analyzed in the previous section are net restricted to the Rind-
ler rotating coordinate (i.i0), but persist for a large class of accelerated coordi-
nates. Following on from the work of Sanchez j6j we generalize 17] the formulation of
quantum field theories based on analytical mappings in two-dimensional space-times
to the massive case in four dimensions by including possible rotation or drifting.
We describe the sub-manifold by the curvilinear or accelerated coordinates,
, defined by the transformation law :

/:z._ t = -~ -

(A. l) = F ' (B.1) Iz = z'+


=
89

where ~ , ~ ) E (~)~/ ~, ~ ) are Minkowskian cylindrical coordinates) ~ a strictly


monotonic function defined on ~ ~-- an angular velocity and@ ~ and~ drifting
parameters. Such a transformation ensures that the accelerated coordinates cover
completely their domain of communication. The horizons associated with such a region
are defined by the singularities of F, the inverse mapping of ~ :

(A and B.2) U--- ~(~oO)

If U+_ =-+ ~O , there is no horizon, but if ~ are finite the accelerated coordina-
tes only cover a limited region of total space-time (Minkowski space-time)

In such coordinates, the metric takes the form

~) ~s~: _ A{~')C~'~J~ ~) . ~ ( J ~ ' ) ~ ~p'~


I t I~ i~

and the whole cinematic of the "observers" inside the static regions :~')-~#~ A(~,~')
or ~ i ~ A(~I ~' ) is determined by the mapping.
However at the quantum level, the two statements (i) and (3) are not sufficient and
/
it is necessary to make clear the behaviour of ~{ioo) so as to define a Q.F.T. in
accelerated coordinates. To illustrate what precedes we limit ourselves to a sub-
manifold with only one horizon : U_= O and U~=+co and to a mapping which has
the asymptotic behaviour :

l )
(A and B.4) ~(_oo)= O

Su=h a class of mappings includes the Rindler mappings : n --


-~(X~,)= e which
describe uniformly-accelerated observers. So as to be able to define a complet set
of wave funcions for global space-time from the wave functions defined in the sub-
manifold we must also prolong the transformation by (see figure)

s3~(~) L, : -~(u')

(A.5) and (B.5)

2 --S'
90

Figure

The Fock space associated with the quantization of a massive scalar field is
built up from a bas~s of wave functions which are solutions of

and which have a positive "charge" defined by scalar product

Relative to the global coordinates ~ff, ~) , the Fock space is built upon the cyl~n-
drical waves of positive energy :

and ~ is a Bessel function.


or upon the plane wave functions

The creation-annihilation operators O~ I O~ define the global vacuum I0> :

With accelerated coordinates, we must define the quantum states which can be associa-
ted with accelerated "observers" in the region ~ . These wave functions do not
make up a complete basis for global space and thus are not sufficient to build a
Fock space. In order to form a complete basis from these states we use the PCT sym-
metry. The wave functions ~ ~ relative to the region ~ T are defined by certain
Cauchy data on -~" whose support is included in --~I~ --~----n~-~ . Under the-
se conditions, ~ are always null on ~-~- (but not on F and P). Each ~
-- is
il.

associated with a state ~ defined as


91

The ~ are null throughout the region R I.


Consequently, for ~ # ~ ~ to constitute a complete basis for global space, it is
sufficient for ~A to be a complete basis for the class of wave functions which
possess null Cauchy data on ~ ~ -- ~ ri~R~- . This can be shown by decomposing
l~j ~% ~ on the basis of the "Rindler states" defined in the previous section of
this paper.
The Fock space is thus built upon the creation-annihilition operators C~ t C~
and O~ t~t..~, relative to ~ and ~ j we have

and [ ~1_, (l),] : 0

The operators C~, Cll~ define the accelerated vacuum I0'2


c~:lo'> : d~ t o / > = 0

The PCT construction ensures that the theory in accelerated coordinates is complete-
ly determined by its formulation in the region R I. Indeed~we have
@J
where ~ is the anti-unitary PCT operator. The Bogoliubov transformation between
the two representations of the Fock space is written as

(l and B. 7 )

It is desirable to note that the canonical quantization is achieved first of all in


the global space-time ~ . Otherwise the operator PCT could not be built up. The
Bogoliubov transformation is simply the unitary transformation linking two choices
of possible base states for the Fock space.
In coordinates ('~--#j I~#) the wave equation takes the form :

[-'~/+ "~,< - Ac~',~')~ ~ ] ~A,c~',~') : o

wit,, t~<~ <x'~,,c~,~),~ ~'),(s-~$ "< +


92

and ~@= ~+ ~@

and ~-'[~'=~"~'~-4-~: + ~:

The asymptotic condition (5) implies that the effective mass A[X~I~/) N 4 is null at
the horizons and that it is infinite at infinity so that no particle can escape. So
we can choose as the base functions, the functions i +:
- ) which satisfy =
V.

,~

UI--> +
where explicitly : u)

The functions ~ are orthonormalized with respect to the scalar product o n R I.


Moreover, OO_-~j_~'~-~ in the first case and 6 0 : ~ _ ~ _ ~ in the second one
are interpreted as the asymptotic frequencies on the past-horizon.
Another choice of the base states is possible, imposing asymptotic conditions on the
future horizon so that :

V~+ee

In the first case, the vacuum is denoted I0' ; in > whereas in the second the vacu-
um will be called I0' ; out >. In general, the non-stationary character makes the
two vacuums inequivalent (only for the Rindler mapping is 10';in> = 0';out>). From
here on, we write I0'> for 10';in> unless explicitly stated.
With respect to the region RI, we note that, by construction, the states defined by
d~ are not observable. The commutator, [ ~Ej ~ ~= O expresses the absence of
a causal relationship between R I and R~. So, relative to the region RI, the pure
state I0> which corresponds to the global vacuum is described by the density matrix
obtained by tracing-out the states ~ :
A

This matrix is completely determined by the population functions :


93

An e x p l i c i t calculation gives =

Eb4k, oZ~:lu e "~


+~' _D,,V'iu) - -~ ,~'+ I,-, u

a 'AX' -o CU - h~t+ ; E )e
.o. _ i~,,v (u)_.il,~v~L,9
a ~&&' e
with

So the Bogoliubov coefficients ~g$ and ~ ) ~ are not the same as in the non-mas-
sive case but ~(~j~l) and ~ ( ~ i ) are not dependent on the mass as the asympto-
tic condition imposes a total redshift on the past horizon (see dispersion relation).
Thus it is the asymptotic behaviour which determines the thermal properties. Indeed
the results already obtained by N. Sanchez can be extended.
p
i) The relation between the mapping and ~(~, ~') is reciprocal and we can
invert the relation

du/L a
where N l is defined by

A.9)

ii) The above relation makes it possible to show that the Rindler mapping,
--4~(UI)=eX~(tltl/),- is the only one which satisfies the global thermal balance

~y~) is the population function for a unity of volume and, in the Rindler c a s e ,
we obtain :
94

(A. I0) ~(~)= ~ /i@;~--(6~-~j2") -- ~']

and (B. I0) ~/'(~) -- ~ ~"~

where'~=-~/~ appears as the temperature


and .~L~ ~ ) ~ play the role of chemical potentials.
iii) The thermic properties are defined by the asymptotic behaviour of the map-
ping. For an asymptotic Rindler mapping, ~(u')=e~(~_U p) when LI/----'~ --4"
the population function behaves according to the law

Wil-~ =

andthere is a simple analogous expression for the case B.


Here, the asymptotic temperature "~+ can be written as

(A and B. II) --~+ - ~ X ~- ILn~(~l)]I f

Contrary to the previous case, there is no global thermal equilibrium but only an
asymptotic thermal equilibrium in the region where the coordinates and tend
towards infinity.
Moreover, in order to extend the analogy between the examination of the thermal
properties linked to these mappings (but in flat space-time) and those that can
exist in curved space-time, it is useful to introduce the surface gravity'. ~ can be
defined by the ratio of the proper acceleration, a', to the temporal compenent,'1) ~j
of the speed of the observers that follow the flux lines defined by the normals to
the hypersurfaees, t' = constant. Then the asymptotic temperatures are

H'I ='
I Vl= *
This relation can also be interpreted as a generalisation of the Unruh-Hawking tem-
perature T=o/~ for uniformly and linearly accelerated observers. The asymptotic
character of the thermal effect, and the link between flat space-time and curved spa-
ce-time effects are clearly shown. In particular, near the horizon of a Kerr black
hole the transformation between the Kruskal coordinates (JI~ Vk. ) and the "tortoise"
coordinates ( II / r~.-/S ) f ~)
95

is basically of type (i) :

with~L=~Li~ the angular velocity of the horizon of the black hole a n d ~ the sur-
face gravity of the Kerr-black-hole:

The Hawking temperature follows from this analogy. But the analogy cannot be pursued
further. In particular, the supperradiance effect cannot be reproduced as is shown
by the expression (~o~0)o~ ~{~l~. If one wished to show schematically such an effect
with another mapping, better reflecting the properties of the Kerr metric~ the sta-
tionary character would be lost ; the vacua 10';in> and 10'~out> are then no longer
equivalent. In that case, it is no longer possible to distinguish the effects of
non-stationarity from the effects of superradiance due to a difference between asym-
ptotic frequencies. The same problems would present themselves if one wished to re-
establish the isotropy : the stationary character is destroyed.
This previous study can he extended to mappings with non-constant rotation or drif-
ting unless they becomes constant at the horizons.

Remark i.
In a thermal equilibrium situation at a temperature T, we typically define the ther-
mal average of an observable ~ , by computing the expectation of ~ at the tempe-
rature T and by substracting its value at - ~ = O . i.e. :

In this spirit, the natural definition of the average in an accelerated frame seems
to be

In particular, if ~ is the stress tensor in a two dimensional massless case, this


definition gives a renormalized stress-tensor which takes into account the energy
carried by the "created particles" due to the acceleration. [The meaning of this de-
finition is to give a "physical reality" to the created particles). Namely, for acce-
lerated frames (u~v~ :

the stress tensor reads 181

T.,.,. %;,= v'J


96

where (fY 'f is the schwarzian derivative.


This stress-tensor definition explicitely breaks covariance by coordinate transfor-
mation. Indeed, the choice of the renormalization prescription(~o~i)is not a cova-
riant one because the accelerated vacuum I0'> is frame dependent. At this stage, we
can either abandon the definition (~-~ and find a covariant one or, find a law
which tells us how must transform the vacuum by a frame transformation. The semi-cla-
sscial equation of the back reaction problem :

gives us this transformation law. Explicitely, this equation breaks up 191, in the
two dimensional case, into a geometrical equation and into a set of equations linking
the accelerated frames to the vacuum states. This relation tell us how to transform
the vacuum by frame transformation in order to compensate the non-covariant charac-
ter of the renormalization scheme.

Remark 2.
It will be observed that our study yields a temperature T = o / ~ in the Rindler
case, and not - ~ = O / ~ as t'Hooft suggested recently II01. This ambiguity is due
to the procedure adopted by t'Hooft for the definition of the associated states in
the region R I. In order to define a quantum covariance principle and to secure a one-
to-one correspondance between the global space ~ and the region RI, he identifies
the physics of the left region R I with that of the right region and, he defines a
linear relation between a quantum state in ~ and a density matrix in R I. In order
to describes his proposal, we introduce the P.C.T. symmetry and we link the Fock-spa-
ce W E associated to the operators C~ to the Fock space ~ associated to the

a by, 0 0
where O is the P.C.T. antiunitary operator.

Then, to the state IV> = ~---~ ~ I~ ~ I ~ > ~ ' <OT~ I


is associated the density matrix : ~= ~'--~ ~ I~>
This relation is independant of the choice of the basis in ~ and J O ~ Fock-
spaces. It follows that for the Minkowski vacuum state 10>, the new density matrix :
"" k><.;

is now, a thermal state with a temperature twice the standard one. But the hermitici-
ty condition for the density matrix restrict the~/~ by reality conditions. Therefo-
re, we must restrict ourselves to a real quantum mechanics. More accurately, the
density matrix ~ is a hermitian operator only if the state I ~ > is P.C.T.
invariant. We do ~ot know if such a real formalism has a physical meaning.
Recently, an approach to this problem, based on the construction of symmetric wave
functions has been given Iiii. Even if after identification, the period in the imagi-
97

nary time appears to be half the standard one, the resulting Q.F.T. does not have
a finite temperature at all.

3. THE ASYMPTOTIC CHARACTER OF THERMAL EFFECTS FROM A LOCAL PRINCIPLE.

The previous study has shown the asymptotic character of the thermal effects. The
temperature (8AI) depends only on the behavior of the mapping at the asymptotic re-
gions of the space-time, in particular at the horizon. Recently, Haag and co-workers
1121 have deduced the Hawking temperature from a local principle, implemented outsi-
de and on the horizon. Contrary to the commutator ~(~j~l)= [ ~ ) p ~ t ~ which is
state-independant 1131 (in a globally hyperbolic space-time), the anticommutator
function G (%~)= (x)) (~q) is state-dependent. This function can be used in
order to implement a local criterion satisfied by the "physically allowed" states.
Haag and co-workers have chosen to define their local principle on the tangent-space
of the s p a c e - t i m e ~ .
Let ~ a map from the tangent-space - ~ at ~ to~
such that

The l o c a l principle becomes :

lit can be shown to be independent of the choice of the m a p ~ . ~ G~ )


This local critirion requires just that the singular part of the function
looks like the Minkowski-one.
The aim of this principle is to show that the only thermal-equilibrium-state (with
respect to the Rindler time ~ ) satisfying this critirion on the horizon has the
Hawking temperature T = Q / ~ Indeed,zfor all observables A and B say, a thermal equi-
librium state with temperature ~ = ~ / ~ satisfies K.M.S. condition 1141.

r o =

t
where N~)=U(r)n U(r) ~ and L l ( r ) i s the evolution-operator with respect to
the ~ -time independent hamiltonian.
It is desirable to introduce the commutator IA,BI, and therefore to write the K.M.S.
condition as
98

e~ , ~/[i~(-e), e Joa dr'

In particular for A and B being the field operator at different points we get

In order to analyse the singular part of the ~{I) function, we express the commuta-
tor function G in terms of the solutions of the Klein-Gordon equation in theRindler frame

(3.6) K i._.q2.(ll~"1) QX~(x- ~l.~ l ) ; ~+ ~'~'


o.
and we use :

We take : ~= ~i+ ~Z o

(3.8) ~J:= ~ -+ Sz4"

It is easy to show that the equation ( ~ ) with the local principle (3.2) imposed
outside the horizon allows any value of the temperature, because the Singular
part of the r~h.s, of (3.5) is independent of ~ . But on the horizon ~ ~ I _ _ ~ O j
the behavior of the modified Hankel function becomes

(3.9) Q

Then, the K.M.S. equation (3.5) yields


_ ~ s ~-~k~,~ -~
ko(N e)
4-vc'~p
(3.101
a
> x

Thus the comparison with (3.2) on the horizon shows that the local principle is sa-
tisfied only for the Hawking temperature
99

For quantum fields in accelerated frames, the event horizon plays the role of a ther-
mostat and fixes the temperature.

Remark
This local principle is in fact based on the most singular term of the Hadamard
development I15,161 which postulates the expression of the function as

C311) G (~,~')- ~ li(%~,) V(~,~')~ w,"(~,~'

at least in a small normal neighbourhood. Where O'(~i~-l) is one-half the square


geodesic distance between ~K and ~_i ) and U,V,W some smooth functions. But the
previous principle is too local to ensure the unitarity equivalence between diffe-
rents vacua satisfying this critirion. However, if we impose the Hadamard singulari-
ty up to the order one in ~J/~ in ~{~I) and F(9C~ #) )the unitarity isll81
ensured 1171. The currently favored methods of renormalisation of the stress-tenso~
I t~) 0) I
x---~ ~ J si.~ahr
with 0(.,I//~@ a differential operator~
are applicable only if the G (4) function possess a Hadamard development up to the
order two. Therefore, it seems quite reasonalbe to extend this locality critirion
and to require the asymptotic Hadamard form up to all orders for the anticommutator
function. With this asymption, we are able to renormalize il71 all quantities classi-
cally defined by an expressions like :

~ ' ~...~q
with an arbitrary number of derivatives and an arbitrary number of products of fields.

4. A HAMILTONIAN FORMULATION

A hamiltonian formulation of these effects is crucial for the study of the Wheeler-
De Witt equation for quantum gravitational fields. The equation governing the evolu-
tion of scalar field is built on from the action

by standard procedure. Since, in arbitrary coordinate systems, the metric reads

ds ..- +

the canonical m o m e n t u m S , defined with respect to the time t is :


100

3T'- ~
~()~) _-_ JRi f
(4.~)

and the hamiltonian density is

In particular, the Legendre transformation ( 4 - ~ ) becomes singular if N vanishes.


The submanifold in which the Hamiltonian formulation remains well defined is bounded by
hypersurfaces locally defined by N = 0. They are always null-hypersurfaces in kee-
ping with causal nature of the field propagation. The SehrSdinger equation for the
wave function ~ is

(4.5) ~L

We shall compare the hamiltonian formulation in Minkowski and Rindler frames and in
particular, we analyse the ground states of the differents hamiltonians. For the sa-
ke of simplicity, we restrict ourselves to the two dimensional massless case. The
other cases are similar and can be found in the reference 1191. The Rindler coordi-
nate system is chosen to be ( ~ ) :
op
: e s,~auj

(4.6) OZ : e Oy COSk(Qr) ~e : O~ ~= eY

In a Minkowski coordinate system (t,x), with Diricklet boundary conditions)~--Oat


~-+-~o (the case with Newmann boundary conditions is similar), we expand the wave
function ~ as

fi
Taking q~ as canonical variable, the Schrodinger equation can be written as :
.,1,oo

The ground state of the Minkowski hamiltonian ~rl iS

O
(4.,) ~. (%)=
For the Rindler case, we expand t~ as
101

(4.1o) Y = l a l ) Q~ i1"~
Taking Q~ as canonical variable, the Schr~dinger equation (~- %) is

clz- -~ 1' Sq~ Sq,


The ground state of the Rindler hamiltonian ~ is

i4.1.) ~(%) = ~ ex? _$ dI) i~lqr%


These two ground states can be compared on the hypersurface ~ = ~ = O . Inversing the
mode decomposition (4.7) and (4.10), yields the Bogoliubov transformaiton between
the canonical variable I~ and 2 ~ :

% =

We way now substitute this relation into the expression (49) for the Minkowski
ground state and we obtain in terms of Rindler modes :
-~OO

The two ground state appear clearly different. For high p.momentum, the structure
of ~ approaches that of ~7~ . The difference between ~o and ~ o is signi-
ficant only for small p-momentum, reflecting the infrared (large distance) nature of
the Hawking-Unruh effect. The average of the Rindler number operator for the p-th
momentum mode, ~I~) say, can be calculated directly by analogy with the simple har-
monic oscillator problem :

(4.16)
d
e ~I?I/a _
102

which is the familiar result.


We thus see that the Rindler modes in the Minkowski ground state are populated in a
thermal distribution. But, by construction, the Minkowski state is described by a
coherent state density matrix 1201 and not by mixed one.

Remark

In a non-stationary metric, the choice of the vacuum state as the ground state of
the hamiltonian becomes delicate. In particular, it does not satisfy the local prin-
ciple (911). Therefore it does not satisfy the unitary condition and it does not
possess a well defined renormalized stress-tensor (at least within the standard re-
normalization schemes Ii81).

5. VACUUM FLUCTUATIONS IN ACCELERATED FRAMES.

Hawking radiation is currently interpreted as due to "creation of partlcles' ". T h i s


interpretation, which appears naturally in the formulation based on a field decompo-
sition on the Rindler modes, is supported by a study of particle detection process.
On the other hand, Sciama and co-workers 121,221 have pointed out that this thermal
bath has its origin in the zero-point fluctuations of the quantum fields. The spec-
trum of the zero-point field energy appears to be distorted by the acceleration. He-
re, following ref.(23), we define the particle density and the energy-density from
conserved currents.
For a massless scalar field, the standard density current reads as

Moreover, if ~ is a Killing vector for the background space-time, ~ generates a


transformation which leaves the action invariant. The conserved Noether current
associated to that invariance is

The orbits of the Killing vector can be identified with world lines, ~ ) say, of
some observers. The normalized velocity vector is

(5.3) ~

We opt for the following definition of the density of particles , (n) and the energy
density (e) seen by those observers

(5.4)
103

where < > stands for the vacuum expectation value.


It is convenient to express the vacuum expectation values in terms of the Wightman
functions, W(~,~#=~{~)~(~}>and to introduce the Fourier transform defined with res-
pect to the proper time along these world lines ~f.~(~) :

---l- /" ioas


(5.5) w :] Is

Then, simple calculations give :

(5.7) e-- t~,~ &O _ _ ~/{~).I.W(~/<O)

Now, interpreting ~J as the frequency measured (with respect to the proper time
) by a detector moving along these trajectories, ~-j~) and ~-~ become
the particle and energy densities, respectively.
For inertial observers, we get :

(5.8) -

On the otherhand, for Rindler observers, we have :

(5.9) + ~v~/~o~
e _7
where O~o ~ is the (local) acceleration of these trajectories.
The particles densities are equal in both cases and express that there is "one parti-
cle" in each phase-space cell. But the zero-point energy has, in the Rindler case, an
additional Planckian term. This is the distortion of the zero-point energy due to the
acceleration in agreement with the interpretation of Boyer, Sciama and others.

DISCUSSION

It is quite surprising to note that, despite of its asymptOtic character, the Unruh
effect admits a local description via the detector models 1241. This special feature
has its origin in the high degree of symmetry of the Rindler accelerated frame. The
lines, ~ i = constant, coincide with the world-lines of a system of uniformly accelera-
ted observers and furthermore, every where in the region RI, this coordinate system is
locally the Fermi-Walker coordinate system associated with these hyperbolic trajecto-
ries. For this reason, the Rindler frame appears as the most adapted one to these
104

trajectories. But in general, for a given f~ow of trajectories there is no coordina-


te system which, every where, is locally the Fermi-Walker system associated with them.
(The Rindler frame is the unique one which has this property). Therefore, there does
not exist a coordinate system "naturally" adapted to the flow of trajectories. This
feature has some consequences in non-inertial quantum field theory. In particular,
J. Letaw and J. Pfautsch 125,261 have studied the link between the canonical formu-
lation of Q.F.T. in accelerated frames and the models of quantum detection processes
by non-inertial observers : A rotating detector plunged into the Minkowskian vacuum
responds (the spectrum of the excitations has no simple expression but does not va-
nish) whereas the rotating vacuum defined by a mode decomposition in a rotating fra-
me is equivalent to the Minkowski one. Here, the choice of the non-inertial rotating
frame has been criticized as being highly non-adapted 1271. But, except in the Rindler
case, even if it is possible to consider a single point-like detector by the use of
the Fermi-Walker coordinate system associated with it, it is impossible to take into
account the finite size of the detector. ( In particular, we do not know the effect
of the acceleration on the internal hamiltonian of this detector). Therefore, up to
now, the link between models of detection process and Q.F.T. in accelerated frames
is not really etablished. The study of non-inertial quantum field theory seems better
adapted to analyse the link between the formulations of quantum theory in a global
manifold and in a submanifold respectively, with direct consequences on quantum gra-
vitational effects (section Z and vacuum covariance). The problem of quantum detec-
tion is an old one. We would like to point-out some new problems which appear from
the non-inertial character. First, if we would like to represent the quantum measu-
rement by an observable, without describing the detection process, we must find
a) which element of the observable algebra represents an ideal detection porcess ?
b) A transformation law which tells us how this observable is modified when the same
detector is forced to move along some other world-line ?
Up to now, there is no "quantum covariance principle" that answers these questions.
Therefore, people looked for models of detection processes. Thus to ensure that the
internal hamiltonian of the detector is the same along any trajectories, the current-
ly studied detectors were based on point-like monopoles. By studing the reponse func-
tion of the detectors, these models of detectors propose to analyse the "effective
particle content" of the quatnum state seen along the detector trajectory. Recently,
Hinton 128[ and Davies 1291 have made important remarks concerning these models.
i) How to normalize the function reponse, in particular in curved space-time
ii) Do different models of detectors give the same "effective particle content" to

the same quantum state.


Their studies indicates that the "effective particle content" is detector model depen-
dant. Therefore, these questions are still open : What is really measured by these
detection ? or to which measurements do the detection process correspond ? It would
be interesting to discuss the problem of anisotropy in the detection of the accelera-
tion radiation in connection with these questions 1301. A way to avoid these problems
105

is to go back to the first point of view, and to search to describe measurement by


intrinsic quantities like , ~ or other vector densities. In the Kindler
case, the hyperbolic trajectories are generated by a Killing vector and therefore the
Noether current defined in section V is "naturally" adapted to represent the quantum
measurement. But in general, for an arbitrary trajectory, such symmetry does not exist
and therefore no current is intrinsically defined that answers the questions a) and
b). Formulation of a quantum covariance principle as expressed by the previous ques-
tions is still an open problem.

ACKNOWLEDGMENTS

I am grateful to Norma Sanchez for numerous discussions, advice and encouragments.


I acknowledge Brandon Carter for numerous stimulating discussions and for a critical
reading of the draft manuscript.

REFERENCES

I. S.W. Hawking, Com.Math. Phys., 43, 199, (1975).


P.C.W. Davies, J.Phys., A8, 609 (1975).
W.G. Unruh, Phys.Rev. DI4, 870 (1976).
S.A. Fulling, PHys.Rev. D7, 2850 (1973).
2. H. Rumpf, Phys. Rev. D28, 2946 (1983).
S.M. Christensen and M.J. Duff, Nucl. Phys. B146, ii (1978).
P. Candelas and D. Deutsch, ProC.Roy. Soc. A362, 251 (1978) and Proc. Roy. Soe.
A354, 79 (1977).
3. R.J. Hughes, Preprint CERN T.H. 3670 (1983).
G.L. Sewell, Ann. Phys.N.Y., 141, 201, (1982).
4. C.W. Misner, K.S. Thorne, J.A. Wheeler, Gravitation,(Freeman, SanFrancisco 1973).
5. W.G. Unruh and N. Weiss, Phys.Rev. D29, 1656 (1984).
6.6. C. Itzykson and J.B. Zuber, Quantum field Theory, (Mc Graw Hill, N.Y., 1980).
6. N. S~nchez, Phys.Rev. D24, 2100 (1981).
7. D. Bernard and N. Sdnchez, preprint in preparation.
8. ~ W . Davies, Proc. Roy. Soc. Lond, A354, 529 (1977).
9. R. Balbinot and R. Horeanini~ Phys.Lett., 151B, 401 (1985).
N. Sa~chez, to appear in Nucl. Phys. B.
i0. G.t'Hooft, J. Geometry and PHys., i, 45 (1984).
G. t'Hooft, Utrecht Preprint, December (1984).
Ii. G. Gibbons, in Carg~se Lectures 1985, to appear.
N. S~nchez and B. Whiting, in preparation.
12. K. Haag, H. Narnhofer and U. Stein, Com.Math. Phys., 94, 219 (1984).
13. Lichnerowitz, in Les Houches, 1963, edited by C. de Witt and B.S. de Witt,
Gordon and Beach.
14. R. Kubo, J.Phys.Soc. Japan, 12, 570 (1957).
P.C. Martin and J. Schwinger, Phys.Rev. 115, 1342 (1969).
15. See, for example, F.G. Friedlander, The wave equation on a curved spacetime,
(Cambridge University Press, Cambridge, 1975).
16. S.A. Pulling, F.J. Narcowich, R.M. Wald, Ann. Phys. N.Y., 136, 243 (1981).
17. D. Bernard, Meudon Preprint, 1985.
18. N.D. Birrel and P.C.W. Davies, Quantum fields in curved space (Cambridge Univer-
sity Press, Cambridge, 1982).
106

M.R. Brown and A.C. Ottewill, Proc.Roy. Soc.Lond. A389, 379 (1983).
19. K. Freese, C.T. Hill and R. Mueller, Nucl. Phys. B , (1985).
20. T.D. Lee, Columbia University Preprint (1985).
21. D.W. Sciama, P. Candelas and D. Deutsch, Adv. Phys. 30, 327 (1981).
22. T.H. Boyer, Phys.Rev. D2!, 2137 (1980).
23. S. Hacyan et al., Phys. Rev. D32, 914 (1985).
24. W.G. Unruh, Phys.Rev. DI4, 870 (1976).
B.S. de Witt, in General Relativity : an Einstein centenary survey, edited by
S.W. Hawking and W. Israel, (Cambridge University Press, Cambridge, 19 ).
W.G. Unruh and R.M. Wald, Phys.Rev. D29, 1043 (1984).
25. J.R. Letaw, Phys.Rev. D23, 1709 (1981).
26. J.R. Letaw and J.D. Pfautsch, Phys.Rev. D24, 1491 (1982).
27. N. Myhrwold, Phys. Lett., 100A, 345 ( 1 9 8 ~ .
28. K.J. Hinton, J. Phys. A : Math.Gen., 16, 1937 (1983).
K.J. Hinton, Class. Quantum Gray., i, 27, (1984).
29. P.C.W. Davies, in Essays in Honor of the Sixtieth Birthday of B.S. deWitt ; edi
ted by S. Christensen. (Adam Hilger, Bristol~ 1984).
30. N. S~nchez, to appear in Phys.Lett.A
K. Hinton, P.C.W. Davies, J. Pfautsch, Phys. Lett. 120B, 88 (1983).
W. Israel, J.M. Nester, Phys.Lett., 98A, 329 (1983).
STOCHASTIC DE SITTER (INFLATIONARY) STAGE
IN THE EARLYUNIVERSE

A.A. STAROBINSKY
Landau Institute for Theoretical Thysics,
Moscow, 117334, U.S.S.R.

and

ER 176 C.N.R.S. "D~partement d'Astrophysique Fondamentale"


Observatoire de Meudon
92195 Meudon Principal Cedex
FRANCE

Abstract

The dynamics of a large-scale quasi-homogeneous scalar f i e l d producing the de Sitter


( i n f l a t i o n a r y ) stage in the early universe is strongly affected by small-scale quan-
tum fluctuations of the same scalar f i e l d and, in this way, becomes stochastic. The
evolution of the corresponding large-scale space-time metric follows that of the sca-
lar f i e l d and is stochastic also. The Fokker-Planck equation for the evolution of
the large-scale scalar f i e l d is obtained and solved for an arbitrary scalar f i e l d
potential. The average duration of the de-Sitter stage in the new inflationary sce-
nario is calculated (only partial results on this problem were known e a r l i e r ) . Appli-
cations of the developed formalism to the chaotic inflationary scenario and to quan-
tum i n f l a t i o n are considered. In these cases, the main unsolved problem lies in i n i -
t i a l pre-inflationary conditions.

1. Introduction

In the models of the early universe with an i n i t i a l or intermediate metastable de


S i t t e r ( i n f l a t i o n a r y ) stage with an effective cosmological constant produced both
by quantum gravitational corrections to the Einstein equations 111 and by a scalar
field 12-4 , of extreme importance is the e x i t from this stage that depends on the
way of decay of the effective cosmological constant because i t determines the spec-
108

trum and amplitude of metric perturbations for the subsequent evolution. These per-
turbations break the homogeneity and isotropy achieved earlier at the inflationary
stage and can, in the worst case, destroy all the advantages of i n f l a t i o n . Two ways
of decay of the effective cosmological constant are possible : via (quasi) homogene-
ous classical i n s t a b i l i t y and via inhomogeneous quantum fluctuations. In the f i r s t
case, the amplitude of perturbations of the de Sitter space-time in the modes which
preserve (exactly or approximately) the isotropy and homogeneity of the 3-space in
some frame of reference is much more than the amplitude of other, inhomogeneous per-
turbations. Thus, we have a classical (quasi)-homogeneous perturbation from the very
beginning and the subsequent evolution is deterministic ; the duration of the de
S i t t e r stage is t o t a l l y determined by the i n i t i a l amplitude of this perturbation.
This type of decay takes place, for example, in the author's model i l t for the case
of the closed 3-space section i f the spatial dimension of this section was of the
order of H "4 at the beginning of the de Sitter stage (in the paper, we put
= c = 1 ; a(t) is the scale factor of the Friedmann-Robertson-Walker isotropic
cosmological model ; ~ = J/a).
The existence of a quasi-homogeneous classical scalar f i e l d is also assumed in the
"chaotic" inflationary scenario 151 (for the inclusion of the R2 term where R is the
Ricci scalar, see 161). Here, the term "chaotic" simply means the unspecified depen-
dence of the metric and the scalar f i e l d on space coordinates though this dependence
is weak enough, so that the spatial derivatives of all variables are much less than
the temporal ones.
In the second case, we have no large (quasi) homogeneous perturbation at the begin-
ning of the de Sitter stage. This p o s s i b i l i t y was f i r s t pointed in 171 in connection
with the model i l i . But, in fact, this situation is more typical for the models whe-
re the de Sitter stage arises from the i n i t i a l l y radiation-dominated, "hot" universe
in the course of a non-equilibrium, close to the ~ order phase transition (for exam-
ple, the "new" inflationary scenario). Here, nevertheless, a large quasi-homogeneous
"classical" perturbation with characteristic wavelengths >>H"1 can arise during the
de Sitter stage from small-scale quantum perturbations. In other words, "classi-
cal order" appears from "quantum chaos". In spite of being effectively classical, the
evolution of this large-scale perturbation and the space-time metric as a whole is
essentially stochastic. The duration of the de Sitter stage also becomes a stochastic
quantity in this case.
This is just the process we are interested in. I t belongs to the class of the so-cal-
led "synenergetic" problems which arise in different branches of science and attract
much interest at the present time. We shall consider the new inflationary scenario
where the role of the abovementioned perturbation is played by the non-zero large-
scale scalar f i e l d ~) . I t is assumed that ~ ~ 0 (or s u f f i c i e n t l y small) at the
beginning of the de Sitter stage. We shall obtain the Fokker-Planck equation for the
evolution of the probability distribution of ~ (Sec.2) and calculate the average
duration of the de S i t t e r stage in the new inflationary scenario in Sec.3 (only par-
109

tial results on t h i s problem or order-of-magnitude estimates were obtained e a r l i e r


18-101). A f t e r t h a t , we shall turn to the chaotic i n f l a t i o n a r y scenario (Sec.4)
and discuss the modern state of the Problem of the "creation" of the universe b r i e f l y
(Sec.5).

2. Evolution of a scalar f i e l d in the new i n f l a t i o n a r y scenario.

The de S i t t e r stage in the new i n f l a t i o n a r y scenario is assumed to be produced by


the vacuum energy of some scalar f i e l d with the Lagrangian density

(i)
2,
where the vacuum e f f e c t i v e potential V~) has the f o l l o w i n g properties :

v(@.) : o (2)
I M,~ ~ _v 4 3- ~ ~
M2 can have both signes. ~_.- g o is the flat space-time (true vacuum). ~ - 0 is
the f a l s e vacuum. We include the term in to describe the case of the s o - c a l l e d
"primordial" i n f l a t i o n I l l l simultaneously. At the non-zero temperature T, the po-
t e n t i a l V acquires the additional thermal term which is either small or, with the
sufficient accuracy, has the form ~-B'I "L ~ ' , B(41, T:~a -I.
At the de Sitter stage, H = Ho = const, a = ao exp (Hot), where ~o~= ~a~C~Vo/.~
(the spatial curvature is negligable). In order to have enough long de Sitter stage
and enough small perturbations at the subsequent stages, the following conditions
should be f u l f i l l e d :
?.
IMP! ~ H/Zo ; v/t.t. ~ to" ; ~ ~o'" (3)

The Coleman-Weinberg potential does not e v i d e n t l y meet these requirements, so i t is


usually assumed now that ~ is some weakly i n t e r a c t i n g scalar f i e l d , in particulaG
i t should be the s i n g l e t with respect to SU(5) or any other grand u n i f i c a t i o n group.
In such a way, the s p i r i t , though not the l e t t e r , of the "new" i n f l a t i o n a r y scenario
is maintained.
The de S i t t e r stage begins when T4~V o, I t can be divided into two successive periods:
"hot" and "cold" (vacuum). During the hot period, the temperature T ~H o and quantum-
g r a v i t a t i o n a l effects caused by the space-time curvature are unimportant. The dura-
tion of t h i s period is rather short ; in dimensionless u n i t s ,

~. ,, ~ ,,, z,, ( v~l, 1 H. ) .,. ~ ( c, "~ V." '/~ ), ( ~)


110

that is of the order of 10 t y p i c a l l y . After that, the cold (vacuum) period begins
where T << Hoand, in fact, temperature effects can be neglected (except only for the
calculation of the i n i t i a l dispersion of ~ ; see Eq.(13) below). This period is
the most interesting because quantum-gravitational effects connected with the space-
time curvature play the decisive role here (we denote its beginning by to).
To obtain q u a n t i t a t i v e l y (not only qualitatively) correct results one should not
use such quantities as < ~l~ > or < ~ 2 > (the approaches based on these quantities
have been correctly c r i t i c i z e d in [12, 131), Instead of t h i s , we represent the quan-
tum scalar f i e l d (]~ (the Heisenberg operator) in the form :

't , ~)o
k =

Here, ~ ' ( t , ~)
, oct) :
t
contains only long wavelength modes with k<< Hoa(t),
&:o~Y:.

~(~
4< : .

is the
small correction that can be neglected in the leading order in small parameters
NM21/Ho2, ~/H o, ~ and the second integral term in Eq. (5) satisfies the free
massless scalar wave equation in the de S i t t e r background : [ ~ = O. Thus,

(6)

and ~ and a% are the usual creation and annihilation Bose-operators. The a u x i l i a r y
small parameter 6~ is introduced to refine the derivation, i t w i l l not appear in all
final equations. ~n fact, i t cannot be a r b i t r a r i l y small ; the immediate comparison
of d i f f e r e n t terms in Eq. (5) suggests that ~ >> IMI/H o but more refined treatment
consisting in the substitution of the solution (6) by the solution of the free
massive wave equation O ~ + M2~ = 0 in the de S i t t e r background (that does not
change Eq. (8) below in the leading approximation in [M21/Ho2) shows that the signi-
f i c a n t l y weaker condition i -~n& I << max (Ho2/M2 , Ho/P , ~ - i ) is s u f f i c i e n t .
I t can be also seen immediately that the account of the abovementioned thermal
correction to V ( ~ ) results in the substitution

(7)

in Eq. (6). This gives an effective infrared c u t - o f f that can be important in some
problems.
The scalar f i e l d ~ satisfies the operator equation of motion [ ~ ' + dV /d~ = 0
111

exactly. Using (5, 6) and the conditions of "slow r o l l i n g " (3), one obtains the
following equation f o r ~ in the leading order :

a Wo oL (8)

(zn')#" ,, +

That is the main point : the large-scale scalar f i e l d ~l~ changes not only due to
the classical force d V ( ~ ) / d ~ but also due to the flow of i n i t i a l l y small-scale
quantum fluctuations across the de S i t t e r horizon k = a(t)H o in the process of expan-
sion. Moreover, the evolution of inhomogeneous fluctuations is l i n e a r inside the de
S i t t e r horizon and even in some region outside i t ; on the other hand, the evolution
of ~ is non-linear but here the spatial and second time derivatives o f ~ are small.
Below, we shall omit the bar above ---~ , so ~ w i l l mean the large-scale f i e l d only.
Two important consequences follow from Eq. (8). F i r s t l y , there are no spatial deri-
vatives in Eq (8) at a l l . This means that the evolution of ~ can be studied
l o c a l l y , in the "point" ( t h i s "point" has, in f a c t , spatial dimension ~ Ho" I ) . The
temporal evolution o f ~ ) i s slow as compared to Ho-I ( i f the i n f l a t i o n exists at a l l ) ,
so our time " d i f f e r e n t i a l " dt can be also chosen .,, Ho-1 ; only the processes with
characteristic times ~>> Ho-i w i l l be considered. Secondly, though ~ and f have
a complicated operator structure, i t can be immediately seen that a l l terms in
Eq. (8) commute with each other because ~k and ~k+ appear only in one combination for
each possible ~ ! Thus, we can consider ~)and f as c l a s s i c a l , c-number quantities.
But they are c e r t a i n l y stochastic, simply because we can not ascribe any d e f i n i t e
numerical value to the c o m b i n a t i o n [ ~ ~:C.~(-~ ~ ) - ( ~ ~ ) ] . As
a r e s u l t , the peculiar properties of the de S i t t e r space-time - t h e existence of the
horizon and the appearance of the large " f r i c t i o n " term 3Ho~ in the wave equation-
s i m p l i f y the problem of a non-equilibrium phase t r a n s i t i o n greatly and make i t s solu-
tion possible, in contrast to the case of the f l a t space-time.
I t is clear now that Eq. (8) can be considered as the Langevin equation f o r ~ b ( t )
with the stochastic force f ( t ) . The calculation of the correlation function for f ( t )
is straighforward and gives ( ~ i s the same throughout) :

< = Ho3
(9)

Thus f ( t ) has the properties of white noise. This appears to be the case because
d i f f e r e n t moments of time correspond to d i f f e r e n t k because of the ~ -function
112

in the definition of f, and ~k and ak+ with different'~commute. For spatially


separated points,

< > :
e.,~., 14.0 I~-~1 "(10)

We are interested in the average values <F(~ )> where F is an arbitrary function.
For that case, one can introduce the normalized probability distribution ~ ( ~ ,t)
for the classical stochastic quantity (~

= ) so that
(too

~o,O

By the standard procedure, the Fokker-Planck (or, better to say, Einstein-


Smoluchowski) equation for ~ follows from (8) and (9) :

This equation has to be supplemented by some initial condition for ~ at t = t o. It


should be noted also that Eq. (12) is applicable at the stage of "slow rolling"
(I ~ I << Ha ~ ) only. When this condition ceases to be valid ~hat takes place at
~min (Ho ~ -1/2 , Ha2-}) -1)), the second time derivative o f ~ comes into
play (though spatial derivatives are s t i l l unimportant), the de Sitter stage ends
and ~ reaches its f l a t space-time equilibrium value ~ o during the time interval
less than Ho-1. After that, a number of oscillations around ~o is possible.
Thus, strictly speaking, we can use Eq. (12) i f only i ~l<<min (Ho ~-1/2,Ho2 ~-1).
But just because we are not interested in time intervals ~ t A_sHo-I when calcula-
ting such quantities, as e.g., the average duration of the de Sitter stage, we can
safely substitute V ( ( ~ ) in Eqs. (8,12) by its expansion for I~LT~I << ~ o (the
second line in Eq. (2))and use Eqs. (8, 12) for arbitrary ~ . Then the (stochastic)
moment of time t s when the de Sitter stage ends coincides with the sufficient accu-
racy ( ~ t ~ Ha-l) with the moment when i ~ l reaches infinity according to Eq. (8)
(the stochastic force f ( t ) becomes unimportant at the last stage of evolution).
A note should be added about time reversibility. The microscopic evolution of the
total scalar field operator ~) is, certainly, time-reversible, so the apparent,
diffusion-like irreversibility of the evolution of ~ ( ~ ,t) is due to, as usually,
"coarse-graining" that takes place continuously in the process of neglecting more
and more information contained in separate modes with different k.
113

3. Average duration of the de Sitter stage in the new inflationary scenario.

Now we have to introduce the i n i t i a l condition for ~ at the beginning of the "cold"
part of the i n f l a t i o n : ~ : I o ( ~ a t t = t O. The simplest possible choice would be
te(~) : ~ ) In f a c t , the situation is more complicated and depends on the
initial conditions at the Planckian moment tip = G. I f one assumes thermal e q u i l i -
brium before the de S i t t e r stage, then the contribution of thermal quanta of the
scalar f i e l d ~ with the rest mass m2(T) << T2 (B<<1) to ~ e ( ~ ) is gaussian with
the dispersion

< : L el,)
IT~ o
At T <<Ho, the main contribution to the integral is due to the region k~ko <<aT,
I ~ I < < 1 . Using (6,7), we obtain [I0,141 :
CW~

a r t ~" (.k'+ i ~ : ) :'


This expression is valid i f the modes with k~, ko are inside the horizon at the begin-
ning of the de S i t t e r stage that requires B>>GV~vHo/Mp , where M@= G- is the
Planck mass. In the opposite case, the modes with k < kI = Ho a(t=Ho-1) are never
inside the horizon. For these modes, ~0~., const, and, in fact, nothing definite can
be said about their occupation numbers. The probability distribution needs not be
gaussian either, but i t is independent of time (we do not include the term R~2/12
2
into the Lagrangian (1) because then the fine-tuning between M2 and Ho is necessary
for the i n f l a t i o n to occur). In this case, the reasonablelower l i m i t on the i n i t i a l
dispersion can be obtained by integrating from kI to ~ in Eqs. (13,14) that gives

< <~(t:~.)> >~ HoV. V~ ,,, Ho'/* M~/z >>Ho z . (15)

i f thermal equilibrium is assumed in the whole region inside the horizon at the
beginning of the de S i t t e r stage.
Thus, the i n i t i a l dispersion of ~ , in general, exceeds Ho2 s i g n i f i c a n t l y . Never-
theless, i t appears (see below) that i f

then the i n i t i a l dispersion can be neglected because its effect on the average dura-
tion of the de Sitter stage proves to be small. Therefore, there exists a set of pos-
s i b l e (though not necessary) i n i t i a l conditions at t = t~,for which we can use the
i n i t i a l condition ~)o(~)=~(~)at t = to"
Note that, i f the last term in Eq.(12) can be neglected (that takes place in the be-
114

ning of the "cold" period of i n f l a t i o n ) , then Eq. (12)is the usual diffusion equa-
tion. Thus, the i n i t i a l l y gaussian distribution ~C(~) remains gaussian in the
course of time evolution and its dispersion changes as

~rr {
This is just the result obtained in [9,10,15J. In the presence of the quadratic po-
tential V = M2~2/2, the distribution remains gaussian and the dispersion can be
obtained from the "one-loop" equation 1101

In this case, Eq.(20) below reduces to that of the harmonic o s c i l l a t o r and can be
solved analytically.
In the general case, the solution of Eq.(12) is :
3

a# ~ /
where ~ ( ~ } i s the complete orthonormal set of eigenfunctions of the Schrodinger
equation

i - 2,

I t was explained at the end of Sec.2 that we may set V(eo) : -JV(-~)I = - ~ . Therefo-
re, W(%m) = ~ and Eq.(20) has the discrete spectrum of eigenvalues only. For V(~)
given in Eq.(2), i t is the equation of the anharmonic (or doubly anharmonic) oscil-
lator. The coefficients cn are obtained from the i n i t i a l condition for ~)(!T~, ~ )
at t = t o :
115

The behaviour of J C ~ , I : ) at large times i s , as usually, determined by the lowest


energy level Eo. Eo is s t r i c t l y positive that follows from the "supersymmetric"
form of the potential W(~).
In practice, we are more interested not in ~ ( ~ , i : ) i t s e l f but in w(t s) - the probabi-
l i t y d i s t r i b u t i o n for the stochastic moment t s when the de S i t t e r stage ends'.w(ts)
can be obtained from ,~(~i:) by the following way. ~et the r o l l i n g of the scalar
f i e l d to both sides is possible : V ~ ) = -~. The integral .~ d~)~.--~)"I converges at
] ~ ) ~ - ~ that means
I*
that [ ~ ) J approaches i n f i n i t y in f i n i t e time.~~V'For l~l-," ==, the
evolution of ~ becomes deterministic ; both the stochastic force in Eq.(8) and the
second d e r i v a t i v e with respect to ~ in Eq.(12) can be neglected. Then the solution
of Eq.(12) for ~-~~is, correspondingly,

where g is some unknown function that has to be determined from the previous evolu-
tion. The form of the solution represents the fact that the p r o b a b i l i t y is transpor-
ted without changing along the classical paths

Therefore, one can introduce w(ts)cK.g(ts). The exact c o e f f i c i e n t of p r o p o r t i o n a l i t y


is determined by the condition of p r o b a b i l i t y conservation

along the path (23). I f we do not make difference between r o l l i n g down to the l e f t
and to the r i g h t sides, then the resulting expression for w(ts) is

_ 3Uo

I f the r o l l i n g of the scalar f i e l d is possible to the r i g h t side only (V(-m) = ~ ,


V(~) = - m ; e.g., when~ = 0 in Eq.(2)), the second l i m i t in Eq.(25)hastobeomitted.
The d i s t r i b u t i o n w(ts) is c e r t a i n l y non-gaussian. I t s behaviour for large t s is ex-
ponential and is determined by the lowest energy level Eo. Though w(ts) cannot be
computed a n a l i t i c a l l y , i t is remarkable that the closed e x p l i c i t expressions for a l l
moments <(Ho(ts-to))n> with integer n can be obtained in the form of successive in-
tegrals. The approach used here is s i m i l a r to the Stratonovich's " f i r s t time passage"
method.
Let us consider a set of the functions
116

(26)

Then

(27)

Integrating both sides of Eq.(12) over t from t = t o to t = ~ , we obtain the ordina-


ry d i f f e r e n t i a l equation

H? Q~"+ ~ ( av Q.), __y.(~)

I t s solution, subjected to the boundary conditions Qo(~.~) = 0 (becauseS(m,t)=O),is

C ~, ~ ~

a,:
I f the r o l l i n g is possible to the r i g h t ( l e f t ) side only, then C=O (C=1). For the
symmetric case V(-~) = V(~) and ~ ( - ~ ) = ~ ( ~ ) , C : . Now,

~U--, _,..~ ~. a--- ~ ( ]} ) = O - ~ ) - c _- ~ = . ( ~ d t ( ~ o l


eo
Thus, the p r o b a b i l i t y w(ts) introduced according to Eq.(25) is properly normalized.
By multiplying both sides of Eq.(12) by ( t - t o ) n and integrating over t from t o to
t = ~ , the recurrence relation between Qn can be found. I t has the form (n>zl) :

ga" + 3 ..'o : - ~ ~?-.-. :~:I


The boundary conditions are Qn(~*:) = 0 for a l l n. Then
117

Con~".

(32)

Using Eq.(27), we obtain

~/Ho -

In p a r t i c u l a r , the average dimensionless duration of the de S i t t e r stage is equal to

: (34)

where~is given in Eq.(4) and Qo is presented in Eq.(29).


Let us now consider several p a r t i c u l a r cases. Let ~ = 0 in Eq.(2) (that corresponds
to the original picture of the "new" i n f l a t i o n ) and #o(~ = ~ C ~ Then Eq.(34)
s i m p l i f i e s (C = ) :

(the constant term in the potential may be omitted because i t cancels in Eq.(35)).
After some manipulation, the expression (35) can be represented in the form contai-
ning only one integration : 4

-i-

' o
] (3,)

V~ '~-' o~- ~_n" M ~"

where ~ is the confluent hypergeometric function.


Three more p a r t i c u l a r cases are of special interest.
I) M2 < 0 ; ~ Ho2 << IM21 << Ho2 ; 141 >> i .
Then
-'IG rr 7" I'l 4 ]
(37)
ZlM'I S~, 14o4 "I-Y
118

where *'6"= 0.577 ... is the Euler constant. In this case, one-loop approximation
which consists in the substitution of <~4> by 3(<~2>)2 in the equation for <~2>
gives the result which is correct with the logarithmic accuracy :
H~
4~,,A~'> = No ae~ + 3
o.n,...-Lo,,f ~ I m~'l R 71 ,qo ~- (38)
However, more accurate approach was developed in 1101 for this case which gave the
right answer. I t consists in the observation that in this case the stochastic force
f ( t ) in Eq.(8) is important then and only then when the classical force (-dV(~)/d~)
can be neglected and vice versa. Thus, Eq.(8) can be integrated directly that gives
the following result for the stochastic quantity t s i t s e l f 1101 :

14o(~-~o) = ~~ ~ IM'I, ; (39)

where ~ I is a gaussian stochastic quantity with zero average and the dispersion

:, (40)

(the thermal contribution to <~12>


is neglected here for simplicity). After avera-
ging~4 in Eq.(39) over the gaussian distribution, just the correct result (37)
appears.
2) IM21 ~ Ho2 ; I~I ~ I.
For this case, only one-loop 1101 or order-of-magnitude 191 estimates were known
earlier. It follows from Eq.(36) that

One-loop approximation gives the numerical coefficient in the second term equal to
~2 / ~ ' ~ 6 . 9 8 that is 2.56 times less.
It is intructive to consider the case of a many-component scalar field ~a with the
symmetry group O(N) and see how the one-loop approximation becomes exact in the limit
N - ~ . Let ~ = (~.a~a) . The strightforward application of the developed approach
shows that the corresponding generalization of Eq.(12) to the N{I case is :

.~ = ,~..,. ~,~-, ~.~ ~ .~)(42)

W~ /
o
119
where SN is the area of the N-dimensional sphere (O(N)-symetrical initial condition
for ~ is also assumed). If ~(~,t) =~(~) at t = t o, then, instead of Eq.(35), the
following expression for the average duration of the de Sitter stage results :
~ t ~ ~"-~

For V((): Vo- ~ ~'/"~


< H. C~.-~.) > = Z)
q~w p.,F ( 4h NI~) (44)
< ,. c~,.~.~ >,.~-- ,-,-'-~
Thus, both expressions tend to the same limit .,rc=l~"/~)P.,~
at N..~aO(but from different sides).
Now we return to the N = I case and calculate the dispersion of the quantity Ho(ts-

<H:(t)"BY theb use


t a f
> i ~---
Eqs'(32'33)'
n ~ I I-"~( J~~IbY~
: we
: ) ~ ~,IA.d~~ ~/~(~f%/~/Z_~Y~;tp.

s~ - < " : ( ~ . ~ . ) ~ > - ( 4 , . c~,-~.) > ) ~ = (~')


_ F ~ (e~ ~f, 'I~)~t

where F(~,k) is the elliptic integral of the f i r s t kind. Also interesting is to cal-
culate the change in the result (41) due to the spreading of the initial condition
at t = t o (the "thermal" correction). If .~C~}is the gaussian distribution with
the zero average and the dispersion ~)~- (see Eqs.(14,15), then by applying Eq.(29)
with C = the following result can be found :
z -Z
_ ~n-~- @ T Ho ;-
(46)
q'~'~, << Ho
Thus, i f the condition (16) is satisfied, then the thermal correction is small ; in
the opposite case, the inflationary stage is very short.
3) ~ Ho2 ~ M2 ~ Ho2 ; IX~ i.
In this case, the result (36) simplifies to the form :

C.&rr= M ~ )
~EM" e ~ , . z, 3 Ho ~ (47)
120

The exponent j u s t coincides with the r e s u l t obtained by Hawking and Moss 181 with
the help of the de S i t t e r instanton. Thus, our approach reproduces the instanton re-
sults without using instantons at a l l . Moreover, we have obtained a l i t t l e more -
the c o e f f i c i e n t of the exponential, that corresponds to the summation of a l l one-loop
diagrams on the instanton background in the standard functional integral approach.
The corresponding p r o b a b i l i t y d i s t r i b u t i o n w(ts) is determined by the lowest energy
level Eo of Eq.(20) with the excellent accuracy and, thus, is purely exponential :

4~'~ 4It ~
~oI4; _ ~M" ( _ 2. ~ M ,) "/ , ' ~ , / , ( 4 8 )

I t is clear in our approach that the transition of the scalar f i e l d through the po-
tential barrier takes place only locally, that is, in the volume.~.Ho"3 (in fact,
somewhat larger), but not in the whole 3-space. This fact can be also understood
in the functional integral approach i f one rewrites the de Sitter instanton in the
static, "thermal" form :

_ + (t_ +

~. {.,, M "B""I`
where "~ is periodic with the period 2 ~ H o - l . Then the instanton t e l l s us that
has reached the top of the potential b a r r i e r inside the horizon (r < Ho-1) but gives
us no information about the behaviour of ~ outside the horizon.
That is enough for the case of the "new" i n f l a t i o n . Now we shall turn to the so-cal-
led "primordial" i n f l a t i o n i l l 1 where i t is assumed that P ~ O, ~ = 0 and present
the most i n t e r e s t i n g results b r i e f l y . In t h i s case, the average duration of the de
S i t t e r stage is given by Eqs.(34,29) with C = O. Two l i m i t i n g cases are the most
important and representative.
i ) IM21 ~ Ho4/3 V 2 / 3 .
Then

/ ,~ F'("/s)(_g.)i4"o -/s
(5o)
,.o ( " % , ) v, .
2) Ho4/3 ~ 2/3 ~ M2 ~ Ho2.
In t h i s case,

~,," U2 (51)
M~ q Hoe "v'~
121

Again, the exponent is j u s t the action for the Hawking-Moss instanton which is equal
to the difference between the actions for the de S i t t e r instantons (49) with ~ - m ~ x
: M~/p and ~: ~m{~=O. The third case M2<O, Ho4/3 ~ 2/3C IM21 ~ Ho2 reduces, in
fact, to the second one a f t e r s h i f t i n g the scalar f i e l d : ~ . = ~ E I - |M~Jj/~ )
The quantitative results presented in the Sec. 2,3 were f i r s t published by the au-
thor in the shorter form in Russian in i16,171. Two points should be emphasized,
however.
Firstly, though the quantity Y,~(a(ts)/a(to) ) = H o ( t s - t o ) has the well-defined probabi-
l i t y d i s t r i b u t i o n w(ts), the quantity a ( t s ) / a ( t o ) does not, because EoHo2~l in all
cases. Thus, i t seems that the quantity ~na(t) is more suitable for the description
of the stochastic i n f l a t i o n than the scale factor a(t) i t s e l f .
Secondly, the calculated duration of the de S i t t e r stage gives us the typical size
of causally connected regions. However, only a minor last part of this i n f l a t i o n pro-
duces regions those remain approximately homogeneous and isotropic in the course of
subsequent evolution. This follows from the fact that a f t e r the inflation~the space-
time metric at scales much larger than the cosmological p o s t - i n f l a t i o n a r y particl~ol,~
horizon has the following simple structure in the proper ("ultra-synchronous") gauge.

ds 2 = d t 2 - e x p ( h ( ~ ) ) a 2 ( t ) ( d x 2+dy2+dz 2) ; (52)
#I

h(~) = 2 ~(a(ts(3m))/a(to)),

where h(~) is not assumed to be small and a(t) is the scale factor for the s t r i c t l y
isotropic and homogeneous solution. The quantity h(~~) is essentially stochastic, its
rms value is of the order of its average (see, e.g., Eq.(45)). Thus, the metric (52)
becomes anisotropic and inhomogeneous in the course of the a f t e r - i n f l a t i o n a r y expan-
sion when spatial gradients of h ( ~ (omitted in Eq.(52) in the leading approximation)
come into play. This situation i l l u s t r a t e s the well-known fact that "general" i n f l a -
tion produces neither isotropy nor homogeneity of the present-day universe and, there-
fore, cannot "explain" them without further assumptions. Nevertheless, i f the condi-
tions (3) are f u l f i l l e d , then the last, "useful" part of i n f l a t i o n does produce suf-
ficiently large regions with the degree of homogeneity and isotropy that matches the
observations. I t is important that during this part of i n f l a t i o n the stochastic force
f ( t ) in Eq.(8) becomes small as compared to the classical force ( - d V ( ~ ) / ( d ~ . Then,
for regions those are not too large, h(~) can be represented in the form which was
used in 110,18-211 :

h(~) = const +~h(r~ ; ~h(~) = - 2 H o ~ E ( t , ~ ) / ~ , (53)

where ~ is the small fluctuation of ~ ( t ) produced by f ( t , ~ ) . Here ~h(~) is


r e a l l y small.
The duration ~ of this "useful" part of i n f l a t i o n (when i ~ h I < l ) is easily e s t i -
mated using the expression for perturbations (53) :
122

t4o ~,~:, ",, ~-'H I M'I /_..~ '~4/~ t-I0~ , ,)=o ;


IMP IM'I >> ~,'/~ No= ~ ~=o.
IM'I Ho('
(54)

M tl ~ 1 M'! >) p'~Ho31~ , ~ = o

Ho ~ t I contains no exponentially large multipliers. I f ~ or "~ are fixed, then


A t I is maximal and the amplitude of perturbations at the given present-day scale
is minimal when IM2[ ~ ~ 1/3Ho2 or iM2[ ~ Ho3/2 ; the upper limits on ~ and
presented in Eq.(3), s t r i c t l y speaking, refer just to these cases. I f M does not
satisfy these conditions, the duration of the "useful" part of inflation diminishes ;
however, the numerical restrictions on } and V remain practically unchanged due
to the f i r s t condition in Eq.(3). I t should be pointed also that the case M2>O pre-
sents no more advantages than the case M2<O.

4. Evolution of the scalar f i e l d in the chaotic inflationary scenario.

In the chaotic i n f l a t i o n a r y scenario, i t is assumed that the i n i t i a l value of the


quasi-homogeneous scalar f i e l d ~is non-zero and, in fact, large ; t y p i c a l l y ,
I~I>M~ at t=t~ . The potential V(!~) can be a rather arbitrary function ; the only
condition is that i t should grow less faster than exp(const, l~i) for J ~ i - - ~ . Typi-
cal examples are V(~) = ~ 4 / 4 15I and even V(~)=M2 ~ 2 / 2 with M2>O (the dynamics of
the l a t t e r model was studied in 122-261). Here, the quantity H =~/a cannot be cons-
tant in general, but i f IH] ~H2~then the expansion of the universe is quasi-exponen-
t i a l . Thus, the notion of the quasi-de S i t t e r stage with the slow varying H arises.
The scalar f i e l d should also change slowly during this stage : I~I ~ H ~ . Then,
H2 : 81~ GV(~).
We can now repeat the derivation of Eqs.(8,12) (Sec.2) for this case. Because of the
dependence of H on t , the quantity -~A~a(t)= j H ( t ) d t appears to be more proper and
fundamental independent variable than the time t. Eq.(6) retains its form with the
change : Ho..~H. I t is straightforward to obtain the following equation for the lar-
ge-scale scalar f i e l d :

~A~ 3H ~ ~ ~I (55)
123

Then the corresponding Fokker-Planck equation takes the form (H2can be expressed
through V(~)) :
?)
- 4 { (5e)

I t is worthwhile to note that this equation has just the form one would expect to
follow from quantum cosmology because i t is no longer depends on such classical
quantities as t or H, but contains only fundamental variables ~ a and ~ which
remain in quantum case.
Now, the problem of the initial condition for j O ( ~ a ) arises In the studies
of classical chaotic i n f l a t i o n , i t is usually assumed that = ~Po at t=tp that
corresponds to --
--'%eC~) (. ~ C ~ - ~ ) f o r some . ~ 0 ~ . But such a condition con-
tradicts the whole s p i r i t of quantum cosmology. A natural idea is to consider sta-
tionary solutions (e.g., independent of ~ v ~ ) of Eq.(56). They can be thought of as
being in "equilibrium with space-time foam" which may arise at planckian curvatures.
At f i r s t , we introduce the notion of the probability f l u x j ( ~ j ~ ) by rewriting
Eq.(56) in the form

"a~o.
(57)
S
Then, two types of stationary solutions arise : with no f l u x and with a constant f l u x
Jo
:~9 = const. V- l e x p ( 3 / e G 2 v ) - ~
3 ~Jo(GV)-I exp(3/8 G2 V) J d ~ l exp(-3/8 G2 V(~l)).
- -
(58)

The f i r s t solution (with j = O) is just the envelope of the Hartle-Hawking time-sym-


metric wave function 1271 in the c l a s s i c a l l y permitted region (a2~ (83~GV)-1) ; the
exponent is the action for the de S i t t e r instanton with ~ = const (with the correct
sign). Moreover, we have obtained the c o e f f i c i e n t of the exponent, so the solution
appears to be normalizable. I t iseasy to v e r i f y that the average value of ~ calcula-
ted with th~ use of this solution p r a c t i c a l l y coincides_L..with ~ $ --the value of
for which IHI~H2 and the de S i t t e r stage ends ( q~s,~l~pif V(.~ = "~h~_n/n). This
does not mean that the dimension of the universe a f t e r i n f l a t i o n is small (because
all ~ are equally probable for stationary solutions) but suggests that the "use-
f u l " part of i n f l a t i o n is t y p i c a l l y very small ( i f exists at a l l ) in this case. I t
is possible to obtain the "useful" part of i n f l a t i o n that is long enough, but with
the very small probability ~ exp(-3/eG2V(~s))~ exp(-lolO).
124

I t is interesting that the second solution with j { 0 does not, in fact, contain any
exponential at a l l . For G2V(~)~I that corresponds to curvatures much less than the
planckian one, its form for Jo<O is :

In this case, the stochastic force is unimportant. Thus, we have only two possibi-
l i t i e s : e i t h e r the stationary solution contains the instanton contribution exp(-S)
(where S is the action for the instanton, S<O) or the solution is non-exponential.
We have not obtained the solution proportional to exp(S) =exp(-ISl) which was advo-
cated by several authors (including the author of this paper) some time ago 128-301.
I t seems that the l a t t e r solution describing the process of "quantum creation" of the
universe via quantum tunneling to the de Sitter stage, though possible formally, has
a very small probability also (with the same order of magnitude as above). This con-
clusion is similar to that obtained by Rubakov 1311 thoughwe suppose that his ter-
minology of "catastrophic particle creation in the process of quantum tunneling" is
inadequate ; in fact, no real particle creation takes place at the de Sitter instan-
ton solution.

5. Conclusions and discussion.

We introduced and elaborated the approach consisting in taking into account the chan-
ge in a large-scale scalar f i e l d due to the continuous flow of small-scale quantum
perturbations of the same scalar f i e l d across the de Sitter horizon during the de
Sitter (inflationary) stage. That gave us the p o s s i b i l i t y to find the e x p l i c i t expres-
sions for the average duration of the de Sitter stage (and for any higher moment i f
necessary) in the case when the i n i t i a l probability distribution of the scalar f i e l d
before the beginning of de Sitter stage was known. Certainly, the method used in the
paper (as any other mathematical method) cannot solve the problem of i n i t i a l pre-in-
flationary conditions ; new physical hypothesises (or "principles") are necessary
for this purpose.
What can be said now about the p o s s i b i l i t y of "spontaneous quantum creation of the
universe" which was so extensively discussed in 132-351 ? To make the terminology
more precise, the author proposed some time ago 1361 (see also 130i) to call the
"quantum creation of the universe" the situation when we have a solution for the wave
function of the universe with a non-zero probability flux emerging from the region
of small values of a (or, equivalently, large values of space-time curvature). This
proposal can be used in our stochastic approach also. Then the f i r s t stationary solu-
tion of Eq.(56) (the f i r s t term in Eq.(58)) corresponds to the time-symmetric univer-
se which has no beginning and was not created. This coincides with the Hawking's
interpretation of the Hartle-Hawking wave function in quantum cosmology. In the case
of our f i r s t solution, we encounter the serious d i f f i c u l t y connected (as was explai-
125

ned in Sec. 4) with the very small p r o b a b i l i t y of having the large duration of the
"useful" part of i n f l a t i o n .
The second s t a t i o n a r y solution with the non-zero p r o b a b i l i t y f l u x does correspond to
the "creation" of the universe but t h i s creation has very l i t t l e in common with the
picture that was introduced in 132-35J. In p a r t i c u l a r , no quantum tunneling takes
place, and the evolution of the metric and the scalar f i e l d remains classical up
to the planckian curvatures. This type of creation was called the "classical crea-
t i o n " in ]28J but i t should be clear that the "classical creation" is not a new con-
cept but simply the paraphrase of the standard classical picture of a s i n g u l a r i t y as
a boundary of the space-time through which the space-time cannot be continued ; the
only difference is that now t h i s boundary is assumed to have a f i n i t e thickness o
The d i f f i c u l t i e s with the second solution are connected with our i m p o s s i b i l i t y at
the present time to prove the very existence of such a solution (in other words, to
prove the p o s s i b i l i t y of the quantum change of topology) and to say something d e f i -
n i t e about the value of Jo, i f i t is non-zero. Thus, the problem of the p o s s i b i l i t y
of the quantum creation of the universe remains open.
The author would l i k e to thank Prof. Norma Sanchez for the h o s p i t a l i t y in the Groupe
d'Astrophysique R e l a t i v i s t e de l'Observatoire de Paris-Meudon where t h i s paper was
completed and the Centre National de la Recherche S c i e n t i f i q u e for f i n a n c i a l support.

References

1. A.A. Starobinsky, Phys. Lett. 91B (1980) 99.


2. A.H. Guth, Phys. Rev. D23 (1981) 347.
3. AoD. Linde, Pys. Lett. 108B (1982) 389.
4. A. Albrecht, P.J. Steinhardt, Phys. Rev. Lett. 48 (1982) 1220.
5. A.D. Linde, PismavZhETF 38(1983) 149 ; P h y s . Lett. 129B (1983) 177.
6. L.A. Kofman, A.D. Linde, A.A. Starobinsky, Phys. Lett. 157B (1985) 361.
7. V.F. Mukhanov, G.V. Chibisov, PismavZhETF 33 (1981) 549.
8. S.W. Hawking, I.G. Moss, Phys. Lett. 110B (1982) 35.
9. A.D. Linde, Phys. Lett. 116B (1982) 335.
i0. A.A. Starobinsky, Phys.Lett. II7B (1982) 175.
11. J. E l l i s , D.V. Nanopoulos, K.A. Olive, K.Tamvakis, Phys. Lett. 120B (1983) 331.
12. G.F. Mazenko, W.G. Unruh, R.M. Wald, Phys.Rev. D31 (1985) 273.
13. M. Evans, J.G. Mc Carthy, Phys. Rev. D31 (1985) 1799.
14. A. V i l e n k i n , Phys. Lett. II5B (1982) 91.
15. A. V i l e n k i n , L.H. ford, Phys. Rev. D26 (1982) 1231.
16. A. Starobinsky, in : Proc. 6th Sov.Gravit.conf., Moscow 3-5 j u l y 1984 (MGPI
Press, Moscow, 1984), vol. 2, p. 39.
17. A. Starobinsky, in : Fundamental I n t e r a c t i o n s , ed. V.N. Ponomarev (MGPI Press,
Moscow, 1984) p. 54.
126

18. S.W. Hawking, Phys. Lett. 115B (1982) 295.


19. A.H. Guth, S.-Y. Pi, Phys. Rev. Lett. 49 (1982) 1110.
20. J. Bardeen, P.I. Steinhardt, M.S. Turner, Phys. Rev. D28 (1983) 679.
21. A.A. Starobinsky, Pis'ma Astron. Zh. 9 (1983) 579 ISov.Astron. Lett. 9(1983) 302J
22. L. Parker, S.A. Fulling, Phys. Rev. D__77(1973) 2357.
23. A.A. Starobinsky. Pis'ma Astron. Zh. 4 (1978) 155 ISov.Astron. Lett 4(1978) 82J.
24. S.W. Hawking, Nucl.Phys. B239 (1984) 257.
25. D.N. Page, Class.Quantum Grav. i (1984) 417.
26. V.A. Belinsky, L.P. Grishchuk, I.M. Khalatnikov, Ya.B. Zeldovich, Phys. Lett.
155B (1985) 232.
27. J.B. Hartle, S.W. Hawking, Phys. Rev. D28 (1983) 2960.
28. Ya.B. Zeldovich, A.A. Starobinsky, Pis'ma Astron. Zh. 10 (1984) 323 ISov.Astron.
Lett. 10 (1984) 135J.
29. A.D. Linde, Zh.Eksp.Teor. Fiz. 8_7_7(1984) 369 ; L e t t . NuovoCimento 39(1984) 401.
30. A. Vilenkin, Phys. Rev. D30 (1984) 509.
31. V.A. Rubakov, Pis'ma ZhETF, 39 (1984) 89 ;Phys. Lett. 148 (1984) 280.
32. E.P. Tryon, Nature 246 (1973) 396.
33. P.I. Fomin, Dokl.Akad.Nauk Ukrain. SSR (1975) 831.
34. L.P. Grishchuk, Ya.B. Zeldovich, in : Quantum structure of space-time, eds.
M. Duff and Co Isham (Cambridge U.P., Cambridge, 1982) p. 409.
35. A. Vilenkin, Phys. Lett. 117B (1982) 25 ;Phys. Rev. D27 (1983) 2848.
36. A.A. Starobinsky, talk at the seminar in the P.K. Sternberg State Astronomical
Institute (Moscow, November 1983), unpublished.
SOME MATHEMATICAL ASPECTS OF
STOCHASTIC QUANTIZATION

G.Jona-Lasinlo
Dipartimento dl Fislca - Universit~ "La Sapienza",
GNSM and INFN - Rome

In this report I would like to briefly outline some of the ma-


thematical problems encountered in a rigorous implementation of the
program of stochastic quantization first proposed by Parisi and Wu
Yong-Shl/I/ and then developed in its formal aspects by several au-
thors/2/.

We recall that the basic idea of stochastic quantlzatlon con-


sists in considering the Euclidean measure associated to a quantum
mechanical system with finite or infinite degrees of freedom, as the
stationary state of some stochastic process.

The standard proposal for the construction of such a process is


to solve the following stochastic differential equation

F_c(]) "a
(i)

/
where S(~) is the Euclidean action describing the system and W(x,t)
in the Wiener process characterized by the covarlance

E (W ( x , t ) W (x l j t')) = min (t,t') ~x - x') (1.2)


128

The typical form of the functional S(~) is


I

where V(~) is a local polynomial in of even degree and V is the


dimension of the space. Introducing (1.3) in (i.I) we obtain

(I .4)

is the L a p l a c i a n .

As it is eq.(l.4) is only formal because the Wiener process is


non dlfferentiable. This is a well known difficulty already in the
case of ordinary stochastic differential equations where it is over-
come by integrating with respect to the time and transforming the
equation into an i n t e g r a l equation/3/.

The natural thing to do in the case of (1.4) is to obtain an


integral equation by using the Green function of the linear part,
that is

~0

where G satisfies

~& ~ G + G = ~(t'-t) ~(_x*- x) (1.6)


qt

and Z is the G a u s s l a n process

z(t,_~) =~ dV~/jOat' a(t,t',~,~') ~W(t',x')


~ t'
+ ~o(t,x ) (1.7)
129

i
w h e r e ~= is a solution of the linear homogeneous part of (1.4). The
next step depends crucially on the dimensionality ~ .

= I. - In this case (1.5) is a meaningful equation. This depends


on the circumstance that the typical trajectories of the process
Z(t,x) are continuous functions in both variables. (1.5) can then be
solved for each continuous realization of the input Z(t,x). A rather
complete treatment of this case can be found in ref./4/. The ~ =I
theory covers the stochastic quantization of systems with a finite
number of degrees of freedom, that is the case of Quantum Mechanics.
It is interesting to note however that its relevance, goes beyond
quantum mechanical applications. In recent years in fact the theory

developed in /4/ has been useful in entirely different domains/5/-


/6/

= 2. - At ~ =2 the typical difficulties of Quantum Field Theory


appear. The free field Z(t,x) is not anymore a continuous function
but a distribution. If one evaluates for example the expectation
value Ew (Z2(t,~)) this diverges logarithmically. Ew means expecta-
tion with respect to the W i e n e r process (1.2).
Eq.(l.5) has therefore to be m o d i f i e d by introducing counterterms. In
this way however, since in the end the counterterms become infinite,
the equation itself does not have a mathematical meaning. The w a y out
to this problem taken in the physical literature consists in d e v e l o p -
ing the solution in p e r t u r b a t i o n theory and then adjusting the renor-
malization terms in such a way that expectations of the form

E w (~(t I, X l ) ~ (t 2, x 2 ) - - ~ (t n, xl~))

make sense. For a rigorous mathematical treatment this is not suffi-


cient and one has to r e s o r t to a n o n perturbative approach. Therefore
the problem facing us consists in defining a "solution" of (1.5) in
spite of the fact that the equation does not have a meaning. To solve
this problem we appeal to w h a t prohabilists call a weak solution of a
stochastic differential equation. The basic idea is as follows. Let
us first regularize (1.5) by introducing the counter terms. In the
present situation this means taking the W i c k product of the nonlinear
term V'(~) and then introducing a cut-off ~ in evaluating it. That
l

is the nonlinear term will be: V' ( ~ ) : where ~ is the cut-off


/ -

field. The Wick product can be taken with respect to the covarianee
130

of the free field C (x,y) = ( - ~ + I ) -I. If for example V(~) = ~--


: V'(/): = ~ (/3 -3 C (x,x)~ ). With these modifications (1.5)
becomes

= - C *: V ( ): + Z (t.8)

and as long as ~ this is a m e a n i n g f u l equation which can be solved


for each imput Z. The process Z will now induce a measure ~ on the
solutions of (1.8). If now }.L~ converges to a limiting measure
(in the sense of weak convergence of measures) as~->~ we shall say
that (1.8) has a weak solution when the cut-off in removed. Our goal
therefore consists in implementing non perturbatlvely this idea. In
stochastic calculus there is a well known formula, the Girsanov-Came-
ron-Martin formula, which provides the Radon-Nikodym derivative of
the m e a s u r e / ~ w i t h respect to/~ . This is/3/

= exp I- ~oT(: V / ( Z ~ ) : , d W ) - } ~Tdt' I' : V/(Z.): I/ 2} (1"9)

where

(: V/(Z~):, dW) = S d2x: V/(Z/~(t'x)): dW (t, x)

is a scalar product in the space variables and [[ "[I i s the norm


induced by it. We work in the finite volume A The stochastic inte-
gral appearing in (1.9) is a Ito integral, that is it must be consi-
dered as a limit of sums

SL~.li= (:v/(z~(ti)):, w(ti+ 1) - w (tf)) (1.1o)


131

where the ti represent a partition of [ o , T] . Notice that with this


definition the increment is uncorrelated with the integrand due to
the Markov property of the Wiener process. This integral, very natu-
ral probabilistically, does not obey as it is well known to the usual
rules of d i f f e r e n t i a l calculus/3/.

Our basic problem now consists in showing that when ~ ---~ ~


(1.9) is a good stochastic variable and in p a r t i c u l a r

EZo (d d - - ~ ) = 1 (i.li)

that is, the m e a s u r e ~ i s normalized.

Z o means that the expectation is taken with Z (o, _~x) = Zo(X).

The problem now reminds of constructive field theory, only a


rather special lagranglan is involved, To make the connection even
more explicit we notice that by the rules of the Ito calculus it
follows

f(:V/(Z ):, dW) = ~ I : Z4 (T,~): d2x - ~ ~ :Z4(o,~): d2x +

z% A
(1.12)

+ 0Tdt (Z 3

Using (I.12), the exponential in (1.9) takes a less exotic form.

At this point everything seems to be ready to apply the methods


of constructive field theory, in particular the rather straight
forward methods by which P (~ 72 was constructed/7/. One realizes
immediately however that our problem is more difficult. In fact
neither term in the exponent of (1.97 is a well defined stochastic
variable. For example the expectation of the second term diverges.
The remarkable thing is that this type of d i v e r g e n c e does not show up
in p e r t u r b a t i o n theory! The divergence we just mentioned is c a n c e l l e d
by the square of the first term in the expansion of the exponential
and a similar mechanism operates with higher order contributions. The
reason is the special structure of (1.9) which is such as to insure
132

in any case the normalization condition E ' ~ _d/~


_ ) = I. This reminds
of the cancellation mechanisms in supersymme~r~c theories(*). These
divergences however constitute a difficulty in a non perturbative
approach.

At this stage there are two possibilities. If we insist on the


specific form of e q . ( l . 8 ) as the basis for stochastic quantization we
must conclude that the m e t h o d s devised for P ( ~ ) 2 are not sufficient
to treat its stochastic counterpart. We must in such a case look for
more powerful methods like for example the phase space cell expansion
or more generally the renormalization group methods which led to the
construction of ~$ . In fact the above difficulties of stochastic
P ( / )2 seem of a s i m i l a r nature as those encountered in 3 "

The other possibility consists in modifying eq.(l.8) in such a


way that the usual P ( ~ )2 Euclidean theory still represents its
equilibrium state. This was the w a y followed recently by J o n a - L a s i n i o
and Mitter/9/. There is in fact a whole family of stochastic diffe-
rential equations which admit the same equilibrium measure. The one
considered In/9/ is

d ~(t,x) = dW (t,x) - ~ (C-$~ (t,x) + C I-/ :V/(/(t,x))) dt (1.13)

with O ~ I and

E (W(t,x) W (t', ')) = C I- & (~,E') min (t,t') (1.14)

In/9/ it was shown that for ~ L T ~ the methods used for P ( j )2 are
sufficient to prove the existence of an ergodie weak solution of
(1.13). The previous equation (1.8) corresponds to ~ = i. The
approach to e q u i l i b r i u m is slower for (1.13).

We conclude with some comments. It would eertainly be worth to


push the analysis to treat the ease ~ = I i.e. eq.(l.8) with the m o r e
powerful methods mentioned before.

(*) The c o n n e c t i o n b e t w e e n s t o c h a s t i c c a l c u l u s and supersymmetry


has been c o n s i d e r e d by m a n y a u t h o r s / 8 / -
133

In/9/ only the ultraviolet problem was studied. It would be


interesting to take the limit A--> 6~ . In this connection we remark
that the formalism of the cluster expansion applies also to the study
of (1.13).

In conclusion I would like to express my gratitude to P.K.


Mitter. My understanding of the subject discussed here owes much to
our pleasant and fruitful collaboration.

References

i) G.Parisi, Wu Yong-Shi, Sci. Sin. 24, 483 (1981).

2) For a review see for example B.Saklta, 7th Johns Hopkins Work-
shop, ed. G.Domokos, S . K o v e s i - D o m o k o s (World Scientific, Singa-
pore 1983).

3) See e.g. I.I. Gihman, A.V.Skorohod, "Stochastic Differential


Equations", Springer 1972.

4) W.Faris, G.Jona-Laslnio, J.Phys.A, 15, 3025 (1982).

5) R.Benzi, A.Sutera, J.Phys.A, 18, 2239 (1985).

6) M.Cassandro, E.Olivieri, P.Picco, Ann.Inst. H.Poinear~, in Press.

7) E.Nelson, in "Constructive Quantum Field Theory" Lecture Notes in


Phys. Vol. 25, Springer 1973;
B.Simon, "The P ( / ) 2 Euclidean (Quantum) Field Theory" P r i n c e t o n
NJ, Princeton University Press 1974;
J.Gllmm, A.Jaffe, "Quantum Physics" Springer 1981.

8) S.Cecotti, L.Girardello, Phys.Lett. IIOB, 39 (1982);


G.Parlsi, N.Sourlas, Nuel.Phys. 206B, 321 (1982);
E.Gozzi, Phys.Lett. 129Bn 432 ( i - ~ ) ;
V.de Alfaro, S.Fubinl, G.Furlan, G.Veneziano, Phys.Lett. 142B,
399 (1984).

9) G.Jona-Lasinio, P.K.Mitter, Comm.Math.Phys. I01, 409 (1985).


S U P E R S T R I N G S A N D T H E UNIFICATION O F F O R C E S A N D P A R T I C L E S

Michael B. G r e e n ,
P h y s i c s D e p a r t m e n t , Queen Mary College, U n i v e r s i t y o f L o n d o n , U.K.

The q u e s t i o n of how to r e c o n c i l e t h e c l a s s i c a l d e s c r i p t i o n o f t h e g r a v i t a t i o n a l
force embodied in Einstein's general theory of relativity with the principles of
q u a n t u m t h e o r y is a c e n t r a l i s s u e in t h e o r e t i c a l p h y s i c s . A simple a p p l i c a t i o n of t h e
u n c e r t a i n t y p r i n c i p l e s h o w s t h a t a t a d i s t a n c e , Ax, a r o u n d t h e P l a n c k scale, i.e.

Ax ~ /Gh/c3 ~ 10 -35 m e t e r s (i)

(where G is the gravitational constant) q u a n t u m f l u c t u a t i o n s become so l a r g e t h a t


space-time must be considered to contain a sea of virtual black holes. Since
perturbative calculations in quantum gravity assume that the curvature of
space-time is small on all length scales they are invalid and lead to
non-renormalizable infinities. Non-perturbative methods have not led to calculable
consequences.
I t a p p e a r s likely t h a t s u p e r s t r i n g theories unite gravity and quantum mechanics
in a c o n s i s t e n t m a n n e r . This is a c h e i v e d b y a m o d i f i c a t i o n of g e n e r a l r e l a t i v i t y a t
s h o r t d i s t a n c e s so t h a t E i n s t e i n ' s t h e o r y e m e r g e s a s a l o n g d i s t a n c e a p p r o x i m a t i o n .
Furthermore, the quantum consistency of superstring theories provides very
s t r i n g e n t r e s t r i c t i o n s o n t h e p o s s i b l e u n i f y i n g Yang-Mills g a u g e g r o u p s . As a r e s u l t
g r a v i t y is u n i f i e d w i t h t h e o t h e r f o r c e s a n d p a r t i c l e s in a n almost u n i q u e m a n n e r .
The o n l y p o s s i b l e u n i f y i n g g r o u p s a r e

S0(32) o r E8 x E8 (2)

[$0(32) is a l a r g e o r t h o g o n a l g r o u p while E 8 is t h e l a r g e s t e x c e p t i o n a l Lie g r o u p . ]


The d i m e n s i o n a l i t y of s p a c e - t i m e is a l s o r e q u i r e d to t a k e a special (or "critical")
value

D = 10 (3)

in o r d e r to o b t a i n a c o n s i s t e n t s u p e r s t r i n g quantum theory. Clearly, in o r d e r to


have any chance of describing the observed physics of our (approximately)
four-dimensional world, six dimensions must turn out to be curled-up (or
135

" c o m p a c t i f i e d " ) t o a v e r y small size.


The i d e a o f h i g h e r dimensions arose in m o d e r n p h y s i c s in t h e proposal by
Kaluza a n d Klein 1 in t h e 1920's to u n i f y e l e c t r o m a g n e t i s m w i t h g r a v i t y b y a s s u m i n g
t h e e x i s t e n c e of a f i f t h d i m e n s i o n w h i c h f o r m s a v e r y small c i r c l e . This idea has
b e e n r e v i v e d in the c o n t e x t of s u p e r g r a v i t y t h e o r i e s w h i c h h a v e t r i e d to u n i f y all
the interactions in this manner. [A theory in eleven dimensions has been
particularly popular.] In t h e s e Kaluza-Klein schemes the gauge symmetries of the
effectively four-dimensional theory arise from t h e s y m m e t r i e s of t h e compactified
space. In this respect superstring theories are very different. A l r e a d y in t h e
ten-dimensional theory t h e r e is more g a u g e symmetry than anyone could wish for
since the possible gauge groups (in (2)) a r e so l a r g e . The e o m p a c t i f i c a t i o n o f t h e
e x t r a d i m e n s i o n s is h e r e e x p e c t e d to r e d u c e t h e g a u g e s y m m e t r y d o w n to a s m a l l e r
symmetry group. T h i s s h o u l d lead to s o m e t h i n g like a " G r a n d U n i f i e d " s y m m e t r y in
the effective four-dimensional theory at high energies. Furthermorey to explain
observed accelerator physics, this symmetry must in turn break down to the
standard model w i t h s y m m e t r y g r o u p s SU(3) (for c o l o u r ) a n d SU(2) x U(1) ( f o r t h e
electro-weak forces).
A l t h o u g h a c o m p l e t e l y r e a l i s t i c w a y in which this might happen is n o t yet
understood it is a l r e a d y clear that superstring theories have a good chance of
making contact with observed physics. The p r o g r a m m e is v e r y a m b i t i o u s s i n c e t h e s e
theories contain no f r e e input parameters (although the space of solutions may
logically have free parameters). The t e c h n i q u e s required for analysing both the
phenomenological predictions and the theoretical structure of t h e s e t h e o r i e s i n v o l v e
t h e u s e o f m a n y i d e a s in m o d e r n m a t h e m a t i c s t h a t h a v e n o t b e e n u s e d b y p a r t i c l e
p h y s i c i s t s u n t i l now. Conversely, many a s p e c t s of s u p e r s t r i n g theory raise issues
of i n t e r e s t in p u r e mathematics.

CHIRALITY

A key constraint on any theory is t h a t it m u s t give rise to t h e o b s e r v e d


chirality (i.e. parity violation) of the four-dimensional world due to the weak
interactions. O v e r t h e l a s t f e w y e a r s t h e s t u d y o f t h e Kaluza-Klein m e c h a n i s m h a s
indicated that chiral physics can p r o b a b l y only emerge from a h i g h e r - d i m e n s i o n a l
t h e o r y if two c o n d i t i o n s a r e s a t i s f i e d 2 :
(a) t h e higher-dimensional theory is c h i r a l (which excludes odd-dimensional
t h e o r i e s , s i n c e c h i r a l i t y o n l y e x i s t s in e v e n d i m e n s i o n s ) a n d
(b) t h e r e is a g a u g e group~ G, i n t h e h i g h e r - d i m e n s i o n a l t h e o r y . This seems to
be necessary to avoid losing the chirality property in the process of
compactification. The gauge fields can twist up into a topologically non-trivial
configuration (such as a magnetic monopole) in the internal space - this then
distinguishes the different four-dimensional chiralities.
136

CHIRAL ANOMALIES

A n y c h i r a l t h e o r y is l i k e l y to b e p l a g u e d b y i n c o n s i s t e n c i e s k n o w n a s c h i r a l
gauge "anomalies". These represent the breakdown in the quantum theory of
s a c r o s a n c t c o n s e r v a t i o n laws t h a t w e r e built into the classical theory. Anomalies
may in general arise in the conservation of Yang-Mills currents and in the
c o n s e r v a t i o n of g r a v i t a t i o n a l c u r r e n t s i.e. t h e e n e r g y - m o m e n t u m t e n s o r (as well a s in
the supersymmetry current). I n f o u r d i m e n s i o n s a t h e o r y w i t h Weyl f e r m i o n s also
contains a n t i - f e r m i o n s of the opposite chirality. Only if the fermions lie in a
complex r e p r e s e n t a t i o n of a gauge group (so that the anti-fermions lie in the
complex c o n j u g a t e r e p r e s e n t a t i o n ) is t h e t h e o r y c h i r a l . I n t h a t c a s e t h e r e may b e
Yang-Mills a n o m a l i e s b u t no g r a v i t a t i o n a l a n o m a l i e s s i n c e g r a v i t y is i n s e n s i t i v e to
the gauge group quantum numbers. H o w e v e r , in t e n d i m e n s i o n s ( a n d g e n e r a l l y in
4n+2 d i m e n s i o n s ) a f e r m i o n a n d i t s a n t i - p a r t i c l e h a v e t h e same c h i r a l i t y a n d there
can be both Yang-Mills anomalies (with fermions in any representation) and
g r a v i t a t i o n a l anomalies.
The e x i s t e n c e of anomalies r e n d e r s a t h e o r y i n c o n s i s t e n t b e c a u s e t h e y lead to a
violation o f u n i t a r i t y d u e to t h e c o u p l i n g of u n p h y s i c a l l o n g i t u d i n a l m o d e s of g a u g e
p a r t i c l e s to t h e p h y s i c a l t r a n s v e r s e modes. Up to l a s t s u m m e r it h a d b e e n t h o u g h t
that there were no anomaly-free chiral theories with gauge groups in ten
dimensions. I t was t h e n d i s c o v e r e d 3 t h a t a n o m a l i e s may be a b s e n t f r o m t h e o r i e s
with the gauge groups mentioned earlier.
Superstring t h e o r i e s with t h e s e g a u g e g r o u p s a r e b o t h f r e e f r o m a n o m a l i e s a s
well a s t h e i n f i n i t i e s t h a t p l a g u e q u a n t u m t h e o r i e s of g r a v i t y (as f a r a s h a s b e e n
checked). T h e s e s u c c e s s e s a r e u n p r e c e d e n t e d in a n y q u a n t u m t h e o r y o f g r a v i t y .

WHAT ARE SUPERSTRINGS?

In contrast to usual relativistic field theories, in which the fundamental


constituents are structurless point particles~ t h e constituents of a n y string field
theory have extension in one dimension. This leads to significant differences
between string field theory and conventional "point" field theories such as
Yang-Mills o r g e n e r a l r e l a t i v i t y . A s i n g l e c l a s s i c a l r e l a t i v i s t i c s t r i n g c a n v i b r a t e in
a n i n f i n i t e s e t of normal m o d e s with u n l i m i t e d f r e q u e n c i e s . The s e p a r a t i o n b e t w e e n
t h e f r e q u e n c i e s of t h e s e m o d e s is d e t e r m i n e d b y t h e r e s t t e n s i o n o f t h e s t r i n g , T.
The modes can be quantized so that the quantum mechanics of a single string
d e s c r i b e s a n i n f i n i t e s e t of s t a t e s w i t h m a s s e s w h i c h i n c r e a s e w i t h o u t b o u n d , t h e i r
separation given by

A ( m a s s ) 2 = 2~T (4)
137

These states also have spins which increase without bound since they lie on
straight-line Regge trajectories (with s l o p e c~' - 1/2zyT). This is n o t a n a c c i d e n t -
s t r i n g t h e o r y o r i g i n a t e d in t h e l a t e 1960's w i t h t h e d u a l r e s o n a n c e model 4 w h i c h was
developed to explain h a d r o n i c phenomena. The e a r l i e s t s t r i n g theory (the bosonic
string t h e o r y 5) h a d a c r i t i c a l d i m e n s i o n D - - 2 6 while t h e spinning6 string theory
w h i c h also i n c o r p o r a t e d f e r m i o n s had D - 10. I t was n o t i c e d t h a t t h e s p e c t r u m of
t h e s p i n n i n g s t r i n g t h e o r y c o u l d be t r u n c a t e d to g i v e a s u p e r s y m m e t r i c s p e c t r u m 7
i.e. a t every mass level t h e r e are an equal n u m b e r of b o s o n a n d f e r m i o n s t a t e s .
This gave rise to the explicit construction of theories with space-time
supersymmetry over the last five y e a r s 8. I shall refer to these theories as
superstring theories. The g r o u n d s t a t e s of s u p e r s t r i n g theories are m a s s l e s s (in
contrast to the earlier string theories which were plagued by having tachyonic
ground states i.e. states with negative (mass)2). These massless states form
supersymmetry multiplets corresponding to the familiar massless states in
tan-dimensional super-Yang-Mills and supergravity.
The m a s s s c a l e s e t b y t h e s t r i n g t e n s i o n is s u p p o s e d t o b e t h e P l a n c k s c a l e (in
ten dimensions). T h i s m e a n s t h a t , f o r m a n y p u r p o s e s , w h e n c o n s i d e r i n g momentum
scales much less than the Planck scale the higher mass states are effectively
infinitely massive and t h e y decouple leaving an effective " l o w - e n e r g y " t h e o r y of the
massless ground states. This is just a conventional point field theory such as
supergravity and super-Yang-mills. The f u n d a m e n t a l p a r t i c l e s o b s e r v e d in n a t u r e
(the quarks, leptons, gauge particles,....) should occur among the massless g r o u n d
s t a t e s s i n c e t h e i r m a s s e s a r e n e g l i g i b l e c o m p a r e d to t h e P l a n c k mass. However, t h e
fact that the low e n e r g y theory has arisen from an almost unique superstring
theory suggests t h a t t h e p a r a m e t e r s m e a s u r e d in e x p e r i m e n t s ( s u c h a s t h e m a s s e s
and coupling strengths) s h o u l d b e d e t e r m i n e d w i t h little a m b i g u i t y f r o m t h e t h e o r y .
At m o m e n t u m s c a l e s a r o u n d t h e P l a n c k s c a l e t h e m a s s i v e s t a t e s of t h e s t r i n g c a n
b e e x c i t e d so t h a t superstring theory then d i f f e r s r a d i c a l l y from a n y p o i n t field
theory. T h i s s c a l e is j u s t where the problems with quantum gravity arise. I t is
because they differ from E i n s t e i n ' s t h e o r y (or a n y supergravity field theory) at
these scales that certain superstring theories avoid quantum inconsistencies. In a
space-time picture the strings h a v e a n a v e r a g e s i z e of t h e P l a n c k l e n g t h so t h e y
appear as points when looked at c o a r s e l y b u t their n o n - z e r o e x t e n s i o n is c r u c i a l
w h e n c a l c u l a t i n g q u a n t u m f l u c t u a t i o n s a t small s c a l e s .

SUPERSTRING DYNAMICS

I will g i v e a v e r y s k e t c h y o u t l i n e o f t h e w a y in w h i c h t h e d y n a m i c s of a f r e e
superstring is f o r m u l a t e d .
As a string moves through space-time it sweeps out a (two-dimensional)
world-sheet just as a point particle traces out a world-line. The space-time
138

c o o r d i n a t e o f a n y p o i n t on t h e s t r i n g a t a g i v e n time, X~(O,r), is a f u n c t i o n o f t h e
two p a r a m e t e r s o f t h e w o r l d - s h e e t , a a n d T, a n d /J ( = 0,1,...9) is a s p a c e - t i m e v e c t o r
index. In superstring theories there are additionally one or two anticommuting
coordinates oa(a,T) which are Weyl spinets (which have 16 components in ten
d i m e n s i o n s ) l a b e l l e d b y t h e i n d e x a = 1,2,...16. These spinor coordinates embody the
s u p e r s y m m e t r y of t h e t h e o r y (X~ a n d e a a r e s u p e r s p a c e c o o r d i n a t e s ) .
The classical dynamics of a relativistic string is obtained from an action
principle t h a t g e n e r a l i z e s t h a t of a relativistic point particle. J u s t as the action for
a r e l a t i v i s t i c p o i n t p a r t i c l e is t h e l e n g t h of i t s w o r l d - l i n e , t h e a c t i o n f o r a s t r i n g is
taken to b e p r o p o r t i o n a l to t h e a r e a of t h e w o r l d - s h e e t 9. This is a g e o m e t r i c a l
quantity which does not depend on the way in which the world-sheet is
parametrized. In t h e case of the s u p e r s t r i n g theories the notion of the a r e a is
g e n e r a l i z e d so t h a t , r o u g h l y s p e a k i n g , t h e a c t i o n i s p r o p o r t i o n a l t o t h e a r e a o f t h e
world-sheet in superspace. The fact that the action, S, is i n d e p e n d e n t of the
p a r a m e t r i z a t i o n o f t h e t w o - d i m e n s i o n a l w o r l d - s h e e t m a k e s i t like a t h e o r y o f g r a v i t y
in t h e t w o - d i m e n s i o n a l a - r s p a c e 10

S = I 4m dT ~ n ~ g~ 8~X~ gBXu + e terms (5)

w h e r e gcq~ is a t w o - d i m e n s i o n a l m e t r i c (c%~--O,T) a n d g is i t s d e t e r m i n a n t . This


m e t r i c is a n o n - d y n a m i c a l a u x i l i a r y f i e l d in two d i m e n s i o n s w h i c h c a n b e e l i m i n a t e d
by substituting t h e s o l u t i o n o f i t s e q u a t i o n o f motion b a c k i n t o t h e a c t i o n . The e
t e r m s in (5) a r e d e s i g n e d to e n s u r e the s u p e r s y m m e t r y of the action. The a b o v e
a c t i o n d e s c r i b e s a s t r i n g m o v i n g in flat t e n - d i m e n s i o n a l Minkowski s p a c e w h e r e r ~
is the flat space-time metric. Much work of recent months has concerned
g e n e r a l i z a t i o n s to b a c k g r o u n d s p a c e s w i t h six c o m p a c t i f i e d d i m e n s i o n s . Requiring
the compactified string theory to be consistent puts severe restrictions on the
p o s s i b l e b a c k g r o u n d s p a c e - t i m e s a s I will d e s c r i b e l a t e r .
An i m p o r t a n t f e a t u r e of t h e a c t i o n , in a d d i t i o n to t h e m a n i f e s t r e p a r a m e t r i z a t i o n
i n v a r i a n c e , is i n v a r i a n c e u n d e r rescalings g~ -~ AgxB w h e r e A(a,r) is a n a r b i t r a r y
function. T h e s e s y m m e t r i e s allow t h e c h o i c e of a c l a s s of g a u g e s in w h i c h g0~ = 1
(the conformal gauges} and in which the theory is invariant under (pseudo}
eonformal transformations. This conformal invarianee plays a crucial role in the
c o n s i s t e n c y o f t h e q u a n t u m m e c h a n i c s o f a s i n g l e f r e e s t r i n g in e n s u r i n g t h a t t h e
s t a t e s c r e a t e d b y t h e t i m e - l i k e o c i l l a t i o n s of t h e s t r i n g d e c o u p l e f r o m t h e p h y s i c a l
space of s t a t e s . This is i m p o r t a n t s i n c e t h e t i m e - l i k e m o d e s h a v e n e g a t i v e n o r m a n d
are therefore ghost states. The c h o i c e of s u c h a gauge i s o n l y p o s s i b l e in t h e
q u a n t u m t h e o r y in t h e c r i t i c a l d i m e n s i o n w h i c h is t e n f o r s u p e r s t r i n g theories. It
is conceivable that superstring theories could be obtained in lower dimensions
(D = 3, Lt o r 6) b y t e c h n i q u e s a d v o c a t e d b y P o l y a k o v 11 b u t t h a t i s s u e is s o m e w h a t
murky at p r e s e n t . Only in t e n dimensions are the physical modes of the string
purely transverse just as gauge fields are transversely polarized in Yang-Mills
139

theories.
The solutions of the classical equations, derived from the action, can be
e x p a n d e d in a n i n f i n i t e s e t of normal m o d e s w h i c h c a n then be q u a n t i z e d . The
s p e c t r u m d e p e n d s on the b o u n d a r y conditions.
A string with free e n d p o i n t s can c a r r y internal quantum numbers associated
w i t h a c l a s s i c a l g r o u p 12 (SO(n), U(n) o r U S p ( n ) ) . The c h a r g e s a r e a t t a c h e d t o t h e
e n d s o f t h e s t r i n g ( r a t h e r like t h e old p i c t u r e of a m e s o n a s a s t r i n g w i t h a q u a r k
at one end and an a n t i - q u a r k at the other). It turns out that a string with free
e n d p o i n t s has a massless v e c t o r particle among its massless states. This a p p a r e n t l y
a c c i d e n t a l f e a t u r e is t h e r e a s o n w h y s t r i n g t h e o r i e s r e d u c e to Yang-Mills t h e o r i e s in
t h e l o w - e n e r g y limit 13 ( w h e n all t h e m a s s i v e s t a t e s e f f e c t i v e l y d e c o u p l e ) .
A closed string contains a massless spin-2 particle which is the graviton
associated with the fact that the low energy effective theory contains general
r e l a t i v i t y 14.
When t h e interactions between strings are included there is a remarkable
u n i f i c a t i o n b e t w e e n g r a v i t y a n d Yang-Mills. In the t h e o r i e s c o n t a i n i n g o p e n s t r i n g s
( k n o w n a s t y p e I t h e o r i e s ) two o p e n s t r i n g s interact by joining at their endpoints
to f o r m a s i n g l e o p e n s t r i n g o r a n o p e n s t r i n g s p l i t s i n t o two s t r i n g s .

Fig.(1)
T h i s is a local i n t e r a c t i o n a n d consistency requires t h e same i n t e r a c t i o n to c o u p l e
t h e two e n d s o f a s i n g l e o p e n s t r i n g to f o r m a c l o s e d s t r i n g a s i l l u s t r a t e d b y

Fig.(2)
so t h a t t h e e x i s t e n c e o f o p e n s t r i n g s ( a n d h e n c e t h e Yang-Mills s e c t o r ) r e q u i r e s t h e
existence of closed strings (and hence the gravity sector). The gravitational
140

c o n s t a n t j K, a n d t h e Yang-Mills c o u p l i n g , g, a r e r e l a t e d b y K ~ g2T.
T h e r e a r e also t h e o r i e s w i t h o n l y c l o s e d s t r i n g s . For example, t y p e II t h e o r i e s
describe closed strings which have an orientation i.e. they have excitations
c o r r e s p o n d i n g to waves r u n n i n g a r o u n d the s t r i n g i n d e p e n d e n t l y in e i t h e r direction
( t h e II r e f e r s to t h e f a c t t h a t t h e s e t h e o r i e s h a v e twice a s much s u p e r s y m m e t r y ) .
T h e s e t h e o r i e s may h a v e no n e t c h i r a l i t y ( t y p e IIa) o r may b e c h i r a l ( t y p e IIb).
The l a t t e r t h e o r y is s t r i k i n g s i n c e i t s low e n e r g y limit y i e l d s a p o i n t f i e l d t h e o r y 15
w h i c h is f r e e from all g r a v i t a t i o n a l a n o m a l i e s 16. H o w e v e r , t y p e II t h e o r i e s do n o t
have an internal symmetry group and so do n o t r e d u c e in a n y o b v i o u s way to a
chiral four-dimensional theory.
The m o s t i n t e r e s t i n g k i n d of s u p e r s t r i n g t h e o r y is t h e h e t e r o t i c s u p e r s t r i n g 17,
This d e s c r i b e s c l o s e d s t r i n g s w h i c h c a r r y i n t e r n a l s y m m e t r y ( u n l i k e t h e o t h e r c l o s e d
superstring theories} with charges which are smeared out as densities along the
string. These theories are built from m o d e s of t h e ten-dimensional superstring
running around the string in o n e s e n s e w i t h m o d e s o f t h e 2 6 - d i m e n s i o n a l b o s o n i c
s t r i n g t h e o r y r u n n i n g a r o u n d in t h e o t h e r s e n s e : This a p p a r e n t l y bizarre mixture
o f d i m e n s i o n a l i t i e s is reconciled by the i d e n t i t y 18 26 = 10+16 w h e r e t h e f i r s t ten
d i m e n s i o n s o f t h e r i g h t p o l a r i z e d m o d e s a r e t a k e n to be t h e s p a c e - t i m e d i m e n s i o n s .
The other sixteen dimensions become i n t e r n a l coordinates forming a hypertorus
associated with a sixteen dimensional lattice. The c o n s i s t e n c y of t h e t h e o r y r e q u i r e s
this lattice to be even and self-dual. There are known to be only two such
l a t t i c e s 19 w h i c h a r e r e l a t e d to t h e r o o t l a t t i c e s o f t h e g r o u p s E 8 x E 8 a n d SO(32)
(or, more accurately, the group (Spin 32)/Z 2 which has the same algebra as SO(32)).
Therefore the heterotic string theory is only consistent for the two groups that
were already k n o w n to be selected by requiring the absence of anomalies. In the
heterotic string theory K g/CT.

SUPERSTRING INTERACTIONS

Superstring s c a t t e r i n g a m p l i t u d e s c a n b e c a l c u l a t e d in p e r t u r b a t i o n theory by
c o n s t r u c t i n g a s e r i e s o f d i a g r a m s t h a t g e n e r a l i z e t h e F e y n m a n d i a g r a m s o f familiar
point field theories.

Tree Diagrams

The t r e e a p p r o x i m a t i o n to t h e s c a t t e r i n g a m p l i t u d e of, f o r example, f o u r c l o s e d


s t r i n g s is r e p r e s e n t e d b y a c o n t i n u o u s w o r l d - s h e e t t h a t j o i n s t h e two i n c o m i n g a n d
two o u t g o i n g s t r i n g s .
141

Fig.(3)
This diagram describes two i n c o m i n g c l o s e d s t r i n g s which join together by touching
at a point to form one intermediate closed string which subsequently splits into the
two f i n a l s t r i n g s (time is t a k e n to b e i n c r e a s i n g f r o m l e f t to r i g h t ) . I t is p o s s i b l e
t o d e r i v e t h e a m p l i t u d e f o r d i a g r a m s l i k e fig.(3) e i t h e r b y a s t r i n g g e n e r a l i z a t i o n of
Feynman's path integral approach to q u a n t u m m e c h a n i c s o r f r o m a s e c o n d - q u a n t i z e d
formalism expressed in terms of s t r i n g fields which create and destroy complete
strings. Unfortunately, f o r t h e m o m e n t t h e o n l y c o m p l e t e f o r m u l a t i o n of t h e f i e l d
theory of s t r i n g s 20 is i n a s p e c i a l g a u g e 21 ( t h e light-cone gauge) which is not
satisfactory for understanding the geometric structure of t h e theory but suffices
for perturbative calculations. A string field, ~[X(a),e(a)], is a f u n c t i o n a l of t h e
string configuration. A particularly important aspect of c l o s e d string theories is
that the interactions only involve terms which are cubic in closed-string f i e l d s (as
can be seen by slicing through the world-sheet of fig.(3) a t t h e p l a c e w h e r e o n e of
the interactions takes place) corresponding to t h e local j o i n i n g o r s p l i t t i n g of t h e
strings

Fig.(4)
( w h i c h is t h e c l o s e d - s t r i n g a n a l o g u e of t h e o p e n - s t r i n g i n t e r a c t i o n of fig.(1). There
are no higher order contact interactions whereas in the perturbative treatment of
gravity based on the Einstein-Hilbert action there are an infinite number of
interaction terms involving contact interactions between arbitrary numbers of
gray/tons. All t h e s e contact terms emerge as low e n e r g y effective interactions
arising from the exchange of the massive string states between cubic string
vertices. T h i s is a n a l o g o u s t o t h e w a y i n w h i c h t h e f o u r - F e r m i model of t h e w e a k
i n t e r a c t i o n s i s now k n o w n to e m e r g e a s a n e f f e c t i v e t h e o r y f r o m t h e W e i n b e r g - S a l a m
t h e o r y a t e n e r g i e s m u c h l e s s t h a n t h e W o r Z b o s o n mass.
142

Fig.(3) g e n e r a l i z e s t h e f o u r - g r a v i t o n t r e e d i a g r a m c o n t r i b u t i o n to t h e s c a t t e r i n g
amplitude of E i n s t e i n ' s theory (since the graviton is o n e of t h e massless string
states). T a k i n g t h e e x t e r n a l s t a t e s to b e g r a v i t o n s t h i s a m p l i t u d e is g i v e n b y

T(s,t,u) =rPintlresultfield t h e o r y } x r'(1-8~T)r'(1-8~T)F(1-8~ T) (6)

(where s,t,u are the Mandelstam invariants defined by s = ( p l + P 2 ) 2, t = ( p l + P 4 ) 2,


and u = (pl+P3) 2 where Pl,P2,P3,P4, are the external momenta). In this expression
all t h e string features are contained in the F functions. I n t h e low e n e r g y limit
s,t,u ~ T the expression manifestly reduces to t h e f a m i l i a r r e s u l t based on general
relativity.

Loop D i a g r a m s

Higher order diagrams in the perturbation e x p a n s i o n a r e g i v e n b y a s e r i e s of


loop diagrams analogous to those of point field theories. For example, the
world-surface of t h e o n e - l o o p c o n t r i b u t i o n to t h e s c a t t e r i n g of f o u r c l o s e d s t r i n g s i s
represented b y e i t h e r of t h e d i a g r a m s .

Ca) (b)
Fig.(~)
T h e r u l e s of s t r i n g t h e o r y m a k e t h e s e two d i a g r a m s e q u i v a l e n t b e c a u s e t h e y c a n b e
distorted into each other. This equivalence between different distortions of t h e
world-sheet is a striking property of string theories {"duality") which has no
a n a l o g u e in p o i n t field t h e o r y . Fig. 5(a) l o o k s like a s t r i n g box d i a g r a m while fig.
5(b) looks like a tadpole diagram. In the original bosonie and spinning string
theories there is a n i n f i n i t y 22 i n t h i s a m p l i t u d e . This infinity has a very simple
interpretation in terms of the tadpole diagram configuration (fig. 5(b)). The
propagator i n t h e leg of t h e t a d p o l e is s i n g u l a r s i n c e i t i n c l u d e s ( a m o n g a n i n f i n i t e
set of s t a t e s ) the contribution from the massless scalar partner of the graviton
which has the form 1/k 2 where t h e m o m e n t u m i n t h e leg, k, i s z e r o b y m o m e n t u m
conservation. T h i s d e s c r i p t i o n of t h e d i v e r g e n c e of a loop d i a g r a m a s a n i n f r a - r e d
e f f e c t is u n i q u e to s t r i n g theories. The discovery that the type II s u p e r s t r i n g
t h e o r i e s a r e f i n i t e a t o n e loop 23 was t h e f i r s t i n d i c a t i o n t h a t superstring theories
143

might be consistent quantum field theories (remember that the loop integral is
evaluated in 10 dimensions where ordinary point field theories have terribly
ultra-violet divergences). This result h a s also r e c e n t l y been established for the
heterotic superstring t h e o r y 24.
I t h a s a l s o now b e e n e s t a b l i s h e d t h a t t h e o p e n - s t r i n g one-loop amplitudes with
gauge group SO(32) a r e f i n i t e w i t h f o u r 25 (or m o r e 26) e x t e r n a l s t a t e s a n d i n f i n i t e
for any other gauge group.
It is probable that any possible divergences at higher loops can also be
associated with the emission of massless scalar particles at zero momentum via
generalized tadpoles. F o r example, a t t w o l o o p s t h e d i v e r g e n t t a d p o l e c o n t r i b u t i o n
to a c l o s e d - s t r i n g a m p l i t u d e is r e p r e s e n t e d b y t h e ("E.T.") d i a g r a m .

Fig.(6)
From t h i s it follows t h a t t h e c o n d i t i o n f o r a n a m p l i t u d e to be f i n i t e a t a n y n u m b e r
of loops is t h a t ~ = 0 where ~ is t h e o n - s h e l l c o u p l i n g o f t h e
m a s s l e s s s c a l a r p a r t i c l e to t h e g e n e r a l t a d p o l e . But t h i s c o n d i t i o n is p r e c i s e l y t h e
requirement that there be an u n b r o k e n supersymmetry. Since s u p e r s y m m e t r y can
o n l y b e b r o k e n in p e r t u r b a t i o n t h e o r y in t e n d i m e n s i o n s if t h e r e a r e a n o m a l i e s it
follows t h a t f r e e d o m f r o m a n o m a l i e s *-~ f i n i t e n e s s 27.
I t is i m p o r t a n t to e s t a b l i s h b y e x p l i c i t c a l c u l a t i o n w h e t h e r t h e t h e o r i e s a r e f i n i t e
to all o r d e r s . This is being intensively studied at the present time28, 29 a n d a
complete p r o o f of f i n i t e n e s s (at least for the type II and heterotic superstring
t h e o r i e s ) s h o u l d b e f o r t h c o m i n g in t h e n e a r f u t u r e .

A N O M A L I E S A N D THEIR C A N C E L L A T I O N

The s i g n a l f o r a n a n o m a l y is t h e p r e s e n c e of a n o n - z e r o c o u p l i n g b e t w e e n a n
u n p h y s i c a l l o n g i t u d i n a l mode of a g a u g e p a r t i c l e a n d a n y p h y s i c a l t r a n s v e r s e modes.
Just a s in f o u r dimensions an anomaly can arise from a triangle diagram, in ten
dimensions an anomaly can arise from a hexagon diagram with c i r c u l a t i n g ehiral
fermions and external gauge particles. F o r example, t h e e v a l u a t i o n o f t h e a n o m a l y in
t h e Yang-Mills c u r r e n t in a t y p e I t h e o r y r e q u i r e s t h e c a l c u l a t i o n of
144

Fig.(7)
where the i n t e r n a l lines are o p e n s t r i n g p r o p a g a t o r s and one of the external s t a t e s
is a longitudinal mode o f a Yang-Mills p a r t i c l e while the others are transverse
modes. The r e s u l t of t h i s c a l c u l a t i o n is t h a t t h e anomaly v a n i s h e s 3 w h e n t h e g a u g e
g r o u p is SO(32).
T h i s may a p p e a r p u z z l i n g b e c a u s e t h e h e x a g o n d i a g r a m s o f t h e c o r r e s p o n d i n g
low e n e r g y massless point field theory do not give a vanishing anomaly. The
explanation is b as ed on the fact t h a t c e r t a i n o p e n - s t r i n g hexagon diagrams contain
closed-string bound states in v a r i o u s channels. These bound states include the
massless s t a t e s of the s u p e r g r a v i t y sector which means that the low-energy theory
(in w h i c h all t h e m a s s i v e s t a t e s d e c o u p l e ) g e t s c o n t r i b u t i o n s f r o m t h e s e s t a t e s in
addition to the expected anomalous c o n t r i b u t i o n from the massless hexagon diagrams.
These extra terms have the form of tree diagrams in which the supergravity
particles are exchanged and w h i c h h a v e anomalies w h i c h e x a c t l y c a n c e l t h e usual
a n o m a l y of t h e m a s s l e s s h e x a g o n d i a g r a m s . This explains the a b s e n c e of an anomaly
in t h e l a n g u a g e o f t h e l o w - e n e r g y p o i n t f i e l d t h e o r y a s b e i n g d u e to a c a n c e l l a t i o n
b e t w e e n t h e e x p e c t e d q u a n t u m a n o m a l y ( d u e to t h e u s u a l h e x a g o n d i a g r a m s ) a n d a
new, a n o m a l o u s , t e r m in t h e c l a s s i c a l t h e o r y (associated with these tree diagrams).
The new t e r m c a n b e t h o u g h t of a s a n a d d i t i o n a l (anomalous) t e r m in t h e e f f e c t i v e
p o i n t f i e l d t h e o r y a c t i o n w h i c h is a local polynomial in t h e f i e l d s . This m e c h a n i s m
d e p e n d s o n a d e l i c a t e i n t e r p l a y b e t w e e n g r a v i t a t i o n a l a n d Yang-Mills e f f e c t s .
The a n o m a l i e s a s s o c i a t e d w i t h g r a v i t a t i o n a l c u r r e n t s ( d u e to h e x a g o n d i a g r a m s
with external gravitons) and the mixed a n o m a l i e s (due to h e x a g o n d i a g r a m s with
e x t e r n a l Yang-MiUs p a r t i c l e s t o g e t h e r w i t h g r a v i t o n s ) h a v e n o t y e t b e e n c a l c u l a t e d
in t h e superstring theories. However, t h e a n a l y s i s of t h e low e n e r g y p o i n t field
theory was carried out for these a n o m a l i e s also 3. It turned out that all the
anomalies can be cancelled by adding local a n o m a l o u s terms to t h e a c t i o n if t h e
Yang-Mills g r o u p , G, is s u c h t h a t :
(a) The d i m e n s i o n of t h e a d j o i n t r e p r e s e n t a t i o n of G - 496 (which g u a r a n t e e s
the a b s e n c e of the gravitational anomalies).
(b) An a r b i t r a r y matrix, F, in t h e a d j o i n t r e p r e s e n t a t i o n of G satisfies

T r F 6 _- 1 TrF 2 (TrF 4 _ 1 (TrF2)2) (7)


48 300
145

(which ensures the absence of the Yang-Mills and mixed anomalies). The only
g r o u p s f o r w h i c h t h e s e c o n d i t i o n s a r e s a t i s f i e d a r e S0(32) a n d E8 x E8 ( a p a r t f r o m
t h e p r e s u m a b l y u n i n t e r e s t i n g c a s e s (U(1) 496 a n d E 8 x U(1)248). The t y p e I t h e o r i e s
do n o t a d m i t e x c e p t i o n a l g r o u p s b u t t h e h e t e r o t i c s t r i n g i n c o r p o r a t e s b o t h of t h e m .
F o r c o m p l e t e n e s s i t w o u l d b e d e s i r a b l e f o r t h e a n a l y s i s of p o s s i b l e s u p e r s y m m e t r y
a n o m a l i e s to b e c a r r i e d o u t 30.
The p r e c e d i n g discussion referred to a n o m a l i e s a s s o c i a t e d w i t h i n f i n i t e s s i m a l
gauge transformations. There is still t h e p o s s i b i l i t y of a n o m a l i e s in the "large"
transformations which are n o t c o n t i n u o u s l y c o n n e c t e d to t h e i d e n t i t y . There are,
f o r example, k n o w n to b e 991 t y p e s of l a r g e g e n e r a l c o o r d i n a t e t r a n s f o r m a t i o n s in
ten dimensional spherical space-time. T h e s e h a v e b e e n s h o w n n o t to b e a n o m a l o u s
for the cases in which the gauge group is one of those free of infinitessimal
anomalies31. T h i s r e s u l t h a s a l s o b e e n g e n e r a l i z e d 32 with some a s s u m p t i o n s to more
general spaces than the ten-sphere in t h e c a s e of t h e E 8 x E 8.

COMPACTIFICATION OF EXTRA DIMENSIONS

The structure of s p a c e - t i m e a n d hence the p o s s i b i l i t y o f c o m p a c t i f i c a t i o n to f o u r


d i m e n s i o n s s h o u l d b e d e t e r m i n e d b y t h e s o l u t i o n of t h e e q u a t i o n s of t h e s u p e r s t r i n g
field t h e o r y (which h a s so f a r o n l y b e e n f o r m u l a t e d in t h e l i g h t - c o n e g a u g e ) . This
approach has not yet been productive. However, t h e r e are many o t h e r o b s e r v a t i o n s
t h a t s u g g e s t t h a t r e a l i s t i c f o u r - d i m e n s i o n a l p h y s i c s may well e m e r g e . Many of t h e s e
are based on topological features of t h e t h e o r y . For example, o n e of t h e c r u c i a l
features of t h e a n o m a l y c a n c e l l a t i o n a r g u m e n t d i s c u s s e d a b o v e is t h e p r e s e n c e of
t h e m a s s l e s s s e c o n d - r a n k a n t i s y m m e t r i c t e n s o r f i e l d B~rv w h i c h h a s a f i e l d s t r e n g t h
H~lj9 d e f i n e d b y

___ly + L
H~9 = a l i v e ] 30 ~ e w~ (8)

where [ ] d e n o t e s o n t i s y m m e t r i z a t i o n o f t h e i n d i c e s and t h e Chern-Simons t e r m s a r e


Y
d e f i n e d by ~[o~jjvp]=TrF[o~Fljg] (where t h e Y a n g - M i l l s f i e l d s t r e n g t h F ~ i s a m a t r i x
L
i n t h e a d j o i n t r e p r e s e n t a t i o n o f t h e gauge group) and a[o~/2v@]=trR[o~vp] (where t h e
]nil
Riemann c u r v a t u r e R/2v i s a m a t r i x in t h e t a n g e n t s p a c e w i t h m , n = 0 , 1 , . . . 9 ) . The
c o n d i t i o n t h a t H/2v~ s h o u l d b e s i n g l e - v a l u e d is t h a t t h e i n t e g r a l of i t s c u r l o v e r a n y
f o u r - d i m e n s i o n a l s u b s p a c e s h o u l d v a n i s h 33, i.e.

(trR[o~Rv@] ~0 TrF[~Fv@]) = 0 (9)

T h i s c o h o m o l o g y c o n s t r a i n t h a s two immediate a n d i m p o r t a n t c o n s e q u e n c e s :
(a) It is t h e condition that ensures that the anomalies which were previously
s h o w n to v a n i s h in f l a t t e n - d i m e n s i o n a l s p a c e continue to be a b s e n t when some
146

d i m e n s i o n s are curved.
(b) It indicates that in general w h e n there is non-zero curvature, so that R~aj*0,
the Yang-Mills field strength is also non-zero, i.e. F~#0. This is just what is
needed, since a non-zero field strength will lead to a breaking of the (very large)
gauge group of the ten-dimensional theory to a smaller group in the compactified
theory (which will hopefully be of more direct interest for physics).

TOWARDS FOUR-DIMENSIONAL PHYSICS

A particularly interesting class of possible compactified solutions has been


proposed in ref. 34 in which the ten-dimensional space is the product of
four-dimensional flat Minkowski space and a curved six-dimensional compact space
with vanishing Ricci curvature. The holonomy group of this compact six-dimensional
space is SU(3) i.e. the curvature, R ~ , is an SU(3) matrix in the tangent space - a
condition that leads to a four-dimensional theory possessing an unbroken
supersymmetry. Such compact Ricci-flat spaces have come to be called "Calabi-Yau"
spaces (they were conjectured to exist by Calabi and shown to exist by Yau).
Although the original motivation for suggesting that Ricci-flat spaces may be of
interest was based on analysing the effective point field theory that approximates
the superstring at low energy, it is possible to argue convincingly that they are
also solutions of the full string theory.
The equations of the low energy effective point field theory also suggest that
the curvature and the Yang-Mills field strengths should be proportional (which is
yet another example of the unification of the gravitational and Yang-Mills aspects of
the theory), i.e.

R~, ~ F ~ (i0)

(suppressing the matrix indices) so that the field strength is also non-zero in an
SU(3) subgroup of E 8 x E 8 (the SO(32) case does not appear to hold m u c h prospect
of describing physics). As a result the symmetry is broken down to a subgroup
that commutes with that SU(3), namely

E6 E8 (11)

The E6 factor plays the r61e of a Grand Unified group for the effectively
four-dimensional theory (familiar from t h e P h e n o m e n o l o g y o f t h e m i d - 1 9 7 0 ' s 35) but
with certain novel features. The E 8 f a c t o r d e s c r i b e s a n o t h e r s e c t o r o f m a t t e r ( t h a t
has been dubbed ' s h a d o w m a t t e r ' ) c o n s i s t i n g of p a r t i c l e s w h i c h a r e n e u t r a l u n d e r
t h e E6 f o r c e s a n d t h e r e f o r e u n d e t e c t a b l e e x c e p t via t h e i r g r a v i t a t i o n a l i n t e r a c t i o n s
w i t h t h e m a t t e r t h a t we o b s e r v e . I t i s h i g h l y n o n - t r i v i a l t h a t i t is p o s s i b l e f o r t h e
147

symmetry to break in this manner and yet be consistent with the topological
c o n d i t i o n s implied b y eq.(9).

Consequences
There isn't yet any systematic procedure f o r c l a s s i f y i n g all p o s s i b l e C a l a b i - Y a u
spaces but it has been conjectured by Yau that there are a discrete but large
n u m b e r of s u c h s p a c e s ( a p p r o x i m a t e l y 10000!) a n d s o it i s to b e h o p e d t h a t f u r t h e r
theoretical constraints will restrict the choice. For the moment the particular
Calabi-Yau space used for t h e six compact d i m e n s i o n s is a d i s c r e t e parameter that
can be c h o s e n a r b i t r a r i l y . However, the phenomenological considerations described
below provide such severe restrictions that no single space so far constructed
encompasses all desirable features. Nevertheless, the way in which s o m e of the
known spaces come close to explaining various aspects of "low energy" physics
m a k e s u s e of s u c h r e m a r k a b l e f e a t u r e s t h a t it s e e m s l i k e l y t h a t t h e s c h e m e h a s t h e
potential for making contact with physics. Some of the major features of this
s c h e m e follow.
(a) U n l i k e w i t h c o n v e n t i o n a l k i n d s of u n i f i e d t h e o r i e s , t h e r e a r e f e w a d j u s t a b l e
parameters once the Calabi-Yau space has been selected. For example, t h e n u m b e r
of s p e c i e s of m a s s l e s s p a r t i c l e s (and hence the number of fermion g e n e r a t i o n s ) is
determined by a topological property of t h e s p a c e , n a m e l y , it i s e q u a l to ~ x ( E u l e r
characteristic). An aspect of this scheme that is crucial for describing E6
phenomenology is that the massless chiral fermions in the effectively
four-dimensional theory lie in the complex E6 representation 27 (which is again
consistent with eq.(9)). T h e r e is no f r e e d o m to a d j u s t the particle c o n t e n t so t h i s
is a l r e a d y a notable s u c c e s s for t h e scheme.
(b) T h o s e s p a c e s w h i c h g i v e r i s e to a s m a l l n u m b e r o f f e r m i o n f a m i l i e s ( t h e E6
phenomenology restricts t h e n u m b e r to b e t h r e e o r f o u r ) t u r n o u t to h a v e h o l e s i n
them (they are not simply connected). This in turn p l a y s a k e y rble in allowing t h e
E6 s y m m e t r y to b r e a k d o w n to a r e a l i s t i c low e n e r g y symmetry. In conventional
unified point field theories symmetry breaking can be introduced by arbitrarily
adding Higgs fields. This cannot be d o n e i n s u p e r s t r i n g theories s i n c e t h e y allow
no s u c h adjustable parameters. However, loops of E 6 flux can become trapped in
the holes in t h e compact space, which has the same effect as having an effective
Higgs field in the adjoint representation. With t h i s t o p o l o g i c a l m e c h a n i s m t h e v a l u e s
of t h e e f f e c t i v e Higgs fields a r e d e t e r m i n e d , u p to a d i s c r e t e c h o i c e , w h i c h in t u r n
determines, u p to a d i s c r e t e choice, t h e way in which E6 b r e a k s to a low e n e r g y
symmetry group. A m o n g p o s s i b l e low e n e r g y symmetry groups a r e 36
SU(3) x SU(2) x U(1) x U(1) x U(1), SU(3) x SU(2) x S U ( 2 ) , ... A g e n e r i c f e a t u r e of
t h i s m e c h a n i s m is t h a t t h e r e i s a l w a y s e x t r a low e n e r g y s y m m e t r y in a d d i t i o n to t h e
standard m o d e l ( t h e r e s i d u a l g r o u p h a s to h a v e a t l e a s t r a n k 5).
The m e c h a n i s m of b r e a k i n g t h e s y m m e t r y by flux loops c a n n o t be c o n t i n u o u s l y
s w i t c h e d off. T h i s m e a n s t h a t t h e E6 s y m m e t r y i s n e v e r a n e x a c t s y m m e t r y o f t h e
148

four-dimensional theory, even at high energy (in c o n t r a s t to i t s p r e v i o u s r61e in


so-called Grand Unified schemes}. Although there is unification of the gauge
c o u p l i n g s t h e Yukawa c o u p l i n g s do n o t s a t i s f y t h e E 6 r e l a t i o n s w h i c h a v o i d s some of
t h e bad p r e d i c t i o n s of E6 m a s s r e l a t i o n s .
(c) All t h e c o u p l i n g s of t h e m a s s l e s s p a r t i c l e s a r e d e t e r m i n e d b y t o p o l o g i c a l
c o n s i d e r a t i o n s 37 a n d do n o t depend on detailed knowledge of the metric of the
Calabi-Yau s p a c e {which is j u s t a s welt s i n c e n o n e of t h e m e t r i c s of t h e s e s p a c e s
has ever been constructed!). A possible problem arises with proton stability since
t h e r e is no a p p a r e n t r e a s o n f o r t h e Yukawa c o u p l i n g s r e s p o n s i b l e f o r p r o t o n d e c a y
a t t h e t r e e l e v e l to v a n i s h . N e v e r t h e l e s s , e x p l i c i t c a l c u l a t i o n of t h e s e c o u p l i n g s in
many o f t h e k n o w n s p a c e s s h o w s t h a t f o r most o f t h e m t h e p r o t o n is s t a b l e 38 (up to
the usual considerations about decay caused by radiative corrections}.
{d) T h e r e a r e p o s s i b l e s t a t e s a s s o c i a t e d with s t r i n g s w i n d i n g t h r o u g h h o l e s in
t h e c o m p a c t s p a c e w h i c h would lead to s t a b l e , u n c o n f i n e d p a r t i c l e s w i t h f r a c t i o n a l
e l e c t r i c c h a r g e s w i t h m a s s e s a r o u n d t h e P l a n c k m a s s 39 ( t h e v a l u e o f t h e f u n d a m e n t a l
c h a r g e d e p e n d i n g on w h i c h Calabi-Yau s p a c e is used}. Correspondingly, magnetic
m o n o p o l e s a r e p r e d i c t e d b y t h e t h e o r y w h i c h h a v e c h a r g e s w h i c h a r e a multiple of
t h e u s u a l Dirac v a l u e .
(e) The p o i n t o f v i e w a d o p t e d in t h i s whole s c e n a r i o is t h a t m u c h p h y s i c s c a n
b e o b t a i n e d f r o m t r e a t i n g t h e low e n e r g y e f f e c t i v e p o i n t field t h e o r y in l o w e s t o r d e r
in p e r t u r b a t i o n theory. However, c e r t a i n aspects clearly require a much deeper
understanding of the theory. The effective four-dimensional low-energy classical
theory has the form of a no-scale supergravity t h e o r y 40 a n d it is i m p o r t a n t to
understand how supersymmetry gets broken. One s u g g e s t i o n 41 invokes a
c o n d e n s a t e of t h e g l u i n o s in t h e " s h a d o w " E8 s e c t o r w h i c h t r i g g e r s a b r e a k i n g of
t h e s u p e r s y m m e t r y w i t h o u t l e a d i n g to t h e u s u a l p r o b l e m o f g e n e r a t i n g a cosmological
constant. Despite this virtue, the suggested mechanism requires a non-perturbative
e f f e c t t h a t g o e s o u t s i d e of t h e a p p r o x i m a t i o n s o n w h i c h t h e s c h e m e is b a s e d .
(f) (Mildly) p e s s i m i s t i c n o t e . The c o n s i d e r a t i o n of Calabi-Yau s p a c e s c a n b e
well-motivated provided the radius of the compact dimension (which is a free
p a r a m e t e r t h a t s h o u l d b e a p p r o x i m a t e l y t h e i n v e r s e of t h e u n i f i c a t i o n mass) i s m u c h
more t h a n t h e P l a n c k scale. The c a l c u l a t i o n s also a s s u m e t h a t t h e s t r i n g c o u p l i n g
c o n s t a n t ( w h i c h s h o u l d u l t i m a t e l y b e d e t e r m i n e d b y t h e t h e o r y ) is weak. There are
c o n v i n c i n g a r g u m e n t s 42 t h a t n e i t h e r of t h e s e a p p r o x i m a t i o n s c a n b e v a l i d a n d t h a t
superstring theory is i n t r i n s i c a l l y a s t r o n g l y coupled theory, It is i m p o r t a n t to
e s t a b l i s h , in t h a t c a s e w h e t h e r t h e r e s u l t s b a s e d o n t o p o l o g i c a l c o n s i d e r a t i o n s m i g h t
still h a v e some v a l i d i t y .
Since s u p e r s t r i n g theories are so v e r y d i f f e r e n t f r o m p o i n t f i e l d t h e o r i e s it
would be more s a t i s f y i n g to f i n d a q u a l i t a t i v e l y n e w k i n d o f e x p e r i m e n t a l p r e d i c t i o n
rather than trying to predict details of presently measured accelerator data.
[Examples of s u c h p r e d i c t i o n s a r e t h e e x i s t e n c e of e x t r a low e n e r g y s y m m e t r i e s a n d
t h e i r associated gauge particles, the o c c u r r e n c e of u n c o n f i n e d fractionally c h a r g e d
149

p a r t i c l e s a n d t h e e x i s t e n c e of s h a d o w m a t t e r . ]

THEORETICAL DEVELOPMENTS

(a) S t r i n g T h e o r i e s i n C u r v e d S p a c e - T i m e .
The s t r i n g a c t i o n s o f t h e f o r m o f eq. (5), w h i c h d e s c r i b e t h e motion of a s t r i n g
in a flat Minkowski space background (with metric ~) can be thought of as
t w o - d i m e n s i o n a l field t h e o r i e s of g r a v i t y (in w h i c h t h e " f i e l d s " a r e t h e s u p e r s p a c e
coordinates}. The t r e e d i a g r a m s of t h e c l o s e d - s t r i n g t h e o r y a r e a s s o c i a t e d w i t h a
two-dimensional world-sheet (as in fig. 3) which is a closed surface that is
t o p o l o g i c a l l y e q u i v a l e n t to a s p h e r e . The n - l o o p c o r r e c t i o n s ( i l l u s t r s t e d i n fig. 5)
c o r r e s p o n d to t w o - d i m e n s i o n a l s u r f a c e s w i t h n h a n d l e s . T h e r e f o r e to a n y o r d e r in
string perturbation theory a string theory is equivalent to a two-dimensional
g r a v i t a t i o n a l t h e o r y e v a l u a t e d o n a manifold of p a r t i c u l a r g e n u s . This v i e w p o i n t is
a t h e m e in m a n y i n t e r e s t i n g d e v e l o p m e n t s .
(i) The heterotic s u p e r s t r i n g 17 possesses not only two-dimensional
(world-sheet) coordinate invariance and two-dimensional supersymmetry in common
with the other superstring t h e o r i e s b u t is also c h i r a l in t h e t w o - d i m e n s i o n a l s e n s e
(due to t h e asymmetric t r e a t m e n t of the right and left polarized modes). It is
noteworthy that all these properties are also properties of the theory in the
ten-dimensional sense. C o n s i s t e n c y of string theory requires it to be free of
a n o m a l i e s in t w o - d i m e n s i o n a l c o o r d i n a t e t r a n s f o r m a t i o n s t h a t c a n n o t be C o n t i n u o u s l y
connected to the identity (these "large" coordinate transformations are modular
transformations). This is t h e i n g r e d i e n t in t h e h e t e r o t i c s u p e r s t r i n g theory that
r e s t r i c t e d the possible g a u g e g r o u p s to j u s t t h o s e p r e v i o u s l y o b t a i n e d b y r e q u i r i n g
t h e a b s e n c e o f t h e i n f i n i t e s s i m a l t e n - d i m e n s i o n a l c h i r a l anomalies. Furthermore, this
g e n e r a l i z e s to t h e s i t u a t i o n in w h i c h t h e t e n - d i m e n s i o n a l s p a c e is c u r v e d provided
t h e i d e n t i f i c a t i o n in eq.(10) is made 43. T h i s i d e n t i f i c a t i o n is also r e q u i r e d in o r d e r
to e n s u r e t h a t t w o - d i m e n s i o n a l c h i r a l a n o m a l i e s v a n i s h in t h e c o m p a c t i f i e d t h e o r y 44.
(ii) In o r d e r to d e s c r i b e a s t r i n g p r o p a g a t i n g in a c u r v e d background it is
n e c e s s a r y to r e p l a c e t h e f l a t m e t r i c in eq.(5) b y a c u r v e d m e t r i c , G ~ ( X ) , w h i c h is a
function of X. This gives the action of a non-linear sigma model m a p p i n g the
world-sheet into the curved space-time. T h e r e a r e , h o w e v e r , s t r o n g c o n s t r a i n t s on
formulating consistent string theories in a curved background due to the
r e q u i r e m e n t t h a t t h e t h e o r y c a n be f o r m u l a t e d i n a p a r a m e t r i z a t i o n i n w h i c h it is
c o n f o r m a l l y i n v a r i a n t ( r e c a l l t h a t t h i s w a s n e c e s s a r y to p r o v i d e t h e g a u g e c o n d i t i o n s
r e q u i r e d to decouple t h e n e g a t i v e - n o r m e d states). I n g e n e r a l t h i s will r e q u i r e t h e
presence of o t h e r terms in the two-dimensional action involving, in addition to
G~(X)9 a n a n t i s y m m e t r i c t e n s o r b a c k g r o u n d , B ~ ( X ) , a s c a l a r b a c k g r o u n d , (X), a n d
fermionic t e r m s 45. [These background fields correspond to the massless field
content of the string field theory.] In addition, in order for one of these
150

g e n e r a l i z e d sigma models to c o r r e s p o n d to a compac~ification o f t h e t e n - d i m e n s i o n a l


superstring the algebra satisfied by the generators of t w o - d i m e n s i o n a l c o n f o r m a l
transformations (the "Virasoro" algebra) must have a central extension with a
particular coefficient. The c o n d i t i o n t h a t s u c h a t h e o r y b e c o n f o r m a l l y i n v a r i a n t is
t h a t t h e r e n o r m a l i z a t i o n g r o u p ~ f u n c t i o n s s h o u l d all v a n i s h ( t h e r e is o n e ~ f u n c t i o n
f o r e a c h k i n d of b a c k g r o u n d field). It is k n o w n t h a t a n o n - l i n e a r sigma model on a
group manifold can be made conformally invarlant by the addition of the term
involving B~ if t h i s is n o r m a l i z e d to a s p e c i a l v a l u e 46 ( t h i s c o r r e s p o n d s to a d d i n g a
torsion term that parallelizes the curvature of the group manifold). H o w e v e r , it
s e e m s u n l i k e l y t h a t s u c h s t r i n g t h e o r i e s 47 c a n b e s u p e r s y m m e t r i c in s p a c e - t i m e a n d
t h e y do n o t c o r r e s p o n d to a c o m p a c t i f i c a t i o n of a t e n - d i m e n s i o n a l t h e o r y ( t h e y h a v e
t h e w r o n g v a l u e f o r t h e c e n t r a l e x t e n s i o n t e r m in t h e V i r a s o r o a l g e b r a ) .
(iii) The e q u a t i o n s implied b y t h e v a n i s h i n g of t h e #B f u n c t i o n s , a n d hence by
t h e r e q u i r e m e n t of c o n f o r m a l i n v a r i a n c e , h a v e b e e n s t u d i e d f o r a wide c l a s s of s i g m a
models in p e r t u r b a t i o n t h e o r y ( w h e r e t h e e x p a n s i o n p a r a m e t e r is t h e i n v e r s e s t r i n g
tension). As e x p e c t e d b y g e n e r a l a r g u m e n t s 48 t h e s e e q u a t i o n s a r e just t h e field
equations for the massless components of the s u p e r s t r i n g f i e l d s e x p a n d e d in a p o w e r
s e r i e s in t h e i n v e r s e s t r i n g t e n s i o n i.e. in a low e n e r g y e x p a n s i o n . This a n a l y s i s
p r o v i d e s more e v i d e n c e , a t l e a s t in low o r d e r s in t h i s e x p a n s i o n , f o r t h e f a c t t h a t
R i c c i - f l a t s p a c e s o f SU(3) h o l o n o m y (Calabi-Yau s p a c e s ) a r e s o l u t i o n s of t h e s t r i n g
theory together with t h e i d e n t i f i c a t i o n of t h e c u r v a t u r e with the Yang-Mills field
strength (eq.(lO)).
(iv) T h e r e is now a p r o o f 49 t h a t s u p e r s y m m e t r i c n o n - l i n e a r s i g m a models in
which Gi/ij is the m e t r i c on a Calabi-Yau space are finite to all o r d e r s in this
expansion and so have vanishing /3 f u n c t i o n (and are therefore conformally
invariant) and are suitable candidates for compactified superstring theories. There
is also e v i d e n c e t h a t the R i c c i - f l a t c o n d i t i o n is n o t b y i t s e l f s u f f i c i e n t b u t it is
crucial for the space to be Kahler 50 (which is equivalent to demanding SU(3)
holonomy), t h u s l e n d i n g f u r t h e r s u p p o r t to t h e s c h e m e of ref.(34).
iv) S i n c e no m e t r i c o n a Calabi-Yau s p a c e h a s e v e r b e e n e x p l i c i t l y c o n s t r u c t e d
it is n o t p o s s i b l e to g i v e a n e x p l i c i t s o l u t i o n f o r a s u p e r s t r i n g t h e o r y in w h i c h six
d i m e n s i o n s a r e c o m p a c t i f i e d on s u c h a s p a c e . However, c e r t a i n Calabi-Yau s p a c e s
r e d u c e , in a s i n g u l a r limit, to s i x - d i m e n s i o n a l t o r i w i t h d i s c r e t e i s o m e t r i e s d i v i d e d
out. These singular spaces are called "orbifolds". Superstring theories defined
with orbifold backgrounds can be analyzed explicitly and behave as if the
b a c k g r o u n d w e r e a Calabi-Yau s p a c e ( t h e s i n g u l a r i t i e s of t h e o r b i f o l d a r e i r r e l e v a n t
in t h e s t r i n g t h e o r y ) 51.
(vi) The c o v a r i a n t f o r m u l a t i o n of s u p e r s t r i n g t h e o r i e s can e i t h e r be d e d u c e d b y
s t a r t i n g from t h e s p i n n i n g s t r i n g t h e o r y a n d then truncating to a s u p e r s y m m e t r i c
s u b s e t of s t a t e s (as m e n t i o n e d e a r l i e r ) o r f r o m a m a n i f e s t l y s u p e r s y m m e t r i c a c t i o n in
superspace as implied by eq.(5) 8. The latter formulation has a much more
g e o m e t r i c a l i n t e r p r e t a t i o n 52. It h a s also b e e n g e n e r a l i z e d to c u r v e d gravitational
151

b a c k g r o u n d s f o r t h e t y p e I t h e o r i e s 53 ( a n d , r e c e n t l y , f o r t h e t y p e II t h e o r i e s 5 4 ) .

It s e e m s u n l i k e l y that the p h y s i c s o f two d i m e n s i o n s will d e t e r m i n e all t h e


c o n s t r a i n t s o n t h e t h e o r i e s a l t h o u g h it is r e m a r k a b l e how r e s t r i c t e d t h e p o s s i b i l i t i e s
are for constructing a suitable conformally invarlant sigma model on a
t w o - d i m e n s i o n a l manifold o f a r b i t r a r y genus. A suggestion that non-perturbative
string effects must play an important rSle is t h a t f l a t t e n - d i m e n s i o n a l Minkowski
s p a c e s a t i s f i e s all t h e r e s t r i c t i o n s we k n o w of t h a t follow f r o m t h e t w o - d i m e n s i o n a l
v i e w p o i n t a n d y e t we h o p e to p r o v e it is n o t a p o s s i b l e s o l u t i o n of t h e t h e o r y .

(b) T o w a r d s a g a u g e - i n v a r i a n t field t h e o r y of s u p e r s t r i n g s .
I n o r d e r to a r r i v e a t a more g e o m e t r i c a l u n d e r s t a n d i n g of s t r i n g f i e l d t h e o r y it
is p r o b a b l y n e c e s s a r y to f o r m u l a t e it i n a g a u g e - i n v a r i a n t manner. A preliminary
step was the understanding of the free bosonic string field theory in a
Lorentz-covarlant gauge using the BRS t e c h n i q u e 55. This has led to a gauge
i n v a r i a n t f o r m u l a t i o n o f t h e f r e e s t r i n g field t h e o r y 56 {as well a s some a s p e c t s o f
the interactions57).

(c) O t h e r T o p i c s
(i) Throughout the d e v e l o p m e n t of s t r i n g t h e o r i e s t h e r e has b e e n a parallel
d e v e l o p m e n t o f Kac-Moody a l g e b r a s in m a t h e m a t i c s . The representation theory of
Kac-Moody a l g e b r a s has deep c o n n e c t i o n s with the d y n a m i c s of s t r i n g t h e o r i e s 58,
These infinite-dimensional algebras express the algebra o f local currents in the
string world-sheet. The c o n n e c t i o n b e t w e e n Kac-Moody a l g e b r a s a n d s t r i n g t h e o r i e s
h a s b e e n a c t i v e l y s t u d i e d a n d w a s c r u c i a l in d e v e l o p i n g t h e h e t e r o t i c s u p e r s t r i n g 59,
(ii) A l t h o u g h a p p e a l i n g f r o m a g e o m e t r i c a l p o i n t of v i e w , t h e f o r m u l a t i o n of
superstring theories in terms of a manifestly supersymmetric Lorentz-covariant
a c t i o n like eq.(5) h a s n o t b e e n q u a n t i z e d in a c o v a r i a n t m a n n e r . This s e e m s l i k e l y
to b e s o l v e d b y e x t e n d i n g t h e s y m m e t r i e s o f t h e a c t i o n 60. Furthermore, there has
been progress towards formulating a manifestly supersymmetric first-quantized
theory by directly constructing t h e q u a n t u m o p e r a t o r s o f t h e t h e o r y 61. This may
l e a d to a p r o o f of t h e a b s e n c e of s u p e r s y m m e t r y a n o m a l i e s a t n l o o p s f o r t h e t y p e
II a n d h e t e r o t i c s u p e r s t r i n g t h e o r i e s , a n d h e n c e to t h e i r f i n i t e n e s s .
(iii) I t is now p l a u s i b l e t h a t t h e r e q u i r e m e n t s that a chiral ten-dimensional
t h e o r y w i t h a Yang-Mills g a u g e g r o u p b e s u p e r s y m m e t r i c a n d also f r e e of a n o m a l i e s
l e a d s i n e x o r a b l y to s u p e r s t r i n g theory. The "minimal" t e n - d i m e n s i o n a l f i e l d t h e o r y
of s u p e r - Y a n g - M i l l s c o u p l e d to s u p e r g r a v i t y h a s anomalies. These can be cancelled
by adding extra terms to the theory3 (motivated by the low energy limit of
superstring theory) which spoil its s u p e r s y m m e t r y . The p r o c e s s o f r e s t o r i n g the
supersymmetry b y a d d i n g y e t more t e r m s s h o u l d e v e n t u a l l y r e c o n s t r u c t the infinite
n u m b e r o f t e r m s t h a t c o n s t i t u t e t h e e x a c t e x p a n s i o n of t h e superstring theory in
terms of the massless fields. At t h e l o w e s t n o n - t r i v i a l o r d e r in t h i s e x p a n s i o n t h i s
152

h a s b e e n s h o w n 62 to imply t h e e x i s t e n c e of t e r m s i n t h e low e n e r g y a c t i o n w h i c h
are quadratic in the Riemann curvature and which have the structure initially
c o n j e c t u r e d in r e f . 63.

CONCLUSION

Superstring theories have p a s s e d all t h e t e s t s of c o n s i s t e n c y w i t h q u a n t u m


mechanics that cause problems with conventional theories of gravity, based on
Einstein's theory. Furthermore, this consistency restricts the p o s s i b l e Yang-Mills
g r o u p s almost u n i q u e l y a n d t h e r e f o r e h o l d s t h e e x c i t i n g p r o s p e c t o f a u n i f i e d a n d
c o n s i s t e n t q u a n t u m t h e o r y of all t h e i n t e r a c t i o n s .
T h e s e a r e e a r l y d a y s , h o w e v e r , a n d t h e r e a r e many q u e s t i o n s to be a n s w e r e d
a b o u t how s u p e r s t r i n g t h e o r i e s may make c o n t a c t w i t h o b s e r v e d p h y s i c s . Up u n t i l
now, q u i t e a p a r t from the phenomenologieal issues d e s c r i b e d earlier, t h e s e t h e o r i e s
h a v e n o t p r o v i d e d a n a t u r a l e x p l a n a t i o n f o r s e v e r a l of t h e most a c c u r a t e l y k n o w n
n u m b e r s in p h y s i c s :
- At t h e v e r y l e a s t we m u s t u n d e r s t a n d how it is t h a t t h e s e t h e o r i e s , f o r m u l a t e d
i n i t i a l l y in D-10 d i m e n s i o n s p r e d i c t t h a t , to a v e r y g o o d a p p r o x i m a t i o n ,

D=4

in the world that we s e e at accessible energies which is a n assumption in the


scheme outlined above, lit would at least be satisfying to discover that
t e n - d i m e n s i o n a l Minkowski s p a c e is n o t a s o l u t i o n of t h e t h e o r y . ]
- The f a c t t h a t t h e cosmological c o n s t a n t is z e r o to a n a m a z i n g a c c u r a c y h a s n o t
y e t b e e n e x p l a i n e d in a n a t u r a l way in s u p e r s t r i n g theory. [Although the scheme of
ref.(34) o u t l i n e d e a r l i e r d o e s n o t g e n e r a t e a cosmological c o n s t a n t i n t h e p r o c e s s o f
compactification there is no obvious mechanism that would prevent one being
g e n e r a t e d in t h e s u b s e q u e n t s y m m e t r y b r e a k i n g t r a n s i t i o n s . ]
- Another outstanding q u e s t i o n is w h e r e t h e m a s s s c a l e a s s o c i a t e d w i t h weak
s y m m e t r y b r e a k i n g c o m e s from.

The present theoretical understanding of superstring theories is somewhat


primitive. The t h e o r i e s a r e f o r m u l a t e d in t e r m s o f i n v a r i a n c e p r i n c i p l e s r e l a t e d to
the two-dimensional world-sheet. T h e y c o n t a i n b o t h t h e m a s s l e s s Yang-Mills p a r t i c l e
and t h e m a s s l e s s g r a v i t o n w h i c h is w h y t h e y r e d u c e , a t low e n e r g i e s , to { s u p e r )
Yang-Mills c o u p l e d to (super) gravity. However, t h i s a p p e a r s to b e a f o r t u i t o u s
a c c i d e n t s i n c e t h e t h e o r i e s w e r e n o t e x p l i c i t l y b a s e d o n a n y g e o m e t r i c a l p r i n c i p l e in
space-time. The d i s c o v e r y of s u c h a p r i c i p l e , w h i c h would b e a g e n e r a l i z a t i o n of
the principle of general relativity, would lead to a much more profound
u n d e r s t a n d i n g of t h e b a s i s o f t h e t h e o r y a n d t h e r e f o r e of i t s p r e d i c t i o n s .
153

REFERENCES

(i) Th. Kaluza, Sitzungsber. Preuss. Akad. Wiss. Berlin, Math. Phys. K1
(1921) 966;
O. Klein, Z. Phys. 37 (1926) 895; Nature 118 (1926) 516.
(2) C. W e t t e r i c h , Nucl. Phys. B233 (1983) 109;
S. Randjbar-Daemi, A. Salam, E. Sezgin and J. S t r a t h d e e , Nucl. Phys.
B214 (1983) 491;
E. Witten, Proceedings of the 1983 Shelter Island Conference,
ed. N. Khuri et. al. (MIT press, 1985).
(3) M.B. Green and J.H. Schwarz, Phys. Lett., 149B (1984) 117;
(4) G. Yeneziano, Nuovo Cim. 57A (1968) 190;
M.A. Virasoro, Phys. Hey. 177 (1969) 2309;
J.A. Shapiro, Phys. Lett. 33B (1970) 361.
(5) Y. Nambu, Int. Conf. on Symmetries and Quark Models, Wayne State
University 1969 (Gordon and Breach, 1970) p.269;
H. B. Nielsen, Proceedings Df the 15th Int. Conf. on High Energy
Physics, Kiev 1970;
L. Susskind, Nuovo Cim. 69A (1970) 457; Phys. Rev. D1 (1970) 1182.
(6) P.M.Ramond, Phys. Rev. D3 (1971) 2415;
A. Neveu and J.H. Schwarz, Nucl. Phys. B31 (1971) 86; Phys. Rev. D4
(1971) 1109.
(7) F. Gliozzi, J. Scherk and D.I. Olive, Nucl. Phys. B122 (1977) 253.
(8) M.B. Green and J.H. Schwarz, Nucl. Phys. HI81 (1981) 502; Phys.
Lett. 109B (1982) 444; Nucl. Phys. B198 (1982) 252; B198 (1982)
441; Phys. Lett. 136B (1984) 367; Nucl. Phys. B243 (1984) 285;
M.B. Green, J.H. Schwarz, L. Brink, Nucl. Phys. B198 (1982) 474.
(9) Y. Nambu, Lectures at the Copenhagen Symposium, 1970;
T. Goto, Progr. Theor. Phys. 46 (1971) 1560.
(i0) S. Deser and B. Zumino, Phys. Lett. 65B (1976) 369;
L. Brink, P. Di Vecchia and P.S. Howe, Phys. Lett. 65B (1976) 471.
(ii) A.M. Polyakov, Phys. Lett. I03B (1981) 207; Phys. Lett., I03B
(1981) 211.
(12) J. H. Schwarz, Proc. Johns Hopkins Workshop on Current Problems
in Particle Theory 6, (Florence, 1982) 233;
N. Marcus and A. SaEnotti, Phys. Lett. IIgB (1982) 97.
(13) A. Neveu and J. Scherk, Nucl. Phys. B36 (1972) 155.
(14) T. Yoneya, Nuovo Cim. Lett. 8 (1973) 951; Progr. Theor. Phys.
51 (1974) 951.
J. Scherk and J.H. Schwarz, Phys. Lett. 52B (1974) 347.
(15) M.B. Green and J.H. Schwarz, Phys. Lett. 122B (1983) 143;
J.H. Schwarz and P.C. West, Phys. Lett. 126B (1983) 301;
J.H. Schwarz, Nucl. Phys. B226 (1983) 269;
P. Howe and P.C. West, Nucl. Phys. B238 (1984) 181
(16) L. Alvarez-Gaume and E. Witten, Nucl. Phys. B234 (1983) 269.
(17) D.J. Gross, J.A. Harvey, E. Martinec, and R. Rohm, Phys. Hey. Lett.
54 (1985) 502; Nucl. Phys. B256 (1985) 625.
(18) P.G.O. Freund, Phys. Lett., 151B (1985) 387.
(19) J.-P. Serre, A Course in Arithmetic (Springer-Verlag, 1973);
P. Goddard and D.I. Olive, in Vertex Operators in Maths and Phys.,
ed. J. Lepowsky, S. Mandelstam and I.M. Singer, Math. Sci. Research
Inst. Publication ~3 (Springer-Verlag 1985).
~54

(20) S. Mandelstam, Nucl. Phys. B64 (1973) 205; B69 (1974) 77;
M. Kaku and K. Kikkawa, Phys. Rev. DlO (1974) 1110, 1823.
E. Cremmer and J . - L . Gervais, Nucl. Phys. B76 (1974) 209.
J . F . L . Hopkinson N.W. Tucker and P.A. C o l l i n s , Phys. Rev.
D12 (1975) 1653.
M.B. Green and J.H. Schwarz, Nucl. Phys. B218 (1983) 43.
M.B. Green, J.H. Schwarz and L. Brink, Nucl. Phys. B219 (1983) 437.
M.B. Green and J.H. Sehwarz, Phys. L e t t . 140B (1984) 33;
Nucl. Phys. B243 (1984) 475.
(21) P. Goddard, J. Goldstone, C. Hebbi and C.B. Thorn,
Nucl. Phys.B56 (1973) 109.
(22) J. A. Shapiro, Phys. Hey. D5 (1972) 1945.
(23) M.B. Green and J.H. Schwarz, Phys. L e t t . 109B (1982) 444.
(24) D.J. Gross, J.A. Harvey, E. Martinec, and R. Rohm, " H e t e r o t i c S t r i n g
Theory I I . The I n t e r a c t i n g H e t e r o t i c S t r i n g " , P r i n c e t o n p r e p r i n t
(June 1985).
(25) M.B. Green and J.H. Schwarz, Phys. L e t t . 151B (1985) 21.
(26) P.H. Frampton P. Moxhay and Y.J. Ng, Harvard p r e p r i n t
HUTP-85/A059 (1985);
L. C l a v e l l i , Alabama p r e p r i n t (1985).
(27) M.B. Green, C a l t e c h p r e p r i n t CALT-68-1219 (1984), t o be p u b l i s h e d
in the volume in honour of the 60 t h b i r t h d a y o f E.S. Fradkin.
(28) S. Mandelstum, Proceedings o f t he N i e l s Bohr Cent enni al Conference,
Copenhagen, Denmark (May, 1985);
A. H e s t u c c i a and J. G. T a yl or , King's College p r e p r i n t (1985).
(29) S. Mandelstam, Proceedings o f t he Workshop on U ni fi ed S t r i n g
T h eo r ies , ITP Santa Barbara, 1985.
(30) R. Kallosh, Phys. L e t t . 159B (1985) 111.
(31) E. Witten, "Global G r a v i t a t i o n a l Anomalies", P r i n c e t o n
p r e p r i n t (1985).
(32) E. Witten, Proceedings of t he Cambridge N u f f i e l d Workshop (CUP 1985)
(33) E. Witten, Phys. L e t t . , 149B (1984) 351.
(34) P. Candelas, L. Horowitz, A. Strominger and E. Witten, Nucl. Phys.
B258 (1985) 46.
(35) F. Gursey, P. Ramond and P. Sikivie, Phys. Left. 60B (1976) 177.
(36) E. Witten, Nucl. Phys. B258 (1985) 75;
J.D. Breit, B.A. Ovrut, G. Segr~, Phys. Lett. 158B (1985) 33;
M. Dine, V. Kaplunovsky, M. Mangano, C. Nappi and N. Seiberg,
Princeton preprint (1985).
(37) h. Strominger and E. Witten, Princeton preprint (1985).
(38) A. Strominger, Princeton preprint (1985).
(39) X.G. Wen and E. Witten, Princeton preprint (1985).
(40) E. Witten, Phys. Lett. 155B (1985) 151.
(41) M. Dine, R. Rohm, N. Seiberg and E. Witten, Phys. Lett.
156B (1985) 55;
J . - P . Derendinger, L.E. Ibafiez and H.-P. N i l l e s , Phys. L e t t .
155B (1985) 65.
(42) M. Dine and N. Seiberg, Phys. Hey. L e t t . 55 (1985) 366; P r i n c e t o n
p r e p r i n t (1985);
Y. Kaplunovsky, Phys. Rev. L e t t . 55 (1985) 1033.
(43) E. Witten, Proceedings of t he Argonne-Chicago Symposium on
Anomalies, Geometry and Topology (World S c i e n t i f i c , 1985).
155

(44) R e f . ( 3 4 ) ;
C. Hull and E. Witten, "Supersymmetric Sigma Models and t he
H e t e r o t i c S t r i n g " , MIT p r e p r i n t (1985).
(45) E.S. Fradkin and A.A. T s e y t l i n , Lebedev p r e p r i n t
~261 (October, 1984).
(46) E. Witten, Comm. Math. Phys. 92 (1984) 455;
T. C u r t r i g h t and C. Zakhos, Phys. Rev. L e t t . 53 (1984) 1799.
(47) D. Nemeschansky and S. Yankielowicz, Phys. Rev. L e t t . 54 (1985) 620
R.S. J a i n , H. Shankar and S. Wadia, Tara I n s t i t u t e p r e p r i n t
(January, 1985);
E. B e r g s h o e f f , S. Ranjbar-Daemi, A. Salam, H. Sarmadi and E. Szesgin
ICTP, T r i e s t e , p r e p r i n t (1985).
(48) D. Friedan, Ph.D. T hes i s , LBL p r e p r i n t LBL-11517 (August, 1980);
ref.45;
A. Sen, Fermilab p r e p r i n t s FERMILAB-PUB-85/60-T (A pri l , 1985)
FERMILAB-PUB-85/77-T ( May, 1985);
C.G. Callan, E. Martinec M.J. P e r r y and D. Friedan, " S t r i n g s in
Background F i e l d s " , P r i n c e t o n p r e p r i n t (1985).
(49) L.Alvarez-Gaum4, S. Coleman and P.Ginsparg, Harvard p r e p r i n t (1985)
(50) D. Gross and E. Witten, to appear.
(51) L. Dixon, J.A. Harvey, C. Yafa and E. Witten, " S t r i n g s on O r b i f o l d s "
P r i n c e t o n p r e p r i n t (1985).
(52) M. Henneaux and L. Mezincescu, Phys. L e f t . 152B (1985) 340;
T.L. C u r t r i g h t , L. Mezincescu, C.K. Zachos, Argonne p r e p r i n t
ANL-HEP-PR-85-28 (1985).
(53) E. Witten, " T w i s t o r - L i k e Transform in Ten Dimensions", P r i n c e t o n
p r e p r i n t (May, 1985).
(54) M.T. G r i s a r u , P. Howe, L. Mezincescu, B. Nilsson, P.K. Townsend,
D.A.M.T.P. Cambridge p r e p r i n t (1985).
(55) W. S i e g e l , Phys. L e t t . 149B (1984) 157, 162.
(56) T. Banks and M. Peskin, Proceedings of the Argonne-Chicago
Symposium on Anomalies, Geometry and Topology (World Scientific,
1985); SLAC preprint SLAC-PUB-3740 (July, 1985);
D. Friedan, E.F.I. preprint EFI 85-27 (1985);
A. Neveu and P.C. West, CEHN preprint CERN-TH 4000/85;
K. Itoh, T. Kugo, H. Kunitomo and H. Ooguri, Kyoto preprint (1985);
W. Siegel and B. Zweibach, Berkeley preprint UCB-PTH-85/30 (1985);
M. Kaku and Lykken, CUNY preprint (1985);
M. Kaku, Osaka University preprints OU-HET 79 and 80 (July, 1985);
S. Raby, R. Slansky and G. West, Los Alamos preprint (1985).
(57) A. Neveu and P. West, ref.56.
(58) M. Halpern, Phys. Hey. DI2 (1975) 1684 (appendix B);
J. Lepowsky and R.L. Wilson, Comm. Math. Phys. 62 (1978) 43;
I.B. Frenkel and V. Kac, Inv. Math. 62 (1980) 23;
G. Segal, C o m . Math. Phys. 80 (1981) 301.
(59) P. Goddard and D.I. Olive, ref. 20.
(60) W. Siegel, Berkeley preprint UCB-PTH-85/23 (May, 1985).
(61) D. Friedan, S. Shenker and E. Martinec, E.F.I. preprint EFI 85-32
(April 1985).
(62) L.J. Romans and N.P. Warner, Caltech preprint CALT-68-1291 (1985)
(63) B. Zweibach, Phys. L e t t . 156B (1985) 315.
CONFORMALLY INVARIANT FIELD THEORIES IN TWO DIMENSIONS
CRITICAL SYSTEMS AND STRINGS

J.-L. GERVAIS
Physique Theorlque, Ecole Normale Superieure
24 rue Lhomond 75231 Paris cedex 05

At the present time it is hardly necessary to emphasize


the fundamental importance of string models since super
string theories are the most promising candidates for a
completely unified theory of all interactions. An other key
point is that string concepts have plaid an important role
in the recent developments of theoretical physics and
mathematics by suggesting man~,~new important ideas,such as
in particular supersymmetry--~J%~ ,and have led to very
interesting progress in the related critical models in two
dimensions. In these notes I shall mostly concentrate on
this latter aspect which is not, presently, so directly
aimed at a unified theory of all interactions but is quite
interesting in its own right .
The unifying feature of string theories and critical
systems is that they are both associated with conformally
invariant field theories. We shall first review the
essential features of this connection considering only, for
simplicity, bosonic strings. At the level of the present
discussion, supersymmetric strings are not basically
different. One essentially replaces the oonformal group by
its superconformal generalization.
The position of a string at time ~ is specified by a
field X ~ ( ~ , ~ ) where ~ distinguishes the various points
along the line. Hence one has a two dimensionnal field
theory in parameter space. When ~ varies the string sweeps
out a world sheet and one sets up the dynamics in such a
way that it be invariant under reparametrization of the
corresponding geometrical surface. One can rigourously show
that it is always possible to choose the parametrization
in such a way that the curves ~ =cste and the curves
=care intersect at right angle. This choice is not unique
since this orthogonality condition is left invariant by
all conformal transformations of ~ and ~ .In string
theories the conformal group is thus the residual symmetry
of the system with an orthogonal choice of O~)
parameters. Basically it is the group of all
transformations of the form

where f and g are two arbitrary real functions of one


157

variable. In the present dicussion we stick to the Lorentz


covariant string quantization where conformal invariance
is not explicitely broken.
Since a physical string has a finite length, ~ ' v a r i e s over
a finite range. It is always possible te redefine the
parameters in such a way that the dynamics is periodic in
with period 2 ~ .It is often very convenient to go to
Euclidean time by letting

For real ~ the conformal group can best be described as


the group of analytic transformations of
~ : e~~ (3)
Indeed with this variable, the strip o~ ~ ~ ~ is
represented by the whole complex z plane. In this
pioture, a conformal transformation in given by

= (4)
where F is an arbitrary complex function of one variable.
In genera)~, a quantity ~ ( % . ~ ]
~ ~
is called
l ~ 4
eonformally
.
oovarlant ) if it transforms accordlng to

where ~ and ~ are parameters depending on the quantity


considered which are called conformal weights. This notion
was rediscovered recently (~) and the corresponding fields
were called primary. If we separate the real and imaginary
parts according to

the differential transforms as

where ~ is the rotation matrix with angle 0 and where


is a dilatation factor. ~ and 8 are given by the
differential equations
--
Q9~ z 9~C~ (8)
Formula (5) becomes

(s)

Since ~ and ~ are the local dilatation and rotation


parameters respectively, d is the dimension and J is the
spin of the quantity considered.
For critical systems in two dimensions, OC~and DC$ a r e the
two coordinates. It is well known that a stastistieal
system at a point of transition of second order becomes
158

scale invariant. The corresponding rescaling of 9(:aandgC


is a particular case of (4).Polyakov has proposed @~h~
critical systems are invariant under the full conformal
group (4).
One thus sees thas that both string theories and critical
systems are based on conformally invariant field
theories,with however different descriptions. ~ and
are string parameters while x~ ,x~ are the coordinates of
the critical system.
In a conformally invariant field theory the improved energy
momentum tensor is symmetric traceless and conserved. For
two dimensional field theories in real ~ ) ~ s p a c e this leads
to

)( T: 1 --o (in)

Due to the periodicity in @ one can write


I

~ (Too , T ~ ) = Z L~.~ -~ ; 3 ;
(li)
~ C~-~)

The operators L m a n d L ~ are the infinitesimal generators


of conformal transformations. For a conformally covariant
field which satisfies equation (5) one has

(i2)

- ~ +(~)i] 0
The operators Lm and Lm each satisfy the Virasoro
algebra

where the central charge C depends on the model


considered. Its actual value is a key point. We shall have
more to say about this below, ln general the field theories
we are discussing are characterized by C and by the set of
conformally oovaria~t fields ~ together with the set of
conformal weights h ~ .Since the two V i r a s o r o algebras
(13) have the same properties we only consider explicitely
the algebra of the L m ' s most of the time.
As a first simple example, let us recall the essential
159

features of the standard bosonic (Veneziano) model. In this


case one only considers massless two dimensional free
fields X ~ . W e shall denote by A ~ the associated Virasoro
generator. As it is well known they satisfy eq.(13) with

C = ~) (14)

the vertex for the emission of the lightest string state


is simply given by

. e (15)

where k -~ is the energy momentum of the emitted


particle. ~t is well known that V ~ satisfies condition
(12) with

The present ~ i s c u s s i o n of string is in the covariant


formalim where one has to make sure that the time like
components of X ~ deoouple from the physical S matrix. Such
a ghost killing mechanism requires first of all that ]:,he
vertex V 4 have dimension I. From eq. ( 18 ) this leads
to ~ % - ~ . ~ .the emitted particle is a tachyon with ,,ass
m 2 = --2.
For critical models the ~ and ~ are critical
exponents. Indeed it is easy to see t h a t the global
dilatations, rotations , and translations of x ,x are
genrated by L~ ~= L t and "Ll .The vacuum state
1 % f the
system must therefore annihilated by these operators. As
a result the two point function of any covariant operator
can be computed u p _ to a constant factor by means of
equation (12).If ~ = ~ for instance, one finds

<01
Hence ~ gives the power behaviour of the two point
function at the critical point. It is quite obvious that
must be positive for physical operators since the
correlation f u n c t i o n s must decrease when the separation
increases.
As it is well known (~)(9] ,the derivative of a covariant
operator is not covariant in general. An important
exception is the case of an operator of vanishing
weight, lts derivative with respect to has ~ =i , ~ = Q
Conversely, assume there exists an operator I(z) with
=i Then it is obvious that

=o (18)

In stastistical mechanics such operators are called


marginal. If they exist one has critical lines instead of
160

critical points since they can be added to the action


with arbitrary coefficients without destroying conformal
invariance. For the Veneziano model we have just recalled
that V ~ (z) has weight l.lndeed equation (18) for V~ is
the basic ingredient for the decoupling of ghosts.
The notion of conformally covariant operator was
introduced in string models (~) in order to dicuss posible
generalizations of the Veneziano model. We now recall the
essential points of this approach. Generalized string
models involve other fields besides the free fields X ~ a n d
the corresponding two dimensional field theory may have a
non trivial interaction. Such is the case, for instance, if
we have additional space components which are
oompaotified. Quite generally we can consider that X
remains a free field which does not mix with the
additional two dimensional fields. These will be
characterized by the set of covariant operators ~
together with the set of weights ~ .The ghost killing
condition now requires the emission vertex to have
conformal weight one under the action of the total
Virasoro generator

where L m is the Virasoro generator of the additional two


dimensional dynamics. This is realized by c h o o s i n g

We therefore see t h a t the spectrum of lightest particles

w i l ~ lnvozve no tachyon provided t h a t


~ ~% (22)
and this condition selects the conformally invarian~ field
theories for which all eovariant fields have weights
larger than one. From the view point of critical systems
this condition is unusual since it means that the
corresponding two point function has a Fourier transform
which has at most a logarithmic singularity at zero
momentum. We shall come back to this later.
The last important general point about string theories is
that they make sens only if the central charge of the
total Virasoro algebra is equal to the critical value
28.This can be seen in many ways. The simplest one is to
notice (~) that the central charge of the Faddev Fopov
ghosts is precisely -26 so that, with the above value, the
central charge vanishes when all the fields are included
and there is actually no breaking of conformal
invariance. For the Veneziano model where only the X
field enters this means that ~ =26. In the generalized
models the total central charge is C + ~ where C is the
central charge of the additional dynamics. Hence one must
satisfy
161

~5-_ ~ - C (23)
If C is an integer larger than one, this will effectively
allow to lower the space time dimension.
At this point it is useful to recall some general
properties of representations of the Virasoro
algebra( 13).From the group theory viewpoint it can be
regarded as being in a Weyl Caftan basis,Lo being the
only operator of the commuting subalgebra, and L,t with
n O being step operators. Hence a highest weight vector
will be such that

o (24)
An irreducible representation is characterized by the
values of E and C .The corresponding vector space which
is called a Verma module, is spanned by all vectors of the
form

where J ~ m { are arbitrary positive integers,They are


eigenstates of Le"

All eigenvectors with the same eigenvalues are said to


belong to the same- level N. Kae (E) has considered the
matrix of all inner products in a given module which is
entirely determined :from the V i r a s o r o algebra together
with the hermiticity condition

H e n c e it is purely alebraic and only depends upon the


values of ~ and C It obviously factorizes into products
of finite matrices at each level. Kac (s) has obtained a
closed formula for each finite determinant. Define the
quantity

(28)

where p and q are arbitrary integers, lf we c o n s i d e r a


highest weight representation with ~ = ~(p,q) for some
given p,q both larger than zero, the Kae determinant
vanishes at the level N=pq. T h i s v a n i s h i n g shows that the
metric of the Verma module need not be positive
definite. The unitarity of the representation is thus in
162

question. It is easy to show that the negative values of


the highest weights are all excluded. For p o s i t i v e values
we note that, for C >I, ~ (p,q) is always negative for
p>l ,q>l , i.e. when it corresponds to a zero of a Kae
determinant. Hence, in this region the Kac d e t e r m i n a n t s
never change sign f o r . . p o s i t i v e 6 and one can show by
explicit construction that there exist a unitary
representation for all ~ >O,C>I .For C<I ,on the
contrary, there are Kac zeroes for p o s i t i v e a and the
p o s i t i v i t y of the m e t r i c is not assured. It has been shown
that u n i t a r y r e p r e s e n t a t i o n s only exist for(7)
(29)
C= i

where r,p,q are integers. Hence t h e allowed values of g


p r e c i s e l y coincide with zeroes of Kac determinants.
Going b a c k to c o n f o r m a l l y invariant field thories we
recall that, given a c o v a r i a n t operator with w e i g h t ~ ,it
is easy to see that the state
0(3) o> (30)

is a h i g h e s t weight v e c t o r with ~ = ~ .The spectrum of


highest weights coincides with the set of conformal
weights which, in general, involves more than one values. The
r e p r e s e n t a t i o n of the V i r a s o r o algebra is thus reducible
since the H i l b e r t space is the sum of the c o r r e s p o n d i n g
V e r m a modules. The c o v a r i a n t operators are intertwening
operators between the different irreducible
representations. For a r b i t r a r y ~ and C one can c o n s t r u c t
an infinite family of c o v a r i a n t operators with weights
given by f o r m u l a (28) for all p and q p o s i t i v e or n e g a t i v e
as a natural b y p r o d u c t o;_~he exact quantum solution of
quantum L i o u v i l l e theory ~) .These operators are not all
physical since formula ( 2 ~ is not always positive. ]'his
shows n e v e r t h e l e s s that the set of critical dimensions
must coincide with Kac formula in general.
In v i e w of formula (28), it is clear that one has to
d i s t i n g u i s h three regions for the p o s s i b l e values of C.
I-The region C<I
As we a l r e a d y pointed out ,the Kac zeroes occur for
positive highest weights. A s y s t e m a t i c d i s c u s s i o n has been
given(S) which uses this fact, and shows the e x i s t e n c e of
special values of C
C ="l * 6(~+~)Z (31)

where r and s are integers of opposite signs. The virtue of


this formula is that then
= (32)
163

is the ratio of squares of integers.ln view of formula


( 2 9 ) , u n i t a r i t y is s a t i s f i e d only for r+s=-l. This subseries
remarkably reproduces a whole set of standard critical
models (T) .In p a r t i c u l a r for r=3 and 5 one recovers the
Ising (or 2 state Ports ) model and the 3 state Pobts
model. A s i m p l e c a l c u l a t i o n shows that if we introduce

cc- )
we obtain the correct n u m b e r of spin components i.e. Q=2,3
for the Ising and the three state Potts model
respectively. The Q state Potts model can be d e f i n e d for
continuous values of Q if we t r a n s f o r m it into the random
cluster model. For C<I there exist various equivalent
critical models. In p a r t i c u l a r the Q state critical Ports
model is e q u i v a l e n t to a C o u l o m b gas model. For C<I we have
Q<4 and one is in a C o u l o m b phase. The point Q=4
corresponds to a point of transition of
Kosterlitz-Thouless. Above Q=4 one enters into the p l a s m a
phase and the t r a n s i t i o n becomes first order. We shall come
back to this below.
II-The region C>25
This region has some similarities with the region C<I
since in both cases the square root of formula (28) is
real. A d i f f e r e n t approach is needed, however, since now the
Kac d e t e r m i n a n t s do not vanish for positive .The region
C>l is n a t u r a l l y covered b y the quantum L i o u v i l l e field
t h e o r y since its central charge is given by(9)
c= + (34)
where 4% is the Planck constant. The region C>25
corresponds to ~ <i/8 i.e. to the w e a k coupling regime of
Liouville theory which is c o n n e c t e d to the semi classical
limit ~ n~ O. In the exact quantum solution (9) ,special
values of C were again found

They can be put under the form of equation (31) continued


to r=N and s=l. The s p e c t r u m of weights is again given by
formula (28) with

It is e a s i l y checked that all-this values are larger than


l, and c o n d i t i o n (22) is satisfied. The a s s o c i a t e d string
model has no tachyon. H o w e v e r formula (23) shows that ~ <i!
so that one has not g a i n e d much from this viewpoint. The
existence of conformally invariant field theories for
t h e s e values of C does h o w e v e r suggest that there are new
critical models. These are models of a new type since all
the known critical models have C<l,and a spectrum of
164

conformal weights between 0 and i. .The unusal feature of


the new models is, as we a l r e a d y pointed out, that the two
point functions are at most l o g a r i t h m i c a l l y divergent. The
experimental feature of the transition are thus rather
d i f f e r e n t from the standard ones.
III-The region 1<C<25
In this case formula (28) gives complex values except when
p= ~ q. The choice p=q is unacceptable except for
p=q=l,since it leads to n e g a t i v e ~ .For three special
values of C
C = 7 ; ~ ~ , 19 (37)
local fields have been construoted0~uch that the
spectrum of weights is given by formula (28) for

Condition ,~ again satisfied and the associated


string t h e o r y has no taohyon. Condition (23) leads to
= d9 ~ J3, 7 (39)
and there exist new string models for these values of ~ .
From the view point of statistical model one t h e r e f o r e
predicts (~) isolated points of second order phase
transition for the above values of C. This may be a bit
surprising since for C>I one is outside of the Coulomb
phase. One can d i r e c t l y see, h o w e v e r that these points must
enjoy special properties. F o r m u l a (33) when continued for
i<C<25 leads to Q c o m p l e x in general. For the values (37)
one obtains

and one can v e r i ~ y that three special values are the


only ones for which Q is real even though the argument of
the cosine is complex.
As a conclusion it is clear that the study of
o o n f o r m a l l y invariant field theories from the d o u b l e view
point of string theories and cratioal systems has
unravelled an interesting structure. The string theories
d i s c u s s e d here at not based on free field theories in two
d i m e n s i o n s and, hence, the dual amplitudes are d i f f i c u l t to
determine. The common feature of all the new conformally
invariant field theories d i s c u s s e d here is the a p p e a r e n e e
of operators of d i m e n s i o n l. For the associated string
theories, it corresponds to the e x i s t e n c e of a massless
string state. This fact should play a key role in the
complete understanding of these models. On the other
hand,we must say that, at the p r e s e n t time, the r e l e v a n c e
of these new models to p a r t i c l e physics is not yet clear.
The s u p e r s y m m m e t r i c version of the p r e s e n t d i s c u s s i o n has
been worked out in all details. ~%)
165

REFERENCES
(I) J.-L.GERVAIS, B. SAKIT A Nucl. Phys. B34(1971)832
(2) J.-L.GERVAIS, B. SAKITA Nucl. Phys. B34(1971)477
(3) A.A. Belavin, A.M. Polyakov ,A.B. Zamolodchikov Nucl. Phys.
B241(1980}333
(4) D. Friedan Les Houches Lectures Notes 1982
(5) V. KAC. Proceeding of the International Congress of
Mathematicians Helsinsky 1978;Lecture Notes in Physics
voi.94 p.441 Springer Verlag
(6) J.-L.GERVAIS, A. NEVEU COM. MATH.PHYS. IO0(1985)I5
(7) D.FRIEDAN, Z.QIU, S. SHENKER in Vertex Operabor in
Mathematics and Physics ed.J.LEPOWSKY et al. Springer;
Phys. Rev. lett. 52(1984)1575
(8) J.-L. GERVAIS, A. NEVEU Nucl. Phys. B257 FS14(1985)59
(9) J.-L.GERVAIS, A.NEVEU NucL.Phys. B224(IS83)329;B238(1984)
125
(i0) J.-L. GERVAIS, A.NEVEU Phys. lett. 151B(1985) 271
(ii) J.-F.ARVIS Nucl. Phys. B212(1983) 151; B218(1983) 303.
O.BABELON Nucl. Phys. B258(1985)680
LIOUVILLE MODEL ON THE LATTICE

L.D. Faddeev and L.A. Takhtajan


Steklov M a t h e m a t i c a l Institute
Fontanka 27
Leningrad DII - 191011 - USSR

Abstract :
Liouville e q u a t i o n is put on the lattice in a c o m p l e t e l y integrable
way. The classical version is i n v e s t i g a t e d in details and a lattice
deformation of the V i r a s o r o algebra is obtained. The q u a n t u m v e r s i o n
still lacks a s a t i s f a c t o r y definition of th~ Hamiltonian.

The L i o u v i l l e equation

e. ~. 0
(i)

on the finite interval 0 ( X ~ 2~ has a t t r a c t e d a considerable


attention because of its p o s s i b l e role in the q u a n t i z a t i o n of the
string m o d e l ilL. In particular, the h a m i l t o n i a n treatment of the
corresponding classical dynamical system in cases of p e r i o d i c and "open"
boundary conditions was p e r f o r m e d in 12L - 131. The q u a s i c l a s s i c a l
quantization obtained thereby is not c o m p l e t e l y satisfactory, especially
in the p e r i o d i c case. Indeed, the c o n t r i b u t i o n of the zero modes to
masses of the string excitations is c o n t i n u o u s because of the non-
c ompact n e s s of the c o r r e s p o n d i n g portion of phase space. The use of
singular solutions 141 could give d i s c r e t e spectrum of the zero modes
which is however negative and so the tachyon p r o b l e m becomes even worse.
Another possibility is to abandon the q u a s i c l a s s i c a l approach completely
with hope to get some kind of a n o n e q u i v a l e n t quantization.

W i t h these considerations in mind we have formulated some time ago a


program to put L i o u v i l l e model on the lattice in a g r e e m e n t w i t h the
general spirit of the t r e a t m e n t of the c o m p l e t e l y integrable models in
i+i d i m e n s i o n a l space-time 151. As a first step we o b t a i n e d the exactly
soluble classical version of the model. Because of t e c h n i c a l difficu~ies
we were not able to d e v e l o p e the q u a n t u m version to the end. We
advertized our partial results among our colleagues and r e a l i z e d that
recently some became interested in such a p r o g r a m ]61. In this
167

circumstances we d e c i d e d to p u b l i s h the results on the classical


version before the c o m p l e t i o n of the q u a n t u m one. We think that the
lattice deformation of the V i r a s o r o algebra w h i c h naturally appears
in our t r e a t m e n t is i n t e r e s t i n g by itself.

The paper is o r g a n i z e d as follows. In sec. 1 we remind the results of


L2] - 131 on the c l a s s i c a l continuous model in the form w h i c h allows
the natural lattice generalization. The latter will be d e s c r i b e d in
sec. 2. F i n a l l y in sec. 3 we shall indicate a possible quantum genera-
lization.

One of the authors (L.D.F.) is g r a t e f u l to p r o f e s s o r s B. Diu and


~ de Vega for their kind h o s p i t a l i t y at the LPTHE. We thank P r o f e s s o r s
O. Babelon, J.L. Gervais, M. Jimbo, R. M a r n e l i u s and A. Neveu for
interesting discussions.

1. C l a s s i c a l continuous model.

In w h a t follows only p e r i o d i c boundary conditions

(2)

will be used. The main idea of t2] - 131 is to use the change of
variables

(3)

from the initial data ~(x) ) TrtxJ ---- ~ ~(a) (time v a r i a b l e


is supressed) to new fields w i t h trival equations of m o t i o n

(4)

and some r e a s o n a b l y simple boundary condition. One v a r i a n t of this m a ~


described in detail in 171, can be based on the a u x i l l a r y linear problem
(note the absence of the spectral parameter)

(5)
188

where ~(~r) is a 2x2 m a t r i x p a r a m e t r i z e d by the initial data

Q ..

e~ _~ 1 (6)

Let T(x,~) be a fundamental m a t r i x solution n o r m a l i z e d by


T(~) : I . Then ~ : T(z~, o) is called a m o n o d r o m y matrix. The
particular form of ~ @ ) g u a r a n t i e s that d e t ~ " = 1 and tr ~" > 2, so
that ~ is hyperbolic.

Using the n o t a t i o n

Tc~,o) = To,,) = I
!
gc~))
(7)
I Co-;

We introduce functions 2A ~) and Aj-(~)

(8)
cx) C c-J

with the f o l l o w i n g properties

(9)

and

(i0)

where for any f

(il)

A, B, C, D being the m a t r i x elements of the m o n o d r o m y matrix. They are


n o n t r i v i a l observables so that the form (i0) of the b o u n d a r y conditions
169

seems not very transparent.


The Poisson structure
(12)

where ~ plays the role of the coupling constant, leads to the fun-
d a m e n t a l Poisson bracket relations

(13)

(we use already standard n o t a t i o n from 151) where the ~x ~


m a t r i x r is given by

0 0 o 0

0 -~' :~" o
r" -.-
(14)
0 o -~ o

0 o o 0

It follows from an evident r e l a t i o n

and c o m m u t a t i v i t y of the m a t r i x elements of T(y) and T(~,~) (ultra-


locality) that the f o l l o w i n g relations are true

15
Y [~,.~ - ~e~) ~]

+
16
170

where ~(~) is a sign function and we confine ourselves to the fixed


fundamental domain O < X, y < 2 ~ .

The f o l l o w i n g chain of Ansatze

(18)

-
~ 4 J
4 X~"
,~, '~,' 2.. ,~..~.' (19)

I2o)

and a n a l o g o u s l y for ~4- is now introduced. The final object (the


Schwartz d e r i v a t i v e of ~~- ) is known tO be invariant under the trans-
formation (ii) and so the simple b o u n d a r y conditions hold

It is i n t e r e s t i n g that the variables i n t r o d u c e d in (18) - (20) have


beautiful Poisson brackets of their own. With notations ~(a)--~[~(xj],
(w} ~. y ~ ( K } ] and so on we have the following list of formulae

(22)

so that for ~ the Poisson brackets are field independent. Conti-


ruing the d i f f e r e n t i a t i o n we get

(23)
171

so that the p and ~ fields decouple. F i n a l l y for S we acquire the


brackets

x-~) (24)

and a n a l o g o u s r e l a t i o n for S (~) , w h i c h are c h a r a c t e r i s t i c of the


V i r a s o r o algebra. Thus the phase space of the L i o u v i l l e model is essen-
tially the product of two such algebras.

It can be shown that the h a m i l t o n i a n of our model

(25)

has a simple expression in terms of S[~.] and S~,v-]

Z--

the e q u a t i o n of m o t i o n being linear

This allows to call S ( 9 z, ~,~ ) the angle action variables.


The inverse map is given by the famous L i o u v i l l e formula

(27)

with periodic ~{X} .

We finish this section w i t h a comment on b o u n d a r y conditions {i0). One


can s i m p l i f y them by d i a g o n a l i z i n g the m o n o d r o m y m a t r i x w h i c h is
a c h i e v e d by the t r a n s f o r m a t i o n
172

where ~ , , ~ are the real fixed points of

2" ~ ---4, Z ) ~, ~ ~z (29

The new v a r i a b l e s ~ (~) and ~"(~] have t h e f o l l o w i n g p r o p e r t i e s

~(o) = 4

(30

and satisfy simple boundary conditions

where e-P is'the smallest eigenvalue of T , that means that p >/ O.


It is c l e a r that p is the zero F ~ u r i e r coefficient of the p e r i o d i c
function p [ ~. ~') ]

We h a v e calculated the P o i s s o n brackets of t h e s e new v a r i a b l e s with


the following answer

Ia~, ~} = r Ec.-~)[,~c~)-~,~;] ~ +

(33)
173

~v" ( x ) z _ ,~_ (.))

(34)

and

(35)

Note the invariance of the first relation w i t h respect to t r a n s f o r m a t i o n


(ii) w i t h c o n s t a n t coefficients.

F o r t u n a t e l y the rather c o m p l i c a t e d t r a n s f o r m a t i o n (28) is not really


n e c e s s a r y and the o r i g i n a l v a r i a b l e s ~) , ~(~) can be used to
c o n s t r u c t the g e n e r a t o r s of V i r a s o r o algebra as well as H a m i l t o n i a n
of L i o u v i l l e model. It is for them that we are able to find generali-
zation on the lattice both in c l a s s i c a l and q u a n t u m Case.

2. C l a s s i c a l model on the lattice.

W o r k i n g in a c c o r d a n c e w i t h the general spirit of 151 we g e n e r a l i z e


L i o u v i l l e model on the lattice b e g i n n i n g w i t h the a u x i l l a r y linear
problem : instead of (5) we have now the e q u a t i o n

and L ~ must be c o n s t r u c t e d in terms of v a r i a b l e s qT~ , ~ satis-


fying the discret form of Poisson brackets (12)
174

The matrix L~ is essentially uniquely defined by the requirements :

i. In the continuous limit q]'~ = d,c~rxj ; ~ , , , = "J,,.,.


where ~ is lattice with length ~ L~ behaves as follows

L. = | +. ~ e t ~ 0(~) (38)

2. L,,,. exactly satisfies the fundamental Poisson relation of the


form (13)

{a;a} = [,., L. , L . ] (39)

with the r-matrix (14). The explicit formula is given by

4 + ~#' e Ae
L4~. = ~ ; --~r (40)

Introducing the transport matrix

A~ g~
L (41)

we let

(42)
A~ C~
The essential role of local relation (39) consists in the fact that
it leads the same relation for ~'~
17,5

The latter in its turn gives the f o l l o w i n g Poisson bracket relations


for AA~ ~ ~

(44)

(45)

Here the numbers n, m vary in the "fundamental domain" 1 ~ m,n ~ N


w h e r e N is a length of the lattice and ~ is defined as follows

I i n > m

~ 0 n = m (47)

-i n < m

The f o r t u n a t e p r o p e r t y of our lattice f o r m u l a t i o n is that the rela-


tions (44) - (46) look as the most naive g e n e r a l i z a t i o n of the rela-
tions (15) - (17).

This luck continues in the c o n s t r u c t i o n of analogous of A n s a t z e (18)


(2O).

(48)

We let

(49)

~..~ 4- ~.~.~ ~i _ ,~_, (50)


176

5~, = 0-~..,)0. ~',,-.) = (51)


(~ , - ~ - , ) C c~,,,. _ ~_~ )

and a n a l o g u o u s l y for ~ .

Observe that 5~ is invariant under the t r a n s f o r m a t i o n s (ll),

The most striking property of the new v a r i a b l e s consists in the r e l a t ~ e


simplicity of their Poisson brackets; s t r a i g h t f o r w a r d calculations
give the f o l l o w i n g formulae

t F'., f,..t = i ( r""," - ~ ' . , . . . . ) O - r 2 ) ( ~ - e - ) (53

and

2-

(54

The last formula gives a lattice generalization of the V i r a s o r o algebr~


interesting in its own. In the continuous limit we have

~,,, = '~ S(~.) (551


t.t
so that the h a m i l t o n i a n
177

4
(56)

is a natural g e n e r a l i z a t i o n of (26).

The e q u a t i o n s of m o t i o n

(57)

g e n e r a t e d by the h a m i l t o n i a n are known to be c o m p l e t e l y integrable,


as was shown by S. M a n a k o v and M. Kac - P. van Moerbeke . In fact
they a p p a r e n t l y a p p e a r e d first in the e c o l o g i c a l papers of Volterra.
So the v a r i a b l e s S~ (and S ~ c o r r e s p o n d i n g to ~/~ ) constitute the
first step in c o n s t r u c t i n g the a n g l e - a c t i o n v a r i a b l e s for the L i o u v i l l e
model on the lattice. At this point we stop the d i s c u s s i o n of the
classical lattice model.

3. Partial q u a n t u m results.

C o n t i n u i n g to w o r k in the spirit of 151 we get the q u a n t u m v e r s i o n of


the lattice model via the c o n s t r u c t i o n of the L~-operator in terms
of the o p e r a t o r s ~ and " ~ ' ~ w i t h the usual c o m m u t a t i o n relations

This L~ must turn to (40) in the c l a s s i c a l limit and satisfy the fun-
damental commutation relation

& (L. L. > : ( L . ~ L. II<. {sg)

with a p a r t i c u l a r C - n u m b e r ~ x ~ matrix ~

The formula
178

~/~ I/~ "~/2

60)

/~e ~ e

gives such an object, the R - m a t r i x being

I1 0
0
0
0
e i~"
01
0 (61)
0 ei~" l_e2i~ 0
0 0 0 1

Now we literally repeat what was done before, namely introduce the
transport matrix

= (62)
and the operators

~, = A~ Bf I j v, : c., D I ' (631

chosing a p a r t i c u l a r order of the factors. It is g r a t i f y i n g to check


that these operators satisfy rather simple relations

(65)
179

~ ~d'~ -~- ~ 2d-~ ~- (4 -- ~ z ) ,%1- (66)

w here ~ = ~ 6~ . Note that the c o u p l i n g constant enters only


through ~ in a n a l o g y w i t h the S i n e - G o r d o n case.

Unfortunately we were not able to find natural generalizations of the


~ , p~ and S~, variables in the q u a n t u m c a s ~ so that the quan-
tum v e r s i o n of the lattice Liouville model is still in a n o n s a t i s f a c -
tory state. This was the reason w h y we did not p u b l i s h our r e l a t i v e l y
old results on the c l a s s i c a l version. However some new d e v e l o p m e n t s
and in p a r t i c u l a r new results of M. Jimbo 181 on the tensor products
of r e p r e s e n t a t i o n s of S k l y a n i n algebras 191 w h i c h one of the authors
(L.D.F.) learned during his visit to the LPTHE in april 1985 seem
encouraging. With this o p t i m i s t i c note we finish this paper.

References.

11 A.M. Polyakov, Phys. Lett. IO3B (1981) 207.


12 J.L. Gervais, A. Neveu, Nucl. Phys. B199 (1982) 59 ; B209 (1982)
125.
3 A. Kihlberg, R. Marnelius, ITP - G 6 t e b o r g preprint N82-2 (1982).
4 C.P. Dzhozdrhadze, A.K. Pogrebkov, M.K. Polivanov, Theor. Math.
Phys. 4__00(1979) 221.
51 L.D. Faddeev in Les Houches Lectures,vol.XXXIX North Holland(1984)
61 J.L. Gervais, preprint ENS, 1985.
171A.K. Pogrebkov, M.K. Polivanov in Soviet Science Reviews, Sec. C
(Math. Phys.) ed. S.P. Novikov, vol 3 (1985). Gordon and Breach.
181M. Jimbo, preprint RIMS (1985).
191 E.K. Sklyanin, Funct. Anal. Appl., vo117 (1983), P 34-48.