QUANTUM THEORY OF GRAVITY.

II

where we now use the propagator 8+ in place of G in front of the integral so as to obtain correct external-line wave
functions for the S matrix.
The only vertices which get inserted by the factor in square brackets in (20.36b) are the bare vertices S~ and
Vt;&e. Therefore (O, co $'~0, —co) may be expressed in the compact form (12.20), but with G replaced by S~,
~

provided the symbol 8/5 p is no longer taken literally but is understood to yield GSsG when acting on G and S~+t
when acting on S,
to have no effect on V &„;&s, and to insert (in all possible ways) into any 6ctitious quantum loop
"
merely one more vertex V&„&p Itaving lhe same orientatt'ott as att the other vertices already irt the loop. With this
understanding it is easy to see that (20.35) then yields also Eqs. (12.21), (12.22), and (12.23), with the modi6cation
G ~ N~ applied to all external lines. Chronological product amplitudes de6ned in this way may be used directly
in (13.6) to calculate the S matrix.
The consistency of these simple rules with previously obtained results is readily checked. For example, if non-
causal chains are reinserted into Figs. 2(b) and 3(b) the resulting primary diagrams for the lowest-order radiative
corrections to the one- and two-quantum amplitudes are precisely those obtained by the present prescription. We
note in particular the suKciency of the vertices S„and V~;~ p and the uniform orientation of the latter around any
6ctitious quantum loop.
"It will be noted that the operators S/by', when redefined in this way, are still commutative.

PHYSICAL REVIEW VOLUME j. 62, NUM BER 5 25 OCTOBER &967

Quantum Theory of Gravity. III. Applications of the Covariant Theory*

BRvCE S. DEWITT
Institute for Advanced Study, Princeton, itIeto Jersey

Departnient of Physics, University of North Carolina, Chapel Hill, Iiorth Carol&tat

(Received 25 July 1966; revised manuscript received 9 January 196'I)

The basic momentum-space propagators and vertices (including those for the fictitious quanta) are
given for both the Yang-Mills and gravitational fields. These propagators are used to obtain the cross
sections for gravitational scattering of two scalar particles, scattering of gravitons by scalar particles,
graviton-graviton scattering, two-. graviton annihilation of scalar-particle pairs, and graviton bremsstrah-
lung. Special features of these cross sections are noted. Problems arising in renormalization theory and the
role of the Planck length are discussed. The gravitational Ward identity is derived, and the structure of
the radiatively corrected 1-graviton vertex for a scalar particle is displayed. The Ward identity is only one
of an infinity of identities relating the many-graviton vertex functions of the theory. The need for such
identities may be eliminated in principle by computing radiative corrections directly in coordinate space,
using the theory of manifestly covariant Green's functions. As an example of such a calculation, the con-
tribution of conformal metric fluctuations to the vacuum-to-vacuum amplitude is summed to all orders.
The physical significance of the renormalization terms is discussed. Finally, Weinberg s treatment of the
infrared problem is examined. It is not difBcult to show that the fictitious quanta contribute negligibly to
infrared amplitudes, and hence that Weinberg s use of the DeDonder gauge is justified. His proof that the
infrared problem in gravidynamics can be handled just as in electrodynamics is thereby made rigorous.

1. INTRODUCTION' canonical or Hamiltonian theory and the other on the

'N the Grst two papers of this series' two distinct manifestly covariant theory of propagators and dia-
- mathematical
grams. So far no rigorous mathematical link between
approaches to the quantum theory
the two has been established. In part this is due to the
of gravity were developed, one based on the so-called
kinds of questions each asks. The canonical theory
leads almost unavoidably to speculations about the
*This research was supported in part by the Air Force OfBce meaning of "amplitudes for diferent 3-geometries" or
of Scientific Research under Grant AFOSR-153-64 and in part by
the National Science Foundation under Grant GP7437.
"
"the wave function of the universe. The covariant
ts Permanent address. theory, on the other hand, concerns itself with "micro-
B. S. DeWitt, Phys. Rev. 160, 1115 (1967); preceding paper, processes" such as scattering, vacuum polarization, etc.
ibid. 162, 1i95 (1967). These papers will be referred to as I and
II, respectively. The notation of the present paper is the same as Some of the questions raised by the canonical theory
that of II, which should be consulted for the definition of un- were explored in I. In this third and 6nal paper of the
familiar symbols, e.g. , S„ for the n-pronged bare vertex and
V(;)p for the asymmetric vertex coupling real and fictitious series we examine some of the consequences of the
quanta. covariant theory.
1240 BRYCE S. DrKITT
Armed with the formalism constructed in II one can rather than in momentum space. An example of such
in principle carry out the calculation of any "micro- a calculation is given in Sec. 7, where the contribution
process" to any order of perturbation theory in a of conformal metric Quctuations to the vacuum-to-
manner which is completely invariant and unambigu- vacuurn amplitude is summed to all orders. The calcu-
ous except for the arbitrary high-energy cutoG which lation, which is manifestly covariant throughout,
must be introduced to render divergent integrals finite. makes use of an integral representation for the ampli-
A few of these calculations have actually been per- tude. A resume is given of that part of the mathe-
formed, and the only thing which prevents more of matical theory of covariant Green's functions which is
them from being done is the extreme tediousness of needed.
the algebra involved and the )ack of any experimental Section 8 concludes the paper with a review of
motivation for them. It is a pity that Nature displays Weinberg's treatment of the infrared problem (see Ref.
such indiQerence to so intriguing and beautiful a sub- 37). If Yang-Mills quanta are assumed to be massless
ject, for the calculations themselves are of considerable then, since they can act as their own sources, they give
intrinsic interest. The present paper contains several rise to the special infrared divergences which plague
examples. They are by no means exhaustive but have massless electrodynamics. Weinberg showed that gravity
been selected as useful landmarks in a still largely miraculously escapes these difhculties; its infrared di-
unexplored territory. Not all of these were originally vergences can be handled by the standard methods
carried out by the author, but it is hoped that their familiar in ordinary quantum electrodynamics. His
unified presentation here will make their. results more proofs, however, were incomplete, since he did not
accessible than hitherto. have available a fully elaborated quantum theory. In
Section 2 begins with the rules of calculation in particular he used the DeDonder gauge without taking
momentum space. The basic structural elements of the into account the Gctitious quanta. It is not diQicult to
theory, namely the propagators for real and Gctitious show that the fictitious quanta contribute negligibly
quanta, the vertices S3, S4, U~;)p, and the coupling in the infrared limit. Weinberg's results are therefore
with matter fields, are given for both the Yang-Mills rigorous.
and gravitational fields. The standard Feynman rules
are summarized. The results of a few lowest-order 2. RULES OF CALCULATION IN
scattering calculations based on these rules are given MOMENTUM SPACE
and discussed in Sec. 3. Included are the cross section We begin with the vertex functions for the Yang-
for gravitational scattering of two scalar particles, the Mills Geld interacting with itself. We have seen in II
cross section for scattering of gravitons by scalar that when the standard field variables are used only
particles, the corresponding annihilation cross section, S3 and S4 are nonvanishing for this case. In momentum
and the graviton-graviton cross section. Section 4 is de- space these become (apart from a 8 function expressing
voted to the problem of gravitational bremsstrahlung. conservation of momentum)
The role of the energy quadrupole moment tensor and
the absence of the forward peak at high energies, charac-
teristic of photon bremsstrahlunt„are noted.
~ic.a f(p'» "
p"»)q—
8A „Ri&', 82~",
Section 5 discusses some of the problems which arise
in renormalization theory. Although the Yang-Mills + (p"" p") ~'"+ (p' —p")~»"3 (—
2 ~)
theory looks as if it may be renormalizable (provided S4S
its infrared difficulties can be disposed of), quantum
gravidynamics is definitely not renormalizable in the
c,
— ', („'g"'—g '„"')
usual sense. Tentative proposals for dealing with this —c „c'pg(g "g"—g 'q'")
situation are brieRy described, as is also the evidence
—
that gravity contains its own cutoG at the Planck
"
c-~.c'.~(n"'V'— n""n")—
length. Illustration of the actual details of the re- (2.2)
norrnalization program, by explicit calculation of a
radiative correction, is postponed to Sec. 7. The correspondence of momenta with indices is pap,
The gravitational Ward identity and its implications p'P'»', p"p"a", p'"&"'r"'. All momenta are incoming
for gravitational form factors are derived in Sec. 6. (to the vertex), and momentum conservation implies
The general structure of the radiatively corrected 1- p+p'+p"=0 for S3 and p+p'+p"+p'"=0 for S4.
graviton vertex is displayed in the case of a scalar Indices on the structure constants are raised and
particle. It is emphasized that the gravitational Ward lowered by means of the Cartan metric p p. When all
identity is only one of an infinity of identities relating indices are in the lower position the structure constants
the many-graviton vertex functions of the theory. The are completely antisymmetric.
need for Ward identities can be eliminated by com- In addition to the above vertices, the fictitious
puting radiative corrections directly in coordinate space vertex V~;~p is needed for the calculation of radiative
 QUANTUM THEORY OF GRAVITY. 11I
corrections. For the Yang-MiOs Geld it takes the form 6eld are much more complicated. In this case we shall
employ the momentum-index combinations pq4, p'OT'',
V( T""')p ~ —
ic.p, P"= —
ic.~p(P +P'") . (2.3) p"p'9", p"'4'"z"'. The vertices must not only be sym-
metric in each index pair but must also remain un-
The propagators for the normal and Gctitious quanta
changed under arbitrary permutations of the momen-
are, respectively, turn-index triplets. At least 171 separate terms are
G ~ v'v"/p', (2.4) required in the complete expression for S3 in order to
exhibit this full symmetry, and for 54 the number is
G ~ ~aP/P2 (2 5) 2850. However, these numbers can be greatly reduced
by counting only the combinatorially distinct terms~
with p' being understood to have the usual small and leaving it understood that the appropriate sym-
negative imaginary part. metrizations are to be carried out. In this way S3 is
The corresponding quantities for the gravitational reduced to j. 1 terms and S4 to 28 terms, as follows:

~ Pyv~'Po' r'~ P p" X"

symL l~.(p—
A""n"~ ") lI'4(p -p'~""~ ")+lI'4(p AP ~" ~ ")+lI'4(p A""~"~ ")+~4(P'P"~""~")
'I'4(p'P "-0" 0"")+'I'3(p'P "-0"'6"')+'~4(p'P"0"'0"')+~ (P'P "n'"0"')+f' (P'P'"0"n"")'
I' (p -p'~" ~"~"")j, (26)

ql p" y" ~pc"'x"'

symL —l & (p p'~""0 "~'"~'")—l ~ (p'p'~""~'"~'") —:I' (p'p'"~"—'I'"~'")+lI' (p p'~"'~"'~'"~'")
y ~P4(p pg»g«g pic~*)+ —
'P»(prpr7J»g pcg~ 4)+ 'P, (prp'-pyJvrrj pcg& 4) 4p, (p —
p'gl rqvr. g pc/&v)

+4~24(p P'n""n"n'"n")+'~24(p'P'n"'n""~")+ '&»(P'P'"n"-'n"'n'")+&24(p P"~'"n""~'")

'"~"~""~'")+i~»(p p'0'"9""n'") —2&24(p p'~""~"~"'~")
'~"'I"~""~") k&»(p'p—
kI'»(p p—
p2 (prp P»(p pp Agv cgKP gru') I 24 (p p pg rcrjrpgv x) p~ (p pp cg r priv cc)
rgv pg x cg Kcc)
Alvin

+I'4(p p'~" p~" —

~ "~"") I'»(p'p'~""~ "~"") 2I'»(p p'~" p— ~""~"~'") I'»(p'p'—~'"~"'n"")
P (ppp c~hcc~pcr~vr) P (pap p~rp~vc~ccX) P (pcrpvp~rp~Xc~rv)+2+ (p. p ~vcr~rp~kc~ccp) j (2 7)

The "Sym" standing in front of these expressions indi- him best we shall not shackle him by describing one
cates that a symmetrization is to be performed on each here. V(e also make no attempt to display S& or any
index pair pp, a.r, etc. The symbol P indicates that a higher vertices.
summation is to be carried out over all distinct permu- The vertex V(;)p has the following form for the
tations of the momentum-index triplets, and the sub- gravitational Geld:
script gives the number of permutations required in aIIr" )v'
(p
each case.
Expressions (2.6) and (2.7) can be obtained in a ,'Symrt2P" pP"8„'—
-P"pP'„g"
straightforward manner by repeated functional diGer- +(p.p" p'.p )4'+p'P'4 —'j & (2 g)
entiation of the Einstein action. This procedure, how- where the momentum-index combinations are pp, PY,
ever, is exceedingly laborious. A more eKcient (but p"0"T", and the symmetrization is to be performed on
still lengthy) method is to make use of the hierarchy the index pair o.r. The propagators for the normal and
of identities (II, 17.31). It is a remarkable fact that Gctitious quanta are given by
once 52' is known all the higher vertex functions, and
hence the complete action functional itself, are de- G ~ (~P.n-+~P. ~- n»~-)/P'-(2.9)
termined by the general coordinate invariance of the
G ~n""/P'. (2.10)
theory. It is convenient, in the actual computation of
the vertices via (II, 17.31), to invent diagrammatic ' The choice of termsis not completely unique since momentum
conservation may be used to replace a given term by other terms.
schemes for displaying the combinatorics of indices. We give here what we believe (but have not proved} to be the
Since each reader will devise the scheme which suits expressions containing the smallest number of terms.
1242 BRYCE S. D EW I TT I62
If one wishes to calculate processes involving the the mass shell, to precisely the forms (2.13).and (2.15)
interaction of the Yang-Mills and/or gravitational regardless of the magnitude of the particle spin. This
Geld with matter, additional vertices describing this may be proved in each instance as a straightforward
interaction must be included. As prototypes of such consequence of the gauge invariance of the theory and,
vertices, we shall display those which arise from inter- when extended to the radiatively corrected vertices,
actions with scalar (or pseudoscalar) particles. The constitutes a boundary condition on the Yang-Mills
latter particles contribute to the total action functional and gravitational form factors. ' LSee also Sec. 6. j
an expression of the form It is to be emphasized that the inclusion of addi-
tional fields in no way affects the formal theoretical
1 structure developed in II. The topology and invariance
S = ——g" (p , „y., &+. m'pq)dh, (2.11)
properties of diagrams remain completely unchanged.
2
j,
One simply permits the field indices i, etc. , to extend
where the covariant derivative is defined in Table I over a greater range of values in order to accommodate
of II and where the components of the new fields which have been
= v v, 7Gv'= —G added. The only differences are differences of detail
v— (2.12) such as, for example, the sign modifications due to
statistics which appear when some of the added compo-
y being the matrix which connects the two forms of a
nents are those of fermion fields, or changes in the
self-contragredient representation (of the Yang-Mills
Lie group) generated by the matrices G and — G
structure of the invariance group which arise from
respectively. We Gnd having both the Yang-Mills and gravitational fields
simultaneously present and interacting with each other. 4
The rules for combining vertices and propagators
(2. 13) into transition amplitudes are completely standard.
5g'8qW. "„~ With the notational conventions of the present paper
they may be summarized as follows: (1) An expression

"„"BAt'"'„-
~(G.G&+Gg.) qs", such as (2.1), (2.2), (2.3), (2.6), etc. , for each vertex;
bg'8yhi (2) an expression such as (2.4), (2.5), (2.9), (2.10),
etc. for each propagator; (3) a factor ( — i)/(2~)4 for
O'S„
each independent closed loop; (4) an additional factor
(— 1) for each closed ferrnion or fictitious-quantum
loop, or when necessary to assure antisymmetry of
84S„ fermion amplitudes; (5) an over-all factor t'(2e. )4 times
a 8 function assuring total energy-momentum con-
servation; (6) a wave function u'~ (see Table II of
II) or its complex conjugate evaluated at x=O for each
external line; (7) integration over all the independent
+ (P"P"+O'P'") n"'+ (P"P"+O'P'") ~" momenta.
+(P"P"+P P'")n" +(P"P"+P P'")~"' Gauge invariance may be invoked as a useful con-
"
(P"P'"+P"P'")~— (P'P"+P'P")—
~"".
sistency check in all calculations. However, it must be
applied to the entire amplitude for a given process
and not merely to a single diagram. It is therefore
The corresponding vertices which describe the inter- algebraically more laborious than corresponding checks
action of the gravitational and/or Yang-Mills 6elds in electrodynamics. It is no longer possible to exploit
with particles having spin are obtained by straight- lines
charge conservation by following individual
forward computation from the pertinent action func-
tional. The latter is obtained in each case via the 'These are analogs of the electromagnetic form factors. The
"principle of minimal coupling" (which, in the case gravitational form factors are also sometimes referred to as
of gravity, is nothing but the "strong equivalence stress-energy, mass, or mechanical form factors.
4 These are, in fact, the most important differences. It is worth
principle" ) from the corresponding action functional mentioning that when fermion fields are included it is usually
in the absence of gravitational and Yang-Mills fields, convenient to replace the metric field g„, by a eierbein field.
Otherwise the group transformation laws are no longer linear.
by replacing ordinary derivatives by covariant deriva-
tives, the Minkowski metric p„„by g„„, and the volume
(See B. S. DeWitt and C. M. DeWitt, Phys. Rev. 8?, 116 (1952}.
Ke also mention that the combined vierbein-general-coordinate-
j
element dx by g'~' dx. We do not give here the results transformation group has the structure of a semi-direct product
of such calculations for particles with spin but merely based on the automorphisms of the eierbein group under general
coordinate transformations. In the combined group only the
point out (what is more useful for the reader) that the vzerbein group is an invariant subgroup. The coordinate trans-
three-pronged vertices, when sandwiched between nor- formation group is its factor group. Similar statements apply
to the combined Yang-Mills-general coordinate-transformation
malized wave functions, always reduce, in the limit of group. The analysis of these cases is therefore correspondingly
zero momentum transfer, with particle momenta on coxnplicated.
 QUANTUM THEORY OF GRAVITY. III
through diagrams, for now the conserved quantity— which permit the amplitude to be recast in the form
—
j Too(1) Tpp(2) —4Tpi(1) Toi(2) —4Tos(1) Too(2)
Yang-Mills charge, energy-momentum leaks all over
4
every diagram. Moreover, when Yang-Mills quanta or
gravitons interact with themselves, the closed loops
—Tps(1) Tps (2) —Tpp(1) )Tii (2)+ Tss(2) 7
form traKc jams of spurious charge which can be un- —t T» (1)+ Tss (1)7Too(2) }/iI'+4 ( LTii (1)—Tss(1)7
snarled only by calling the fictitious quanta to the X LT»(2) —T»(2)7+4T»(1) T»(2) }/q'
rescue. +exchange and virtual annihilation terms. (3.9)
3. SCATTERING CROSS SECTIONS The first term yields an instantaneous "Newtonian"
interaction, while the second gives rise to a "delayed"
We now display some of the lowest-order amplitudes
interaction propagated by transverse gravitons. In this
and scattering cross sections which the covariant theory
case the factors which couple separately to the two
yields. One of the simplest is the amplitude for the
states of linear polarization are Tgg
scattering of two identical scalar particles by exchange
respectively.
of a single Yang-Mills quantum. This has the form
From (3.6) it is straightforward to compute the
—
i(2m) 'B(Pi'+Ps' —Pi —Ps) (1) (2)/q' j j differential cross section for gravitational scattering
of identical scalar particles in the center-of-mass frame.
+exchange and virtual annihilation terms, (3.1)
One finds'
where
do. G' E'-( I+3p') (1 —e')+4w'(I+s') cos'(8/2)
pi pl ps ps (3.2)
j „=,'i (E'E) '-"X'tyG, X(p'„+p„), (3.3) dO 16 ps Slils (8/2)
the X's being the internal (Yang-Mills group) states of (I+3vs) (1 —us)+4s'(I+ps) sin'(8/2)
the particles and the remaining notation being con- n' cos (8/2)
ventional. The same form (3.1) also holds for particles
with spin, but the expression for the current „is then j + (3 —vs) (I+e')+2v4 sin'8, (3.10)
more complicated.
Since the initial and final momenta are on the mass where v= p~/E and the gravitation constant has been
shell we have the conservation laws ~

reintroduced through the units convention 16xG=1.

j-(1) q= j-(2) q=o, (3 4) The nonrelativistic and extreme relativistic limits of
this cross section are, respectively,
which permit the scattering amplitude
in the form
j-o(1)j o( ) j-i( )j i(2)+j-s(1)j s(2)
to be reexpressed
/do)
(dQ)Na
do'
GsBP

16
1
. ,
ossinssi8
+, cocos-8
1
+3
2

(3.11)

Q2 (f2
= 4G' E'p cto'-,
' +8tan'-,'8+-', + p,
»n'87'. (3.12)
+exchange and virtual annihilation terms, (3.5) kdnl aR
where a factor i(2m-) '5(pi'+ps' — pi —ps) has been re- In a similar manner one may compute the cross
moved, and the 3-axis has been chosen in the direction section for scattering of gravitons by scalar particles.
of the spatial part q of the space-like 4-vector q. The The relevant diagrams are shown in Fig. 1, the heavy
first term of (3.5) represents the instantaneous "Cou- lines denoting particles and the light lines gravitons.
lomb" interaction of the particles; the second repre- Diagrams (a) and (b) vanish in the rest frame of the
sents a "delayed" interaction propagated by transverse target particle, and one finds for unpolarized gravitons'
quanta, the factors andj j
2 being separately coupled
&

to the two states of linear polarization of these quanta.

der G2m'
Lcoss-', 8+ sin'-,'8
The corresponding amplitude arising from exchange dfI L1+2o sin'-'87' sin'-,'8
of a graviton is
+2o(coso-'8+sinois8) sin'8
s(4 ) '8(p '+p '
p p)T"(1)-- + os(cos~is8+sin4-,'8) sin487, (3.13)
X("-+' ~- ~'-)T.,(2)/q- (da )
+exchange and virtual annihilation terms, (3.6) (cos'-,'8+sin'-'8) (3.14)
sin4-,'8

.
kd II 9 Na
where
T"=!(E'E)-'"Lp.p'. +p.p'. —.(p p'+ )7. (3.7)
(do ) = 4Gsms (cos4-,'8+ sin'-', 8) cot'~s8, (3.15)
EdII i Ea
Again we have conservation laws
~ C. F. Cooke, Ph. D. thesis, University of North Caro1ina, 1964
T~. (1)q"= T, (2) q"= o,. (unpublished),
 B RYCE S. 0 H& I TT
demonstrated only in lowest order, by carrying out a
brute-force computation of the relevant amplitudes.
The tediousness of the algebra involved in obtaining
the graviton-graviton cross section may be inferred
from the complexity of the vertex functions (2.6) and
(a) (b) (c)
(2.7) which are involved in the diagrams which repre-
Fn. 1. Lowest-order diagrams for scattering of a graviton by a sent the amplitude (Fig. 1 with the heavy lines replaced
material particle. The heavy line denotes the particle and the
light lines denote gravitons.
by graviton lines). Fortunately, the presence of the
polarization tensors in the external-line wave func-
tions, and the momentum condition Ps=0 for free
where & is the energy of the graviton measured in units
of m. It will be noted that these cross sections have no quanta, eliminate many of the terms from these ex-
pressions. Nevertheless, a large amount of cancellation
resemblance to those for Compton scattering, but on
between terms still has to be dug out of the algebra,
the contrary, continue to display the sharp forward
and this, combined with the fact that the final results
"Rutherford peak" characteristic of long-range inter-
are ridiculously simple, leads one to believe that there
actions. This feature is due to diagram (d) of Fig. 1
must be an easier way. The cross sections which one
whose presence, as may be readily checked, is essential
finds are
for the gauge invariance of the scattering amplitude.
Owing to the equivalence principle gravitons, like 80' cos'-'8 sin'-'8
de++
photons, are deflected by a gravitational Geld (in par- 4@2~
ticular by the long-range static field of any material sin'-'8 cos'-'0
particle), and the above cross sections are dominated
by this eGect. +4—-,'sin'8 (3.19)
By the well-known "substitution rule" the diagrams
of Fig. 1 yield also the amplitude for annihilation of a do+ cos"-,'8
pair of scalar particles into gravitons. We record here 462+2 (3.20)
dO sin4-, 8
only the low- and high-energy limits of the total
annihilation cross section in the center-of-mass frame: showing again the forward Rutherford peaking.
O'NR = 2m'G m /'v, (3 16) We shall not record here the corresponding cross
sections involving Yang-Mills quanta, since these
o RR —(387r/3)(M'. (3.17) depend, in their finer details, on which Lie group is
chosen as generator of the Yang-Mills group and on
The cross section for the inverse process, namely, the which representations are chosen for the material
production of a scalar pair by colliding gravitons particles. There is also a serious difhculty with the
(again in the center-of-mass frame) is identical with Yang-Mills field in regard to the infrared catastrophe,
(3.17) at high energies. Near threshold, on the other which will be discussed in Sec. 8. Since our primary
hand, it is given by interest in this article is the gravitational field, we
o = 2mG'm'(s' —1)"'/s, s& 1. (3.1&) refer the reader with a special interest in Yang-Mills
cross sections to the dissertations of Remler' and
The only elastic process which remains to be con- Dotson. 7 It is, however, perhaps worth remarking that
sidered is the scattering of one graviton by another. in the case of the scattering of one Yang-Mills quan-
This process has some unusual features. It turns out tum by another the phenomenon of helicity conserva-
that the helicity of the colliding gravitons is imdi~idgal/y tion is again found to hold, with or without the in-
conserved. That is, there is no spin Qip, in spite of the clusion of graviton exchange forces in the total ampli-
presence of derivative coupling. If both gravitons are tude, and regardless of the choice of the Lie group.
right (left) handed before collision then both are right Moreover, an extension of the helicity conservation
(left) handed after the collision. If one is right handed rule to processes involving real gravitons in interaction
and the other left handed then they maintain this with Yang-Mills quanta apparently exists. Thus indi-
relationship also, through the collision. vidual helicities remain unchanged when a graviton
The helicity of extremely relativistic particles, and and a Yang-Mills quantum collide elastically. If the
of massless quanta in particular, is notoriously rigid. diagrams contributing to this process are turned on
In the classical theory, for example, the spin of such their sides so as to yield the amplitudes for annihilation
a particle seers no precession under geodetic motion of two Yang-Mills quanta into a pair of gravitons (or
in an external gravitational held but remains always the reverse process) further selection rules emerge. One
pointing parallel or antiparallel to the trajectory. How-
ever, no general principle has yet been discovered ~E. A. Remler, Ph. D. thesis, University of North Carolina,
1964 (unpublished).
which implies that helicity conservation must hold to 'A. C. Dotson, Ph. D. thesis, University of North Carolina,
all orders of perturbation theory. It has so far been 1964 (uupublished).
 QUANTUM THEORY OI" C RA V I TY, III

6nds that it is impossible to produce two gravitons and simply multiplies the original amplitude. This
having opposite helicities by annihilation of Yang-Mills limiting form actually holds for all external lines, re-
quanta, or conversely to produce a Yang-Mills pair gardless of the spin character of their associated
having opposite helicities by the reverse process. The particles. It even holds when the external line is a
quanta in both the initial and anal states must have graviton line, provided the emission vertex is inserted
identical helicities if the amplitude is to be nonvanish- not merely into a single diagram but into the sum of
ing. Helicity selection rules exist even for the process all diagrams contributing to the original amplitude.
in which two Yang-Mills quanta coalesce to produce This may be verified in a straightforward manner by
a single Yang-Mills quantum and a graviton. If both plugging in the 3-pronged graviton vertex (2.6) and
initial quanta have the same helicity the final quanta eliminating the terms involving q. Of the remaining
must have this helicity too; if the initial helicities are terms only those survive which yield a net contribution
opposite the 6nal helicities must be opposite. The same of the form (4.3); the rest disappear in virtue of the
obviously holds for the reverse process. gauge invariance of the total original amplitude.
The rnultiplicative factor (4.3) exhibits the well-
4. GRAVITATIONAL BREMSSTRAHLUNG known infrared divergence and can be obtained from
a purely classical model. We note that the infrared
Since the problem of gravitational radiation from divergence shows up only when the emission takes
accelerating masses has bedeviled classical relativists place from lines on the mass shell; it does not occur
for years it is a pleasant surprise to discover that its when the emission is from internal lines of a scattering
treatment within the quantum framework is quite diagram. The external lines therefore dominate the
simple. ' Consider a scattering diagram in which one of soft graviton emission. This means that the precise
the lines represents a scalar particle (real or virtual) details of the scattering process have little relevance
of momentum p. Let the diagram be modified by the in the limit q — +0, and that the long-wavelength end
emission of a graviton of momentum q from this line. of the emission spectrum is determined primarily by
If the momenta of all lines subsequent to the inserted the asymptotic trajectories of the incoming and out-
graviton vertex are held fixed while those prior to the
going particles, just as in the case of photon brems-
vertex are adjusted in such a way as to conserve strahlung. For wavelengths large compared to the
momentum and keep external lines on the mass shell, space-time region in which the collision takes place

|
then the only additional eGect of the graviton emission
(the size of this region is determined by the magnitudes
is to introduce into the corresponding amplitude, a of typical energies exchanged in the collision) the effec-
factor tive graviton source is a stress tensor of the form
e ~*e "*
T~"(x) = g zl„m„v„~v„" 8(x —V r)dr, (4.4)
(2zr)@' gq'
P.(P.+q.)+P.(P.+q, ) n;L~'+P +q) 3
(P— which idealizes the particles to classical points colliding
X , (4.1)
(P+ q) z+ mz —z0 at the coordinate origin. Here nz and V are, respec-
tively, the mass and 4-velocity of the eth particle,
which follows from Eq. (2.15) and Table II of II. and the sign factor q„ tells whether the particle is in-
Alternatively, if the momenta prior to the vertex are coming or outgoing. The summation is over all the
held fixed we get a factor which differs from (4.1) by external lines, and the velocities are subject to the
the replacement q — + q. — energy-momentum conservation law:
If the graviton is emitted from an external line these
factors reduce to g &„m„V„=0. (4 S)

where
ey" ey"
gq'
(2~)zl'
1 ppp +2'g(ppq

—
q'+2rlP q
zi=+1 or 1 according as
+p qp
i0—ffp

the external line is

p'q)
(4.2)
The classical emission spectrum is obtained by pro-
jecting (4.4) onto the graviton wave functions I„„+(x,q)
(see Table II of II). The corresponding quantum
amplitude is
outgoing or incoming, and p is held fixed on the mass
shell. In the long-wavelength limit q~
0 (4.2) itself —,z N„„g*(x,q) T~" (x)dx
reduces to

2(2zr)z"Qq' P
(e~* p)'
iq0— (4.3)
&.m„(e,* V.)'
2 (2zr)'"Qq'
dS dr e '~ *b(x Vr)'—
q
The comparison is actually unfair. The questions which
g„m„(eg* V„)'
classical relativists ask are usually quite detailed —e.g. , the precise ~ 2(2zr)'"Qq' V~ q
—zg~0 (4 6)
damped motion of radiating sources, or the nonlinear properties
of coherent waves of large amplitude —
and are inevitably much QThe gauge invariance holds in every order of perturbation
more dif5cult. theory.
 B RYCE S. D EK I T T
which, in view of the relation p =m„V„, is just (4.3) resemble bundles of plane waves having momenta con-
summed over all the external lines. fined to narrow cones. These bundles (particularly
When the collision is nonrelativistic (4.6) reduces to their outer regions) have diKculty readjusting to the
altered particle trajectories arising from the collision
—
s'(2e. q') "'eg* AZ eg', (4.7) and hence partly escape as radiation.
where the graviton gauge is chosen so that the compo- In the gravitational case the sharp forward emission
nents e+' of the polarization vectors e+ vanish, and
"
is absent. In fact for an extremely relativistic collision
hg is the change in the spatial integral of the total (~y„~ =E„) which is confined to a plane (e.g. , 2-
3-stress dyadic as a result of the collision: particle scattering) it is easy to verify that the total
sum (4.6) yields an amplitude which vanishes for
"
emissionie the plane. This implies that, unlike photon
AX=A Tdx= g r)„y„v„, emission, graviton emission is a cooperative phenome-
non which cannot be traced to the individual particle
T= P 8(r)„x')p„v„b(x—v„x'),
fb
(4.9) fields. Indeed the real gravitational 6eld of a particle,
namely the Riemann tensor, falls oR as the inverse
—V„/V '=p /E. =p /m„.
v„= (4. 10) cube rather than the inverse square of the distance,
and hence its outer regions contribute negligibly to the
Now it is well known" that energy-momentum con- emission. This has obvious implications for investiga-
servation permits the integral of the 3-stress dyadic tions of classical 2-body radiation as well as for at-
to be reexpressed as one half the second time derivative tempts to introduce Weizsacker-Williams approxima-
of the second moment J'xxTMdx of the energy density. tion schemes into quantum calculations.
Moreover, since e+* e+*—0, the trace of hZ may be
removed from (4.7).u Therefore the emission amplitude S. RENORMALIZATION A5'D THE
may be written in the alternative form: PLAJf CK LENGTH
'(2e.q') —s"e„* a(d'Q/dP)
—, e~* (4.11) In lowest-order perturbation theory the formal rules
of the manifestly covariant theory yield results which
where 4(d'Q/dP) denotes the change in the second agree with the classical theory in the correspondence
time derivative of the energy quadrupole moment principle limit. In higher orders, divergences appear,
tensor just as they do for other 6eld theories, and almost
nothing is known about how to extract Gnite and
Q= (xx —-', 1x') Tpgdx, (4.12) physically meaningful radiative corrections from the
results. In the case of quantum gravidynamics the
severity of the divergences is such that the theory is
showing that soft gravitons are emitted predominantly not, by standard criteria, renormalizable. This is due
in the quadrupole mode. to the quadratic momentum dependence of the vertices
It is of interest to examine the angular distribution S„(e~&3), which in turn may be traced to the de-
of the emitted radiation. From (4.6) one sees that each pendence of the light cone on the background field,
external line makes a contribution to the emission i.e., to the 6eld dependence of the coeKcients of the
amplitude, which has an angular distribution of the second time derivatives appearing in S2. Thus by
'fol m counting momentum powers one Gnds for the super-
sin28 6cial degree of divergence of any diagram
(4. 13) D= 2L;+2 Q V„+—
1. —e cos8 4K, (5.1)

where 8 is the angle between v and q, and y is a helicity where I.; denotes the number of internal lines, V„ the
phase angle. In the case of photon bremsstrahlung the number of m-pronged vertices, and E
the number of
sine appears linearly instead of quadratically in the independent momentum integrations. Now it is not
numerator, with the consequence that for relativistic diQicult to show that
collisions (v=1) the emission is concentrated sharply
in the forward directions of all the particles (initial as %=I.—Q V +1. (5.2)
well as final). This peaking may be attributed to the
individual I.orentz-contracted Coulomb fields, which 12This was erst pointed out by R. P. Feynman in a mimeo-
graphed letter to V. F. V/eisskopf dated January 4 to February
11, 1961 (unpublished).
o See, for example, L. D. Landau and E. M. Lifshitz, The ' Introducing unit vectors 0 and 0 in the directions of q and
Classical Theory of Fields, translated by M. Hammermesh p„, respectively, one may write the amplitude in this case in the
(Addison-Wesley Publishing Company, Inc. , Reading, Massa- form
chusetts, 1962), rev. 2nd Ed. " =constX
(ttxa„)s
~ In view of the nonrelativistic energy conservation lair constX Pe& Pe
S
Z (1+0 0),
& 8
e (m +-', y„v ) =0, this trace is just twice the rest mass lost
in the collision and already vanishes for elastic collisions. which vanishes by energy-momentum conservation.
162 QUANTUM THEORY OF GRA VI TY. III 1247

Therefore light on the ultraviolet problem. Faced with its brutal

D=2(K+1), E~&1, (5.3) consequences there are several paths one may try to
follow to make life bearable. One of these is to soften
which increases without limit as the order of the
the degree of divergence by abandoning S-matrix uni-
diagram increases. tarity (i.e. , positive de6niteness of Hilbert space)
In the case of the Yang-Mills field the situation is through the introduction of field equations of the fourth
better. Here we have
—Vs+4K
D= 2L+ (5 4) terms of the form f f
differential order. This may be accomplished by adding
g'"&"R'dx and g'i'R &&"dx to
the Einstein action, which changes Eqs. (5.1) and (5.3)
which, in combination with the readily verified com- to
binatorial law
D= —4L;+4 Q V„+4K=4 (5.7)
L,+2L;= Q nV„=3Vs+4V4, (5.5)
for all diagrams of order greater than zero. Such a
yields procedure in eGect introduces a separate unit of mass
a=4 —I„ (5.6) (or length) into the theory, and if this mass is chosen
where L, is the number of external lines. In order to suQiciently big, the S matrix will be nearly unitary,
compensate the divergences one may introduce counter significant departures from unitarity occurring only
terms into the original Lagrangian in the conventional under extreme conditions, when collision energies ap-
manner. The most divergent counter term is always the proach that of the unit of mass.
simplest. It is necessarily of the form const)&F „„F &", Nevertheless, it would be nice if any breakdown in
provided the divergence has been handled in a mani- conventional ideas which may be necessary were to
festly group-invariant manner. But such a counter emerge from the quantized Einstein theory in its un-
term can be detected by as many as four external mutilated form. There is already a unit of mass in the
lines. Hence D is never actually greater than zero. theory: the absolute unit (hc'16sG)'i'=3. 07X10 'g
This is quite analogous to the situation which occurs 10" BeV, and one is loath to introduce another.
with vacuum polarization diagrams in quantum elec- One might try to use the extra mass merely as a
trodynamics: Although L„=2, D is reduced from 2 to regulator, in a spirit similar to that of the $-limiting
0 by gauge invariance. One therefore expects that, proposal of Lee and Yang" for the charged vector
with proper handling of the overlapping divergences, boson. Equation (5.7) suggests that the regulated
a careful analysis will show that none of the ultraviolet theory may be renormalizable, requiring only three
divergences of the Yang-Mills theory is worse that counter terms, respectively, quartically, quadratically,
logarithmic, and that only a single counter term, of and logarithmically divergent. If this is so one might
the above mentioned type, is needed for each group- attempt to let the regulator mass become infinite
invariant set of diagrams. If that is true it is then not after the renormalization has been performed. How-
difficult to show that renormalization merely rescales ever, there is no guarantee that the renormalized am-
the structure constants c p~. Mathematically, this is plitudes will themselves remain finite in the limit, nor
equivalent to a rescaling of coordinates in the group that unitarity, which is violated by the regulator, will
manifold; physically, it corresponds to a change in the then be restored. If unitarity stays violated one is not
strength of the coupling of the Yang-Mills field to sure whether this represents a fundamental feature of
itself and to other Yang-Mills-charge bearing fields. the quantized Einstein theory or is merely a conse-
It should be remarked that although the above quence of the regulator approach; one would be inclined
results )Eqs. (5.3) and (5.6) $ have been stated for the to suspect the latter. Halpern" has criticized unortho-
case in which each field interacts only with itself, they dox uses of regulators in handling nonrenormalizable
hold also when other fields bearing stress-energy or field theories, and has shown that they often lead to
Yang-Mills charge are present, provided the spins of illegal modifications of analytic properties.
the added fields are not greater than and their other
~~
If regulators are to be excluded then perturbation
mutual interactions are renormalizable. '4 Unfortunately, theory cannot be used except in a formal way. One
in the case of the Yang-Mills field the ultraviolet must necessarily sum infinite classes of diagrams and
divergences are not the whole story; infrared difficulties hope that the increasingly strong divergence of the
of a special type also make their appearance. These successive terms of the series, as expressed by Kq.
will be discussed in Sec. 8. (5.3), will lead to high-energy damping and a finite
result for the total amplitude. The author has shown'~
In the case of gravity there are no infrared problems
that this hope is actually fulfilled for at least one
beyond those which can be handled by conventional
methods. Equation (5.3), however, casts a rather dismal rs T. D. Lee and C. N. Yan8, Phys. Rev. 128, 885
(1962); 128,
899 (196').
'4 In the case of gravity additional Gelds of spin 1 are allowed "M. B. Halpern, Phys. Rev. 140, 81570 (1965).
r7 3.
if they are massless. S. DeWitt, Phys. Rev. Letters 13, 114 (1964).
 B RYCE S. D EW I TT
simple class of diagrams, namely, those which represent integral equations, or otherwise simplify the computa-
two scalar particles exchanging gravitons in the ladder tional labor. It is clear that the results can give at
approximation. It turns out that the "leading terms" best only a qualitative insight into the true analytic
(i.e. , the most divergent) of the Bethe-Salpeter ampli- structure of the theory.
tude can be summed exactly, and, owing to certain
remarkable cancellations, the sum of the ladder-type 6. THE GRAVITATIONAL WARD IDENTITY
contributions to the gravitational self-energy can be
Although the computational difhculties involved in
expanded in a power series in the bare mass, with no
whatever. The method can also be extracting physical information from quantum gravi-
approximations
dynamics are formidable, the theory has a redeeming
extended to the case of charged scalar particles, with
feature in its general covariance, which serves as a
one or more of the graviton ladder rungs replaced by
cross check on the consistency of various calculations
photons, and a simple expression can be obtained for
and imposes constraints on the permissible forms of
the lowest-order electromagnetic self-energy. The self-
various amplitudes. One of these constraints has
energies and renormalization
way are all finite
constants found in this
recently been discussed by Brout and Englert. These "
authors derive a generalized Ward identity relating
The finiteness of these quantities may be traced to
the gravitational vertex function of a scalar particle
the behavior of the particle-particle scattering arnpli-
tude. In the limit of very high momentum transfer the
to the self-energy function arising from all its inter-
actions. Their derivation is easily generalized to the
singularity of the gravitational interaction kernel is
case of a particle of arbitrary spin.
displaced off the light cone in coordinate space and onto
Denote the field of the particle by y~. In addition
a hyperboloid lying at a distance X= (4AG/s c')'I'= 1.82
&&10 " cm in spacelike directions. This is roughly
to the functions R' (or, in expanded notation, R„„;)
characterizing the coordinate transformation behavior
equivalent to endowing the scalar particles with the
of the gravitational field (see II) we now have corre-
properties of hard spheres of diameter X, and may be
sponding functions R~ for y~. The explicit structure
regarded as a manifestation of the smearing out of the
light cone due to quantum Quctuations.
of these functions may be inferred from Table I of II:
"
Similar results have been found for spin-~ electrons R" .= —y" h(x x')+G" "squib, „(x,x'). (6.1)
by Khriplovich, and there seems to be no reason
why, with enough labor, they may not also be extended We note that E~„. vanishes in the limit q~ — + 0, and
to particles of higher spin, including the graviton in that its functional derivative has the momentum-space
interaction with itself. Thus gravity may indeed prove form
to be the universal regulator which renders all field
theories finite. R"„.g" + ib"np"„+-i—
G"„"np'„, (6.2)
It should be remarked that the self-energy functions
which are obtained by summing ladder graphs appear
to correspond to "good" spectral functions, which do
a minimum of violence to unitarity. This suggests that
j
in which the association of mom. enta with indices is
pA, p' ', p"B" (p p'+p"=0).
Let us denote the full (radiatively corrected) propa-
gator for the particle by S~~. It is the sum of the bare
no illegal analytic operations have inadvertently crept
propagator G~~ and a function obtained by applying the
into the summation procedure. An improved calcula-
operator G"~5/b&pn twice to the vacuum-to-vacuum am-
tional method, which insures analytic legality in gen-
eral, has been developed by Halpern. "He sums first
plitude. Since the vacuum-to-vacuum amplitude is an
invariant the propagator S"~, like G~, transforms in
the absorptive parts of any amplitude and then obtains
the full amplitude by a dispersion integral. The tech-
the manner indicated by the position of its indices. Its "
inverse must transform contragrediently:
nique is applicable to gravity theory as well as to other
nonrenormalizable theories, and is amenable to E/D S gs, ~R gg+S AB, cR u
approximation schemes. It is probably the safest S'cpR~, ~ S—
'~cRo, g. .3)
(6—
method currently available, but it is very complicated
to apply. Equation (6.3) is the gravitational Ward identity. To
Although the 6nite results which have been obtained get it into more familiar form one must reexpress it in
thus far are very suggestive, one must remember that momentum space, with al) the background fields set
they derive from restricted classes of diagrams. They equal to zero. In this limit S '~~, ; becomes the negative
are therefore not y-invariant but depend on the par- of the gravitational vertex function, which is conven-
ticular gauge chosen for the internal graviton lines. So
far calculations have been restricted to those gauges "R. Brout and F. Englert, Phys. Rev. 141, 1231 (1966). See
which avoid "dangerous" singularities in the resulting also K. Justand K. Rossberg, Nuovo Cimento 40, 1077 (1965).
~Thiswill be true even if y" possesses a gauge group of its
own, provided the gauge conditions which determine G are
~8 I. Q. Khriplovich, report, Siberian Section, Academy of covariant. Note that the "background Geld" now includes q~ in
Science, USSR, Novosibirsir, 1965 (unpublished). addition to the metric field.
 QUANTUM THEORY OF GRAVITY. III 1249

tionally denoted by I"&", the particle indices being The cancellation of divergences which is implied by
suppressed and the index i being replaced by the more (6.12) applies only to the leading term of the vertex
explicit iru. Making use of (6.2) and the momentum function, in the limit P' — + p, and only on the mass
',
space form of E. which is given in Table II of II, one shell. In order that no divergences occur in the remain-
readily finds ing terms, or oG the mass shell, the interactions which
»."(P',P)q =S '(P')P.
——
S'—
(P)P'.
—
the field q" experiences with other fields must be of
the renormalizable type (or else summable to finite
[s '(P')G" G"u s '(P))q„, (6.4) values). The example of the scalar particle provides an
where p and p' are, respectively, the incoming and adequate illustration of the conditions which must be
outgoing particle momenta and q=p' P is the in- — satisfied. In this case we have"
coming graviton momentum. This, with the spin terms
involving G"„omitted is the equation given by Brout
G '(P) = P'+m' (6.13)
and Englert. It holds, as a simple consequence of v" (O' P) = s[PuP'. +P P'. n"(P—
P'+m')3 (6 14a)
general covariance, no matter how many other fields
are coupled to the field q ~ and involved in the structure
= 'r'u. (P', P) .~m',
kr)u— (6.14b)
of the vertex function. where the index 0 refers to the bare mass, and we may
Now introduce the vertex and wave-function renor- write
malization constants Z& and Z&. They are defined by
s- (p) = p+;+z(p),
t(P)r„.(P, P)N(P) =z,—'I (P)y„„(P,P)N(P), 6ms = m' ms' —
Z( —
m'), —
(6.15)
P'= —m',
r..(p', p) =~'..(p', p)+&:(p',p) (6.16)
s '(P)=z '[G '(P)+&(P)3 The functions Z and are related by the Ward identity
]„=„=0, (6.6)
A.
[aZ(P)/t)P as follows:
where y„„and 6 are the bare vertex and propagation
functions, respectively, m is the particle rest mass,
»"(P', P)q"=z(P")P, —z(P') p'„. (6.17)
and N(P) is a particle wave function satisfying It is not hard to show that the general solution of
(6.17) is
s '(P)N(P)=G '(P)N(P)=0 (6.7)
I z(p")-z(p')
on the mass shell. From (6.6) we may infer Jt"(P', P) = ,
P"-P'
-(P.P'+P P'.+sq.q.)
[as-(p)/ap j, „=z;[aG (p)/~p g„. .
P'-+
(6.8)
2
-![z(p")+z(p')7~..
On the other hand, (6.4) yields, in the limit
2r" (P,P)= Pu~S '(P)/~P" nu. S '(P)—
P,
+P(P" P' q') (A . qq.) (6 18—
)
—s- (P)G„„—G„„s-(P), where F is an arbitrary function. Therefore the graviton
(6.9) vertex of a scalar particle is characterized on the mass
whence, in virtue of (6.7), shell, by a single function of q'. This is the gravita-
2 (p)r„, (p, p) (p) = p„ '(p)[t)s '(p)/t)p"j (p), tional form factor.
Now introduce the renormalized self-energy function
P'= —ms. Z, defined by
Now, since (6.4) is a consequence simply of general
z(P') = bm'+ (zs ' —1) (P'+m')+zs 'z(P')
covariance, it holds also if I'„, and S ' are replaced by (6.19)
p„, and 6 ', respectively. Therefore we have Z( —m') =0, [dZ(P )/dPsj =0.
2 .
'(P)~. (P, P) (P) =P. '(P)[~G-'(P)/~P
P'=
j (P),
— m'.
„, In terms of this function Eq. (6.18) takes the form
.
A. (P', P) =(z. '-lb. .(P', P) !-~ ",. -
prom (6.5), (6.8), and (6.11) it follows that
Zl
When both vertex and wave-function radiative correc-
(6.12) +-', z ',
z(p")-z(p')
P"—P'
(P„P'„+P„P'„+', q„q„)-
tions are taken into account the two renormalizations
—4zs 'Lz(P")+z(P') jv
cancel, and there remains only the graviton renormali-
" '(z; —1)j
+[p(P', P, q) —,
zation Z3 arising from vacuum polarization,

~'
which
has the eQ'ect of modifying the gravitation constant.
The polarization of the vacuum by a gravitational Geld is of
X (A„.
'Equation (6.14a) is obtained from (2.1$) by maiting tbe
quq„), 20)—
(6.

the quadrupole type. Examples of renormalization terms to which replacement p'~ —

p', since p' is here an outgoing and not an
it leads are given in Sec. 7. incoming momentum.
 $250 BRYCE S. DEKITT 162

which suggests that we also introduce a renormalized magnitude of the gravitation constant G, in terms of
form factor E, defined by arbitrarily chosen (e.g. , international mks) mass stand-
F(P" P' q') =-'(Zs ' —1)+Zs 'F(P" P' q'). (6 21)
ards, to be determined by experiment. "
It is clear that the gravitational Ward identity is
Combining (6.14b), (6.16), and (6.20) we then get only one of an infinity of identities, derivable from Eq.
(17.31) of II, which relate vertex functions involving
r„„(P',P) —
= ZsI'„„(P',P) (6.22a) e gravitons to those involving n-1 gravitons. Such
identities become superQuous if calculations are per-
»(P")-~(P') formed in coordinate space rather than in momentum
vo (P P)+ (P P +P P o+2qoq )
pj2 ps space, for then the general covariance of the theory
can be kept constantly manifest. That such calcula-
.
!f~-(P")+~(p')3~. ++F(p",P', q') tions are actually feasible will be demonstrated in the
)( (q'rt„, —q„q„), (6.22b)
next section.

which reduces, on the mass shell, to 7. RENORMALIZATION IN COORDINATE SPACE.

CONFORMAL VACUUM FLUCTUATIONS
r„„(P',P) = y„„(p',p)+F ( m', —n—t', (q'rto„q„q„), —
q')
p" = p'= —ms. (6.23) The chief tool for studying quantum gravidynamics
directly in space-time is the theory of Green s functions
The Zs factor in (6.22a) takes into account the wave- in hyperbolic Riemannian manifolds developed by
function renormalization arising from self-energy in- Hadamard. '4 The basic structural element of this
sertions in the external lines. theory is the geodetic ie/enal, denoted by g-, which is "
If the scalar particle is coupled to other fields through defined as one half the square of the distance along the
nonrenormalizable interactions then the functions 5 geodesic between any two space-time points x and x'.
and F will diverge in perturbation theory. In particular, The geodetic interval is a symmetric function of x and
they will diverge if virtual gravitons are permitted to x' which transforms as a biscalar, i.e., as a scalar
contribute to the vertex function. Thus unless an separately at x and x'. lt satisfies the differential
arbitrary cuto6 is used, or someone discovers a way equation"
to sum gravitational interactions to all orders, the
gravitational Geld must be allowed to act only through (7.1)
the external graviton line. Although the identity (6.12) and the boundary condition
continues to hoM formally in the nonrenormalizable
case it is then of no utility. Because of the divergence (7.2)
which remains in F, Eq. (6.23) will yield an infinite
cross section for the scattering of the particle in an In a general Riemannian manifold g is not single-
external gravitational field. valued, except when x and x' are suKciently close to
In the renormalizable case Z and II' are Gnite, and one another. '7 The geodesics emanating from a given
expression (6.23) has a well-de6ned limit as q — +0, point will often, beyond a certain distance, begin to
namely) cross over one another. The locus of points at which
r, .(p, p) = p.p p'= -m' (6.24) the onset of overlap occurs forms an envelope of the

More generally, with particles of arbitrary spin one "The necessity of measuring G disappears if absolute units are
finds adopted, with A=c=16m-G=1. However, the masses of the ele-

Nt(p)1'„„(p p)N(p) = (2') 'Pp Prl2& P ™,. (6—

25)
mentary particles must then be measured in absolute units, which
is operationally the same thing as measuring both G and the
masses in mks units.
~ $. Hadamard, Lectures on Cauchy's Problem in Linear Partial
when the wave functions st(p) are chosen to correspond Dgererttial Ettttatiorts (Yale University Press, New Haven,
to 5-function normalization with respect to 3-momen-
"
tum. As Brout and Englert point out, the universality
Connecticut, 1923).
2' B. S. DeWitt and R. W. Brehrne, Ann. Phys. (N. Y.
) 9, 220
(1960l. See also J. I . Synge, Relaiieity: The Gertera/ Theory (North-
of (6.25) implies that the equivalence principle relating Holland Publishing Company, Amsterdam, 1960). Synge calls
gravitational and inertial mass holds in the quantum this function the world function and denotes it by the symbol Q.
~' The semicolons denote covariant differentiation. For a scalar
theory as well as the classical theory. In particular the this is the same as ordinary differentiation. 0.;„is a vector of length
motion of a nonrelativistic particle in a slowly varying equal to the distance along the geodesic between x and x', tangent
gravitational Geld is independent of its mass. to geodesic at x, and oriented in the direction x ~ x. o;„ is a
vector of equal length, tangent to the geodesic at x', and oriented
If a high-energy cutoff is permitted then the Ward in the opposite direction.
jdentity may be applied to gravity itself, i.e., to the "In some manifolds (e.g. , some compact manifolds) every pair
three-graviton vertex. In this case the wave function of points may be linked by more than one geodesic. It is always
possible, however, to deane a single-valued function 0. in the
renormalization constants Z~ and Z3 coincide, and Eq. neighborhood of x by starting at x and following each geodesic
—
(6.12) tells us that Zi=Zs Zs. This leaves only the emanating from x until it hits a caustic.
162 QUANTUM THEORY OF GRA VI TY. III

family of geodesics, known as a caustic surface. The the factors g+'/' being inserted to insure the covanance
equation for the caustic surface relative to a given of operator functions. '
point is D '=0, where Taking matrix elements of (7.11) one obtains
D= —
det( o— ,) .
„„. (7.3)
g" G(x x')g" = z (x,s x', 0)ds, (7.12)
D is a bidemsityof unit weight at both x and x', which i

satisfies the boundary condition

where
lim D=g. (7.4) —(x[exp(ig '7 Fg
(x,s)x', 0)= '~
s) (x'), (7.13)
It is convenient to replace D by the biscalar satisfying

~= i, 8
Z=g-»2Dg'-~/2, lim
s'-+s
(7.5) —i — (x,s~ x', 0)=(x,s~ x', 0)., „~ (7.14)
8$
whose values differential
at given points are independent of the
choice of coordinate system. By covariantly differen- and the boundary condition
tiating Eq. (7.1), one can derive the
(x,0 x'0) = (x x') = (x, x') .
i i lN (7.15)
equation"
The "Schrodinger equation" (7.14) is solved by the
g-t(Z&. ~). =4 or o.~„=4—~.~(ink), » (7.6)
, , ansatz
which shows that 6 increases or decreases along each
geodesic from x' according as the rate of divergence of (x,s~ x', 0) = i(4zr) 'D— "s 'e&'"'&' P—(z„(is)", (7.16)
the neighboring geodesics from x', which is measured
by 0,.I'„, is less than or greater than 4, the rate in Rat co=i' (7.17)
space-time. If the divergence rate becomes negatively
infinite a caustic surface develops and d blows up.
We shall illustrate the use of 0. and 6 in the theory
time. "
which is suggested by its known solution in Bat space-

Inserting (7.16) into (7.14) and making use of (7.1)

of Green's functions by considering the Feynman prop- and (7.6), one finds
agator of the simplest of all Gelds: the massless scalar
Geld. The defining equation is" o, ~a„., „+rza.=h '~2(h'~srz t),.„~,
g»2G , „«(x,x') = o.(x,x'), — (7.7)
22=1, 2, 3 ~ ~
. (7.18)
These recursion relations may be solved by successive
together with appropriate boundary conditions. The
quadratures along each geodesic emanating from x'.
introduction of boundary conditions is most easily Hadamard" and Riesz" have shown that the solutions
accomplished in the abstract formalism which replaces
as well as the series (7.16) converge up to the ftrst
(7.7) by caustic. Formally we may write
IiG= —i, (7.8)
where
F= p.g'"g""p, — (7 9) (z = Q P(io, (P„+m) 'P '~'g '~ Fg '~4g'~sj 1, (7.19)
m 1
G(x, *') =(x G) x'), (7.10)
where ip„and g"'Fg 'I' a-re now to be understood as
~

the ~x') being eigenvectors of a commuting set of

Hermitian operators x& in a fictitious Hilbert space, " the gradient and I aplace-Beltrami operators, respec-
tively. Setting
and the p's being Hermitian "momenta" "canonically
conjugate" to the x's. The formal solution of (7.8)
which incorporates the Feynman boundary conditions (i~, ~p„+m). Lexp(zo., &p„+m)t'„,ddt', (7.20)
Ls
and making the variable transformation
gl/46gl/4—
Py= t'q,
g t/4Fg Il4+z0 t 2=4 —ti, ~2 ~1+Z2y (7.21)
exp(ig '~
Fg "s)ds, (7.11) $a=f 1+12+ ' '+$ I&

"In Eq. (7.7) G(x)x') is to be understood as a biscalar and the

B function B(x,x') as a bidensity

"
weight at x'.
of unit weight at x and zero
Cf. , J. Schwinger, Phys. Rev. S2, 664 (1951).
80 For example

(
"M. Riesz,
Ih ") I ) dG"gc

Acta Math. Sl (1949).

"'fG( '")g'"'-"G("", )d ".
 1252 8 R YCE S. D I~WITT

we may recast (7.19) into the form and where the symbol T indicates that the operators
&n-1 G, (t) appearing in the exponential, via (7.25), are to
dt2 ~ ~
dt„C}(t,) C, (t„).1, (7.22) be "chronologically ordered" with respect to the pa-
rameters t.
where Now the chronological ordering operation commutes

& (t) —
= exp/(ia. "p„+1)tj 6
,
' 'g ' '&g ' 6' ' with diBerentiation or integration with respect to the
parameter s. Hence Eq. (7.24) may be inserted directly
X exp( —io., "p„t) . (7.23) into (7.12). The result is a formal generalization of a
Substitution into (7.16) then yields well-known expression

(x,si x', 0) gl/2 -Q (2) ((2Ro)1/2)-

= —j(4/r) —2D'/2s —2TLexpi(Rs+o/2s)g 1, (7.24) G(x, x') =— T ~
1, (7.26)
8x (2Ro)'"
where

G. (t)dt, (7 25)
8 (» being
1 the Hankel function of the second kind of
order 1. This formula has the series expansion

zd'"
G(x, x') = T —Rr 2y —ln2+ln( —R —i0)+ln(o+i0)) — + (2Ro)' -',

82r2 o+i0 22. 22. 42. 4 g

-2R — (1+!)+
—:
2Ro (2Ro)'
(1+l+-:)- " 1, (7.27)

y = 0.5772. (7.28)

where the instructlions "a+i0" and R i— 0" in—dicate " Table I of II). Because of the coordinate invariance
what is evident from (7.24), namely, that G(x, x') is of the theory the functional integration is redundant
the boundary vaue of a function of o and R which and ambiguous, and since no one has yet discovered
is analytic in the upper-half a plane and the upper-half an analytically accessible nonredundant subspace for
R plane. The singularity structure in o. reflects the the integration, we are forced to accept Eq. (20.12) of
usual' behavior of the Feynman propagator on the II as the effective de6nition of the integral. However,
light cone (o=0). The remaining singularity structure there is an inconlplete nonredundant subspace which is
symbolized by the logarithm of R i0, on the other —— easily accessible, namely, the subspace of all conformally
hand, is far from simple owing to the presence of the equivalent geometries. One may simply set
chronological ordering operation.
In the perturbative approach to quantum gravi-
4»=Xg», q»+4»=g» r/~» g»=—(1+X)g»v (730)
dynamics we must deal not with the scalar propagator and integrate over X, to obtain the partial contribution
(7.27) but with the vector and tensor propagators to (0, ~ ~0, —
~) arising from conformal fluctuations
G ~ and G'&. However, the latter have structures in the vacuum geometry. The special interest of this
closely similar to (7.27); the only difference is that the integration is that it can be performed exactly, giving
operators K(t) out of which R is built are slightly the conformal contribution to all orders of perturbation
more complicated, and. the "1"
standing on the right theory. The only "Qy in the ointment" is that this is
of Eqs. (7.19), (7.22), (7.24), (7.26), and (7.27) is the one contribution for which high-energy damping
replaced by the geodetic parallel disP/acescent function cannot be expected to produce a finite cutoB. There is
Therefore we can gain a qualitative understanding of no smearing out of the light cone, because conformal
the renormalization program in coordinate space al- metric fluctuations leave the light cone invariant.
ready by studying the scalar propagator. Moreover, It is easy to show that
there is an interesting nonperturbative treatment of
gl/2 (4)g —(1+X)gl/2 (4)g
the vacuum-to-vacuum amplitude in which the scalar
propagator itself directly enters:
—3g'"(1+X) 'X. X.& 3g"X /'
(7.31)— .
Consider the Feynman functional integral, Eq. and hence
(20.33) of II, which may be rewritten in the form SEv+4 j Sgv j S,gqg(t"— —
expilof(/2 j Lgl/2 (4)r(v gl/2 (4)g+gl/2(glvl g»v(4)g)y
=Z expi (St &+vt)j SL&'j , (7.29—
S„P~&4)')(t(I)— )
= —3 g'12 i X -~X., „X,~ dh. (7.32)
where S js the Einstein action and (o»=g» —r/» (see
 QUANTUM THEORY OF GRA VI TY. III

The following change of variables then suggests itself: Several comments are now in order. First we remark
that although the final result is divergent, the degree of
x=y+-2y2 1+x= (1+-.2e)2. (7.33) divergence is boueded. The singularity at x'=x is there-
This change not only simplifies expression (7.32) but fore not an essential one as one might have expected
at the same time guarantees the integrity of the signa- on the basis of Eq. (5.3). As a matter of fact (7.39) is
ture of space-time. We may allow p to range from —co identical in structure with the contributions which the
to ~ without danger of encountering unphysical geo- propagators 0'& and G t' of the full theory make in
metries and at the trivial cost of counting each distinct lowest perturbation order (i.e., the single closed loops
geometry twice at every point instead, of only once. of Wfil).
Thus we write It may be conjectured that inclusion of the non-
conformal vacuum Quctuations will eliminate the di-
eXP (ZZ//conformal)
vergences altogether, and. that a rough approximation
to the exact vacuum-to-vacuum amplitude can be ob-
=Z exp~ 3z —g"'@,„P "dx , ~
dy. , (7.34) tained simply by making the replacement o(x,x) & —
A ' in (7.39), where A is a high-energy cutoff of the
—
—2,

from which we immediately obtain order of unity in absolute units. The "i0" attached to
each o in (7.39) reflects the presence of unremoved
=
Wconformal— ff/conformalfgg CUconformalLO] noncausal chains. In passing from W to W these
imaginary infinitesimals should be discarded. We ob-
det (g'"Ggi/4)
= —-,'i ln (7.3S) tain, therefore, the estimate
detGo

The formal determinant

W= 2'dx+ constant, (7.40)
may be evaluated by a
variational technique with the aid of Eq. (7.11). Under
a change in the background field (7.35) suffers the g1/2
T 4cV+A'++i in'
( —Ã —20)
varlatlon
8m
HVconformal= 2i tr—
fg" Gg'"b(g "'Fg-'") j
—O. s4s" (7.41)
=-,' tr exp(zg "'Fg '/'s)0( "Fg "') ds
0
estimates of W can be obtained by
Still cruder
finding approximations to the complicated quantity Q.
s 'exp(zg "I'g "s) ds& (7.36) By repeated covariant differentiation of Eqs. (7.1) and,
(7.6), and use of the conlmutation laws for the indices
which may be iTnmediately integrated, to yield thereby induced. , one can show that

lim g 1/2g 1/4Fg 1/4/1—

/2.
rv nonformal= —gZ1' tl s ' exp{ig "Fg "s)ds s'~s
1

—lim Q 1/2+ 1/2 w 1(4)g —

(7 42)
+constant. (7.37) x'-+x

The trace symbol here means "integrate the diagonal

matrix element over space-time. Hence, making use " This quantity raised to the eth power can be extracted
from expression (7.19) or (7.22) for fz„. Moreover, it is
of (7.12), (7.13), (7.16) and (7.27), we find clear that the operator Q. (t) has the dimensions of the
curvature sca1ar and in the limit x'~
x, is a kind, of
Wconformal = ZconformaldX+COnstant & (7.38) nonlocal, or mean curvature averaged over a certain
neighborhood of x. If we represent the purely nonlocal
where
part schematically by AR, we may write
~co&formal
G, (t) —f."(-' &4&g+ ag)
x' —+x (7.43)
—2 1z
— s '(x)s x, O)ds= gi/2/i BG(x,x')/Off /;, ' f4&X+A%.
R Sl —
~

(7.44)
j.i2- j.
+2 R The crudest approximation to 2' is then obtained
Szr' (0+i0)
(0+zO)2
+L2y —ln2 — ', +ln( —R —20)
—, »The manifestly covariant occurrence of three distinct types
of divergences: quartic, quadratic, and logarithmic, already in
lowest order, implies that the conjecture of Srout and Knglert
+in(o+z0))M2 1 (Ref. 19) that quantum gravidynamics is conventionally re-
(7.39) normalizable is unfounded.
 B RYCE S. 0 EWITT
simply by omitting 6% altogether: basically . a local phenomenon, and global conditions
should have little relevance here.
A4 h.2
gl/2+ gl/2 (4)g+ gl/2 (4)E2
2/ 48m' 288m' 8. THE INFRARED PROBLEM
4)R
„(— i—
(h The most important contributors to the gravita-
&( ln ~
—0.545 ~ ~
. (7.45) tional polarization of the vacuum, and to the mod. i-
6/1, 2 fications in Einstein s equations which this polariza-
tion produces, are the massless 6elds, including gravity
Expression (7.45) is prototypical of the contributions itself. These are also the 6elds which most read. ily
which aQ fields make to the geometrical part of the yield, real quanta. The e8ect of real quantum produc-
vacuum-to-vacuum amplitude. (The only deviations
from it occur with massive fields, for which —, &4)E gets ' tion on the vacuum-to-vacuum
account by the " " amplitude is taken into
20—attached to —
— Q in the logarithm
replaced by rN2+s (4)R, and with fermion fields, for of (7.41). Owing to the complexity of Q, however, the
which the sign of each term is reversed. ) These con- branch-point behavior of the logarithm is very in-
tributions originate in the vacuum polarization which volved, and. it is not easy to investigate directly in-
the background geometry induces, and give rise to coordinate space whether or not serious infrared diK-
nonobservable renorrnalizations as well as physically culties lie hidden in this expression.
real rad, iative corrections. In a closed finite world such difhculties cannot arise
The first term in (7.45) is a "cosmological" term since there is a natural low-energy cutoff; troubles
representing the zero-point vacuum energy which every occur only in infinite worlds. Let us for simplicity
field, including the gravitational field itself, possesses.
It is eliminated by redefining the zero point.
con6ne our attention to Rat backgrounds. "
It is then
appropriate to revert to momentum space to study
The second term in (7.45) renormalizes the gravita- the problem. The analysis for this case is straight-
tional interaction strength. The relation between the forward and has been carried out by Vfeinberg ~ we
renormalized and "bare" gravitation constants (G and shall surrimarize his results.
Gp, respectively) is The amplitude for a single soft graviton to be pro-
G= ZGp, (7.46) duced in a given process has already been derived.
—'. LEq. (4.6)$. The corresponding amplitude for the
Z (1+A2/484rs) (7.47) emission of N soft gravitons in all possible ways from
In the theory of the pure gravitational 6eld Z is the a given diagram is just the product of N single-graviton
constant which occurs (provided, amplitudes. The form of these amplitudes is such that
orily renormalization
of course, the exact theory is really finite. ) Because of an infrared divergence arises in the computation of the
the manifest covariance of (7.45) it is clear that the rate at which any physical process takes place when
same renormalization applies to all vertex functions
arbitrary numbers of soft gravitons having total energy
less than E are simultaneously emitted. This divergence
no matter how many graviton prongs they possess. No
Ward identity is needed. disappears if the contributions from virtual soft gravi-
The third term in ('7.45) is the only one having tons are also included. YVeinberg shows that the correct
observable physical consequences. "
In the classical total rate is given by an expression of the form
limit of long wavelengths and large coherent ampli- I'(E) = Fo(E//1) ~b (8), (8.1)
tudes it may be regard, ed as a correction to the Einstein 1 " sino. 1 gsQ0'
Lagrangian. Hill~ has applied such a correction term b(B)= exp~ 8—— dhp do (8.2)
to the problem of gravitatiorial collapse of the Fried- p
mann universe, with encouraging results. He 6nds that = 1 ——'s.sP+
12
if the sign of the coeKcient in front is negative, as
would, be the case for the contribution from a fermion where l'p is the rate without graviton emission and, h.
is a parameter marking the dividing line between
field, this term succeeds in turning the collapse cycle
around before infinite curvature is reached, . It may " "soft" and "hard" virtual gravitons. If A is chosen to
be objected, that in applying the correction to the be of the order of the typical energies involved, in the
Friedmann model one violates the boundary conditions physical process, Eq. (8.1) gives a fair estimate of the
of asymptotic Qatness which were assumed, to get it rigorous value which would be obtained for F(E) if the
in the first place. However, vacuum polarization is contributions from ultraviolet virtual gravitons were
also includ, ed. and appropriate renormalizations per-
8'The nonlocal part of the A' term, which has been omitted formed. The soft gravitons make appreciable contribu-
from P.45), also has observable consequences.
34 T. %. Hill, Ph. D. thesis, University tions only if attached to the external lines of the I'p
of North Carolina, 1965
(unpublished) .
" Not, however, until a density of the order of unity in absolute
units is reached. At this density all the matter in. the visible
"Conclusions reached for this case are presumably valid also
for inhnite worlds having other background geometries.
universe has been comprcgst;1 &a y., gegion the size of a nucleon. "S. Weinberg, Phys. Rev. 140, B516 (1965}.
162 QUANTUM THEORY OF GRAVITY. II I 1255

diagrams, and hence the only radiative corrections which permits (8.4) to be decomposed as follows:
which should be included in Fo are those which involve
internal lines and, vertices. The quantity is given by 8 8= —(2G/v) Q g,qbp pb ln( —2p pb)
G

2~ na, n
1+v„' gym
(1 v 2)1/2
mm
In,
—
1+vnm

1 V~~
(8.4)
—(4G/v. ) p g.r/ p. p ln( —2p, .p /m )

= $1 —(m
v„— m /p. p )'J/2; (8.5) G 1+vnwP r/nr/mmnmta 1+vnm
ln
it depend. s only on the parameters of the external lines. 2v te, n(1 vnwP) vnm 1 vnm
These results are completely standard. Except for .
the detailed form of the quantity 8 they are identical +(2G/v) Q g.gbp. pb ln(m. mb)
a, b
with the corresponding results in quantum electro-
dynamics. The question which now must be asked is: (4G/v) Q r/, r/ P, .P lnm, . (8.7)
+
What happens when emission takes place from particles
which are themselves massless' In quantum electro- With the aid of the energy-momentum conservation
dynamics such emission is known to give rise to a new law
and more serious kind of infrared divergence which
cannot be removed. in any simple or completely natural p. +Er/ p =o, En. (8.8)
way. This circumstance has been invoked as the
"reason" why massless charged particles do not occur the last two terms of (8.
7) may be combined into
in Nature. In the case of the Yang-Mills and gravita-
tional fields the di8liculty presents itself in a peculiarly (2G/v)Pr/, gbP, Pb (lnmb — lnm ),
acute form since these fields are themselves both
"
massless and "charged. Moreover, although there is which vanishes by symmetry. The masses m, mb thus
no experimental evidence for the existence of the Yang- disappear from (8.7), and since p, pb ln( — 2p, .pb)
Mills field, gravity is an established fact, as is also its vanishes when either a=b or p is parallel to yb, it is
interaction with photons. evident that B is completely free of divergences.
Since the Yang-Mills field is a vector field its diverg- The only uncertainty which remains in Weinberg's
ence difFiculties are similar to those of massless electro- analysis, and which he himself points out, concerns his
dynamics and hence are difFicult if not impossible to use of the DeDonder gauge for the virtual gravitons.
remove. Because of the noncommutativity of the emis- Except when the stress-energy tensor is conserved. at
sion vertices for Yang-Mills quanta it is not possible each virtual graviton vertex it is not easy to see that
to sum the effect of arbitrary numbers of real and, the choice of gauge is immaterial. But stress-energy
virtual quanta into a closed expression like (8.1). conservation of this simple type holds only when the
Moreover, the situation is further complicated by the particle lines on both sides of the vertex are on the mass
fact that there is a non-negligible amplitude for the shell. LSee Eq. (3.8)j. Since only the external lines
soft quanta themselves to emit soft quanta. However, satisfy this condition Weinberg must appeal to the fact
there is no evidence whatever that the situation would that the other lines are only slightly og the mass shell
improve if one could find a way to take all these extra and hence violate the conservation conditions only
minimally. '8
complications rigorously into account.
In the case of the gravitational field, , on the other Weinberg's act of faith on this question can be
hand, the difficulties miraculously disappear. This rigorously justified within the framework of the com-
happy state of affairs is a consequence of the detailed plete theory developed in II. We known from this
structure of expression (8.4), which in turn derives theory that the choice of gauge for internal lines is
from the special form of the graviton emission vertex: irrelevant provided: (a) it is applied consistently and
—
Fp&„~ p&p„as q 0. We shall now show how it comes
& (b) all diagrams contributing to a given process are
about. included. Now Weinberg omits the diagrams which in-
Let us use indices from the first part of the alphabet volve infrared fbctitiols quanta. But it is not hard, to
to distinguish the massless particles from the others. show that the contributions of these diagrams all vanish
We shall continue to use the symbols m, mb, etc. but as the infrared, momenta go to zero, and hence may be
with the understanding that these masses ultimately neglected. This is a consequence of the fact that the
tend to zero. We may then write fictitious quanta always occur in closed loops containing
uniformly oriented vertices V~;~ p. Because of the special
1+Vab '//dl/mamb 1+Vab
ln 3'In electrodynamics it is not difBcult to show that gauge
(1 v 2)1/2 Vab Vab
invariance holds when every vt:rtex along each charged particle
/' line is taken into account. In gravidynamics, however, every line
-4&.„,p. p, in~ — (8.6) "
is '-'charged, and the "charge" splits up or recombines at every
m, mb )~, vertex.
1256 B RYCE S. DEW I TT 162

form (2.8) which these vertices possess, the uniform fourth order in the velocities. This expression can be
orientation guarantees that at least one of the vertices greatly simplified with the aid of the energy-momentum
in each infrared, loop is proportional to an infrared conservation laws
momentum.
We conclude this section by repeating Weinberg's P g„m. (1+-,'v„'+-,'v.'+ ) =0,
8
calculation of in the nonrelativistic limit and correct-
ing a minor mistake in his result. The quantity e„ is g rt„m„v„(1+-,'v„'+ ) =0,
6rst expand, ed in the form
e „'= (v„—v )' —v 'v '+2(v„'+v ')v„v and one 6nally obtains the compact formula
—3 (v„v„)'+ . , (8.9) 8= (4G/5x)
~

tr(ad'Q/dP)', (8.11)
where v„= p„/E„. This expansion is then inserted into
where Ad'Q/dP is the d.yadic previously de6ned by
1+&nm rtertmmnmm 1+&em
ln Eqs. (4.11) and (4.12), having the explicit traceless
(1 —r „„')'" ~ 1—v„„ form39

hd'Q/dP=Q (v„v„—xelv '). (8.12)

1+—v„'+—
11 63 rt m
=2rt rt„m m i
e„„'+ i
(8.10)
40
39
By inadvertently dropping a term Weinberg obtains a
to obtain a lengthy expression for 8 correct to the dyadicwhichisnot traceless.

PH YSI CAL REVIEW VOLUME 162, NUM BER 5 25 OCTOBER 1967

Quantum Statistics of Coupled Oscillator Systems

B. R. MOLLow*
Physics Department, Brandeis University, 8'altham, Massachusetts
iReceived 10 April 196'I)

The statistical properties of systems of coupled quantum-mechanical harmonic oscillators are analyzed.
The Hamiltonian for the system is assumed to be an inhomogeneous quadratic form in the creation and
annihilation operators, and is allowed to have an explicit time dependence. The relationship to classical
theory is emphasized by expressing pure states in terms Of the coherent-state vectors, and density operators
by means of the P representation and an analogous representation involving the Wigner function. The
state which evolves from an initially coherent state of the system is found, and equations governing the time
evolution of the Wigner function and the weight function for the P representation are derived, in differential
and integral form, for arbitrary initial states of the system. The results remain valid for couplings which do
not preserve the vacuum state, and for cases in which the time dependence of the coupling parameters gives
rise to large-scale amplification of the initial field intensities. The analysis is performed by first treating
general linear inhomogeneous canonical transformations on the oscillator variables, and then specializing to
the case in which these transformations represent the solutions for the Heisenberg operators in terms of their
initial values. The results are illustrated within the context of a model of parametric amplification.

I. IÃTRODUCTIOÃ classical system. The time-dependent expectation values

' 'N a wide variety of physical processes, all of the dy- of dynamical operators for a given initial state of the
system may be evaluated straightforwardly with the
- e namical elements which enter into the description
aid of the solutions to the Heisenberg equations of mo-
of the state of the system may be treated formally as tion and the commutation relations for u; and u;~, and
quantum-mechanical harmonic-oscillator modes. The
some indication is thereby provided of the way in which
coupling between the modes typically takes the form of
quantum fluctuations inhuence the time development of
a quadratic expression in the annihilation and creation the oscillator system. '—'
operators a, (t) and a, 1(t), in which the coupling param-
eters are time-dependent in the general case. In addition, 'R. Serber and C. H. Townes, in Quantum E/ectronics A —
driving terms linear in the oscillator variables may be Symposium, edited by C. H. Townes (Columbia University Press,
present. The operators a, (t) and a, t(t) then obey linear New York, 1960), p. 233.
'W. H. Louisell, A. Yariv, and A. E. Siegman, Phys. Rev. 124,
inhomogeneous equations of motion, and the solutions 1646 (1961).
to these equations. take the same form as the solutions 3 H. A. Haus and J. A. Mullen, Phys. Rev. 128, 2407 (1962).

for the c-number complex amplitudes in the analogous

4
J. P. Gordon, W. H. Louisell, and L. R. Walker, Phys. Rev.
129, 481 (1963).
~ National
~
J. P. Gordon, L. R. Walker, and W. H. Louisell, Phys. Rev.
Science Foundation postdoctoral fellow. 130, 806 (1963).