Volume 236, number 2 PHYSICS LETTERS B 15 February 1990

EXACTLY SOLVABLE F I E L D T H E O R I E S O F C L O S E D S T R I N G S

E. BRI~ZIN and V.A. K A Z A K O V 1

Laboratoire de Physique Statistique, Ddpartement de Physique de l'Ecole Normale Supdrieure,
24 rue Lhomond, F-75231 Paris Cedex 05, France

Received 7 November 1989

Field theories of closed strings are shown to be exactly solvable for a central charge of matter fields c= 1 - 6 / m ( m + 1), m= 1,
2, 3, .... The two-point function Z(2, N), in which 2 is the cosmologicalconstant and N - ~is the string coupling constant, obeys a
scaling law Z(2, N) = N -~,,+ 1/2~f((2c - 2 )N ,,/~m + ~/2)) in the limit in which N -~ goes to zero and 2 goes to a critical value 2c; we
have determined the universal non-linear differential equation satisfied by the function f From this equation it is found that a
phase transition takes place for some finite value of the scaling parameter (2c-2)N m/~m+~/2);this transition is a "'condensation
of handles" on the world sheet, characterized by a divergence of the averaged genus of the world sheets. The cases m = 2, 3 are
elaborated in more details, and the case m= 1, which corresponds to the embedding of a bosonic string in - 2 dimensions, is
reduced to explicit quadratures.

1. Introduction ditional variables, such as spins, which lie at the ver-

tices of the triangulation; exactly solvable examples
One of the most tantalizing goals of m o d e r n quan- of this approach are the bosonic Polyakov string [ 2 -
tum field theory is the non-perturbative investiga- 5], the fermionic string [6,7], O ( n ) - s p i n s and ADE
tion of various string theories [ 1 ]. This goal seems to face models [8,9 ], (ii) or by means o f a multicritical
be almost inaccessible from standard coordinate ap- p h e n o m e n o n for dynamical graphs of mixed type
proaches which involve integrations over the moduli [ 10 ] (i.e. made of triangles, squares, pentagons, etc. ),
space and the fixing of a gauge (a coordinate system) which can exhibit different types of critical points,
for the internal metric. Even the first few lowest or- thereby reproducing different kinds of matter fields.
ders of the string loop-expansion (in the n u m b e r of A natural framework for the investigation of these
holes in the world sheet of a string, for open strings, models is the 1 I N expansion of various simple field
or in the n u m b e r of handles of the world sheet for theoretic models involving NX N matrices [ 11,12 ].
closed strings) require considerable calculational Many of the a b o v e - m e n t i o n e d models have been
efforts. solved exactly from this approach in the planar
On the other hand in the past few years a coordi- approximation.
nate-free approach has been developed for q u a n t u m Recently one of the authors has investigated by
strings [2,3] which is based on a lattice regulariza- these methods a very simple field theory of open
tion of the world sheet in terms of dynamical trian- strings [ 13] (2D gravity without any matter field).
gulations of the corresponding planar manifolds. The A non-perturbative p h e n o m e n o n (with respect to the
integral over the metrics is represented as the sum string coupling constant) was found there, consisting
over all possible triangulations with a fixed topology of the " s p o n t a n e o u s tearing" of the world sheet of a
[2]. The i n t r o d u c t i o n of matter fields for these string, and some new critical exponents were com-
models of two-dimensional q u a n t u m gravity may be puted.
done explicitly in two ways: (i) one introduces ad- It was clear from the beginning that this method of
the 1 / N expansion of matrix field theories provided
Permanent address: Cybernetics Council and Academy of an elegant n o n p e r t u r b a t i v e definition of closed
Sciences, ul. Vavilova40, SU-117 333 Moscow, USSR. strings. For instance a basic quantity, such as the

144 0370-2693/90/$ 03.50 © ElsevierSciencePublishers B.V. ( North-Holland )

Volume 236, number 2 PHYSICS LETTERS B 15 February 1990

string susceptibility Z(2, N) (the two-point function the general structure of the equation for the scaling
of the string theory, see refs. [2-5] for definitions), function for any m, (iv) the simplest case m = 1, i.e.
which is a function of the bare cosmological constant c = - 2, or the ( - 2 )-dimensional embedding o f a bo-
2 and of the string coupling constant 1/N, may be sonic string, for which the scaling function f m a y be
represented by the formal expansion computed easily.
The large-x expansion of (6) will reproduce the re-
Z(2, N ) = ~ N-ZgZg(2 ), (1) sults (2) and ( 3 ), and the singular behaviour near xc
g=0 (a double pole for m = 2) will exhibit non-perturba-
in which Zg(2) is the susceptibility for a fixed topol- rive effects of the 1/N expansion. This phenomenon
ogy of given genus g. Near the critical point 2c of 2, is related to the instability of the world sheet with re-
the susceptibility Z has a singular behaviour charac- spect to the creation of handles (or to the formation
terized by a critical exponent Ystr(g) of the type of closed strings) in a regime in which N is large but
x is finite near Xc.
Zg(2) ~ (2c-2)- 7sir(g) • (2)
The exponent ~str(0) has been computed in the lat-
tice approach in some particular examples [2-10 ], 2. Multicritical models in the scaling limit
and later by elegant continuum methods [ 14,15 ]
which yield the general formula for ~str(0 ), in the case
In this section we shall use the approach of ref. [ l 0 ]
of a manifold with the topology of a sphere coupled
in which the matter field is generated from pure grav-
to a conformal matter field of central charge c~< 1.
ity at a multicritical point, and the mathematical
This continuum approach was then generalized to in-
framework and notations of ref. [ 12].
clude any topology [ 16,17 ], and verified on O (n)
In a general one-matrix model of this type a two-
lattice models [ 8 ]; it yields
point function can be defined as
Zstr(g, C ) = -- 1/m+ (2+ 1/m)g, (3)
in which the central charge c is Zm.,(2, N)=[ l f dN2M tr M2m tr M 2n

c= 1-6~re(m+ 1) . (4)
In general m is a continuous positive parameter.
In this article we shall use the generalized one-ma-
trix models [ 1 1 ] in which the multicriticality yields in which M is an N×N hermitian matrix, Z is the
matter fields out of pure gravity [10], and investi- partition function, and 2 plays the role of the cosmo-
gate in the scaling limit the singular point 2 near 2c, logical constant (see ref. [10] ); the potential V(x)
at which surfaces of arbitrary genuses acquire a large is given as
average size (the invariant area is proportional to 1/
V(x)= y~ gpx 2p. (8)
( 2 c - 2 ) ) . In the scaling limit N--+or, A= (2c - - 2 ) ~ 0 p>~l
and x finite in which
As usual in these problems one expresses in the inte-
X:zIN 2m/(2m+ l) , (5)
gral ( 7 ) the matrices in terms of their eigenvalues xa,
the singular part of the susceptibility is expressed in and of the unitary transformations which diagonalize
terms of a universal scaling f u n c t i o n f ( x ) as M. It is then convenient to introduce the orthogonal
polynomials Pn (x):
Z(2, N) = N - 2 / ( 2 m + l)f(x) . (6)

In the following we shall demonstrate this scaling law f dl2(x)P.(x)Pm(x) =hn~.... (9)
and find a non-linear differential equation for the
scaling function f, for any matter central charge c < 1. with respect to the measure
We shall investigate in more details (i) the case of d p ( x ) = d x exp[ - 2 - ' V(x) ] (lO)
pure gravity ( m = 2 ) , (ii) the Ising case ( m = 3 ) , (iii)

145
Volume 236, number 2 PHYSICS LETTERS B 15 February 1990

(Po(x) = 1 ); these polynomials satisfy the recursion

w'(Rc)=W"(R~) . . . . . w ~.... ' ) ( R ~ ) = 0 (19)
formula
are fulfilled.
x P . ( x ) = P,,+, (x) + R.P._, (x) , ( 11 )
From (19), we find the large-N multicritical be-
in which haviour of Z(2, oc):
R,=h,/h,_~. (12) Z(;t, ~ ) ~ ( 2 - 2 c ) ~/m , (20)

Itzykson and Zuber have obtained [ 12 ] a general re- which corresponds to the appearance of a conformal
lation for the coefficients R,: matter field with central charge c = 1 - 6 ~ r e ( m + 1 )
coupled to gravity.
2n
U-2R,, ~ (p+l)gp+~ Z Rm...R,~p, (13) For N finite we can perform a Taylor expansion of
p >/0 staircases the coefficients Rk as follows: settingx= n/N, eq. ( 13 )
in which the coefficient of 2 ( p + 1 )gp+ ~is a sum over determines R , ( 2 ) as R (2x). For x near one, we have
the ( 2 p + 1 ) ! / p ! ( p + 1)! paths along a "staircase"
starting from the height ( n - 1 ), ending at height n in R.(2)=R(2)+2~-R'(2)+ ~ R" (2)
( 2 p + 1 ) unit steps, ( p + 1 ) up, p down. A down step
from k to ( k - 1 ) generates a factor Rk; an up step + .... (21)
gives a factor one.
It is now more instructive to discuss simple examples
By the same technique which leads to 13 ) one can
first, before getting to an arbitrary (even) potential.
easily obtain the relation
N--I
(N-'tr(m2)) =2 Z R,. (14)
0 3. A first example. Pure gravity

If we define a particular two-point function We start from a simple quartic potential which will
a give an ordinary critical point:
3/(2, N) = 2 2 ~ ( N - 1 t r ( m 2)) , (15)
V( x ) =g~ x2 + gzx 4 , (22)
we obtain in the scaling limit from which we get
Z(2, N) : R N W O ( N - ' ) , (16) W ( R ) = 2 g ~ R + 12g2 R2. (23)
in which This potential has a critical point Rc for which the
R N - R c = N -Uf(N~A) . (17) equation W ( R ) = 2 c has Rc as a double root. The
equation which determines RN(2 ) is here
We shall now determine R N - Re in the scaling limit,
from the functional recursion relation (13), as the 2 = 2 R N [ g , +2g2(Ru+~ +RN+RN_, ) ] , (24)
solution of a universal differential equation for the and in the large-N limit, using the Taylor expansion
function f a t the multicritical point. (21 ) we obtain the differential equation
Let us recall that the usual large-Nlimit, with 2 fixed
(i.e. x going to infinity), is obtained from (13) by
2=w(R)+N-24g2~.ZRR"(J.)+O(N -4) . (25)
setting all the R k ( 2 ) = R ( 2 ) ; R is then given by the We work now in the vicinity of2~, w(R) - 2 c is thus
equation proportional to ( R - R ~ ) 2, with the scaling ansatz
( 17 ). It is immediate to verify that 2 - 2 c , w(R) - 2 c ,
2=w(R)=2 o~ ( 2(pp+!l)) ~! gp+~R p+~ . (18) N-2R" (2) are of the same order of magnitude and
that the terms of higher order in 1 / N i n (25) are neg-
A multicritical point 2 ¢ = w ( R ¢ ) of order m, is ob- ligible, iff
tained when the coefficients gp which are defined by
/t=~, 4 (26)
the potential (8) are such that the conditions

146
Volume 236, number 2 PHYSICS LETTERS B 15 February 1990

This is of course in agreement with the result (6) verges near the critical value x = k with an index z:
stated in the introduction (see also eq. (16) for the
<g)~(x-k) -z, withz=2. (31)
case m = 2, which corresponds to pure gravity since
c=0. One may speculate that this condensation of handles
The scaling f u n c t i o n f ( x ) satisfies (after a simple is likely to be generic and thus to take place in more
rescaling) the non-linear differential equation realistic string field theories.

x=f2+f " , (27)

whose solution is a Painlev6 transcendental of the first 4. The Ising case
kind [ 18 ].
This second-order differential equation should be We will see now explicitly that a potential with a
supplemented by two "initial" conditions. The most tricritical point generates an Ising conformal matter
obvious way to impose these conditions is to use the field, c = ½or m = 3, coupled to gravity. We thus take
expansion ( 1 ), in which the coefficients for any fixed a potential
genus g can be computed independently by the stan-
dard 1/N expansion. Indeed, in the large-x limit the
V(x) = g l x2 "~g2x4 +g3 x6 , (32)
topological (i.e. 1/Nexpansion), we deduce from ( 1 ) from which we get
and (6)
w(R ) =2g~R + 12gzR2 +60g3 R3 , (33)
f(x)=xt/Z(Ao-~Aix-Z+A2x-9/2+ ...) , (28)
which will exhibit a tricritical point (w' = w" = 0)
in which the coefficients A, correspond to genii 0, 1,
2, ...; the first two are our initial conditions and all Rc=-gz/15g3, if2g~=5gtg3. (34)
the others are determined by the universal equation The equation which determines RN(2) is now in the
(27). However the non-linear equation (27) has also large-N limit, using the Taylor expansion (21 ),
some non-perturbative solutions for finite values of
2 = w(R) +N-2),2w" ( R ) R " ( 2 ) / 6 + 3 0 N - 2j.2R'2g3
x. Painlev6's transcendentals are characterized [ 18 ]
by their properties that the only "movable" singular- + N-4RR(a)24(g2+33g3R)/3+O(N-6). (35)
ities (i.e. depending upon the initial conditions) are
poles. Eq. (27) exhibits a solution of the form In the vicinity Of2o w ( R ) - 2 c is now proportional
to ( R - R e ) 3. In the scaling limit (17) all the terms
f i x ) ,.~ - 6 / ( x - k) 2 , for x in the vicinity of k , that we have kept in (35) are of the same order of
(29) magnitude and higher orders in 1/N in (25) are neg-
ligible, iff
in which the parameter k has to be determined by the
large-x initial conditions. We do not know the value /t= 2, v=-67, (36)
of k for this problem, and whether it is a universal
number. However, if we assume that it is indeed a in agreement with the result (6) for the case m = 3 ,
finite positive constant, we must conclude that our which corresponds to c = ½, the conformal charge at
model has a singularity at x = k , and we should thus the Ising critical point.
expect some kind of phase transition, such as a con- The scaling function f ( x ) satisfies here the non-
densation of handles, when x goes below k. linear differential equation
More precisely we can calculate the average genus x=f 3+ff "+ (f')2/2+f(4)/lO. (37)
g near this transition point from the relation
A similar discussion on the integration of this equa-
NO 2 tion is required; four initial conditions are required
220<g)__ 20-NZ ( , N )
here. We do not know whether all the movable sin-
[ (2_2~)N4/5_k]-3, (30) gularities of this equation are poles, however one can
verify that, remarkably enough, a double pole is again
from which we obtain immediately that <g> di- allowed by (37 ); a priori one could expect more than

147
Volume 236, number 2 PHYSICS LETTERS B 15 February 1990

one constant analogous to the constant k that we elegant expression for the free energy, valid to all or-
found in the case of pure gravity, and the phase dia- ders o f the topological expansion [20]. This model
gram could be richer. appears to be essentially equivalent to the simple
gaussian unitary ensemble of N × N random matri-
ces. For the study of this model we shall use a repre-
5. The general case sentation o f the generating function for the expecta-
tion values of the powers of the random matrices, in
The potential V(x) is now an even polynomial of terms of a double integral [21 ]; N appears there as
order 2m, and the corresponding w(R) has his first an explicit arbitrary parameter.
( m - 1 ) derivatives which vanish at Rc. Expanding The connection between the two problems may be
as before (21) for n large, the general term will con- seen as follows: the derivative with respect to the cos-
tain k-factors involving derivatives of R with a total mological constant of the partition function of the
of 2p differentiations acting upon them (example: "( - 2 )-dimensional" bosonic string is given by
R ( 4 ) : k = l p = 2 , (R')2:k=2p=l); by definition
2p>k. In the scaling limit (17) a simple counting of
0~ (N,)-) = Z
g = 0 n = l G(n,g) N 2g
powers of N shows (i) that the scaling exponents (6)
are indeed correct, (ii) that all the terms such as
p+k=m have to be kept (for m = 4 , for instance, this
leaves us with R (6), R'R"', (R")2), (iii) that all the (8>

terms such as p + k < m should have a vanishing coef- (39)

ficient at multicriticality for consistency. For in- in which E(;(n,g) is the sum over 03-graphs, drawn on
stance R" has p + k = 2 , but its coefficient is w" (R) a surface of genus g and made of n vertices. The vari-
which vanishes at criticality for any in> 2; for m = 4 ables xi, -~,, i = 1, 2 ..... n, are complex grassmannian
one should check that the coefficients of R (4) and variables (corresponding therefore to an embedding
( R ' ) 2 vanish at Re, etc. The verification of the con- space of dimension - 2 ); Z <,j) denotes the sum over
sistency of these conjectures for general in is in prog- the neighbours defined by the triangulation. In ref.
ress [ 19 ] ; assuming that all this is indeed right, we [20] it was found that (39) could be expressed as an
obtain an equation for the scaling function f o f the integral over N × N hermitian matrices as
following type:
x = f " +,f "fm--2 + a f (2m--2) OF
02 - ! f dN2M[exp{-(½NtrM2)}]
+ Z Cp.f(P)f(2m-p-4) × tr{f(M)[M-f(M)]}, (40)

+ ~, Cp,qf(P!f(q)f(:'-p-q-6) + .... (38) in which f ( M ) is the solution of the quadratic

p,q>~O equation
in which the constants G, Cp,q,etc., are all fixed by the -2f2+j~M. (41)
multicriticality condition. Note that for any in there
is a finite number of terms on the RHS of (38), which Therefore, using a contour integral representation for
are all characterized in the above discussion. Need- the prefactor of the exponent in (40), one can ex-
less to say that the analysis of the singularities of this press F' (2) as
general equation remains to be done. OF N ~
02 -- 222 dz Q(z, N)
6. Field theory of a bosonic string embedded in a X [ 3 2 z - 1+ ( l - 2 z ) x / l - 4 2 z ] , (42)
(-2)-dimensional space
in which Q(z, N) is the average Green function of the
gaussian unitary ensemble:
The planar limit of this theory has been first inves-
tigated in ref. [4]. Kostov and Mehta have found an Q ( z ) = ( N -1 t r ( z - M ) - 1 ) (43)

148
Volume 236, number 2 PHYSICS LETTERS B 15 February 1990

(the imaginary part o f Q is p r o p o r t i o n a l to the den- in a range o f extension N - 2 / 3 near the critical point
sity o f eigenvalues in this ensemble). zc; the scaling function is related to an Airy function.
Since ref. [21 ] might not be readily accessible let (This result has been known for a long time, from
us give an indication o f the m a i n steps. A simple the consideration o f the asymptotic form o f H e r m i t e
gaussian integral in terms of N complex c o m m u t i n g polynomials at large order [ 22 ], but the present der-
variables ua and 2 N g r a s s m a n n i a n variables Va, va, al- ivation is simpler and self-contained.) Substituting
lows one to write this scaling form into (42) it is straightforward to
verify that F ' (2) satisfies the scaling ansatz (6) for
(z--m)ab I = --i f l-I (dUa du* d/7a dVa)l,t*aUb m = 1; the calculation o f the scaling function is re-
duced to simple quadratures. The singularity struc-
× exp{i[u*(z--M)abUb+Oa(Z--M)abVb]}. (44)
ture of this function will be given elsewhere.
F r o m this representation one can perform easily the
gaussian average over the matrix M, integrate then
over the G r a s s m a n n variables f, v and one obtains Note added.
then, after a few simple algebraic steps, a representa-
tion for Q(z) which is valid for any finite N, in terms After this work had been submitted for publication
o f a double integral we learnt that D.J. Gross and A.A. Migdal in Prince-
ton and M. Douglas and S. Shenker in Rutgers had
Q(z)=-i-~-\-f~j dx dy(x--iz)Ny N done simultaneously interesting work along similar
-- oo 0
lines.

× ( l + x y ) exp[-N(½x2+~y2-izy)]. (45)

The large-N limit is governed by the saddle point Acknowledgement

x0 =Yo = ½( i z + ~ ) (46) One of us (V.A.) thanks D.V. Boulatov and A.A.

Migdal for useful discussions.
(the saddle point Xo = ½[ i z - ( 4 - z 2 ) 1/2 ] is sub-lead-
ing) and one finds that in the large-N limit Q(z)-,
- ixo. Wigner's semi-circle law is o f course recovered
References
after taking the imaginary part o f F ( z ) . However near
the critical point Zc= 2 (at which Xo =Yo = i) the two [ 1] M.B. Green, J.H. Schwarz and E. Witten, Superstring theory
saddle points merge and one has to e x p a n d the ex- (Cambridge U.P., Cambridge, 1987).
ponential in the integrand up to cubic order. Defining [2] F. David, Nucl. Phys. B 257 (1985) 45, 543;
V.A. Kazakov, Phys. Len. B 150 ( 1985 ) 28;
Z=Zc+~O, (47) J. Ambjorn, B. Durhuus and J. Fr61ich, Nucl. Phys. B 257
(1985) 433.
and expanding the integrand as [3] V.A. Kazakov, I.K. Kostov and A.A. Migdal, Phys. Left. B
157 (1985) 295.
xo=i+a, yo=i+fl [4] D.V. Boulatov, V.A. Kazakov, I.K. Kostov and A.A. Migdal,
Nucl. Phys. B 275 (1986) 641.
for small a and fl, we obtain at leading order for large
[5] V.A. Kazakov and A.A. Migdal, Nucl. Phys. B 311 (1988)
N, in the vicinity o f zc 171.
[6] V.A. Kazakov, Phys. Lett. A 119 (1986) 140.
iN f dadflafl
Q ( z ) = - ~-~ [ 7 ] D.V. Boulatov and V.A. Kazakov, Phys. Len. B 186 ( 1987 )
379.
[ 8 ] I.K. Kostov, Mod. Phys. Lett. A 4 ( 1989 ) 217;
×exp{N[ko(fl-a)+½i(fl3-a3)]}. (48) B. Duplantier and I.K. Kostov, Saclay preprint SphT/89-
F r o m this representation we obtain readily the scal- 055.
[9] I.K. Kostov, Saclay preprint SphT/89-056.
ing form [ 10] V.A. Kazakov, Niels Bohr Institute preprint 89-25 ( 1989 ).
Q(z) =N - 1/3f((Z_Zc)N2/3) [ 11 ] E. Br6zin, C. Itzykson, G. Parisi and J.B. Zuber, Commun.
Math. Phys. 59 (1978) 35.

149
Volume 236, number 2 PHYSICS LETTERS B 15 February 1990

[ 12 ] C. Itzykson and J.B. Zuber, J. Math. Phys. 21 (1980) 411. [ 19 ] E. Br6zin and V. Kazakov, in preparation.
[ 13 ] V.A. Kazakov, Ecole Normale Sup6rieure preprint ( 1989 ). [ 20] I.K. Kostov and M.L. Mehta, Phys. Lett. B 189 ( 1987 ) 118.
[ 14] A.M. Polyakov, Mod. Phys. Lett. A 2 (1987) 899. [21 ] E. Br6zin, in: Applications of field theory to statistical
[ 15 ] V.G. Knizhnik, A.M. Polyakov and A.A. Zamolodchikov, mechanics, Springer Lecture Notes in Physics, Vol. 216
Mod. Phys. Lett. A 3 (1988) 819. (Springer, Berlin, 1984) p. 115.
[ 16 ] F. David, Mod. Phys. Lett. A 3 ( 1988 ) 207. [ 22 ] M.L. Mehta, Random matrices (Academic Press, New York,
[ 17 ] J. Distler and H. Kawai, Nucl. Phys. B 231 (1989) 509. 1967).
[ 18 ] See for instance, E.L. Ince, Ordinary differential equations
(Dover, New York, 1956).

150