Lecture II-4: The large N limit of the C PN 1 model

Edward Witten

Notes by Dan Freed

Remark. The lecture treated the -model into projective space, but these notes cover the general-
ization to a Grassmannian, as requested in Problem Set 3.

In this lecture we discuss the large N behavior of the two dimensional  -model into the Grass-
mannian Gr(k; N ) of k dimensional subspaces of C N . Here k is xed as N ! 1. We also consider
the real Grassmannian. Since Grassmannians have positive Ricci curvature these eld theories are
asymptotically free, but in any case our task is to investigate the infrared behavior.
The Euclidean action of the  -model is
Z Z
(1) 1
S [] = g2 d x jdj i ();
2 2
 


where  is a Riemann surface,  :  ! Gr(k; N ) a map into the Grassmannian,  2 H 2 Gr(k; N ); Z
a generator of the cohomology, and g;  are parameters of the theory. We specify the metric
on Gr(k; N ) shortly. The second term is a topological term;1 for  closed the integral is integer-
valued. Thus a shift  !  + 2 does not aect the model. The parameter  is also a parameter of
the quantum theory, but renormalization exchanges the dimensionless coupling constant g with a
mass parameter .
The rescaled coupling constant g~, dened by

(2) g2 = g~2=N;

is more natural in the large N limit, as we will see.

1The factor of i is present in the Euclidean action so that the action conjugates under orientation reversal. In
this way its continuation to Minkowski space is real. See Freed's notes Actions and Reality for more details.
1
x4.1. The Questions
We ask specic questions about the behavior of the model.
1. Is there a mass gap?
2. What is the  dependence of the partition function?
Remark. For  small the partition function
Z 1 R d2 x jdj2 R
Z () = D e 2
geff
ei   ()


can be studied using perturbation theory. Here ge is the eective coupling, which varies
with the distance scale in the theory set by the size of the surface . (Classically the model
is conformally invariant, so does not depend on the size of , but quantum mechanically this
is no longer true.) By asymptotic freedom this coupling is small at small distances, hence
the assertion that perturbation theory applies in this regime. Now the classical solutions
to (1) are harmonic maps  ! Gr(k; N ). Note that the only  dependence is through the
degree of , and so we split the integral as a sum over maps  of varying degrees. In degree 0
we obtain constant maps and in general for degree n some moduli space Mn of harmonic
maps, which are the instantons of this model. Degree 1 instantons are (anti)holomorphic
maps. The perturbation expansion around these solutions has the rough form


(3)
X Vol M1 (1 +


) + higher instantons
Z ()  VolpGr(k; N ) (1 +


) + ei e cN=geff2 p
det(
)  det(
)

The constant c in the exponential is the action of a 1-instanton, which is independent of N .

The factor of N comes from (2). If ge << 1 we see that the  dependent term of Z ()
vanishes exponentially as N ! 1. As the area of  increases the eective coupling ge also
increases. Equation (3) is the answer for  compact and of small area, but we will nd a
vastly dierent result for  = R2.
3. Symmetry Breaking. Symmetry breaking in two dimensions is possible for a discrete sym-
metry, and in this model we have the parity symmetry

P: 7 ! ( is  with the opposite orientation)

(4)
7 ! 

For  = R2 we implement the orientation reversal by an orientation-reversing isometry, i.e.,

a re ection. Then for  = 0 and  =  the parity symmetry P acts on a xed theory and
we ask if it is broken in the quantum theory.
2
 4. The group PSU (N ) of isometries of the Grassmannian acts in the classical theory. Since
continuous symmetry groups are unbroken in two dimensions, this symmetry acts in the
quantum theory as well. However, it is possible that a realization of the theory has a
symmetry group which is a cover of PSU (N ), the latter being the group which acts on the
operator algebra. Does that happen here?
x4.2. An Equivalent Formulation
To study the model we rewrite it, that is, we construct an action with the same classical and quan-
tum physics. As a preliminary we recall some basic geometry of the Grassmannian. Over Gr(k; N )
lies a canonical sequence of vector bundles

(5) s Gr(k; N )  C N ! Q ! 0;
0 !S !

where the ber of S at a k-plane  is simply  viewed as a subspace of C N . Fix the standard
metric on C N . It induces a metric on S and identies Q 
= S ? . There is a canonical connection r
on S , obtained by projecting the natural connection on the trivial bundle Gr(k; N )  C N . We
easily compute

(6) r = d sds;
s Gr(k; N )  C N . Then
where s is the inclusion S !

(7) rs : T Gr(k; N )
! Hom(S; S ?)

is an isomorphism. We use it to induce a metric on Gr(k; N ), the metric needed to write down the
-model action (1).
Now if  :  ! Gr(k; N ) we pullback (5) to obtain a sequence of bundles over , and by (7) the
lagrangian density of  is

(8) jdj2 = j(r)sj2:

Note  s :  S !   C N determines . The idea is to replace  by such a bundle map, and so rst
to replace  S by a xed bundle. Note deg( S ) = deg() so that the topology of  S determines
the cohomology class  (), which appears in the second term of the action (1). Hence x a vector
bundle E !  of rank k and degree d. Also x a hermitian metric on E . We introduce a new eld

^ : E !   C N
3
which we constrain to be an isometric immersion:

(9) ^^ = idE :

The image of ^ determines a map  :  ! Gr(k; N ) which is unchanged if we shift ^ by a unitary
gauge transformation of E . To rewrite (8) in terms of ^ we need a connection on E , and as there
is no natural choice we introduce a variable unitary connection A. Using ^ we identify E with
a subbundle of the trivial bundle   C N , so can dierentiate ^ using the usual derivative d.
Writing A as a 1-form plus this trivial connection we nd
jdA^j2 = jd^ + ^Aj2
(10)
= jd^j2 + 2 Re(d^; ^A) + jAj2 ;
since ^ ^ = idE . This expression is quadratic in A, so if (10) is a classical lagrangian for A we can
use the equations of motion to obtain

(11) A0 = ^ d^:

Comparing with (6) we see that A0 is the pullback of the canonical connection on S , and so by (8)

jdj2 = jdA0 ^j2:

In other words, the lagrangian (10) is equivalent to jdj2 for elds which satisfy the constraint (9).
We impose the constraint via a lagrange multiplier eld

 :  ! HermitianEnd(E ):
The  term in the original action (1) can be computed using the (skew-Hermitian) curvature FA
via Chern-Weil theory. Altogether we obtain for our new action2
Z Z Z
(12) S [^; A; ] = g12 jdA^2 j i Tr (^^ idE ) + 2 Tr FA :
  
2Some explanation about the lagrange multiplier term is in order. It is based on the general formula
Z
ei(;x) d = (x);
where  lies in the dual space to x, the measure d is suitable normalized, and (x) is the -distribution supported
at the origin. In the Minkowski space lagrangian the lagrange multiplier term is
Z
Tr (^ ^ idE );

and  should be interpreted as a 2-form (or density); it lies in the dual space to the function ^ ^ idE . Rotation
to Euclidean space yields the second term of (12).
4
As we have explained, the classical equations of motion (and other classical constructs) computed
from (12) are equivalent to those computed from the original action (1). The classical computation
which led to (11) is valid quantum mechanically since the dependence of (12) on A is quadratic and
the Hessian is the identity operator (see (10)). (We ignore the constant determinant factor which
we obtain from the A integral.) Thus the quantum physics is the same as well.
x4.3. The Large N Eective Theory
The argument here is almost identical to that for the large N  -model into a sphere (see lec-
ture II-3). So we will be brief.
First, the ^ integral is Gaussian, so the partition function is
XZ DA DD^ e
Z= S [^;A;]
E vol
(13) XZ DA D exp 
d dA  Z
 Z 
= A
Tr ln g 2 i + i Tr  + 2 Tr FA :
E vol  
P
Here is the sum over bundles E . We take  = R2. Then the sum over E is irrelevant, as is
E
the topological term. In a rst approach to this problem, the topology of the bundle will not be
important, since it is spread over an innite volume. Once we get a basic understanding of what
the quantum vacuum looks like, it will not be hard to go back and see how the topology enters.
We evaluate the leading behavior in large N by evaluating at a stationary point of the exponential
in the integrand, i.e., a classical solution of the eective action.
The only Poincare invariant (that is, Euclidean invariant) possibility for the gauge eld is A = 0.
In that case we rewrite minus the integrand in the last line of (13) as
d  Z
d
Se[A; ] = N Tr ln g2 i + i Tr ;

where the operator in the rst term acts on sections of E . Now the only Poincare invariant
possibility is  constant, and so we diagonalize  and pass to k one dimensional problems for an
eigenvalue. At this point we rescale
 = N ~
(14)
g2 = g~2=N;
and so for each eigenvalue we have exactly the problem we had for the large N  -model in to a
sphere. Thus the solution ~0 has all eigenvalues equal and is specied by

ig~2~0 = 2e 4=g~2 = M2

5
for  an ultraviolet cuto and now g~2 the running coupling constant. We dene M 2 to be this
dynamically generated mass squared.
So in the large N eective action ~ acquires a mass. Note that the A eld has no transverse
degrees of freedom|it is a gauge eld in two dimensions|so does not enlarge the spectrum of
the model (though, of course, it aects the Hamiltonian as we will see, and in fact diminishes the
spectrum). Thus the large N limit has a mass gap. This is the answer to the rst in our list of
questions.
To answer the other questions we need to compute something more precise, namely the leading
approximation to the large N eective action. This means that we do perturbation theory for the
action (12) about the point ^0 = 0, A0 = 0, ~0 = iM 2 =g~2. So shifting ~ by ~0 and rescaling ^ we
have from (12)
Z Z Z Z
(15) S 0[^; A; ~] = jdA^j2 + M 2 j^j2 ig~2 Tr ~ ^^ + 2 Tr FA :
   

Here we omit a constant term (which only shifts the partition function by a constant) and a linear
term in ~ (since we are expanding around a solution of the eective action).

Figure 1: The inverse ~ propagator

The eective action is computed in perturbation theory using one particle irreducible Feynman
diagrams with external lines for A and  and with internal ^ lines. So the inverse propagator for ~
is computed from the diagram shown in Figure 1, where the solid line represents ^ and the dotted
line represents  . (Note that (15) has no quadratic term in ~ , else Figure 1 would be a correction
to such a term.) We evaluate the diagram in momentum space as
Z 4 Nk Z 1 Z
g~4Nk d2q 1 1 g~ 2 1
(2 )2 q 2 + M 2 (p q )2 + M 2 = 4 2 0 d d q [q 2 + (M 2 + (1 )p2 )]2
g~ 4 Nk Z 1 d
(16) = 
0 M 2 +  (1 )p2
4
= g~M Nk + O(p2) as p ! 0:
2
6
This corresponds to a term
C j~j2
in the eective action, with C > 0. (The minus sign comes since in the Euclidean theory diagrams
compute negative contributions to the eective action.) This is the dominant term in the infrared,
which means that ~ is massive and does not aect the long range behavior of the theory.

ν
µ

µ ν

Figure 2: The inverse A propagator

k1

k3

k2

µ ν

I II

Figure 3: The interaction vertices for the A field

The inverse propagator for A is computed by the diagrams in Figure 2, which come from the rst
term in (15). Here the wavy line represents A. To compute these diagrams we need the Feynman
rules for the vertices indicated in Figure 3, which correspond to the terms
2 Re(d^; ^A)
j^Aj2
in the action (15). The indices refer to a standard orthonormal basis for R2. The Feynman rule
(in momentum space) for the second vertex is easy:
II =  :
7
(There is a minus sign since the Euclidean functional integral involves e S .) For the vertex I we
must remember that ^ is complex and that A is skew-Hermitian:
2 Re(@ ^; ^A ) = @ ^A ^ + ^A @ ^ :
Thus the vertex is
I = i(k1 k2):
Note that one of the solid lines in the vertex represents ^ and the other solid line represents ^ . So
the sum of the diagrams in Figure 2 is
Z Z
(17) d2q (p 2q ) (p 2q ) 2
2 (2dq)2 (q 2 +1 M 2 ) :
(2 ) (q + M 2 )((p q )2 + M 2 )
2 2

The factor of 2 in the second term is from the two ways of attaching the A lines to the external
vertices; the corresponding factor of 2 in the rst term is canceled by the symmetry which exchanges
the two internal vertices in the rst diagram. (That is, there is a factor of 1=2 from the expansion
of the exponential, since we have two triple vertices.) Each term in (17) is divergent, but the
divergences cancel in the dierence, and after some computation similar to (16) the answer to
leading order in p is
N 2
(18) 12M 2 (p p p  ):
This corresponds to a term
(19) N jF j2
24M 2 A
in the eective action. (Again we must recall that diagrams contribute negatively to the eective
action.) In fact, (18) and (19) correspond precisely if k = 1 (the  -model into projective space),
since then A is an abelian connection. In the nonabelian case there are cubic and quartic terms,
but by gauge invariance their leading contribution must be as in (19). We remark that (19) is the
lowest order gauge invariant term we can write, and so its appearance can be predicted from gauge
invariance alone, but of course we must do a computation to determine the coecient.
So, nally, the long distance behavior of the two dimensional  -model into the Grassman-
nian Gr(k; N ), in the large N limit, is equivalent to the long distance behavior of a two dimensional
gauge theory with gauge group U (k) and charged massive scalars. The action is
Z Z Z Z
N 
2e2  jFA j + 2  Tr FA +  jdA j + M  jj :
(20) 2 2 2 2

The rst term in (20) was just computed, where the coupling e2 summarizes the numerical factor
in (19). The second term is the  term in (15). The last two terms are the rst two terms in (15),
except we now drop the `^' for convenience. So  is a section of (E )N .
8
x4.4. Real Grassmannians
From the beginning we can replace the complex Grassmannian with the real Grassmannian. In
that case the bundle E is real, ^ : E !   RN , and there is no other change except in the repre-
sentation of the  term. For simplicity we consider the  -model into the Grassmannian GrR0 (k; N )
of oriented k-planes in RN . (It is a double cover of the Grassmannian of unoriented k-planes.)
Now H 2 GrR0 (k; N ); Z 


= Z=2Zand the  term is meant to detect this class. Thus the second term
of (1) is replaced by
i deg2 ;
where deg2 is the mod 2 degree (0 or 1) and  = 0 or  =  . In the reformulation of the problem E is
an oriented real k-plane bundle over , and the topological term in (12) and subsequent formulas
is

(21) iw2(E )[];

where w2 is the second Stiefel-Whitney class. For  = R2 we can rewrite (21) in terms of a Wilson
line operator. Namely, a eld conguration with nite action (12) has A essentially at at innity.
Since R2 is contractible we can lift the SO(k) connection A to a Spin(k) connection A~, and the
holonomy on a large loop C  R2 is approximately 1 depending on the Stiefel-Whitney class of the
induced bundle on S 2 . (We let ` 1' denote the nontrivial element of Spin(k) covering 1 2 SO(k).)
Choose a representation R of Spin(k) and consider

(22) holC (A) :

WR(C ) = TrRdim R
In the limit where the loop C becomes large, this computes the exponential of (21), where  = 0 if
R is a representation of SO(k) and  =  if R is a representation of Spin(k) but not of SO(k).
The generalization to an arbitrary connected compact Lie group G is clear. A G bundle over 
has a characteristic class in H 2(; 1G). It pairs with a homomorphism ei : 1 G ! T to give
a term in the exponentiated action. Here T = U (1) is the circle group. On the other hand a
representation of the simply connected covering group induces a homomorphism ei : 1 G ! T, so
we can use the Wilson operator (22) to write the  term on R2.
x4.5. Pure Gauge Theory
We still must determine the quantum behavior of the theory with eective action (20). For this,
we will rst practice by analyzing the pure gauge theory in two dimensions, whose action is the
sum of the rst two terms of (20). We may as well consider an arbitrary connected compact Lie
group G. The norm in the rst term of (20) is dened via a bi-invariant metric on G. We quantize
9
the theory on the circle SV1 of circumference (=volume) V . In general, to quantize a gauge theory
in n dimensions on an n-manifold Y , we consider connections on R  Y in temporal gauge and
take solutions to the equations of motion up to time-independent gauge transformations. (See
Kazhdan's lectures on the quantization of gauge theories.) For pure gauge theory without the
 term the resulting space is the (co)tangent bundle of the space of connections on Y modulo gauge
transformations. For Y = SV1 we rst x a basepoint; then a connection up to gauge equivalence is
specied by the holonomy, an element of G. A change of basepoint conjugates the holonomy, and
so the quantum Hilbert space is

Hgauge ( =0) = L2(G)G;

the space of conjugacy invariant functions on G. A basis for this space is the set of characters of
irreducible representations.
Next we compute the Hamiltonian. Let At = At (x)dx be a connection on R  SV1 (with coordi-
nates t; x) in temporal gauge, relative to some trivialization, and let gt 2 G be the corresponding
path of holonomies. In a gauge where At is constant, we have gt = eV At . Then the rst term in
the action (20) is
 
(23) N Z Z  dA  dtdx = N Z
jg_ j2 dt:
2e2  dt  2e2 V

This is the lagrangian for a classical particle of mass N=e2 V moving on the Riemannian manifold G;
the corresponding quantum Hamiltonian is
2
(24) Hgauge = e2NV
G;

where
G is the laplacian on G. The eigenfunctions are the characters of the irreducible represen-
tations with eigenvalues proportional to the Casimir.
Now consider the  term in (20). For G = U (k) we have a natural closed imaginary 1-form  2
1
i
G which is the trace of the Maurer-Cartan form. Then in terms of the path gt of holonomies the
 term in the action is

(25)  Z g:
2

This is a topological term|it is invariant under reparametrizations of the path gt. More geometri-
cally,  is a at connection on a topologically trivial hermitian line bundle L over G, and up to
10
equivalence it is given by an element in H 1 (G; T). Then (25) is parallel transport in this at bun-
dle.3 More generally, we can \twist" our mechanical system by any hermitian line bundle L with
connection. Physically this describes a particle moving in an electromagnetic eld. The quantum
Hilbert space is the the space of sections of L with Hamiltonian the laplacian for such sections. In
our case we obtain the space
Hgauge () = L2(G; L)G
of invariant sections with Hamiltonian (24). For arbitrary G an element ei 2 H 1 (G; T) corresponds
to a homomorphism ei : 1 G ! T: Recall that 1 G is a subgroup of the center of G. Then the
eigenfunctions of the laplacian acting on Hgauge () are the characters of representations of the
simply connected cover of G whose restriction to 1 G is ei ; the eigenvalue is again proportional
to the Casimir.
For the unitary group G = U (k) we identify  2 R=2Zas before. The smallest Casimir occurs
for a representation det~=2 , where ~ 2 R is a representative of  of smallest absolute value. If
 6=  there is a unique such ~, but for  =  there are two possibilities: ~ =  and ~ = .
For G = SO(k) the simply connected cover is G~ = Spin(k). For  = 0 there is a unique lowest
representation, the trivial representation of SO(k). For  =  the lowest representation is the spin
representation if k is odd, and the two half spin representations if k is even. In the large volume

limit only the lowest eigenvalue survives,

so we have two vacuua for  =  in the complex U (k)
case and for k even in the real SO(k) case. Observe that the two vacuua correspond under the
involution g 7! 1
g in these cases. (Notice also that the bundle L=0 is real.) We will see that
this vacuum structure persists when we add matter. Thus parity symmetry (4) is spontaneously
broken at  =  for G = U (k) and G = SO(2`). This answers the third of our questions.
As the volume V ! 1 the eigenvalues of (24) become widely separated. In particular, in the
innite volume limit there is no state of nite energy above the vacuum (or vacuua). So the physical
Hilbert space in innite volume consists only of the vacuum (or vacuua)|pure gauge theory on R2
is trivial.
Specialize to G = U (1). Then the vacuum energy density , which is the minimum eigenvalue of
the Hamiltonian (24) divided by the volume V , is
2
(26) Evac() = 2eN min
n2Z
(n  )2:
2
Notice that the derivative of Evac has a discontinuity at  =  .
For G = U (k) formula (26) is simply multiplied by k.
The partition function Z () of the pure gauge theory for  = [0; T ]  SV1 has the Hilbert space
interpretation
Z()  h
je TH () j
i as T ! 1;
3In general, the classical eld theory action on a manifold with boundary|here the interval|is not a number,
but we do not stress that point here.
11
where H () is the Hamiltonian and
the vacuum. For the pure gauge theory, we obtain from (26)
for G = U (1)

(27) Z()  exp cTV e2 TV min(n  )2
as T ! 1:

2N n2Z 2
Here c is a constant which represents the indeterminacy of the path integral, or equivalently the
fact that we are free to add a constant (independent of ) to the Hamiltonian. This is quite dierent
than the prediction (3) from the instanton sum. Note that due to the classical conformal invariance
of the  -model, the instantons used in deriving (3) do not have a denite size; they can be rescaled.
For that reason the instanton sum is not reliable at large distances. In any case (27) answers
question 2 for pure gauge theory.
x4.6. Classical Electromagnetism in Two Dimensions
Before analyzing the quantum gauge theory with bosonic matter|the  eld in the action (20)|
we discuss the classical physics. For the classical analysis we work in the G = U (1) theory. The
classical equations|Maxwell's equations|for pure gauge theory are

(28) N d F = 0:
e2 A A
Since we are in two spacetime dimensions, this implies that the electric eld f = fA = FA is a
constant. Let `x' denote the coordinate on space, which is simply a copy of R. Add a point charge
of charge 1 at x = x0 . Then the electric eld f as a function of x satises (28) with a right hand
side due to the charge:
N df = (x x ):
e2 dx 0
Thus the value of the electric eld jumps by e2 =N across a charge. (See Figure 4.) Allowing for
many charges, and assuming all of them are multiples of the basic charge (which is a conclusion of
the quantum theory), we see that

(29)  = 2Nf
e2 2 R=2Z
is a constant. So there is an angle in the classical theory (assuming charge quantization).

e2
f " f N f # f + eN
2

Figure 4: Change in electric field across a positive or negative charge

12
 The classical energy density of an electric eld f is computed from the action (23) (where
f = dA=dt) to be 2Ne2 f 2 . As in the quantum theory, for a xed value of  in (29) there is a unique
minimum obtained at some f = f0 as long as  6=  . Any valid conguration must have nite
energy compared to the vacuum energy. Thus if  6=  we must have f (x) ! f0 as x ! 1. Using
the formula above for the jump of the electric eld across a charge, we see that the total charge
of a nite energy conguration is zero. This means that there is connement |it is impossible to
have a single charged particle or any other isolated collection of charges with nonzero total charge.
On the other hand, for  =  we have minimum energy density at f0 = e2 =2N , and so for any
nite energy conguration f (x) ! e2 =2N as x ! 1. Thus there is a nite energy conguration
with a single particle: the electric eld satises f ( 1) = e2 =2N , f (1) = e2 =2N . (See Figure 5.)
In this case there is no connement. Also, in this case there are four components of nite energy
congurations, depending on the value of f at 1.

e2 " e2
2N 2N

Figure 5: A one particle state for  = 

Notice that whereas in three space dimensions the Coulomb potential between point charges
separated by distance r is proportional to 1=r, the Coulomb potential in one space dimension is
proportional to r. This means the potential energy grows as the charges separate, which is another
way of understanding that connement occurs.
x4.7. Quantum theory with matter
Now we want to show that there is connement in the quantum theory of the lagrangian (20) as
long as  6=  . This is the assertion that every nite energy conguration in the quantum theory
has total charge zero.
As a preliminary, consider the theory of matter only (no gauge eld). In the simplest case k =
N = 1 there is a single free complex scalar eld  with mass M . The Hilbert space H of this
theory is the completed symmetric algebra of W  W , where W is the scalar representation of
Poincare with mass M . There is a global U (1) symmetry which rotates . The corresponding
quantum operator Q , the Noether charge, has value 1 on W and 1 on W , so value p q
on Symp W
Symq W . A state in this subspace represents p positively charged particles and
q negatively charged particles. For arbitrary k; N we have kN copies of this picture. In particular,
for k = 1 there is a global SU (N ) symmetry (beyond the U (1) symmetry already discussed.)
Now add the gauge eld. For k = N = 1 we have a U (1) gauge theory with a single charged scalar
eld. The global U (1) symmetry of the preceding paragraph is now a local symmetry. Consider
13
rst the case  = 0. For small coupling e2 we construct the quantum Hilbert space of the joint
system by quantizing the symplectic manifold of classical solutions to (20). Thus take  to be
the cylinder R  SV1 . Then the space of classical solutions is a vector bundle over the cotangent
bundle T  A to the space A of connections on SV1 ; the ber of this vector bundle at the trivial
connection A = 0 is the real symplectic vector space underlying W
W . (Note we quantize by a
complex polarization, which leads to the Hilbert space H described above.) To implement gauge
symmetry we must take the symplectic quotient by the group G of gauge transformations. Fix
a basepoint (innity) on SV1 . Then the subgroup G0 of gauge transformations which equal the
identity at the basepoint acts freely on A, and the quotient is identied with U (1) via holonomy.
In the pure gauge theory


the symplectic quotient of T A by G0 is dieomorphic to T  U (1); its
quantization L2 U (1) was discussed previously. As the length V ! 1 recall that the only state
which remains of nite energy is the vacuum state. Classically, the vacuum corresponds to the
zero section, a lagrangian submanifold of T  U (1). The subgroup U (1)  G of constant gauge
transformations acts trivially on A|hence trivially on T  A|so does not enter into pure gauge
theory. In the theory with matter the symplectic quotient by G0 is a vector bundle over T  U (1)
with ber W  W at the identity. Now the constant gauge transformations act nontrivially in
the bers by scalar multiplication. To implement the symmetry we have two choices: we can
quantize the symplectic quotient or we can consider the subspace of Hilbert space annihilated by
the corresponding quantum operator Ne . Pursuing rst the latter, we see that in the innite volume
limit the quantum Hilbert space before implementing the U (1) is simply H , since the gauge eld
contributes only the vacuum state. The operator Ne is simply equal to Q . Thus the Hilbert space
of the theory is the subspace of states with total charge Q = 0. Therefore, just as in the classical
theory we have connement. There is a mass gap, and the smallest mass is 2M . In the theory with
small coupling, the qualitative picture is the same.
If instead we take the symplectic quotient of W  W by U (1), we are led to a singular space.
The moment map is (w; w 0 ) = jwj2 jw 0 j2 , and  1 (0) is singular at (0; 0). In the quantization
this singular point corresponds to the vacuum, and it is not hard to see heuristically that we are
led to the same description as before.
Now allow  6= 0. In the language of geometric quantization the prequantum line bundle
over T  U (1) is now twisted by the pullback of L ! U (1). Quantum states correspond to \Bohr-
Sommerfeld" leaves of the given polarization, which are equally spaced parallel circles in the cylin-
der T U (1). More precisely, if p is the (momentum) coordinate in the cotangent space, the circles
occur at p = n =2 for integral n. The energy of the corresponding quantum state (in pure
gauge theory) is given in (26), and as V ! 1 we only keep the closest circle(s) to p = 0, which
corresponds to the vacuum state. If  6=  there is a unique such circle, and we have the same
picture of connement as above. For  =  there are two such circles, corresponding to the two
vacuum states. Thus in the theory with matter, before dividing out by the constant gauge trans-
formations, we must quantize two disjoint copies of W  W to obtain H  H . If, as before, we
14
assign zero charge to each of the vacuum states, then the operator Ne which corresponds to the
U (1) symmetry is Q  Q. The subspace of H  H annihilated by Q  Q is simply two
copies of the system seen previously, each copy with a vacuum. These are two \realizations" of
the quantum theory, each with a mass gap of size 2M . It turns out|and this is hard to explain
from this viewpoint|that we can also assign dierent charges to the two vacuua. In that case the
operator Ne is (Q 1)  Q or Q  (Q + 1). The kernel in each case has one particle states
of mass M , and there is no connement. There is no vacuum state in either realization. The four
realizations correspond to the classical picture of the previous section.
To justify these heuristic pictures we compute. From (20) we see that the Hamiltonian is
Z 1
dx 21 jj2 + jdAj2 + M 2jj2 + 2Ne2 jfAj2 ;


(30) H=
1
where  is a eld on R. Here  is the conjugate momentum to , and f = fA is the Hodge star
of the curvature as before. We derive an eective Hamiltonian for  by plugging in the equation of
motion of A. The latter is obtained by varying the lagrangian (20) with respect to A:

(31) N df = j;
e2 dx
where the current is
j =   :
We integrate (31) to obtain
e2 Z 1 
(32) f (x) = N dy G(x y)j (y) + c
1
for some constant c, where 
G(x y) = 1; x  y;
0; x < y:
Note that
2
(33) lim f (x) = eNc :
x! 1

Plugging into (30) we see

H = H0 +
H;
where H0 is the Hamiltonian for the free scalar particle (if we compute at A = 0), and
2 Z1 Z 1 2

H = 2eN dx dy G(x y)j (y) + c :
1 1
15
The perturbation term
H is nonlocal and singular.
Now we must determine the subspace of states on which
H is nite. Consider rst c = 0.
Then h j
H j i is nite if and only if

lim h jf (x)j i = 0;
x!1

where f (x) is dened by (32). Now

Z 1 Z 1
xlim
!1 dy G(x y)j (y) = dy j (y) = Q
1 1

is the charge operator in the Hilbert space H . So if c = 0 we have connement:

h jQj i = 0:

In general, c is related to  through (33) and (29):

 
jcj = min 
n 2 + n :

This is the value of the electric eld at 1. Also, we subtract an (innite) constant from
H to
account for the nonzero energy at 1:
"Z 2 #
e2 Z1 1
(
H )normalized = 2N dx dy G(x y)j (y) + c c2 :
1 1

This is nite on states which satisfy

(34) Q (Q + 2c)j i = 0;

i.e., Q j i = 0 or Q j i = 2c. For  6=  the only possibility is Q j i = 0 and we have conne-
ment. For  =  we have either c = 1=2, and (34) is satised by states with Q j i = 0 or Q j i =
1. As in the classical theory, this gives four sectors. We denote them H++; H+ ; H + ; H ac-
cording to the value of the electric eld at 1 and +1. There is a vacuum state and connement
in H++ ; H ; there is neither a vacuum nor connement in H+ ; H + .
In the conning cases the symmetry group PSU (N ) acts. In the sectors of the  =  theory
with no connement the covering group SU (N ) acts.
The story for k > 1 is similar.
16
 Finally, we make a remark about the electric charge. In the quantum picture it is an operator Qe ,
and from Noether's theorem applied to (20) we compute the relation to Ne :
2
Qe = eN (Ne + 2 ):
2
The eigenvalues of Ne are integral, but those of Qe are shifted from eN Zif  6= 0. Note the ow
(monodromy) in the eigenvalues of Qe as  runs around the circle from  = 0 to  = 2 . We will
encounter this phenomenon again in four dimensional gauge theory.

17