a

r
X
i
v
:
m
a
t
h
-
p
h
/
0
6
0
4
0
4
5
v
1


2
0

A
p
r

2
0
0
6
Random Point Fields for Para-Particles
of Any Order
Hiroshi Tamura

Department of Mathematics, Kanazawa University,
Kanazawa 920-1192, Japan
Keiichi R. Ito

Department of Mathematics and Physics, Setsunan University
Neyagawa, Osaka 572-8508, Japan
February 7, 2008
Abstract
Random point elds which describe gases consisting of para-particles of any
order are given by means of the canonical ensemble approach. The analysis for
the cases of the para-fermion gases are discussed in full detail.
1 Introduction
Where do the statistics of random point elds come from? We examine what kind of
random point elds follow from the para-statistics of particles.
In the previous paper [9], the boson and/or fermion point elds were derived by
means of the canonical ensemble approach. That is, quantum mechanical thermal sys-
tems of nite xed number of bosons and/or fermions in the bounded box in R
d
were
considered. By taking the thermodynamic limit to the position distribution of con-
stituents, random point elds for boson and/or fermion gases of positive nite densities
and temperatures on R
d
were obtained. There, the method was applied to construct the
random point elds which describe gases consist of para-bosons (resp. para-fermions)
of order 2. In the recent proceeding article [11], the argument for para-particle gases of
order 3 is developed.
In this paper, we pursue the project to the general case: we apply the method to
statistical mechanics of gases which consist of para-particles of any order p N. We
will see that the random point elds obtained in this way are those of = 1/p given

tamurah@kenroku.kanazawa-u.ac.jp

ito@mpg.setsunan.ac.jp
1
in [6]. Our main result in this paper is
Theorem The random point eld for gas of para-fermions (resp. para-bosons of low
density) of order p is equal in law to the convolution of p independent copies of the usual
fermion (resp. boson) point eld for any p N.
We use the representation theory of the symmetric group (cf. e.g. [2, 5, 7]). Its
basic facts are reviewed briey in Sect.2, along the line on which the quantum theory
of para-particles are formulated. We state our main results in Sect.3. Sections 4 and
5 devoted to the full detail of the discussions on the thermodynamic limits for para-
fermions and a few remarks on those for para-bosons, respectively. Some discussions are
given in Sect.6.
2 Brief review on Representation of the symmetric
group
We say that (
1
,
2
, ,
n
) N
n
is a Young frame of length n for the symmetric group
o
N
if
n

j=1

j
= N,
1

2

n
> 0.
We associate the Young frame (
1
,
2
, ,
n
) with the diagram of
1
-boxes in the rst
row,
2
-boxes in the second row,..., and
n
-boxes in the n-th row. A Young tableau on
a Young frame is a bijection from the numbers 1, 2, , N to the N boxes of the frame.
Let M
N
p
be the set of all the Young frames for o
N
which have lengths less than
or equal to p. For each frame in M
N
p
, let us choose one tableau from those on the
frame. The choices are arbitrary but xed. T
N
p
denotes the set of all tableaux chosen
in this way. The row stabilizer of a tableau T is denote by 1(T) , i.e., the subgroup of
o
N
consists of those elements that keep all rows of T invariant, and ((T) the column
stabilizer whose elements preserve all columns of T.
Let us introduce the three elements
a(T) =
1
#1(T)

R(T)
, b(T) =
1
#((T)

C(T)
sgn()
and
e(T) =
d
T
N!

R(T)

C(T)
sgn() = c
T
a(T)b(T)
of the group algebra C[o
N
] for each T T
N
p
, where d
T
is the dimension of the irreducible
representation of o
N
corresponding to T and c
T
= d
T
#1(T)#((T)/N!. As is known,
a(T
1
)b(T
2
) = b(T
2
)a(T
1
) = 0 (2.1)
2
hold for any o
N
if T
2
T
1
. The relations
a(T)
2
= a(T), b(T)
2
= b(T), e(T)
2
= e(T), e(T
1
)e(T
2
) = 0 (T
1
,= T
2
) (2.2)
also hold for T, T
1
, T
2
T
N
p
. For later use, let us introduce
d(T) = e(T)a(T) = c
T
a(T)b(T)a(T) (2.3)
for T T
N
p
. They satisfy
d(T)
2
= d(T), d(T
1
)d(T
2
) = 0 (T
1
,= T
2
) (2.4)
as is shown readily from (2.2) and (2.1). The inner product , ) of C[o
N
] is dened
by
, ) =

for , o
N
and the sesqui-linearity.
The left representation L and the right representation R of o
N
on C[o
N
] are dened
by
L()g = L()

S
N
g() =

S
N
g() =

S
N
g(
1
)
and
R()g = R()

S
N
g() =

S
N
g()
1
=

S
N
g(),
respectively. Here and hereafter we identify g : o
N
C with

S
N
g() C[o
N
].
They are extended to the representation of C[o
N
] on C[o
N
] as
L(f)g = fg =

,
f()g() =

f(
1
)g()
_

and
R(f)g = g

f =

,
g()f()
1
=

g()f()
_
,
where

f =

f() =

f(
1
) =

f()
1
.
The character of the irreducible representation of o
N
corresponding to the tableau
T T
N
p
is obtained by

T
() =

S
N
, L()R(e(T))) =

S
N
,

e(T)).
We introduce a tentative notation as in [9]

g
()

S
N
, L()R(g)) =

,S
N
,
1
)g() =

S
N
g(
1
) (2.5)
for g =

g() C[o
N
]. Then
T
=
e(T)
holds.
3
We consider representations of o
N
on Hilbert spaces. Let H
L
be a certain L
2
space
which will be specied in the next section and
N
H
L
its N-fold Hilbert space tensor
product. Let U be the representation of o
N
on
N
H
L
dened by
U()
1

N
=

1
(1)

1
(N)
for
1
, ,
N
H
L
,
or equivalently by
(U()f)(x
1
, , x
N
) = f(x
(1)
, , x
(N)
) for f
N
H
L
.
Obviously, U is unitary: U()

= U(
1
) = U()
1
. We extend U for C[o
N
] by linearity.
Then U(a(T)) is an orthogonal projection because U(a(T))

= U(

a(T)) = U(a(T)) and

(2.2). So are U(b(T))s, U(d(T))s and
P
pB
=

TT
N
p
U(d(T)). (2.6)
P
pB
is the projection operator to the subspace for para-bosons of order p. Note that
Ran U(d(T)) =Ran U(e(T)) because d(T)e(T) = e(T) and e(T)d(T) = d(T).
For para-fermions, we consider the transposed tableau T

of T T
N
p
by exchanging
the rows and the columns of the Young tableau T. The transpose

of frame is dened
in the same way. Then T

lives in

if T lives in . Clearly
((T

) = 1(T), 1(T

) = ((T) (2.7)
and we dene
P
pF
=

TT
N
p
U(d(T

)) (2.8)
which is the projection operator to the subspace for para-fermions of order p.
3 Para-statistics and Random Point Fields
3.1 Para-fermions of Order p N
We rst consider the quantum system of N para-fermions of order p in the box

L
= [L/2, L/2]
d
R
d
. We refer the literatures [3, 1, 8] for quantum mechanics of
para-particles. (See also [4].) These literatures indicate that the state space of our
system is given by H
pF
L,N
= P
pF

N
H
L
, where H
L
= L
2
(
L
) with Lebesgue measure is
the state space of one particle system in
L
. We need the heat operator G
L
= e

L
in

L
, where
L
is the Laplacian in
L
with periodic boundary conditions at
L
. Then
spec G
L
= exp[
d

j=1
q
2
j
]; q
j
2Z/L,
1
L
d
Tr G
L
=
1
L
d

nZ
d
exp
_
(2[n[/L)
2
_
.
4
It is obvious that there is a CONS of H
pF
L,N
which consists of the vectors of the form
U(d(T

))
(L)
k
1

(L)
k
N
, which are the eigenfunctions of
N
G
L
. Then, we dene a
point eld of N free para-fermions of order p in the box as in section 2 of [9] and its
generating functional is given by
E
pF
L,N
_
e
f,

TT
N
p
Tr

N
H
L
_
(
N

G)U(d(T

))

TT
N
p
Tr

N
H
L
_
(
N
G)U(d(T

))
,
where f is a nonnegative continuous function on
L
and

G
L
= G
1/2
L
e
f
G
1/2
L
. We rst
prove:
Lemma 1
E
pF
L,N
_
e
f,

TT
N
p

S
N

T
()Tr

N
H
L
[(
N

G
L
)U()]

TT
N
p

S
N

T
()Tr

N
H
L
[(
N
G
L
)U()]
(3.1)
=

TT
N
p
_

N
L
det
T

G
L
(x
i
, x
j
)
1i,jN
dx
1
dx
N

TT
N
p
_

N
L
det
T
G
L
(x
i
, x
j
)
1i,jN
dx
1
dx
N
. (3.2)
Remark 1 : The state space H
pF
L,N
= P
pF

N
H
L
is determined by the choice of
the tableaux Ts. The dierent choices of tableaux give dierent subspaces of
N
H
L
.
However, they are unitarily equivalent and the generating functional given above is not
aected by the choice. In fact,
T
() depends only on the frame on which the tableau
T is dened.
Remark 2 : det
T
A =

S
N

T
()

N
i=1
A
i(i)
in (3.2) is called immanant.
Proof : These expressions are derived by the relations
Tr

N
H
L
_
(
N
G
L
)U(d(T

))

= Tr

N
H
L
_
(
N
G
L
)U(e(T

))

and

S
N

g
()Tr

N
H
L
_
(
N
G
L
)U()

= N!Tr

N
H
L
_
(
N
G
L
)U(g)

, (3.3)
with g = e(T

). These relations can be shown by the use of (2.5), the cyclic property of
the trace and the commutativity of U() with
N
G. For details, see [9].
Now, let us consider the thermodynamic limit
L, N ,
L
N/L
d
> 0. (3.4)
In the following, f is a nonnegative continuous function on R
d
which has a compact
support, and is xed through the thermodynamic limit
L
R
d
. We identify the
restriction of f to
L
as f in Lemma 1. We get the limiting random point eld on R
d
.
5
Theorem 2 The nite random point elds for para-fermions of order p dened above
converge weakly to the point eld whose Laplace transform is given by
E
pF

_
e
f,

= Det
_
1
_
1 e
f
r

G(1 + r

G)
1
_
1 e
f

p
in the thermodynamic limit (3.4), where r

(0, ) is determined by

p
=
_
dp
(2)
d
r

e
|p|
2
1 + r

e
|p|
2
= (r

G(1 + r

G)
1
)(x, x).
3.2 Para-bosons of Order p N
We next consider the quantum system of N para-bosons of order p in the box
L
=
[L/2, L/2]
d
R
d
. It is obvious that there is a CONS of H
pB
L,N
which consists of the
eigenfunctions for
N
G
L
of the form U(d(T))
(L)
k
1

(L)
k
N
. Then, we dene a point
eld of N free para-bosons of order p as in section 2 of [9] and its generating functional
is given by
E
pB
L,N
_
e
f,

=
Tr

N
H
L
[(
N

G
L
)P
pB
]
Tr

N
H
L
[(
N
G
L
)P
pB
]
,
where f is a nonnegative continuous function on
L
and

G
L
= G
1/2
L
e
f
G
1/2
L
. Then, we
have:
Lemma 3
E
pB
L,N
_
e
f,

TT
N
p

S
N

T
()Tr

N
H
L
[(
N

G
L
)U()]

TT
N
p

S
N

T
()Tr

N
H
L
[(
N
G
L
)U()]
(3.5)
=

TT
N
p
_

N
L
det
T

G
L
(x
i
, x
j
)dx
1
dx
N

TT
N
p
_

N
L
det
T
G
L
(x
i
, x
j
)dx
1
dx
N
. (3.6)
We again consider the thermodynamic limit (3.4). We get the limiting random point
eld on R
d
for the low density region:
Theorem 4 The nite random point elds for para-bosons of order p dened above
converge weakly to the random point eld whose Laplace transform is given by
E
pB

_
e
f,

= Det
_
1 +
_
1 e
f
r

G(1 r

G)
1
_
1 e
f

p
in the thermodynamic limit, where r

(0, 1) is determined by

p
=
_
dp
(2)
d
r

e
|p|
2
1 r

e
|p|
2
= (r

G(1 r

G)
1
)(x, x),
if

p
<
c

_
R
d
dp
(2)
d
e
|p|
2
1 e
|p|
2
.
6
Remark : The high density region p
c
is related to the Bose-Einstein condensation.
We need a dierent analysis for the region. See [10] for the case of p = 1 and 2.
4 Proof of Theorem 2
It is enough to show the convergence of the generating functionals. In the rest of this
paper, we use the results in [9] frequently. We refer them as, e.g., Lemma I.3.2 for Lemma
3.2 of [9]. Although those in [9] are results for p = 1, their arguments hold for general
p N with obvious changes. Let
T
be the character of the induced representation
Ind
S
N
R(T)
[1], where 1 is the one dimensional representation 1(T) 1, i.e.,

T
() =

S
N
, L()R(a(T))) =
a(T)
().
Since the characters
T
and
T
depend only on the frame on which the tableau T lives,
not on T itself, we use the notation

and

( M
N
p
) instead of
T
and
T
,
respectively.
Let be the frame (p1, , 2, 1, 0) M
N
p
. Generalize

to those = (
1
, ,
p
)
Z
p
which satises

p
j=1

j
= N by

= 0 for Z
p
Z
p
+
and

for Z
p
+
and o
p
such that M
N
p
,
where Z
+
= 0 N. Then the determinantal form [2] can be written as

Sp
sgn
+
. (4.1)
Let us recall the relations

T
() = sgn
T
(),
T
() = sgn
T
(),
where

T
() =

, L()R(b(T

))) =
b(T

)
()
denotes the character of the induced representation Ind
S
N
C(T

)
[ sgn ], where sgn is the
representation ((T

) = 1(T) sgn . Then we have a variant of (4.1):

Sp
sgn

()
. (4.2)
Now let us consider the denominator of (3.1). Let T T
N
p
live on = (
1
, ,
p
)
M
N
p
. Thanks to (3.3) for g = b(T

), we have

S
N

T
()Tr

N
H
L
_
(
N
G
L
)U()

= N!Tr

N
H
L
_
(
N
G
L
)U(b(T

))

7
= N!
p

j=1
Tr

j
H
L
_
(

j
G
L
)A

,
where A
n
=

Sn
sgn()U()/n! is the anti-symmetrization operator on
n
H
L
. In the
last step, we have used
b(T

) =
p

j=1

R
j
sgn
#1
j
,
where 1
j
is the symmetric group of
j
numbers which lie on the j-th row of the tableau
T. Now (4.2) yields

S
N

()Tr

N
H
L
_
(
N
G
L
)U()

Sp
sgn

S
N

()
()Tr

N
H
L
_
(
N
G
L
)U()

= N!

Sp
sgn
p

j=1
Tr

j
j+(j)
H
L
_
(

j
j+(j)
G
L
)A

j
j+(j)

. (4.3)
Here we understand that Tr

n
H
L
_
(
n
G
L
)A
n
_
= 1 if n = 0 and = 0 if n < 0 in the last
expression. Applying the Cauchy integral formula to
Det[1 + zJ] =

n=0
z
n
Tr

n
H
[(
n
J)A
n
]
where J is a trace class operator, we obtain that
Tr

n
H
[(
n
G
L
)A
n
] =
_
Sr(0)
dz
2iz
n+1
Det[1 + zG
L
], (4.4)
where S
r
() = z C; [z [ = r. Note that r > 0 can be chosen arbitrary and
the right hand side equals 1 for n = 0 and 0 for n < 0. Then we have the following
expression of the denominator of (3.1):

M
N
p

S
N

()Tr

N
H
L
[(
N
G
L
)U()]
= N!

M
N
p

Sp
sgn
_

_
Sr(0)
p
p

j=1
Det(1 + z
j
G
L
) dz
j
2iz

j
j+(j)+1
j
.
= N!

M
N
p
_

_
Sr(0)
p

p
(z
1
, , z
p
)
_
p
j=1
Det(1 + z
j
G
L
)dz
j

p
j=1
2iz

j
+pj+1
j
, (4.5)
where
p
(z
1
, , z
p
) is the Vandermondian given by

p
(z
1
, , z
p
)

1i<jp
(z
i
z
j
) =

z
p1
1
z
p1
2
z
p1
p
.
.
.
.
.
.
.
.
.
.
.
.
z
1
z
2
z
p
1 1 1

. (4.6)
8
In the following, we simply write
p
(z) for (z
1
, , z
p
) when there is no danger of
confusion.
To make the thermodynamic limit procedure explicit, we take a sequence L
N

NN
which satises N/L
d
N
as N . In the following, we set r = r
N
[0, ) to be
the unique solution of
Tr
rG
L
N
1 + rG
L
N
= k (4.7)
where
k =
_
N
p
+
p 1
2
_
(4.8)
is the averaged length of the rows in the Young tableau and represents the integer
part.
The existence and the uniqueness of the solution follow from the fact that the left-
hand side of (4.7) is a continuous and monotone function of r. For details, see Lemma
I.3.2 [for = 1]. We also put
v
N
= Tr
_
r
N
G
L
N
(1 + r
N
G
L
N
)
2
_
. (4.9)
We will suppress the N dependence of v
N
and L
N
. Since r
N
r

in the thermodynamic
limit, we have k/(2 + r

) v
N
k for large enough N. See Lemma I.3.5. [There r
N
and r

are written as z
N
and z

respectively.]
Put
(4.5) =
N! Det[1 + rG
L
]
p
(

2)
p
(

v)
1+p(p1)/2
r
N
J
p
. (4.10)
Then we have:
Lemma 5
lim
N
J
p
=
1
p!
_
R
p
[(y
1
, , y
p
)[
_
p

j=1
y
j
_
p

j=1
e
y
2
j
/2
dy
j
> 0. (4.11)
Proof : We set

j
=
j
+ p j k.
Then we have
= (
1
, ,
p
) Z
p
, (4.12a)
p

j=1

j
=
0
N +
p(p 1)
2
pk [0, p), (4.12b)

1
>
2
> >
p
k. (4.12c)
The parametrization
z
j
= r exp(ix
j
/

v) (j = 1, , p)
9
yields
Det[1 + z
j
G
L
] = Det[1 + rG
L
]Det
_
1 + (z
j
r)G
L
(1 + rG
L
)
1

, (4.13a)
z
j
r = r(e
ix
j
/

v
1) = r
_
i sin(x
j
/

v) 2 sin
2
(x
j
/2

v)
_
, (4.13b)
dz
j
= ire
ix
j
/

v
dx
j
/

v, (4.13c)

p
(z) = r
1+2++(p1)

p
(e
ix/

v
) (4.13d)
and
J
p
=

_
p

j=1
_

v
dx
j

2
_

p
(e
ix/

v
)(

v)
1p+p(p1)/2

_
p

j=1
e
i(
j
+k)x
j
/

v
Det
_
1 + (e
ix
j
/

v
1)
rG
L
1 + rG
L
_
_
, (4.14)
where the summation on is taken over all satisfying (4.12).
We consider two regions of x (

v,

v]
1. small x region: [x[ v
1/12
,
2. large x region: [x[ > v
1/12
.
In the large eld region, we have

Det
_
1 + (z r)
G
L
1 + rG
L
_

2
= Det
_
1 4 sin
2
_
x
2

v
_
rG
L
1 + rG
L
_
1
rG
L
1 + rG
L
__
Det
_
1
4
1 + r
sin
2
_
x
2

v
_
rG
L
1 + rG
L
_
exp
_

4
1 + r
sin
2
_
x
2

v
_
Tr
rG
L
1 + rG
L
_
exp(const N
1/6
),
using 0 G
L
1 and v = O(k) = O(N) and the boundedness of r = r
N
> 0 uniformly
in N.
In the small eld region, we have the convergent expansion
Det
_
1 + (z r)
G
L
1 + rG
L
_
= exp
_
(e
ix/

v
1)k
1
2
(e
ix/

v
1)
2
(k v)
+

=3
(1)
1

(e
ix/

v
1)

Tr
_
rG
L
1 + rG
L
_

_
=
_
1 +
n1

=3
c

+ R
n
(x)
_
exp
_
ikx

v

1
2
x
2
_
, (4.15)
10
where
[c

[ const N
/6
, ( = 3, , n 1 ) (4.16)
and
[[R
n
[[

sup
|x|v
1/12
[R
n
(x)[ = O
_
N
n/12
_
hold. We put
n1

=3
c

= (x).
We choose n in (4.15) so large that

v)
1p+p(p1)/2
[[
p
[[

[[R
n
[[

= o(1) (4.17)
holds, i.e., n > 3(p 1)(p 2).
These arguments show that it is enough to consider the contribution from the small
x region, and we have
J
p
=
_

_
p

j=1
_
v
1/12
v
1/12
dx
j

2
e
i
j
x
j
/

vx
2
j
/2
(1 + (x
j
))
_

p
(e
ix/

v
)(

v)
1p+p(p1)/2
_
+ o(1),
=
_

_
p

j=1
_
1 +
_
i

j
__
_

dx
j

2
e
i
j
x
j
/

vx
2
j
/2
_

p
(e
ix/

v
)(

v)
1p+p(p1)/2
_
+ o(1). (4.18)
Thanks to the multi-linearity of the determinant
p
, we have
_
p

j=1
_

dx
j

2
e
i
j
x
j
/

vx
2
j
/2
_

p
(e
ix/

v
)
= det
__

dx
j

2
e
i(l
j
)x
j
/

vx
2
j
/2
_
p 1 0
1 j p
= det
_
e
(
j
)
2
/2v
_
p 1 0
1 j p
=
_
e
/v

_
exp
_

j=1

2
j
2v

p1

=0

2
2v
_
. (4.19)
Since (

v)
p(p1)/2

p
(e
/v
) =
p
(

v(e
/v
1)), the summation over all satisfying
the condition (4.12) yields
lim
N

v)
1p+p(p1)/2
(4.19)
11
=
_
y
1
>>yp
(
p

j=1
y
j
)(y
1
, , y
p
)
p

j=1
e
y
2
j
/2
dy
j
. (4.20)
Here we have put y
j
=
j
/

v and regarded (

v)
1p

as the integral of the suitable

step function. Then, (

v)
1p

_
dy (

j
y
j
) is derived by the use of the dominated
convergence theorem. The limit of the main term of (4.18) is given by (4.20), which is
equal to (5). We may see that the limit of the remainder vanishes from (4.16) and the
convergence of

_
p

j=1
(

v)
a
j

a
j

a
j
j
_

dx
j

2
e
i
j
x
j
/

vx
2
j
/2
_

p
(e
ix/

v
)(

v)
1p+p(p1)/2
=

v)
1p
_
p

j=1
(

v)
a
j

a
j

a
j
j
_

p
_

v(e
/v
1)
_
exp
_

j=1

2
j
2v

p1

=0

2
2v
_

_
y
1
>>yp
(
p

j=1
y
j
)
_
p

j=1

a
j
y
a
j
j
_

p
(y
1
, , y
p
)e

p
j=1
y
2
j
/2
_
p

j=1
dy
j
_
. (4.21)
We obtain this convergence by performing the dierentiations in the second and the
third members of (4.21) and applying the dominated convergence theorem.
The numerator is obtained just in the same way. That is, we replace G
L
by

G
L
=
G
1/2
L
e
f
G
1/2
L
and introduce r = r
N
and v = v
N
by
Tr
r
N

G
L
N
1 + r
N

G
L
N
= k, Tr
r
N

G
L
N
(1 + r
N

G
L
N
)
2
= v
N
.
G
L


G
L
is a positive operator of trace class such that Tr (G

G) = O([[1 e
f
)[[
1
)
and max spec G
L
max spec

G
L
= O(L
d
). Then it follows from the denitions of
r
N
, r
N
, v
N
and v
N
that
0 r
N
r
N
= O(N
1
), [ v
N
v
N
[ = O(1). (4.22)
See Lemma I.3.5 and Lemma I.3.6 for detail.
We dene

J
p
similarly ( v
N
is used instead of v
N
) and we get
lim
N

J
p
=
1
p!
_
R
p
[
p
(y
1
, , y
p
)[
_
p

j=1
y
j
_
p

j=1
e
y
2
j
/2
dy
j
> 0
by the very same argument as in the proof of Lemma 5.
12
Thus we have (writing r = r
N
, r = r
N
and L = L
N
)
(3.1) =
_
Det[1 + r

G
L
]
Det[1 + rG
L
]
_
p_
r
r
_
N
_
v
v
_
p(p1)/4+1/2
J
p
J
p
=
_
Det[1 + r

G
L
]
Det[1 + rG
L
]
_
p
Det
_
1 +
(r r)
1 + r

G
L

G
L
_
p

_
1 +
r r
r
_
N
_
v
v
_
p(p1)/4+1/2
J
p
J
p
. (4.23)
Here
_
Det[1 + r

G
L
]
Det[1 + rG
L
]
_
p
= Det
_
1 +
r
1 + rG
L
(

G
L
G
L
)
_
p
= Det
_
1
_
1 e
f
rG
L
1 + rG
L
_
1 e
f
_
p
Det
_
1
_
1 e
f
r

G
1 + r

G
_
1 e
f
_
p
holds. For details, we refer Proposition I.3.9 (and the argument on (c) in the proof of
Theorem I.3.1). The remaining factor of (4.23) tends to 1 as N since v/ v
1,

J
p
/J
p
1 and
N log
_
1 +
r r
r
_
p log Det
_
1 +
(r r)
1 + r

G
L

G
L
_
= N
r r
r
p
r r
r
Tr
r

G
L
1 + r

G
L
+ O(N
1
)
=
r r
r
(N pk) + O(N
1
) = O(N
1
).
Finally, we also get

p
= lim
N
1
L
d
N
Tr
r
N
G
L
N
1 + r
N
G
L
N
= (r

G(1 + r

G)
1
)(x, x)
from (4.7), (4.8), N/L
d
N
and Proposition I.3.9. This completes the proof of Theorem
2.
5 Proof of Theorem 4
In the case of para bosons, immanants (permanents)

S
N

N
j=1
G
L
(x
j
, x
(j)
) and

S
N

N
j=1

G
L
(x
j
, x
(j)
) appear in the denominator and the numerator, respectively.
We can represent them by the Fredholm determinants by making use of Vere-Jones
formula [12, 6]:
Tr

n
H
[(
n
G
L
)S
n
] =
_
Sr(0)
dz
2iz
n+1
Det[1 zG
L
]
1
,
13
where S
n
=

Sn
U()/n! is the symmetrization operator on
n
H
L
and r (0, 1) in
this case.
Using (4.1), we get the following expression of the denominator of (3.5):

M
N
p

S
N

()Tr

N
H
L
[(
N
G
L
)U()]
= N!

M
N
p

Sp
sgn
_

_
Sr(0)
p
p

j=1
dz
j
2iz

j
j+(j)+1
j
Det[1 z
j
G
L
]
= N!

M
N
p
_

_
Sr(0)
p

p
(z
1
, , z
p
) dz
1
dz
p
_
p
j=1
2iz

j
+pj+1
j
Det[1 z
j
G
L
]
_, (5.1)
where
p
(z
1
, , z
p
) is the Vandermondian introduced in the previous section.
We choose a sequence L
N

NN
which satises N/L
d
N
as N . In this case,
r
N
(0, 1) denotes the unique solution of
Tr
rG
L
N
1 rG
L
N
= k (5.2)
where
k =
N
p
+
p 1
2
(5.3)
as in (4.8). We put
v
N
= Tr
_
r
N
G
L
N
(1 r
N
G
L
N
)
2
_
. (5.4)
The remaining parts are almost the same as those in the para-fermion case. The reader
may complete the proof of Theorem 4, following the previous arguments with the obvious
changes.
6 Discussion
We have shown that
the generating functional of the gas of para-fermions (resp. para-bosons of low density)
of order p is equal to the p-th power of the generating functional of fermions (resp.
bosons).
The random point elds which we have obtained in this paper are a subset of those
in [6], where various properties of the point elds are examined. On the other hand, the
authors of [6] obtained the point elds which do not follow from the representation theory
of the symmetric groups which we discussed in this paper. Therefore it is interesting
to consider physical interpretations of the point elds which do not follow from the
representation theory of the symmetric groups. See e.g. [13].
Acknowledgements. We would like to thank Professors Y.Takahashi and T. Shirai for
useful discussions. H.T. is grateful to the GrantinAid for Science Research No.17654021
14
from MEXT. K.R.I. would like to thank the GrantinAid for Science Research (C)15540222
from JSPS.
References
[1] J.B. Hartle and J.R. Taylor, Quantum mechanics of paraparticles, Phys. Rev. 178
(1969) 20432051.
[2] G. James and A. Kerber, The Representation Theory of the Symmetric Group (
Encyclopedia of mathematics and its applications 16 (Addison-Wesley, London,
1981)
[3] A.M.L. Messiah and O.W. Greenberg, Symmetrization postulate and its experi-
mental foundation, Phys. Rev. 136 (1964) B248B267.
[4] Y. Ohnuki and S. Kamefuchi, Wave functions of identical particles, Ann. Phys. 51
(1969) 337358.
[5] B.E. Sagan, The Symmetric Group (Brooks/Cole, Pacic Grove, CA, 1991).
[6] T. Shirai and Y. Takahashi, Random point elds associated with certain Fredholm
determinants I: fermion, Poisson and boson point processes, J. Funct. Anal. 205
(2003) 414463.
[7] B. Simon, Representations of Finite and Compact Groups (A. M. S., Providence,
1996).
[8] R.H. Stolt and J.R. Taylor, Classication of paraparticles, Phys. Rev. D 1 (1970)
22262228.
[9] H. Tamura and K.R. Ito, A Canonical Ensemble Approach to the Fermion/Boson
Random Point Processes and its Applications, Commun. Math. Phys., 263 (2006)
363380.
[10] H. Tamura and K.R. Ito, A Random Point Field related to Bose-Einstein Conden-
sation, available via http://arxiv.org/abs/math-ph/0509071
[11] H. Tamura and K.R. Ito, Random Point Fields for Para-Particles of order 3, avail-
able via http://www.setsunan.ac.jp/mpg/confs/rims05/proc/tamura.pdf. To ap-
pear in RIMS lecture notes No.1482 (2006).
[12] D.Vere-Jones, A Generalization of Permanents and Determinants, Linear Algebra
Appl., 111 (1988) 119-124
[13] F.Wilczek, Fractional Statistics and Anyon Superconductivity (World Scien-
tic, Singapore, 1990); F.D.M.Haldane, Phys.Rev.Letters, 67 (1991) 937940;
F.Ferrari, H.Kleinert and E. Lezzizzera, Field Theory of N Entangled Polymers,
arXiv:cond-mat/0005300
15