a

r
X
i
v
:
m
a
t
h
/
0
4
0
1
3
2
5
v
2


[
m
a
t
h
.
R
T
]


2
7

F
e
b

2
0
0
4
Standard Young tableaux for nite root systems
Arun Ram

Department of Mathematics
Princeton University
Princeton, NJ 08544
[email protected]
Preprint: March 26, 1998
Abstract. The study of representations of ane Hecke algebras has led to a new notion of shapes
and standard Young tableaux which works for the root system of any nite Coxeter group. This
paper is completely independent of ane Hecke algebra theory and is purely combinatorial. We
dene generalized shapes and standard Young tableaux and show that these new objects coincide
with the classical ones for root systems of Type A. The classical notions of conjugation of shapes,
ribbon shapes, axial distances, and the row reading and column reading standard tableaux, have
natural generalizations to the root system case. In the nal section we give an interpretation of
the shapes and standard tableaux for root systems of Type C which is in a form similar to classical
theory of shapes and standard tableaux.
0. Introduction
In my recent work on representations of ane Hecke algebras [Ra1] I have been led to a
generalization of standard Young tableaux. These generalized tableaux are important in the context
of representation theory because the standard tableaux model the internal structure of irreducible
representations of the ane Hecke algebra. In fact, most of the time the number of tableaux of
a given shape is the same as the dimension of the corresponding irreducible representation of the
ane Hecke algebra.
In this paper I introduce and study generalized shapes and standard tableaux purely combina-
torially. The main theorem is that the generalized standard tableaux of a given shape describe the
connected components of a certain graph, the calibration graph. It is this graph which is intimately
connected to the structure of representations of ane Hecke algebras.
In the Type A case the generalized shapes can be converted into placed congurations of
boxes. This conversion is nontrivial and is the subject of Section 3. In the cases where this placed
conguration of boxes is a placed skew shape the generalized standard tableaux coincide with the
classical standard tableaux of a skew shape. The generalized skew shapes play a major role in the
results on representations of ane Hecke algebras which are obtained in [Ra1].
In Section 1 I give denitions of
(a) skew shapes,
(b) ribbon shapes,
(c) axial distances,
(d) conjugation of shapes, and

Research supported in part by National Science Foundation grant DMS-9622985 and a Postdoctoral Fellowship at
the Mathematical Sciences Research Institute.
2 arun ram
(e) row reading and column reading tableaux,
in the generalized setting. In Section 4 it is shown that these denitions yield the classical versions
of these objects in the Type A case. The last section of this paper explains how one can convert
the generalized shapes and standard tableaux for the Type C case into congurations of boxes and
llings. In this form the shapes and standard tableaux for Type C look similar to the classical
standard Young tableaux.
It is my hope that others will also take up the study of the generalized shapes and standard
tableaux introduced in this paper. There are many more problems than there is time for solving
them and every combinatorial fact which can be proved about these objects says something about
the structure of representations of ane Hecke algebras. One hopes that everything that is known
about classical standard Young tableaux will have an analogue in this more general setting. Al-
though I have uncovered some of these generalizations, there are many facets of classical tableaux
theory which still need to be generalized.
From a representation theoretic point of view, one expects that there might exist generaliza-
tions of
(a) the Robinson-Schensted-Knuth correspondence,
(b) the Littlewood-Richardson coecients,
(c) the Kostka-Foulkes polynomials,
(d) major index and descents of tableaux,
(e) charge of tableaux,
(f) Jacobi-Trudi formulas.
Any solutions to these problems would be extremely helpful for understanding the underlying
representation theory. It is possible that some of the generalizations might be obtained simply by
understanding how to do them for the Type C tableaux given in the last section of this paper.
This approach is attractive since the form of the Type C tableaux given in the last section looks
so similar to classical tableau theory.
Remarks on the results in this paper
(1) Recently, it has become clear to me that one of the reasons that skew Schur functions play
such a important role in the classical theory of symmetric functions is because the skew shapes
describe particularly well behaved irreducible representations of the ane Hecke algebra of
type A. Since these nice representations exist in all types it seems reasonable that the many
wonderful identities involving skew Schur functions should have general type analogues in
terms of the generalized skew shapes dened in this paper. An example of a skew Schur
function identity that has a particularly nice generalization to all types is the identity in [Mac
I 5 Ex. 21b].
(2) A remark similar to (1) can be made concerning ribbon shapes. This special class of shapes
has a good generalization to all types and the representation theory associated to ribbon
shapes has special features [Ro], [Mat, 4.3.5]. This fact seems to give some philosophical
reason why there is such an amazing theory of ribbon Schur functions. The theory of
ribbon Schur functions has been developed in the last decade by Lascoux, Leclerc, Thibon,
Krob, Reutenauer, Malvenuto and others [GK] [La1-2]. I am sure that there is much to learn
about the representation theory of ane Hecke algebras from what is already known in the
ribbon (and noncommutative) Schur function theory.
(3) The theory of generalized shapes gives rise to some strange new shapes even in the type A
case, see Section 3.6. To my knowledge these shapes have not been studied before but they do
retain many of the combinatorial properties that skew shapes have. In particular, standard
standard young tableaux for root systems 3
tableaux make perfectly good sense for these shapes and these strange standard tableaux do
have representation theoretic meaning.
(4) There seem to be strong connections between the combinatorics in this paper, the combina-
torics of the Shi arrangements (see [St1-2], [AL], [He], [ST]) and the combinatorics of sign
types developed by Shi [Sh3]. These connections need to be better understood.
(5) I do not think that there is a connection between the generalized standard Young tableaux
introduced in this paper and the generalized tableaux of P. Littelmann [Li1-2]. Littelmanns
analogue of tableaux are really a generalization of column strict tableaux not of standard
tableaux. Column strict tableaux give information about the representations of GL
n
(C) and
standard tableaux give information about the representations of the symmetric group S
n
.
There is an analogous dichotomy in the generalized case; Littelmanns generalized column
strict tableaux model the representations of complex semisimple Lie groups and my generalized
standard tableaux model the representations of ane Hecke algebras. At the moment, I do
not believe that there is any connection between the representation theories of the complex
Lie groups and the ane Hecke algebras which would allow one to transfer information from
one side to the other (except in the type A case, where one has a Schur-Weyl type duality.)
Acknowledgements
This paper is only a part of a large project [Ra1-3] [RR1-2] on representations of ane Hecke
algebras which I have been working on intensely for about a year. During that time I have beneted
from conversations with many people. To choose only a few, there were discussions with S. Fomin,
M. Vazirani, L. Solomon, F. Knop and N. Wallach which played an important role in my progress.
There were several times when I tapped into J. Stembridges fountain of useful knowledge about
root systems. G. Benkart was a very patient listener on many occasions. R. Simion, T. Halverson,
H. Barcelo, P. Deligne, and R. Macpherson all gave large amounts of time to let me tell them my
story and every one of these sessions was helpful to me in solidifying my understanding. I thank
C. Krilo for her amazing proofreading.
I single out Jacqui Ramagge with special thanks for everything she has done to help with
this project: from the most mundane typing and picture drawing to deep intense mathematical
conversations which helped to sort out many pieces of this theory. Her immense contribution is
evident in that some of the papers in this series on representations of ane Hecke algebras are
joint papers.
A portion of this research was done during a semester stay at Mathematical Sciences Research
Institute where I was supported by a Postdoctoral Fellowship. I thank MSRI and National Science
Foundation for support of my research.
1. Generalized shapes and standard Young tableaux
(1.1) Notations. Let W be a nite Coxeter group and let R be the root system of W. The
root system R spans a real vector space which we shall denote R
n
. There is an inner product on
R
n
via which W is a group generated by reections. Fix a system R
+
of positive roots in R and
write > 0 if R
+
. If w W the inversion set of w is
R(w) = > 0 [ w < 0 .
Let
1
, . . . ,
n
be the simple roots in R
+
and let s
1
, . . . , s
n
denote the corresponding simple
reections in W. The positive roots determine a fundamental chamber C = x R
n
[ x, ) >
0 for all R
+
. An element x R
n
is dominant if x C, the closure of the chamber C.
4 arun ram
(1.2) Placed shapes and standard tableaux. If C dene
Z() =
_
R
+
[ , ) = 0
_
and P() = R [ , ) = 1 .
The set Z() Z() is a (parabolic) root subsystem of R and is generated by the simple roots
that it contains. The stabilizer W

of in W is the Weyl group of the subsystem Z() and the

set P() is stable under the action of W

.
Remark. The sets Z() and P() appear in Heckman and Opdam [HO, Denition 2.1] and im-
plicitly in other works [K1, Theorem 2.2].
A placed shape is a pair (, J) where C and J P(). A standard tableau of shape (, J)
is an element w W such that
R(w) Z = and R(w) P = J.
In Theorem 3.5 we shall show that this is a generalization of the classical notion of standard Young
tableau. Let
T
(,J)
= standard tableaux of shape (, J).
Examples
(1) If is a generic element of R
n
then Z() = P() = . In this case the only possibility for J
is J = and T
(,)
= W.
(2) Let =
1
2

>0

, where

= 2/(, ). Then Z() = and P() =

1
, . . . ,
n
. If
J
1
, . . . ,
n
we have
T
(,J)
= w W [ D(w) = J, where D(w) =
i
[ ws
i
< w
is the right descent set of w W.
(1.3) A nonemptiness condition. I feel that the following conjecture should have a simple
slick proof. The = direction is easy and it may be possible to prove the other direction by using
[Bou, Ch. VI 1, Prop. 22]. For Type A the conjecture is true and is a consequence of Theorem
4.5 of section 4.
Conjecture. Let (, J) be a shape. The set T
(,J)
is non-empty if and only if J satises the
condition
If J, Z and R
+
then J.
(1.4) Placed skew shapes. If R
n
view as the function on the root system R given by
: R R
, )
The element is regular if , ) , = 0 for all R, and is integral if , ) Z for all R. For
each subset K
1
, . . . ,
n
let R
K
be the root system generated by K and let W
K
be the Weyl
group of R
K
. Let

K
denote the function restricted to R
K
.
A placed shape (, J) is a placed skew shape if for all w T
(,J)
,
(a) for each simple root
i
, w

{
i
}
is regular, and
standard young tableaux for root systems 5
(b) for each pair of simple roots
i
,
j

either w

{
i
,
j
}
is regular
or w

{
i
,
j
}
is in the W
{
i
,
j
}
-orbit of
the function given by ,
i
) = 1 and ,
j
) = 0,
where
i
is long and
j
is short.
Remark. In the type A case, a denition of skew shape similar to the one given above has also been
given by Fomin [Fo] in connection with his approach to the theory of seminormal representations
of the symmetric group.
(1.5) Ribbon shapes. A placed shape (, J) is a placed ribbon shape if Z() = . All placed
ribbon shapes are placed skew shapes.
Example. Suppose that (, J), is a ribbon shape and is integral. Then Z() = , P()

1
, . . . ,
n
and
T
(,J)
= w W [ L(w) P = J,
where, as in Example 2 of (1.2), L(w) is the left descent set of w.
(1.6) Although we shall not dene the ane Hecke algebra or discuss its representation theory
in this paper it is important to note that the denition of placed skew shape is motivated by the
following theorem of [Ra1].
Theorem. Let R be the root system of a nite Weyl group and let

H be the corresponding
ane Hecke algebra. There is a one-to-one correspondence between placed skew shapes (, J)
and irreducible really calibrated representations

H
(,J)
of the ane Hecke algebra

H. Under this
correspondence
dim(

H
(,J)
) = ( # of standard tableaux of shape (, J)).
(1.7) Conjugation. Let (, J) be a placed shape and let W

be the stabilizer of in W.
Dene the conjugate placed shape to be
(, J)

= (u, u(P() J)),

where u is the minimal length coset representative of w
0
W

W/W

and w
0
is the longest element
of W.
It will be useful to note the following:
(a) P(u) = uP(), (b) Z(u) = uZ(), (c) R(u) = R
+
Z().
The proofs are as follows:
(a) Since is dominant, u = w
0
is dominant and thus u, u) = 1 only if u > 0.
With this in mind P(u) = uP() follows from the equation u, u) = 1 , ) =
1.
(b) Let v W

such that w
0
= uv. (By [Bou IV 1 Ex. 3], v is unique.) Then R
+
w
0
Z() =
uvZ() = uZ(), and it follows that
Z(u) = R
+
R [ u, ) = 0 = R
+
(uZ() uZ()) = uZ().
6 arun ram
(c) Let R

= R
+
be the set of negative roots in R. Let v W

such that w
0
= uv. Then v is
the longest element of W

and R(v) = Z(). Thus

R(u) = R [ R
+
, w
0
v R

= R [ R
+
, v R
+
, since w
0
R

= R
+
,
= R
+
R(v) = R
+
Z().
(1.8) Proposition. Conjugation (, J) (, J)

is a well dened involution on placed shapes.

Proof.
(a) The weight u = uv = w
0
is dominant (i.e. u C) and u(P()J) P(u)
since uP() = P(u). This shows that (, J)

is well dened.
(b) Write w
0
= uv where where v is the longest element of W

. Similarly, write w
0
= u

where
u

is the minimal length coset representative of w

0
W
w
0

and v

is the longest element in W

w
0

.
Conjugation by w
0
is an involution on W which takes simple reections to simple reections
and W
w
0

= w
0
W

w
0
. It follows that v

= w
0
vw
0
. This gives
u

u = (w
0
v

)(w
0
v) = w
0
w
0
vw
0
w
0
v = 1.
Then, using fact (1.7a) above,
u

(P(u)(u(P()J))) = u

(uP()(u(P()J))) = P()(P()J) = J.
In (4.4) we shall show that this involution is a generalization of the classical conjugation operation
on partitions.
Remark. In type A, the conjugation involution seems to coincide with the duality operation for
representations of p-adic GL(n) dened by Zelevinsky [Ze]. Zelevinskys involution has been been
studied further in [MW] and [KZ] extended to general Lie type by Kato [K2] and Aubert [Au].
(1.9) Proposition. The conjugation of shapes involution extends to an involution on standard
tableaux given by
T
(,J)
11
T
(,J)

w wu
1
where u is the minimal length coset representative of the coset w
0
W

.
Proof. Let w T
(,J)
and let w

= wu
1
. We must show that R(w

) Z(u) = and
R(w

) P(u) = u(P( J)). Let R

= R
+
be the set of negative roots. We shall use the
facts in (1.7) freely.
(a) Since R(w) Z() = , we get
u
1
R(wu
1
) Z() = R [ u R(wu
1
), Z()
= R [ u R
+
, wu
1
u R

, Z()
= R [ u
1
R
+
, w R

, Z()
= R [ u
1
R
+
, R(w), Z(), since Z() R
+
,
= R [ u
1
R
+
, R(w) Z()
= ,
standard young tableaux for root systems 7
and it follows that
R(w

) Z(u) = R(wu
1
) uZ() = u
_
u
1
R(wu
1
) Z()
_
= .
(b) Assume that w W is such that R(w) P() = J. Then
u
1
R(wu
1
) P() = R [ u R(wu
1
), P()
= R [ u R
+
, wu
1
u R

, P()
= R [ u R

, w R
+
, P()
= R [ R(u), R
+
R(w), P(), since P() R
+
= R [ R
+
Z(), R
+
R(w), P()
= R [ R
+
Z(), P()J, since R(w) P() = J,
= P()J, since Z() and P() are disjoint.
This yields
R(w

) P(u) = R(wu
1
) uP() = u
_
u
1
R(wu
1
) P()
_
= u(P()J) .
(1.10) Axial distances. Let (, J) be a placed shape and let w T
(,J)
be a standard
tableaux of shape (, J). Let R be a root. The -axial distance for w is the value
d

(w) = w, ).
These numbers are crucial to the construction of irreducible representations of the ane Hecke
algebra which is given in [Ra1]. In (4.1) we shall see that they are analogues of the axial distances
used by A. Young [Y] in his constructions of the irreducible representations of the symmetric group.
(1.11) Row reading and column reading tableaux. If w W let R(w) be the inversion
set of w. The weak Bruhat order is the partial order on W given by
v w if R(v) R(w).
This denition is not the usual denition of the weak Bruhat order but is equivalent to the usual
one by [Bj, Prop. 2].
Let (, J) be a placed shape. A column reading tableau of shape (, J) is a minimal element
of T
(,J)
in the weak Bruhat order. A row reading tableau of shape (, J) is a maximal element of
T
(,J)
in the weak Bruhat order.
A set of positive roots K is closed if , K, + R
+
implies that + K. The
closure K of a subset K R
+
is the smallest closed subset of R
+
containing K.
The following conjecture says that if is integral and T
(,J)
is nonempty then T
(,J)
contains
a unique column reading tableau and a unique row reading tableau. A proof of the conjecture in
the type A case is given in Theorem 4.5.
Conjecture. Let (, J) be a placed shape such that is dominant and integral. If T
(,J)
,=
then
R(w
min
) = J, R(w
max
) = (P() J) Z()
c
, and T
(,J)
= [w
min
, w
max
],
where K
c
denotes the complement of K in R
+
and [w
min
, w
max
] denotes the interval between w
min
and w
max
in the weak Bruhat order.
2. Calibration graphs
8 arun ram
The following graphs arise naturally in the study of representations of ane Hecke algebras,
see [Ro], [Rg, Prop. 3.5], and [Ra1]. Let R
n
. The calibration graph () is the graph with
Vertices: W
Edges: w s
i
w, if w,
i
) , = 1.
(2.1) Theorem. Assume that R
n
is dominant and let () be the corresponding calibration
graph. The connected components of () are described by the sets
T
(,J)
such that J P() and T
(,J)
,= ,
where T
(,J)
is the set of standard tableaux of shape (, J) dened in (1.2).
(2.2) This theorem will become almost obvious once we change our point of view. The root
system R determines a (central) hyperplane arrangement
/ = H

[ R, where H

= x R
n
[ x, ) = 0.
The set of chambers (connected components) of R
n
(

) is
( = wC [ w W, where C = x R
n
[ x, ) > 0 for all > 0
is the fundamental chamber. A chamber

C ( is on the positive side (resp. negative side) of the
hyperplane H

, R
+
, if x, ) > 0 (resp. x, ) < 0) for all x

C. The following Proposition
allows us to view the calibration graph () in terms of chambers in R
n
. Example 2.7 at the end
of this section illustrates the conversion.
Proposition. Assume that R
n
is dominant. Let Z() and P() be as dened in (1.2) and let
() be the calibration graph containing .
(a) The map
W
11

C ( [

C is on the positive side of H

for every Z()

_
w w
1
C
is a bijection.
(b) Two vertices u and v in () are connected by an edge if and only if the corresponding
chambers u
1
C and v
1
C share a face and the hyperplane H

containing this face satises

, P().
(c) The map
T
(,J)
11

_
_
_

C (

C is on the positive side of H

for Z(),

C is on the positive side of H

for P()J,

C is on the negative side of H

for J,
_
_
_
w w
1
C
is a bijection.
standard young tableaux for root systems 9
Proof. (a) Since is dominant the stabilizer W

of is a parabolic subgroup of W. It is generated by

the simple reections s
i
for
i
Z(). The statement will follow from the following identications:
W
11
W/W

11
minimal length coset representatives of W/W

11
wC [ w minimal length coset representative of W/W

11
wC [ (ws
i
) > (w) for every
i
Z()
11

_
wC [ w
1
C is on the positive side of H

for every Z()

_
11

_
w
1
C [ w
1
C is on the positive side of H

for every Z()

_
,
where C = x E [ x, ) > 0 for all > 0 is the fundamental chamber. Let us give a step by
step explanation.
(1) There is a natural bijection between W and W/W

.
(2) Since each coset in W/W

has a unique coset representative of minimal length (see [Bou, IV.1,

Exercise 3] or [Hum, Proposition 1.10c]) we may identify W with the set of minimal length
coset representatives of W/W

.
(3) The chambers of R
n
(

) are the regions wC, w W and thus we can identify W with

the set wC [ w is a minimal length coset representative of W/W

.
(4) An element w W is a minimal length coset representative of W/W

if (ws
i
) > (w) for
every i such that
i
Z().
(5) So w is a minimal length coset representative if and only if, for any x C, w
1
x,
i
) > 0
for every
i
Z(). Since Z() Z() is a root subsystem generated by the simple roots
that it contains we have that w is a minimal length coset representative if and only if, for
any x C, w
1
x, ) > 0 for every Z(). Thus, we may identify W with the set
_
wC [ w
1
C is on the positive side of H

for every Z()

_
.
(6) The nal step is to replace each chamber wC by the chamber w
1
C.
(b) Recall that u and v are connected by an edge in () if u = s
i
v and v,
i
) , = 1.
If u and v are minimal length coset representatives of W/W

, then the chambers u

1
C and v
1
C
should be connected by an edge if u = s
i
v and , v
1

i
) , = 1. The condition that u = s
i
v means
that u
1
C = v
1
s
i
C. Since C and s
i
C share a face, it follows that u
1
C = v
1
s
i
C and v
1
C
share a face. This face is contained in the hyperplane H
v
1

i
= H
u
1

i
since the face that C and
s
i
C share is contained in the hyperplane H

i
. Finally, the condition , v
1

i
) , = 1 is the same
as saying that v
1

i
, P() and u
1

i
, P().
(c) If w W, > 0, x C and w < 0 then w
1
x, ) = x, w) < 0. Thus
R(w) =
_
> 0 [ w
1
C is on the negative side of H

_
.
So the condition R(w) Z() = is equivalent to the condition that w
1
C is on the positive side
of H

for all Z(). Similarly the condition R(w) P() = J is equivalent to the condition
that w
1
C be on the negative side of H

for all J and on the positive side of H

for all
P() J.
Parts (a) and (b) of Proposition 2.2 allow us to view the calibration graph in terms of chambers
in R
n
(

). The vertices correspond to the chambers on the positive side of the hyperplanes
H

, Z(). The edges of the graph are now the walls between the chambers. The only time that
a wall between two chambers does not form an edge of the graph connecting the two chambers
(vertices) is when that wall is contained in a hyperplane H

with P().
10 arun ram
(2.3) Proof of Theorem 2.1 From parts (a) and (b) of Proposition 2.2 we get that the connected
components of the graph () correspond to the connected components of the intersection of
_

C [

C is on the positive side of H

for Z()
_
with R
n

_
_
_
P()
H

_
_
.
Proposition 2.2 (c) says that the standard tableaux of shape (, J) correspond to the chambers
which are on the positive side of H

for Z() (P() J) and on the negative side of H

for J. The points in the chambers which satisfy these conditions form a connected subset of
R
n

P()
H

_
since the conditions describe them as the points in the intersection of half-
spaces in R
n
. Finally, one only needs to note that the connected components R
n

P()
H

_
are determined by which subset of hyperplanes in P() they are on the negative side of. This
completes the proof of Theorem 2.1.
(2.4) Invariance properties for calibration graphs.
Invariance property (1): As unlabeled graphs, () = (w) for all w W.
Invariance property (2): If Z() = Z() and P() = P() then () = ().
Property (1) is immediate from the denition of (). From the denition of the calibration graph
one sees that the edges of () are controlled by the set P(). Since the vertices W can be
identied with the set W/W

, where W

is the stabilizer of , it follows that the graph ()

depends only on W

and the set P(). Since W

is the group generated by the reections s

for
Z(), it follows that the structure of () is dependent only on the sets Z() and P(). This
establishes invariance property (2).
(2.5) Intersections and shapes. Let / be the arrangement of (ane) hyperplanes given by
/

= H

, H

, H
+
[ R
+
, where
H

= x R
n
[ x, ) = 0,
H
+
= x R
n
[ x, ) = 1,
H

= x R
n
[ x, ) = 1.
The intersection lattice L(/

) is the set of intersections, I =

B
H

, B /

, partially ordered
by inclusion (as subsets of R
n
). Since /

is symmetric under the Weyl group L(/

) also carries a
Weyl group symmetry. The quotient L(/

)/W is constructed by identifying intersections I

1
and I
2
if there is a w W such that wI
1
= I
2
. It follows from the invariance properties of the calibration
graphs that the distinct calibration graphs are in one-to-one correspondence with the elements of
the quotient L(/

)/W. In particular, the number of distinct calibration graphs is nite.

Let (, J) and (, K) be two placed shapes. We shall say that (, J) and (, K) have the same
underlying shape if there is a w W such that
Z(w) = Z(), P(w) = P(), and wJ = K.
When (, J) and (, K) are placed shapes with the same underlying shape then the calibration
graphs () and () are the same and there is a natural bijection
T
(,J)
11
T
(,K)
.
standard young tableaux for root systems 11
(2.6) It would be interesting to determine the size of L(/

)/W, i.e. the number of distinct

calibration graphs. A. Postnikov has explained that in the Type A case the numbers f(n) =
Card(L(/

)/S
n
) have the following generating function

k1
(1 q
k
)
2
k1
=

n1
f(n)q
n
.
One can prove this by counting the number of ways of constructing the congurations of boxes
described in Section 3.
The arrangement /

is very similar to the Shi arrangement

/

= H

, H
+
[ R
+
.
C. Athanasiadis has told me that, for the type A case, the number of elements of the intersection
lattice L(/

) which contain a dominant weight is equal to

n

k=1
_
n 1
k 1
_
F
2k1
where F
1
= 1, F
3
= 2, F
5
= 5, . . . are the odd Fibonacci numbers. The Shi arrangement has been
an object of intense recent study, see [Sh3], [St1], [AL]. There are many indications [Sh3], [Xi,1.11,
2.6] that there should be a strong connection between the work in this paper and the combinatorics
of the Shi arrangement.
(2.7) Example. Consider the root system of type C
2
where R
+
=
1
,
2
,
1
+
2
,
1
+2
2
.
Realize this root system in R
2
by letting
1
= 2
1
and
2
=
2

1
, where
1
,
2
is an orthonormal
basis of R
2
. Let R
2
be given by ,
1
) = 0 and ,
2
) = 1. Then is dominant (i.e. in C)
and integral and
Z() =
1
and P() =
2
,
1
+
2
.
H

1
H

2
H

1
+
2
H

1
+2
2
H

1
+2
2
+
H

1
+2
2

1
+
2
+
H

1
+
2

2
+
H

1
+

s
2

s
1
s
2

s
2
s
1
s
2

............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...........................................................................................................................................................................................................................................................................................................................................................................................................................................
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
The arrangement /

12 arun ram
J =
J =
2

J =
2
,
1
+
2

s
2

s
1
s
2

s
2
s
1
s
2

.............................................................................................................................................................................................................................................................................................
The calibration graph ()
When we convert to regions in R
2
(

) we get the picture

H

1
H

2
H

1
+
2
H

1
+2
2
J =
J =
2

J =
2
,
1
+
2

C
s
2
C
s
2
s
1
C
s
2
s
1
s
2
C
............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
The dashed line is the hyperplane corresponding to the root in Z() and the solid lines are the
hyperplanes corresponding to the roots in P().
3. Type A and congurations of boxes
(3.1) The root system. Let
1
, . . . ,
n
be an orthonormal basis of R
n
so that each sequence
= (
1
, . . . ,
n
) R
n
is identied with the vector =

i
. The root system of type A
n1
is
given by the sets
R =
j

i
[ 1 i, j n and R
+
=
j

i
[ 1 i < j n .
The Weyl group is W = S
n
, the symmetric group, acting by permutations of the
i
.
standard young tableaux for root systems 13
(3.2) Partitions, skew shapes, and (classical) standard tableaux. A partition is a
collection of n boxes in a corner. We shall conform to the conventions in [Mac] and assume that
gravity goes up and to the left.
Any partition can be identied with the sequence = (
1

2
. . .) where
i
is the number
of boxes in row i of . The rows and columns are numbered in the same way as for matrices. In
the example above we have = (553311).
If and are partitions such that
i

i
for all i we write . The skew shape /
consists of all boxes of which are not in . Let / be a skew shape with n boxes. Number the
boxes of each skew shape / along major diagonals from southwest to northeast and
write box
i
to indicate the box numbered i.
See Example 3.4 below. A (classical) standard tableau of shape / is a lling of the boxes in the
skew shape / with the numbers 1, . . . , n such that the numbers increase from left to right in
each row and from top to bottom down each column. Let T
/
be the set of standard tableaux of
shape /. Given a standard tableau t of shape / dene the word of t to be permutation
w
t
=
_
1 n
t(box
1
) . . . t(box
n
)
_
where t(box
i
) is the entry in box
i
of t.
(3.3) Placed skew shapes Let / be a skew shape with n boxes. Imagine placing / on
a piece of innite graph paper where the diagonals of the graph paper are indexed consecutively
(with elements of Z) from southeast to northwest. The content of a box b is the number c(b) of
the diagonal that b is on. Let
=
n

i=1
c(box
i
)
i
,
which we identify with the sequence = (c(box
1
), c(box
2
), . . . , c(box
n
)). The pair (, /) is a
placed skew shape. It follows from the denitions in (1.2) that
Z() =
j

i
[ j > i and box
j
and box
i
are in the same diagonal, and
P() =
j

i
[ j > i and box
j
and box
i
are in adjacent diagonals.
Dene
J =
_
_
_

j

i

j > i
box
j
and box
i
are in adjacent diagonals
box
j
is northwest of box
i
_
_
_
,
where northwest means strictly north and weakly west.
14 arun ram
(3.4) Example. The following diagrams illustrate standard tableaux and the numbering of
boxes in a skew shape /.
1
2 3
4
5
6
7
8
10
9
11
12 13 14
11
6 8
2
7
1
13
5
3
14
10
4 9 12
/ with boxes numbered A standard tableau t of shape /
The word of the standard tableau t is the permutation w
t
= (11, 6, 8, 2, 7, 1, 13, 5, 14, 3, 10, 4, 9, 12)
(in one-line notation).
The following picture shows the contents of the boxes in the placed skew shape (, /) with
= (7, 6, 5, 2, 0, 1, 1, 2, 2, 3, 3, 4, 5, 6).
-7
-6 -5
-2
0
1
1
2
3
2
3
4 5 6
Contents of the boxes of (, /)
In this case J =
2

1
,
6

5
,
8

7
,
10

8
,
10

9
,
11

9
,
12

11
.
(3.5) The following theorem shows how the generalized standard tableaux dened in (1.2)
reduce to the classical standard Young tableaux in the type A case.
Theorem. Let (, /) be a placed skew shape and let J be as dened in (3.3). Let T
/
be the
set of standard tableaux of shape / and let T
(,J)
be the set of generalized standard tableaux
of shape (, J) as dened in (1.2). Then the map
T
/
11
T
(,J)
t w
t
,
where w
t
is as dened in (3.2), is a bijection.
Proof. If w = (w(1) w(n)) is a permutation in S
n
then
R(w) =
j

i
[ j > i such that w(j) < w(i) .
The theorem is a consequence of the following chain of equivalences:
The lling t is a standard tableau if and only if for all 1 i < j n
(a) t(box
i
) < t(box
j
) if box
i
and box
j
are on the same diagonal,
(b) t(box
i
) < t(box
j
) if box
j
is immediately to the right of box
i
, and
standard young tableaux for root systems 15
(c) t(box
i
) > t(box
j
) if box
j
is immediately above box
i
.
These conditions hold if and only if
(a)
j

i
, R(w(t)) if
j

i
Z(),
(b)
j

i
, R(w(t)) if
j

i
P() J,
(c)
j

i
R(w(t)) if
j

i
J,
which hold if and only if
(a) , R(w(t)) if Z(), (b) , R(w(t)) if P() J, and (c) R(w(t)) if J.
Finally, these are equivalent to the conditions R(w(t)) Z() = and R(w(t)) P() = J.
(3.6) Placed congurations We have described how one can identify placed skew shapes
(, /) with pairs (, J). One can extend this conversion to associate placed congurations of
boxes to more general pairs (, J). The resulting congurations are not always skew shapes.
Let (, J) be a pair such that = (
1
, . . . ,
n
) is a dominant integral weight and J P().
(The sequence is a dominant integral weight if
1

n
and
i
Z for all i.) If J satises
the condition
If J, Z(), and R
+
then J
then (, J) will determine a placed conguration of boxes. As in the placed skew shape case, think
of the boxes as being placed on graph paper where the boxes on a given diagonal all have the
same content. (The boxes on each diagonal are allowed to slide along the diagonal as long as they
dont pass through the corner of a box on an adjacent diagonal.) The sequence describes how
many boxes are on each diagonal and the set J determines how the boxes on adjacent diagonals
are placed relative to each other. We want
=
n

i=1
c(box
i
)
i
,
and
(a) If
j

i
J then box
j
is northwest of box
i
, and
(b) If
j

i
P()J then box
j
is southeast of box
i
,
where the boxes are numbered along diagonals in the same way as for skew shapes, southeast means
weakly south and strictly east, and northwest means strictly north and weakly west.
If we view the pair (, J) as a placed conguration of boxes then the standard tableaux are
llings t of the n boxes in the conguration with 1, 2, . . . , n such that for all i < j
(a) t(box
i
) < t(box
j
) if box
i
and box
j
are on the same diagonal,
(b) t(box
i
) < t(box
j
) if box
i
and box
j
are on adjacent diagonals and box
j
is southeast of box
i
,
and
(c) t(box
i
) > t(box
j
) if box
i
and box
j
are on adjacent diagonals and box
j
is northwest of box
i
.
As in Theorem 3.5 the permutation in T
(,J)
which corresponds to the standard tableau t is
w(t) = (t(box
1
), . . . , t(box
n
)). The following example illustrates the conversion.
Example. Suppose = (1, 1, 1, 0, 0, 0, 1, 1, 1, 2, 2, 2) and
J =
4

1
,
4

2
,
4

3
,
5

2
,
5

3
,
7

5
,
7

6
,
8

6
,
10

9
,
10

8
,

10

7
,
11

9
,
11

8
,
11

7
,
12

9
.
16 arun ram
The placed conguration of boxes corresponding to (, J) is as given below.
-1
-1
-1
0
0
0
1
1
1
2
2
2
1
2
3
4
5
6
7
8
9
10
11
12
2
9
10
1
6
11
5
7
12
3
4
8
contents of boxes numbering of boxes a standard tableau
(3.7) Books of placed congurations. The most general case to consider is when =
(
1
, . . . ,
n
) is an arbitrary element of R
n
and J P(). This case is handled as follows. First
group the entries of according to their Z-coset in R. Each group of entries in can be arranged
to form a sequence
+C

= + (z
1
, . . . , z
k
) = ( +z
1
, . . . , +z
k
), where 0 < 1, z
i
Z and z
1
z
k
.
Fix some ordering of these groups and let
= (
1
+C

1
, . . . ,
r
+C

r
)
be the rearrangement of the sequence with the groups listed in order. Since is a rearrangement
of , the calibration graphs () and () are the same (see Invariance Property (1) in (2.4)). This
means that it is sucient to understand the standard tableaux corresponding to .
The decomposition of into groups induces decompositions
Z() =
_

i
Z

i
, P() =
_

i
P

i
, and, if J P(), then J =
_

i
J

i
,
where J

i
= J P

i
. Each pair (C

, J

) is a placed shape of the type considered in the previous

subsection and we may identify (, J) with the book of placed shapes ((C

1
, J

1
), . . . , (C

r
, J

r
)).
We think of this as a book with pages numbered by the values
1
, . . . ,
r
and with the placed
conguration determined by (C

i
, J

i
) on page
i
. In this form the standard tableaux of shape
(, J) are llings of the n boxes in the book with the numbers 1, . . . , n such that the lling on each
page satises the conditions for a standard tableau in (3.6).
Example. If = (1/2, 1/2, 1, 1, 1, 3/2, 2, 2, 1/2, 1, 1, 1, 1/2, 1/2, 0, 0, 0) then one possi-
bility for is
= (2, 2, 1, 1, 1, 0, 0, 0, 1, 1, 1, 1/2, 1/2, 1/2, 1/2, 1/2, 3/2).
In this case
1
= 0,
2
= 1/2,

1
+ C

1
= (2, 2, 1, 1, 1, 0, 0, 0, 1, 1, 1) and
2
+ C

2
= (1/2, 1/2, 1/2, 1/2, 1/2, 3/2).
standard young tableaux for root systems 17
If J = J

1
J

2
where J

2
=
14

13
,
17

16
and
J

1
=
3

2
,
4

2
,
5

2
,
6

3
,
6

4
,
6

5
,
9

7
,
9

8
,
10

7
,
10

8

then the book of shapes is

-1
-1
-1
0
0
0
-2
-2
1
1
1
0
0
0
-1
-1 1
0 1/2
where the numbers in the boxes are the contents of the boxes. The lling
4
5
9
2
12
1
13
15
8
11
17
6
10
16
3
7 14
0 1/2
is a standard tableau of shape (, J). This lling corresponds to the permutation
w = (2, 12, 4, 5, 9, 1, 13, 15, 8, 11, 17, 3, 7, 6, 10, 16, 14) in T
(,J)
S
16
.
4. Skew shapes, ribbons, conjugation, etc. in Type A
As in the previous section let R be the root system of Type A
n1
as given in (3.1). For clarity,
we shall state all of the results in this section for placed shapes (, J) such that is dominant and
integral, i.e. = (
1
, . . . ,
n
) with
1

n
and
i
Z. This simplication is mathematically
unimportant, the reason for it is that it allows us to avoid the notational diculties which arise
when one wants to use books of placed shapes with several pages.
(4.1) Axial distance Let (, J) be a placed shape such that is dominant and integral. Let
w T
(,J)
and let t be the corresponding standard tableau as dened by the map in Theorem 3.5.
Then it follows from the denitions of and w
t
in (3.2) and (3.3) that
w,
i
) = , w
1

i
) = c(box
w
1
(i)
) = c(t(i)),
18 arun ram
where t(i) is the box of t containing the entry i.
In classical standard tableau theory the axial distance between two boxes in a standard tableau
is dened as follows. Let be a partition and let t be a standard tableau of shape . Let 1 i, j n
and let t(i) and t(j) be the boxes which are lled with i and j respectively. Let (r
i
, c
i
) and (r
j
, c
j
)
be the positions of these boxes, where the rows and columns of are numbered in the same way
as for matrices. Then the axial distance from j to i in t is
d
ji
(t) = c
j
c
i
+r
i
r
j
,
(see [Wz]). This may seem confusing at rst but it is simpler if we rewrite it in terms of the
corresponding placed shape (, J) where is the sequence in R
n
determined by some placing of
on innite graph paper. Then one gets
d
ji
(t) = c(t(j)) c(t(i)) = w,
j

i
) = d

i
(w),
where w T
(,J)
is the permutation corresponding to the standard tableau t (see (3.2)) and d

(w)
is the generalized axial distance dened in (1.10). This shows that the axial distance dened in
(1.10) is a generalization of the classical notion of axial distance.
(4.2) Skew shapes. The following proposition shows that, in the case of a root system of type
A, the denition of generalized skew shape coincides with the classical notion of a skew shape.
Proposition. Let (, J) be a placed shape such that is dominant and integral. Then the
conguration of boxes associated to (, J) is a placed skew shape if and only if (, J) is a skew
shape as dened in (1.4).
Proof. =: We shall show that if the placed conguration corresponding to the pair (, J) has any
2 2 blocks of the forms
a b
c .
.
.
.
.
.
.
.
a
b
c
.
.
.
.
.
.
.
.
a
b
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Case (1) Case (2) Case (3)
then w

K
is not regular for some appropriate w T
(,J)
and subsystem R
K
in R. This will show
that the placed conguration must be a placed skew shape if (, J) is a generalized skew shape. In
the pictures above the shaded regions indicate the absence of a box and, for notational reference,
we have labeled the boxes with a, b, c.
Case (1): Create a standard tableau t such that the 2 2 block is lled with
i 1 i
i + 1
.
.
.
.
.
.
.
.
by lling the region of the conguration strictly north and weakly west of box c in row reading
order (sequentially left to right across the rows starting at the top), putting the next entry in box
c, and lling the remainder of the conguration in column reading order (sequentially down the
standard young tableaux for root systems 19
columns beginning at the leftmost available column). Let w = w(t) be the permutation in T
(,J)
which corresponds to the standard tableau t. Let t(i) denote the box containing i in t. Then, using
the rst identity (4.1),
w,
i
+
i+1
) = w,
i+1

i1
) = c(t(i + 1)) c(t(i 1)) = 0,
since the boxes t(i + 1) and t(i 1) are on the same diagonal. It follows that w

{
i+1
,
i
}
is not
regular.
Case (2): Create a standard tableau t such that the 2 2 block is lled with
i 1
i i + 1
.
.
.
.
.
.
.
.
by lling the region weakly north and strictly west of box c in column reading order, putting the
next entry in box c, and lling the remainder of the conguration in row reading order. Using this
standard tableau t, the remainder of the argument is the same as for case (1).
Case (3): Create a standard tableau t such that the 2 2 block is lled with
i 1
i
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
by lling the region strictly north and strictly west of box b in column reading order, putting
the next entry in box b, and lling the remainder of the conguration in row reading order. Let
w = w(t) be the permutation in T
(,J)
corresponding to t and let t(i) denote the box containing i
in t. Then
w,
i
) = w,
i

i1
) = c(t(i)) c(t(i 1)) = 0,
since t(i) and t(i 1) are on the same diagonal. It follows that w

{
i
}
is not regular.
=: Let Z
n
and / describe a placed skew shape (a skew shape placed on innite
graph paper). Let (, J) be the corresponding (generalized) placed shape as dened in (3.3). Let
w T
(,J)
and let t be the corresponding standard tableau of shape /. Consider a 2 2 block
of boxes of t. If these boxes are lled with
i
j
k
then either i < j < k < or i < k < j < . In either case we have i < 1 and it follows that
1 and are not on the same diagonal. Thus
w,

) = c(t()) c(t( 1)) ,= 0,

20 arun ram
and so w

}
is regular.
The same argument shows that one can never get a standard tableau in which and 2
occur in adjacent boxes of the same diagonal and thus it follows that w

{
1
,

}
is regular for
all w T
(,J)
.
Thus (, J) is a placed skew shape in the sense of (1.4).
(4.3) Ribbon Shapes. In classical tableaux theory a border strip (or ribbon) is a skew shape
which contains at most one box in each diagonal. Although the convention, [Mac, I 1 p. 5], is to
assume that border strips are connected skew shapes we shall not assume this.
Recall from (1.5) that a placed shape (, J) is a placed ribbon shape if is regular, i.e.
, ) , = 0 for all R.
Proposition. Let (, J) be a placed ribbon shape such that is dominant and integral. Then
the conguration of boxes coresponding to (, J) is a placed border strip.
Proof. Let (, J) be a placed ribbon shape with dominant and regular. Since = (
1
, . . . ,
n
)
is regular,
i
,=
j
for all i ,= j. In terms of the placed conguration
i
= c(box
i
) is the diagonal
that box
i
is on. Thus the conguration of boxes corresponding to (, J) contains at most one box
in each diagonal.
Example. If = (6, 5, 4, 0, 1, 3, 4, 5, 6, 7) and J =
2

1
,
5

4
,
7

6
,
9

8
,
10

9

then the placed conguration of boxes corresponding to (, J) is the placed border strip
-6
-5 -4
0
1
3
4 5
6
7
where we have labeled the boxes with their contents.
(4.4) Conjugation of Shapes. Let (, J) be a placed shape with dominant and integral
(i.e. = (
1
, . . . ,
n
) with
1

n
and
i
Z) and view (, J) as a placed conguration
of boxes. In terms of placed congurations, conjugation of shapes is equivalent to transposing the
placed conguration across the diagonal of boxes of content 0. The following example illustrates
this.
standard young tableaux for root systems 21
Example. Suppose = (1, 1, 1, 0, 0, 1, 1) and J = (
4

2
,
4

3
,
6

5
,
7

5
). Then the
placed conguration of boxes corresponding to (, J) is
-1 0
-1
-1
1
1
0
.
.
.
.
.
.
.
.
in which the shaded box is not a box in the conguration.
The minimal length representative of the coset w
0
W

is the permutation
u =
_
1 2 3 4 5 6 7
5 6 7 3 4 1 2
_
.
We have u = w
0
= (1, 1, 0, 0, 1, 1, 1) and
u(P() J) = u
4

1
,
5

1
,
5

2
,
5

3
,
6

4
,
7

4

=
3

5
,
4

5
,
4

6
,
4

7
,
1

3
,
2

3

=
5

3
,
5

4
,
6

4
,
7

4
,
3

1
,
3

2
.
Thus the conguration of boxes corresponding to the placed shape (, J)

is
0
1
-1
1
-1
0
1
.
.
.
.
.
.
.
.
(4.5) Row reading and column reading tableaux. Let (, J) be a placed shape such that
is dominant and integral and consider the placed conguration of boxes corresponding to (, J).
The minimal box of the conguration is the box such that
(m
1
) there is no box immediately above,
(m
2
) there is no box immediately to the left,
(m
3
) there is no box northwest in the same diagonal, and
(m
4
) it has the minimal content of the boxes satisfying (m
1
), (m
2
) and (m
3
).
There is at most one box in each diagonal satisfying (m
1
), (m
2
), and (m
3
). Thus, (m
4
) guarantees
that the minimal box is unique. It is clear that the minimal box of the conguration always exists.
The column reading tableaux of shape (, J) is the lling t
min
which is created inductively by
(a) lling the minimal box of the conguration with 1, and
(b) if 1, 2, . . . , i have been lled in then ll the minimal box of the conguration formed by the
unlled boxes with i + 1.
22 arun ram
The row reading tableau of shape (, J) is the standard tableau t
max
whose conjugate (t
max
)

is the
column reading tableaux for the shape (, J)

(the conjugate shape to (, J)).

Recall the denitions of the weak Bruhat order and closed subsets of roots given in (1.11).
Theorem. Let (, J) be a placed shape such that is dominant and integral (i.e. = (
1
, . . . ,
n
)
with
1

n
and
i
Z). Let t
min
and t
max
be the column reading and row reading tableaux
of shape (, J), respectively, and let w
min
and w
max
be the corresponding permutations in T
(,J)
.
Then
R(w
min
) = J, R(w
max
) = (P() J) Z()
c
, and T
(,J)
= [w
min
, w
max
],
where K
c
denotes the complement of K in R
+
and [w
min
, w
max
] denotes the interval between w
min
and w
max
in the weak Bruhat order.
Proof. (a) Consider the conguration of boxes corresponding to (, J). If k > i then either
c(box
k
) > c(box
i
), or box
k
is in the same diagonal and southeast of box
i
. Thus when we create
t
min
we have that
If k > i then box
k
gets lled before box
i
if and only if box
k
is northwest of box
i
,
where the northwest is in a very strong sense: There is a sequence of boxes
box
i
= box
i
0
, box
i
1
, . . . , box
i
r
= box
k
such that box
i
m
is either directly above box
i
m1
or in the same diagonal and directly northwest of
box
i
m1
. In other words,
If k > i then t
min
(box
k
) < t
min
(box
i
) box
k
is northwest of box
i
.
So, from the formula for w
t
in (3.2) we get
If k > i then w
min
(k) < w
min
(i)
k

i
J,
where w
min
is the permutation in T
(,J)
which corresponds to the lling t
min
and J is the closure
of J in R. It follows that
R(w
min
) = J.
(b) There are at least two ways to prove that R(w
max
) = (P() J) Z()
c
. One can mimic
the proof of part (a) by dening the maximal box of a conguration and a corresponding lling.
Alternatively one can use the denition of conjugation and the fact that R(w
0
w) = R(w)
c
. The
permutation w
min
is the unique minimal element of T
(,J)
and the conjugate of w
max
is the unique
minimal element of T
(,J)

. We shall leave the details to the reader.

(c) An element w W is an element of T
(,J)
if and only if R(w)P() = J and R(w)Z() =
. Thus T
(,J)
consists of those permutations w W such that
J R(w) (P() J) Z()
c
.
Since the weak Bruhat order is the ordering determined by inclusions of R(w), it follows that T
(,J)
is the interval between w
min
and w
max
.
standard young tableaux for root systems 23
Example. Suppose = (1, 1, 1, 0, 0, 1, 1) and J =
4

2
,
4

3
,
6

5
,
7

5
. The
minimal and maximal elements in T
(,J)
are the permutations
w
min
=
_
1 2 3 4 5 6 7
1 3 4 2 7 5 6
_
and w
max
=
_
1 2 3 4 5 6 7
1 5 6 2 7 3 4
_
.
The permutations correspond to the standard tableaux
1 2
3
4
5
6
7
.
.
.
.
.
.
.
.
and
1 2
5
6
3
4
7
.
.
.
.
.
.
.
.
.
5. Standard tableaux for type C in terms of boxes
(5.1) The root system. Let
1
, . . . ,
n
be an orthonormal basis of R
n
so that each sequence
= (
1
, . . . ,
n
) R
n
is identied with the vector =

i
. The root system of type C
n
is
given by the sets
R = 2
i
,
j

i
[ 1 i, j n and R
+
= 2
i
,
j

i
[ 1 i < j n .
The simple roots are given by
1
= 2
1
,
i
=
i

i1
, 2 i n. The Weyl group W = WC
n
is
the hyperoctahedral group of permutations of n, . . . , 1, 1, . . . , n such that w(i) = w(i). This
groups acts on the
i
by the rule w
i
=
w(i)
, with the convention that
i
=
i
.
(5.2) Rearranging . The analysis in this case is analogous to the method that was used in
(3.7) to create books of placed congurations in the type A case. For clarity, we recommend that
the reader compare the machinations below with the case done in (3.7).
Let R
n
. By applying an element of the Weyl group to we can rearrange the entries of
in increasing order (0
1

2

n
). Then, if
i
(x, x +1/2) for some integer x, replace

i
with
i
. Next group the elements of in terms of their Z-cosets and rearrange each group to
be in increasing order. There are three kinds of groups which can occur:
+ C

= ( + z
1
, +z
2
, . . . , +z
k
), with (x + 1/2, x + 1) for some x Z,
z
i
Z
0
and z
1
. . . z
k
,
C
0
= (z
1
z
2
z
k
), with z
i
Z
0
and z
1
. . . z
k
,
1/2 +C
1/2
= (1/2 + z
1
1/2 +z
2
1/2 + z
k
), with z
i
Z
0
and z
1
. . . z
k
.
Choose some ordering on the groups and let
= (
1
+ C

1
, . . . ,
r
+C

r
).
Because these changes are obtained by applying elements of the Weyl group, the calibration graphs
corresponding to and the are the same. Thus it is sucient to study the standard tableaux
corresponding to .
24 arun ram
(5.3) Books of placed congurations. As in the type A case, we have set partitions
Z() =
_

i
Z

i
, P() =
_

i
P

i
, and,
for any J P(),
J =
_

i
J

i
, where J

i
= J P

i
.
Each pair ( +C

, J

) is a placed shape and we may identify (, J) with the book of placed shapes
((
1
+C

1
, J

1
), . . . , (
r
+ C

r
, J

r
)) .
We think of this as a book with pages numbered by the values
1
, . . . ,
r
and with a placed
conguration determined by (
i
+C

i
, J

i
) on page
i
. One determines the placed congurations
as follows.
(5.4) Page , ,=
1
2
, 0: As in the type A case we will place boxes on a page of innite graph
paper which has the diagonals numbered consecutively with the elements of Z, from bottom left
to top right. For each 1 i n, place box
i
on diagonal
i
. The boxes on each diagonal are
arranged in increasing order from top left to bottom right. Then
P

=
j

i
[ j > i and box
i
and box
j
are in adjacent diagonals and
Z

=
j

i
[ j > i and box
i
and box
j
are in the same diagonal .
If J

arrange the boxes on adjacent diagonals according to the rules

(a

) if
j

i
J
r
place box
j
northwest of box
i
, and
(a

) if
j

i
P
r
J
r
place box
j
southeast of box
i
.
A standard tableau t is a lling of the boxes with distinct entries from the set n, . . . , 1, 1, . . . , n
such that if i appears then i does not appear and
(a) if j > i and box
j
and box
i
are in the same diagonal then t(box
i
) < t(box
j
),
(b) if j > i, box
i
and box
j
are in adjacent diagonals and box
j
is northwest of box
i
then t(box
i
) >
t(box
j
),
(c) if j > i, box
i
and box
j
are in adjacent diagonals and box
j
is southeast of box
i
then t(box
i
) <
t(box
j
).
Example. Suppose +C

= + (0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3) and
J

=
4

1
,
4

2
,
4

3
,
5

2
,
5

3
,
7

5
,
7

6
,
8

6
,
10

9
,
10

8
,

10

7
,
11

9
,
11

8
,
11

7
,
12

9
.
The placed conguration of boxes corresponding to ( +C

, J

) is as given below.
0
0
0
1
1
1
2
2
2
3
3
3
1
2
3
4
5
6
7
8
9
10
11
12
-9
-7
-5
-12
-8
3
-2
1
4
-11
-10
6
contents of boxes numbering of boxes a standard tableau
standard young tableaux for root systems 25
(5.5) Page
1
2
: We will place boxes on a page of innite graph paper which has the diagonals
numbered consecutively with the elements of
1
2
+ Z, from bottom left to top right. For each
1 i n such that
i

1
2
+C1
2
, place box
i
on diagonal
i
and box
i
on diagonal
i
. The boxes
on each diagonal are arranged in increasing order from top left to bottom right. With this placing
of boxes we have
P1
2
=
_
_
_

j

i
, such that j > i and box
i
and box
j
are in adjacent diagonals

j
+
i
, such that box
j
and box
i
are both in diagonal 1/2
2
i
, such that box
i
is in diagonal 1/2
_
_
_
,
Z1
2
=
j

i
, such that j > i and box
i
and box
j
are in the same diagonal .
If J1
2
P1
2
arrange the boxes on adjacent diagonals according to the rules:
(a1
2
) If
j

i
J1
2
place box
j
northwest of box
i
and box
i
northwest of box
j
.
(a

1
2
) If
j

i
P1
2
J1
2
place box
j
southeast of box
i
and box
i
southeast of box
j
.
(b1
2
) If
j
+
i
J1
2
(j i) place box
j
northwest of box
i
and box
i
northwest of box
j
.
(b

1
2
) If
j
+
i
P1
2
J1
2
(j i) place box
j
southeast of box
i
and box
i
southeast of box
j
.
A standard tableau t is a lling of the boxes with distinct entries from the set n, . . . , 1, 1, . . . , n
such that
t(box
i
) = t(box
i
)
and
(a) If j > i and box
j
and box
i
are in the same diagonal then t(box
i
) < t(box
j
),
(b) If j > i, box
i
and box
j
are in adjacent diagonals and box
j
is northwest of box
i
then t(box
i
) >
t(box
j
),
(c) If j > i, box
i
and box
j
are in adjacent diagonals and box
j
is southeast of box
i
then t(box
i
) <
t(box
j
).
Example. Suppose
1
2
+C1
2
=
_
1
2
,
1
2
,
1
2
,
1
2
,
3
2
,
3
2
,
3
2
,
3
2
,
5
2
,
5
2
,
7
2
_
and
J1
2
=
11

10
,
10

8
,
9

7
,
9

8
,
7

3
,
7

4
,
6

2
,
6

3
,
6

4
,
5

4
,

5

3
,
5

2
,
1
+
2
,
1
+
3
,
1
+
4
, 2
1
.
The placed conguration of boxes corresponding to (
1
2
+C1
2
, J1
2
) is as given below.
-
7
2
-
5
2
-
5
2
-
3
2
-
3
2
-
3
2
-
3
2
-
1
2
-
1
2
-
1
2
-
1
2
1
2
1
2
1
2
1
2
3
2
3
2
3
2
3
2
5
2
5
2
7
2
-11
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
1
2
3
4
5
6
7
8
9
10
11 -3
-5
8
-7
-2
9
10
-6
-4
-1
11
-11
1
4
6
-10
-9
2
7
-8
5
3
contents of boxes numbering of boxes a standard tableau
26 arun ram
(5.6) Page 0: We will place boxes on a page of innite graph paper which has the diagonals
numbered consecutively with the elements of Z, from bottom left to top right. For each 1 i n
such that
i
C
0
,
(1) place box
i
on diagonal 0 if
i
= 0, and
(2) if
i
,= 0 place box
i
on diagonal
i
and box
i
on diagonal
i
.
The boxes on each diagonal are arranged in increasing order from top left to bottom right. With
this placing of boxes we have
P
0
=
_

j

i
, such that j > i and box
i
and box
j
are in adjacent diagonals

j
+
i
, such that box
j
is in diagonal 1 and box
i
is in diagonal 0
_
,
Z
0
=
_
_
_

j

i
, such that j > i and box
i
and box
j
are in the same diagonal

j
+
i
, such that j > i and box
j
and box
i
are in diagonal 0
2
i
, such that box
i
is in diagonal 0
_
_
_
.
If J
0
P
0
then arrange the boxes on adjacent diagonals according to:
(a
0
) if
j

i
J
0
place box
j
northwest of box
i
and if box
i
is not on diagonal 0 place box
i
northwest of box
j
,
(a

0
) if
j

i
P
0
J
0
place box
j
southeast of box
i
and if box
i
is not on diagonal 0 place
box
i
southeast of box
j
,
(b
0
) if
j
+
i
J
0
place box
i
northwest of box
j
,
(b

0
) if
j
+
i
P
0
J
0
place box
i
southeast of box
j
.
A standard tableau t is a lling of the boxes with distinct entries from the set n, . . . , 1, 1, . . . , n
such that
t(box
i
) = t(box
i
), if box
i
is not on the zero diagonal,
t(box
i
) > 0, if box
i
is on the zero diagonal,
and
(a) if j > i and box
j
and box
i
are in the same diagonal then t(box
i
) < t(box
j
),
(b) if j > i, box
i
and box
j
are in adjacent diagonals and box
j
is northwest of box
i
then t(box
i
) >
t(box
j
),
(c) if j > i, box
i
and box
j
are in adjacent diagonals and box
j
is southeast of box
i
then t(box
i
) <
t(box
j
).
Example. Suppose C
0
= (0, 0, 0, 1, 1, 1, 2) and
J
0
=
4

1
,
4

2
,
4

3
,
5

1
,
5

2
,
5

3
,
6

1
,
6

2
,
6

3
,
7

6
,

6
+
1
,
5
+
1
,
5
+
2
,
4
+
1
,
4
+
2
.
The placed conguration of boxes corresponding to (C
0
, J
0
) is as given below.
-2
-1
-1
-1
0
0
0
1
1
1
2
-7
-6
-5
-4
1
2
3
4
5
6
7
4
2
5
6
1
3
7
-6
-5
-2
-4
contents of boxes numbering of boxes a standard tableau
standard young tableaux for root systems 27
(5.7) Using the above rules one produces a book of placed congurations corresponding to
(, J) = ((
1
+C

1
, J

1
), . . . , (
r
+C

r
, J

r
)). A standard tableau t for this book of congurations
is a lling of the boxes with distinct elements of n, . . . , 1, 1, . . . , n such that the lling on each
page satises the conditions for a standard tableau for that page. Let
T
((C

1
,J

1
),...,(C

r
,J

r
))
denote the set of such llings.
The proof of the following Theorem is similar to the proof of Theorem (3.5).
Theorem. Given a standard tableau t for the book of congurations ((C

1
, J

1
), . . . , (C

r
, J

r
))
dene w
t
WC
n
by w
t
(i) = t(box
i
). Then the map
T
((C

1
,J

1
),...,(C

r
,J

r
))
T
(,J)
t w
t
is a bijection.
28 arun ram
References
[AL] C. Athanasiadis and S. Linusson, A simple bijection for the regions of the Shi arrangement
of hyperplanes, Discrete Math., to appear.
[Au] M. Aubert, Dualite dans le groupe de Grothendieck de la categorie des representations lisses
de longueur nie dun groupe reductif p-adique, Trans. Amer. Soc. 347 (1995), 21792189.
[Bj] A. Bj orner, Orderings of Coxeter groups, Combinatorics and Algebra (Boulder, Colo. 1983),
Contemp. Math. 34, Amer. Math. Soc., Providence 1984, 175195.
[Bou] N. Bourbaki, Groupes et alg`ebres de Lie, Chapitres 4,5 et 6, Elements de Mathematique,
Hermann, Paris 1968.
[Fo] S. Fomin, Lecture notes, MIT, Fall 1996.
[GK] I.M. Gelfand, D. Krob, A. Lascoux, B. Leclerc, V. Retakh, J.-Y. Thibon, Non-
commutative symmetric functions, Adv. Math. 112 (1995), 218348.
[He] P. Headley, On a family of hyperplane arrangements related to the ane Weyl groups, J.
Algebraic Combinatorics 6 (1997), 331338.
[HO] G.J. Heckman and E.M. Opdam, Harmonic analysis for ane Hecke algebras, Current
Developments in Mathematics, 1996, International Press, Boston.
[Hu] J.E. Humphreys, Reection Groups and Coxeter Groups, Cambridge Studies in Mathematics
29, Cambridge Univ. Press 1990.
[K1] S. Kato, Irreducibility of principal series representations for Hecke algebras of ane type, J.
Fac. Sci. Univ. Tokyo sec 1A 28 (1981), 929943.
[K2] S. Kato, Duality for representations of a Hecke algebra, Proc. Amer. Math. Soc. 119 (1993),
941946.
[KZ] H. Knight and A.V. Zelevinsky, Representations of quivers of type A and the multisegment
duality, Adv. in Math., to appear.
[La1] A. Lascoux, B. Leclerc, J.-Y. Thibon, Ribbon tableaux, Hall-Littlewood functions, quan-
tum ane algebras, and unipotent varieties, J. Math. Phys. 38 (1997), 10411068.
[La2] A. Lascoux, P. Pragacz, Ribbon Schur functions, European J. Combin. 9 (1988), 561574.
[Li1] P. Littelmann, Paths and root operators in representation theory, Ann. of Math. (2) 142
(1995), 499525.
[Li2] P. Littelmann, The path model for representations of symmetrizable Kac-Moody algebras,
Proceedings of the International Congress of Mathematicians, Vol. 1 (Z urich, 1994), 298308,
Birkhauser, Basel, 1995.
[Mac] I.G. Macdonald, Symmetric functions and Hall polynomials, Second edition, Oxford Math-
ematical Monographs, Oxford Univ. Press, New York, 1995.
[Mat] H. Matsumoto, Analyse harmonique dans les systems de Tits bornologiques de type ane,
Lect. Notes in Math. 590, Springer-Verlag, 1977.
standard young tableaux for root systems 29
[MW] Moeglin and Waldspurger, Linvolution de Zelevinski, J. Reine Angew. Math. 372
(1986), 136177.
[Ra1] A. Ram, Calibrated representations of ane Hecke algebras, in preparation.
[Ra2] A. Ram, Irreducible representations of rank two ane Hecke algebras, in preparation.
[Ra3] A. Ram, Skew shape representations are irreducible, in preparation.
[RR1] A. Ram and J. Ramagge, Jucys-Murphy elements come from ane Hecke algebras, in
preparation.
[RR2] A. Ram and J. Ramagge, Calibrated representations and the q-Springer correspondence, in
preparation.
[Ro] F. Rodier, Decomposition de la serie principale des groups reductifs p-adic, in Noncommu-
tative Harmonic analysis and Lie Groups, Lect. Notes in Math. 880, Springer-Verlag, 1981.
[Rg] J. Rogawski, On modules over the Hecke algebra of a p-adic group, Invent. Math. 79 (1985),
443465.
[Sh1] J.-Y. Shi, The number of -sign types, Quart. J. Math. Oxford Ser. (2) 48 (1997), 93105.
[Sh2] J.-Y. Shi, Left cells in ane Weyl groups, T ohoku Math. J. (2) 46 (1994), no. 1, 105124.
[Sh3] J.-Y. Shi, Sign types corresponding to an ane Weyl group, J. London Math. Soc. (2) 35
(1987), 5674.
[ST] L. Solomon and H. Terao, The double Coxeter arrangement, preprint 1997.
[St1] R. Stanley, Hyperplane arrangements, interval orders, and trees, Proc. Nat. Acad. Sci.
U.S.A. 93 (1996), 26202625.
[St2] R. Stanley, Hyperplane arrangements, parking functions and tree inversions, in Mathemat-
ical Essays in Honor of Gian-Carlo Rota, Birkhauser, Boston, 1998.
[Xi] N. Xi, Representations of ane Hecke algebras, Lect. Notes in Math. 1587, Springer-Verlag,
Berlin, 1994.
[Wz] H. Wenzl, Hecke algebras of type A
n
and subfactors, Invent. Math. 92 (1988), 349383.
[Y] A. Young, On quantitative substitutional analysis (sixth paper), Proc. London. Math. Soc.
(2) 34 (1931), 196230.
[Ze] A. Zelevinsky, Induced Representations of p-adic groups II: On Irreducible Representations
of GL(n), Ann. Scient.

Ec. Norm. Sup. 4
e
s`erie, 13 (1980), 165210.