Rigged Configurations and Kashiwara Operators
For types An⁽¹⁾ and Dn⁽¹⁾ we prove that the rigged configuration bijection intertwines the classical Kashiwara operators on tensor products of the arbitrary Kirillov-Reshetikhin crystals and the set of the rigged configurations.
Збережено в:
| Опубліковано в: : | Symmetry, Integrability and Geometry: Methods and Applications |
|---|---|
| Дата: | 2014 |
| Автор: | |
| Формат: | Стаття |
| Мова: | Англійська |
| Опубліковано: |
Інститут математики НАН України
2014
|
| Онлайн доступ: | https://nasplib.isofts.kiev.ua/handle/123456789/146830 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | Rigged Configurations and Kashiwara Operators / R. Sakamoto // Symmetry, Integrability and Geometry: Methods and Applications. — 2014. — Т. 10. — Бібліогр.: 37 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1859902933411299328 |
|---|---|
| author | Sakamoto, R. |
| author_facet | Sakamoto, R. |
| citation_txt | Rigged Configurations and Kashiwara Operators / R. Sakamoto // Symmetry, Integrability and Geometry: Methods and Applications. — 2014. — Т. 10. — Бібліогр.: 37 назв. — англ. |
| collection | DSpace DC |
| container_title | Symmetry, Integrability and Geometry: Methods and Applications |
| description | For types An⁽¹⁾ and Dn⁽¹⁾ we prove that the rigged configuration bijection intertwines the classical Kashiwara operators on tensor products of the arbitrary Kirillov-Reshetikhin crystals and the set of the rigged configurations.
|
| first_indexed | 2025-12-07T15:58:23Z |
| format | Article |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications SIGMA 10 (2014), 028, 88 pages
Rigged Configurations and Kashiwara Operators?
Reiho SAKAMOTO
Department of Physics, Tokyo University of Science, Kagurazaka, Shinjuku, Tokyo, Japan
E-mail: reiho@rs.tus.ac.jp, reihosan@08.alumni.u-tokyo.ac.jp
URL: https://sites.google.com/site/affinecrystal
Received April 02, 2013, in final form February 28, 2014; Published online March 23, 2014
http://dx.doi.org/10.3842/SIGMA.2014.028
Abstract. For types A
(1)
n and D
(1)
n we prove that the rigged configuration bijection in-
tertwines the classical Kashiwara operators on tensor products of the arbitrary Kirillov–
Reshetikhin crystals and the set of the rigged configurations.
Key words: crystal bases; rigged configurations; quantum affine algebras; box-ball systems
2010 Mathematics Subject Classification: 17B37; 05E10; 82B23
Dedicated to Professor Anatol N. Kirillov with admiration
Contents
1 Introduction 3
2 Background on crystals and tableaux 5
2.1 Crystal bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Kirillov–Reshetikhin tableaux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.1 Reviews on tableau representations . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.2 Filling map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Rigged configurations and the main bijection 11
3.1 Definition of the rigged configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 Convexity relations of the vacancy numbers and their applications . . . . . . . . . . . . . 13
3.3 Rigged configuration bijection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4 Rigged configurations and the Kashiwara operators 21
4.1 Statement of the main result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Preliminary steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.3 Reduction steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.4 Statements to be proved . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.5 Spin cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5 Proof of Proposition 4.7 30
5.1 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.2 Proof for Case A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.3 Proof for Case B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.3.1 Case 1: f̃i creates a non-singular string . . . . . . . . . . . . . . . . . . . . . . . . 34
5.3.2 Case 2: f̃i creates a singular string . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.4 Proof for Case C: preliminary steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.4.1 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.4.2 A common property for Case C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.5 Proof for Case C (1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
?This paper is a contribution to the Special Issue in honor of Anatol Kirillov and Tetsuji Miwa. The full
collection is available at http://www.emis.de/journals/SIGMA/InfiniteAnalysis2013.html
mailto:reiho@rs.tus.ac.jp
mailto:reihosan@08.alumni.u-tokyo.ac.jp
https://sites.google.com/site/affinecrystal
http://dx.doi.org/10.3842/SIGMA.2014.028
http://www.emis.de/journals/SIGMA/InfiniteAnalysis2013.html
2 R. Sakamoto
5.6 Proof for Case C (2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.6.1 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.6.2 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.6.3 Proof for the case (5.4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.6.4 Proof for the case (5.5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.7 Proof for Case C (3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.8 Proof for Case D: preliminary steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.8.1 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.8.2 A common property for Case D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.9 Proof for Case D (1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.9.1 The case ` < j . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.9.2 The case ` = j . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.10 Proof for Case D (2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.11 Proof for Case D (3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.11.1 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.11.2 Proof of Proposition 5.30 for the case `(i−1) < ` . . . . . . . . . . . . . . . . . . . . 55
5.11.3 Proof of Proposition 5.30 for the case `(i−1) = ` . . . . . . . . . . . . . . . . . . . . 57
5.12 Proof for Case D (4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.12.1 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.12.2 Proof of Proposition 5.39 for the case `(i−1) < ` . . . . . . . . . . . . . . . . . . . . 60
5.12.3 Proof of Proposition 5.39 for the case `(i−1) = ` . . . . . . . . . . . . . . . . . . . . 61
5.13 Proof for Case E: preliminary steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.13.1 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.13.2 A common property for Case E . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.14 Proof for Case E (1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.15 Proof for Case E (2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.15.1 Case 1: f̃i creates a non-singular string . . . . . . . . . . . . . . . . . . . . . . . . 65
5.15.2 Case 2: f̃i creates a singular string . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.16 Proof for Case E (3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.16.1 Proof for Case 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.16.2 Proof for Case 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.16.3 Proof for Case 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.17 Proof for Case E (4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.17.1 Preliminary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5.17.2 Proof for Case 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.17.3 Proof for Case 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5.18 Proof for Case E (5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6 Proof of Proposition 4.8: f̃i version 69
6.1 Proof for (1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.2 Proof for (2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.3 Proof for (3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
6.4 Proof for (4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7 Proof of Proposition 4.8: ẽi version 77
7.1 Proof for (5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
7.2 Proof for (6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.3 Proof for (7) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.4 Proof for (8) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
8 Comments on Deka–Schilling’s work 86
References 87
Rigged Configurations and Kashiwara Operators 3
1 Introduction
A search of a natural presentation of the basis for representations of infinite dimensional algebras
sometimes reveals an unexpected and intriguing connection with other models of mathematical
physics. Take the Feigin–Fuchs representation of the Virasoro algebra [7] as an example. In
this case, a cerebrated integrable quantum many body problem called the Calogero–Sutherland
model [35, 36] plays the role. Indeed, it is conjectured in [30] and partially proved in [3] that the
excited states of the Calogero–Sutherland model (usually called the Jack symmetric functions)
behave particularly nicely as a basis for the Feigin–Fuchs module. For example, matrix elements
of the Virasoro generators with respect to the Jack basis become factorized rational functions
whose major part consisting of degree one polynomials of positive integer coefficients which
admit a transparent combinatorial interpretation. This phenomenon seems not accidental since
the theory includes Mimachi and Yamada’s finding [22] which claims that the singular vectors
of the Virasoro algebra coincide with special cases of the Jack symmetric functions. See, for
example, [1, 2, 5, 6, 8] for recent developments on the subject.
The purpose of the present paper is to pursue a parallel investigation for the settings of the
crystal bases [14] of an important class of finite dimensional representations (called the Kirillov–
Reshetikhin modules) of the quantum affine algebras. We call the corresponding crystal the
Kirillov–Reshetikhin (KR) crystal. In this case, the corresponding physical model is the box-
ball system. The box-ball system is a prototypical example of the ultradiscrete (or tropical)
soliton system whose simplest example was discovered by Takahashi and Satsuma [37]. For
our purpose, an important aspect of the box-ball system is that the dynamics of the model is
determined by the combinatorial R-matrices for the KR crystal [9, 10]. This formalism allows us
to define the box-ball system for any tensor products of the KR crystals for arbitrary quantum
affine algebras (see, for example, [11, 12]). See [29] for an introductory review on studies on the
box-ball systems.
Recently, a complete set of the action and angle variables of the type A
(1)
n box-ball systems
is found to be identical to the set of the rigged configurations [20]. The rigged configurations
are combinatorial objects discovered by an insightful analysis of the Bethe ansatz for quantum
integrable models [17, 18]. Originally the rigged configurations are used to prove a combinatorial
identity for the Kostka–Foulkes polynomials called the Fermionic formula as well as to give
a description of the branching coefficients of finite dimensional representations of the quantum
affine algebras with respect to the corresponding finite dimensional subalgebras. In such context
the rigged configurations are labels of representations and thus naturally correspond to the
classically highest weight tensor products of the KR crystals. On the other hand, since the box-
ball system is defined for not necessarily highest weight tensor products, we need to consider
generalized rigged configurations corresponding to any tensor products.
Therefore it is natural to introduce the Kashiwara operators on the set of the rigged confi-
gurations. Let I be the set of the Dynkin nodes of a quantum affine algebra g following Kac’s
convention [13] and set I0 = I \{0}. In [32] Schilling gave a definition of the classical Kashiwara
operators ẽi and f̃i (i ∈ I0) for simply laced algebras and proved that the crystal structure on
the rigged configurations is actually isomorphic to the one on the tensor products of the KR
crystals by using Stembridge’s local characterization of crystals. In Section 3.2 of the present
paper, we provide a direct proof of the fact that the classical Kashiwara operators on the set of
the rigged configurations of types A
(1)
n and D
(1)
n indeed satisfy the axiom of the crystals.
The crux of the rigged configuration theory is the so called rigged configuration bijection
which gives an one to one correspondence between the elements of tensor products of the KR
crystals and the rigged configurations:
Φ : rigged configurations 7−→ tensor products.
4 R. Sakamoto
The algorithm of Φ is a rather complicated combinatorial procedure which is explained in Sec-
tion 3.31. Nevertheless it is known that the map Φ has a lot of nice properties and seems to have
a deep mathematical origin. For example, one of the most important properties of the rigged
configuration bijection is that we can regard the algorithm Φ as a general form of the combi-
natorial R-matrices ([19] for type A
(1)
n case and [31, 34] for type D
(1)
n case). See Remark 3.19
for more precise meanings. This property is the basis for providing the action and angle variab-
les of the box-ball systems [20, 21]. Moreover, as an application of the relationship with the
box-ball system we have a representation theoretical interpretation of the algorithm of Φ−1 for
type A
(1)
n [28], though the full understanding of the rigged configurations themselves is yet to
be seen.
The main result of the present paper is the proof of the following compatibility of the rigged
configuration bijection and the classical Kashiwara operators
[ẽi,Φ] = 0, [f̃i,Φ] = 0, i ∈ I0 (1.1)
for g = A
(1)
n and D
(1)
n (see Theorem 4.1) under the assumption that the bijection Φ is well-
defined. Here some remarks are in order.
• In Appendix C of the preprint version of [4], Deka and Schilling asserted the relation (1.1)
for type A
(1)
n . However, according to the author’s opinion, they did not finish their proof
(see Section 8). Given the fundamental importance of the problem, it seems appropriate
to prove the result completely for type A
(1)
n also. Indeed, the proof for type A
(1)
n is a part
of the proof for type D
(1)
n and easy to extract from it2.
• The proof of (1.1) is direct and does not depend on other results.
Since our approach is direct, we expect that our result paves the way to deal with the rigged
configuration bijection for arbitrary quantum affine algebras. Currently the following cases are
established:
• All cases of type A
(1)
n [17, 18, 19].
• Special cases of type D
(1)
n [24, 31, 34] (see the comments after Conjecture 3.17).
• Tensor products of the vector representations of arbitrary non-exceptional quantum affine
algebras [27] as well as of type E
(1)
6 [26].
In all cases the algorithms for the rigged configuration bijection share many common features.
We expect that the rigged configuration bijection as well as all major properties will extend to
the arbitrary quantum affine algebras.
We expect that such an extension of the rigged configuration theory will provide meaningful
insights for both mathematical physics side and mathematical side which are not easily seen
by other methods. One of the supporting evidences for our expectation is a bijection con-
structed in [23]. In that paper, we constructed a combinatorial bijection between the rigged
configurations for arbitrary non-exceptional quantum affine algebras (under certain restrictions
on ranks) and the set of pairs of a type A
(1)
n rigged configuration and a Littlewood–Richardson
tableau. Remarkably, the construction is uniform for all types of algebras and the Littlewood–
Richardson tableaux naturally appear as the recording tableaux of the algorithm. We expect
1For the most general tensor products Br1,s1⊗Br2,s2⊗· · ·⊗BrL,sL of type D
(1)
n , a proof of the well-definedness
of the bijection Φ is the subject of [25]. However we will not assume this fact in the present paper since the
paper [25] is yet not available for the public. After Conjecture 3.17, we summarize the known cases of this result.
2Ignore all arguments related with the parameter `(a) by formally setting `(a) = ∞. See Section 3.3 for the
notation.
Rigged Configurations and Kashiwara Operators 5
that the bijection coincides with (and generalizes) a canonical Dynkin diagram automorphism
(see Remark 3.2 of [23]).
Another supporting evidence is a numerical comparison of the rigged configuration bijec-
tion [27] for the vector representations of arbitrary non-exceptional quantum affine algebras
and the corresponding combinatorial R-matrices (see Section 3 and subsequent sections of [20]).
The conjecture provided there claims that seemingly complicated algorithms presented in [27]
with many case by case definition possess uniform description in terms of the corresponding com-
binatorial R-matrices. In fact, we expect that the rigged configuration bijection gives a general
form of the combinatorial R-matrices for general quantum affine algebras and thus provides ac-
tion and angle variables for the corresponding box-ball systems. Finally, it should be mentioned
that in [24] we see that the affine crystal structure of type D
(1)
n is essentially governed by the
rigged configurations. Therefore we expect that the construction of the rigged configuration
bijection for general cases should be an intriguing and meaningful future research topic.
This paper is organized as follows. In Section 2, we collect necessary facts from the crys-
tal bases and the tableau representations. In Section 3, we give the definition of the rigged
configurations and the rigged configuration bijection and explain their basic properties (con-
vexity relation of the vacancy numbers). We also introduce the classical Kashiwara operators
on the rigged configurations and show that they satisfy the axiom of the classical crystals. In
Section 4 we give the statement of the main result and clarify what are the essential points to
be shown. Main proofs are given in Sections 5, 6 and 7. During the proof we have tried to
avoid unnecessarily long arguments. Nevertheless we have to deal with many subtle cases and
some of them look very special. Thus some readers might wonder whether such case does exist
and is worthy of careful analysis. In order to clarify that point, we include several examples for
seemingly subtle cases.
2 Background on crystals and tableaux
2.1 Crystal bases
Let us briefly recall some definitions about the crystal bases theory of Kashiwara [14]. For
more detailed introduction, see, for example, [15]. Let g be an affine Kac–Moody algebra, g′ be
its derived subalgebra, g0 be the corresponding finite dimensional simple Lie algebra obtained
by removing the 0 node of the Dynkin diagram of g. Here we follow Kac’ convention [13] for
the labeling of the Dynkin nodes and denote by I the vertex set of the Dynkin diagram of g
and I0 := I \ {0}. Let Uq(g), U ′q(g) and Uq(g0) be the quantized universal enveloping algebra
corresponding to g, g′ and g0, respectively. In this paper we will mainly focus on the case
g = D
(1)
n which has the following Dynkin diagram
1
0
c
c
c c2 3 p p p p p p p p p p p p p p p cn− 2
c
c
n− 1
n
Let αi, hi, Λi (i ∈ I) be the simple roots, simple coroots and fundamental weights of g, Λ̄i
(i ∈ I0) be the fundamental weights of g0. In the present case g0 = Dn, the simple roots are
αi = εi − εi+1, 1 ≤ i ≤ n− 1, αn = εn−1 + εn, (2.1)
and the fundamental weights are
Λ̄i = ε1 + · · ·+ εi, 1 ≤ i ≤ n− 2, (2.2)
Λ̄n−1 = (ε1 + · · ·+ εn−1 − εn)/2, Λ̄n = (ε1 + · · ·+ εn−1 + εn)/2,
6 R. Sakamoto
where εi ∈ Zn is the ith standard unit vector. Let Q, Q∨, P be the root, coroot and weight
lattices of g. Let 〈·, ·〉 : Q∨ ⊗ P → Z be the pairing such that 〈hi,Λj〉 = Λj(hi) = δij . Note that
we have 〈hi, αj〉 = αj(hi) = Aij where Aij is the Cartan matrix of g = D
(1)
n .
Now we give the axiomatic definition of U ′q(g)-crystals.
Definition 2.1. U ′q(g)-crystal is a nonempty set B equipped with maps wt : B → P , εi, ϕi :
B → Z ∪ {−∞} and the Kashiwara operators ẽi, f̃i : B → B ∪ {0} for all i ∈ I such that
(1) 〈hi,wt(b)〉 = ϕi(b)− εi(b),
(2) wt
(
ẽi(b)
)
= wt(b) + αi if ẽi(b) ∈ B,
(3) wt
(
f̃i(b)
)
= wt(b)− αi if f̃i(b) ∈ B,
(4) εi
(
ẽi(b)
)
= εi(b)− 1, ϕi
(
ẽi(b)
)
= ϕi(b) + 1 if ẽi(b) ∈ B,
(5) εi
(
f̃i(b)
)
= εi(b) + 1, ϕi
(
f̃i(b)
)
= ϕi(b)− 1 if f̃i(b) ∈ B,
(6) for b, b′ ∈ B, f̃i(b) = b′ if and only if ẽi(b
′) = b,
(7) if ϕi(b) = −∞ for b ∈ B, then ẽi(b) = f̃i(b) = 0.
In this paper, we only consider the case when the maps εi, ϕi are defined by
εi(b) = max
{
m ≥ 0
∣∣ ẽmi (b) 6= 0
}
, ϕi(b) = max
{
m ≥ 0
∣∣ f̃mi (b) 6= 0
}
for b ∈ B. In particular, we have 0 ≤ εi(b), ϕi(b) < ∞ for all i ∈ I and b ∈ B. Note that the
crystals with these conditions are called the regular crystals. If we have f̃i(b) = b′ for b, b′ ∈ B,
we write an arrow with color (or label) i from b to b′. In this way, the crystal B can be regarded
as the colored oriented graph whose vertices are the elements of B. We call such graph the
crystal graph.
One of the nice properties of the crystal bases is that it behaves nicely with respect to the
tensor product. Let B2 ⊗ B1 be the tensor product of two crystals B1 and B2. As the set, it
coincides with the Cartesian product B2 ×B1. The action of the Kashiwara operators ẽi, f̃i on
an element b2 ⊗ b1 ∈ B2 ⊗B1 is given explicitly as follows:
ẽi(b2 ⊗ b1) =
{
ẽi(b2)⊗ b1 if εi(b2) > ϕi(b1),
b2 ⊗ ẽi(b1) if εi(b2) ≤ ϕi(b1),
f̃i(b2 ⊗ b1) =
{
f̃i(b2)⊗ b1 if εi(b2) ≥ ϕi(b1),
b2 ⊗ f̃i(b1) if εi(b2) < ϕi(b1),
where the result is declared to be 0 if either of its tensor factors are 0. We also have wt(b2⊗b1) =
wt(b2) + wt(b1).
Remark 2.2. We use the opposite of the Kashiwara’s tensor product convention.
In order to deal with multiple tensor products, it is convenient to use the signature rule. Let B
be the tensor product of crystals B = BL⊗· · ·⊗B2⊗B1. For i ∈ I and b = bL⊗· · ·⊗b2⊗b1 ∈ B,
the i-signature of b is the following sequence of the symbols + and −
+ · · · · · · · · ·+︸ ︷︷ ︸
ϕi(bL) times
− · · · · · · · · · −︸ ︷︷ ︸
εi(bL)times
· · · · · ·+ · · · · · · · · ·+︸ ︷︷ ︸
ϕi(b1) times
− · · · · · · · · · −︸ ︷︷ ︸
εi(b1) times
.
The reduced i-signature of b is obtained by removing subsequence −+ of the i-signature of b
repeatedly until it becomes the following form
+ · · · · · · · · ·+︸ ︷︷ ︸
ϕi(b) times
− · · · · · · · · · −︸ ︷︷ ︸
εi(b)times
.
Rigged Configurations and Kashiwara Operators 7
By using this information the action of ẽi and f̃i on b can be described as follows. If εi(b) = 0,
then ẽi(b) = 0. Otherwise
ẽi(bL ⊗ · · · ⊗ b1) = bL ⊗ · · · ⊗ bj+1 ⊗ ẽi(bj)⊗ bj−1 ⊗ · · · ⊗ b1,
where the leftmost − in the reduced i-signature of b comes from bj . Similarly, if ϕi(b) = 0, then
f̃i(b) = 0. Otherwise
f̃i(bL ⊗ · · · ⊗ b1) = bL ⊗ · · · ⊗ bj+1 ⊗ f̃i(bj)⊗ bj−1 ⊗ · · · ⊗ b1,
where the rightmost + in the reduced i-signature of b comes from bj . Finally the actions of ẽi
and f̃i on each tensor factor bj are described by tableau representations of bj described in the
sequel.
2.2 Kirillov–Reshetikhin tableaux
In order to perform explicit manipulations on the crystal bases, it is convenient to use an explicit
realization of the elements of crystals. For the present purpose, it is more convenient to use a new
kind of tableau representation which we call the Kirillov–Reshetikhin (KR) tableaux [24] than
the usual Kashiwara–Nakashima (KN) tableaux [16].
2.2.1 Reviews on tableau representations
Let us briefly recall tableau representations of crystals. Let B(Λ) be the crystal associated
to the highest weight representation of highest weight Λ of Uq(g0). Then in the present case
g = D
(1)
n , the U ′q(g) crystal Br,s has the following decomposition under the restriction to its
finite dimensional subalgebra Uq(g0). If r < n− 1, we have
Br,s '
⊕
λ
B(λ) as g0 crystals, (2.3)
and if r = n− 1, n, we have
Br,s ' B(sΛ̄r) as g0 crystals. (2.4)
Here in the first case r < n − 1, λ in the summation are determined as follows. For this we
identify λ =
∑
i
kiΛ̄i with the Young diagram which has ki columns of height i for all i. Then one
of λ in decomposition (2.3) is shape (sr) rectangular diagram and the remaining λ are obtained
by removing the vertical dominoes from (sr) rectangular diagram in all possible ways. Then
if b ∈ Br,s belongs to some B(λ) in the right hand side of the decomposition (2.3) or (2.4), then
the KN tableau representation of b has shape λ.
KN tableaux for Br,s when r ≤ n− 2. For J ⊂ I, we say b is J-highest element if we
have ẽi(b) = 0 for all i ∈ J . Suppose that uλ ∈ Br,s (r ≤ n−2) is the I0-highest element of B(λ)
which appears in the decomposition (2.3). Then the corresponding KN tableau is obtained by
filling the letter j to all the boxes at jth row of the Young diagram λ. In order to obtain the
other tableaux, we define the action of the Kashiwara operators on tableaux as follows. The
crystal graph of B1,1 for g0 = Dn is as follows
1 -1
2 -2 p p p p p p p p -n− 2
n− 1
�
���n− 1
@
@@Rn
n
�
���
@
@@R
n
n− 1
n− 1
n̄
-n− 2 p p p p p p p p -2
2̄ -1
1̄
8 R. Sakamoto
Weight of each vertex is wt
(
i
)
= εi and wt
(
ī
)
= −εi, respectively. For a given tableau
t ∈ Br,s (r ≤ n− 2),
t =
t1
t2pppppppp
tr
tr+1
tr+2pppppppp
t2rpppp
tr′
p p p p p p p p p p p p p p p p p p p p p p p
ppppppppp
p
tN p p p p p p p p p p p p p p p p p p p p p p p p p
p p p
we introduce the so-called Japanese reading word and regard t as the element of (B1,1)⊗N as
follows
t 7−→ tN ⊗ · · · ⊗ t2 ⊗ t1 ∈
(
B1,1
)⊗N
.
Via this identification we can introduce the Kashiwara operators on tableaux and obtain the
KN tableau representation for the general case.
KN tableaux for Bn,2s+1. Let us consider the basic case Bn,1 ' B(Λ̄n). This is the
so-called spin representation of dimension 2n−1 whose basis are (s1, s2, . . . , sn) where si = ±
and s1s2 · · · sn = 1. The classical Kashiwara operators act on each basis as follows:
ẽi(s1, . . . , sn) =
(s1, . . . ,+,−, . . . , sn), if i 6= n, (si, si+1) = (−,+),
(s1, . . . , sn−2,+,+), if i = n, (sn−1, sn) = (−,−),
0, otherwise,
(2.5)
f̃i(s1, . . . , sn) =
(s1, . . . ,−,+, . . . , sn), if i 6= n, (si, si+1) = (+,−),
(s1, . . . , sn−2,−,−), if i = n, (sn−1, sn) = (+,+),
0, otherwise.
(2.6)
The weight of each basis is
wt
(
(s1, . . . , sn)
)
=
1
2
(s1ε1 + · · ·+ snεn). (2.7)
Thus we can represent each element εi by half width tableau with entry i. We arrange the entries
within a column according to the following order on the letters (smaller ones at higher places)
1 ≺ 2 ≺ · · · ≺ n− 1 ≺ n
n
≺ n− 1 ≺ · · · ≺ 2 ≺ 1.
Here we do not introduce the order between n and n.
For the general case, we start by the highest weight element uλ ∈ Bn,2s+1 where λ = (2s +
1)Λ̄n. Then represent uλ by the spin column (+, . . . ,+) and s columns of 1, 2, . . . , n such as
u(2s+1)Λ̄n
=
1
2ppp
n
1
2ppp
n
1
2ppp
n
1
2ppp
np p p
p p p
p p pp p p
Rigged Configurations and Kashiwara Operators 9
In order to obtain the other tableaux, we embed uλ into Br,1⊗ (B1,1)⊗rs by using the Japanese
reading word for the right s columns and apply the Kashiwara operators f̃i in all possible ways.
We can obtain the KN tableaux for the case Bn,2s similarly if we ignore the leftmost spin
column in Bn,2s+1 case.
KN tableaux for Bn−1,2s+1. In the case Bn−1,1 ' B(Λ̄n−1), we can represent the basis
as (s1, s2, . . . , sn) where si = ± and s1s2 · · · sn = −1. We can use the same formulae (2.5)
and (2.6) for the Kashiwara operators and (2.7) for the weight. For the general case Bn−1,2s+1,
we represent the highest weight element u(2s+1)Λ̄n−1
∈ Bn−1,2s+1 by
u(2s+1)Λ̄n−1
=
1
2ppp̄
n
1
2ppp̄
n
1
2ppp̄
n
1
2ppp̄
np p p
p p p
p p pp p p
Here we obtain the tableau for u(2s+1)Λ̄n−1
by replacing all n of the bottom row of u(2s+1)Λ̄n
by n̄. Then we can use the Kashiwara operators to obtain all the tableaux of Bn−1,2s+1. We
can obtain tableaux for the case Bn−1,2s if we ignore the leftmost spin column.
2.2.2 Filling map
The purpose to introduce the KR tableaux is to give a rectangular presentation of elements of
crystals as opposed to the KN tableaux whose shapes depend on the classical decompositions
given in (2.3) and (2.4). The definition of the KR tableaux is given explicitly for the highest
weight element of Br,s and extended to the general case by using the Kashiwara operators f̃i as
in the KN tableaux case.
To begin with we give a definition of the KR tableaux for the case Br,s (r ≤ n− 2). Let uλ
be the I0-highest element of Br,s with weight λ = krΛ̄r + kr−2Λ̄r−2 + · · · . We denote by fill(uλ)
the KR tableau representation of uλ. Now we describe the algorithm to obtain fill(uλ) which
we call the filling map. Let kc be the first odd integer in the sequence kr−2, kr−4, . . . . If there
is no such kc, set kc = k−1, that is, c = −1. Let t be the KN tableau representation of uλ. We
place t at the top left corner of the r by s rectangle. In order to fill all the empty places, we
start from the leftmost column of t and move rightwards by the following procedure.
1. We do nothing for the height r columns. If c ≥ 0, remove one column from the height c
columns and move the columns that are shorter than c to left by 1.
2. For the columns whose heights are equal to or larger than c, put the following stuff to the
empty places as much as possible. If the height of the corresponding columns is h, put the
transpose of
r r − 1 . . . h+ 1
h+ 1 h+ 2 . . . r
If c = −1, we stop here since all empty places are filled by this procedure.
3. For the remaining columns except for the rightmost one, put the following stuff recursively
from left to right with recursively redefining an integer x. As the initial condition, set
x = c+ 1.
If the corresponding column has height h, put the transpose of
y . . . r − 1 r r . . . x+ 1 x
to the empty place. Since the number of the empty places is r−h, we have y = r−(x−h−2).
We redefine x = y and do the same procedure to the next column.
10 R. Sakamoto
4. By using the final output of x from the last step, we put the transpose of the following to
the rightmost column
1 2 . . . y y . . . x+ 1 x
Recall that the rightmost column is empty by Step 1 whenever c ≥ 0. Then in order to
achieve the above pattern, we have y = (r + x− 1)/2.
Example 2.3. Let us consider the I0-highest element uλ of B8,7 where λ = Λ̄8 + 2Λ̄6 + Λ̄4 +
2Λ̄2. In this case, we have (k6, k4, k2, k0) = (2, 1, 2, 1), thus c = 4. We put the KN tableau
representation of uλ into the 8 by 7 rectangle. Then
1 1 1 1 1 1
2 2 2 2 2 2
3 3 3 3
4 4 4 4
5 5 5
6 6 6
7
8
Step 1−−−−−→
1 1 1 1 1
2 2 2 2 2
3 3 3
4 4 4
5 5 5
6 6 6
7
8
Then we fill the empty places as
5 1
6 2
7 5 7 3
8 6 8 4
8̄ 7 8̄ 5
7̄ 8 7̄ 6
8̄ 7 6̄ 8̄ 6̄ 6̄
7̄ 8 5̄ 7̄ 5̄ 5̄
thus, fill(uλ) =
1 1 1 1 1 5 1
2 2 2 2 2 6 2
3 3 3 7 5 7 3
4 4 4 8 6 8 4
5 5 5 8̄ 7 8̄ 5
6 6 6 7̄ 8 7̄ 6
7 8̄ 7 6̄ 8̄ 6̄ 6̄
8 7̄ 8 5̄ 7̄ 5̄ 5̄
In order to define the KR tableaux for the spin cases, we start from the following expression
for the highest weight elements (collection of s spin columns)
usΛ̄n−1
=
1
2ppp̄
n
1
2ppp̄
n
1
2ppp̄
np p p
p p p
p p pp p p
, usΛ̄n
=
1
2ppp
n
1
2ppp
n
1
2ppp
np p p
p p p
p p pp p p
and apply the Kashiwara operators f̃i (i ∈ I0) for all possible ways to obtain the rest of the
elements of Bn−1,s and Bn,s.
From Remark 5.2 and Remark 5.4 in [24], it follows that the new Kirillov–Reshetikhin (KR)
tableaux carry a crystal structure obtained by their reading word which gives a natural embed-
ding of B(λ) into B(Λ1)⊗rs. Since each classical factor in Br,s occurs with multiplicity 1, we can
use the Kirillov–Reshetikhin tableaux representation in place of the usual Kashiwara–Nakashima
tableaux representation.
For the later purposes, we introduce several operations on the KR tableaux representation of
crystals.
Rigged Configurations and Kashiwara Operators 11
Definition 2.4. We use the KR tableaux representation for the element b = bL⊗bL−1⊗· · ·⊗b1 ∈
BrL,sL ⊗BrL−1,sL−1 ⊗ · · · ⊗Br1,s1 . Then we define the following operations.
(1) Suppose that BrL,sL = B1,1. Then we define the operation called left-hat by
lh(b) = bL−1 ⊗ · · · ⊗ b1 ∈ BrL−1,sL−1 ⊗ · · · ⊗Br1,s1 .
(2) Suppose that BrL,sL = BrL,1 and rL > 1. Thus bL has the form bL =
t1
...
trL−1
trL
. Then we
define the operation called left-box by
lb(b) = trL ⊗
t1
...
trL−1
⊗ bL−1 ⊗ · · · ⊗ b1 ∈ B1,1 ⊗BrL−1,1 ⊗BrL−1,sL−1 ⊗ · · · ⊗Br1,s1 .
(3) Suppose that sL > 1. Let ci be the ith column of bL such that bL = c1c2 · · · csL . Then we
define the operation called left-split by
ls(b) = c1 ⊗ c2 · · · csL ⊗ bL−1 ⊗ · · · ⊗ b1 ∈ BrL,1 ⊗BrL,sL−1 ⊗BrL−1,sL−1 ⊗ · · · ⊗Br1,s1 .
3 Rigged configurations and the main bijection
3.1 Definition of the rigged configurations
Our strategy to define the rigged configurations of type g = D
(1)
n has three steps. First we define
the highest weight rigged configurations. Next we define the classical Kashiwara operators ẽi
and f̃i for i ∈ I0 on the rigged configurations. Finally, we define the general rigged configurations
by all possible applications of the operators f̃i on highest weight rigged configurations.
Let us prepare several notation to describe the rigged configurations. The rigged configura-
tions are combinatorial objects made up with partitions ν(a) and integers J (a) for a ∈ I0. ν(a) =(
ν
(a)
1 , . . . , ν
(a)
k
)
is a partition called the configuration and J (a) =
(
J
(a)
1 , . . . , J
(a)
k
)
is a sequence
of integers called the rigging which are associated with rows of the corresponding configuration.
We should regard that the pairs
(
ν(a), J (a)
)
are located on the vertices I0 of the Dynkin diagram
of D
(1)
n . We denote by m
(a)
l (ν) the number of length l rows of ν(a). We will abbreviate the pair(
ν(a), J (a)
)
by (ν, J)(a) and call the components of (ν, J)(a) =
{(
ν
(a)
1 , J
(a)
1
)
, . . . ,
(
ν
(a)
k , J
(a)
k
)}
the
strings.
The rigged configurations also depend on the shape of the tensor product B = BrL,sL ⊗
· · · ⊗ Br1,s1 . Let L
(a)
l be the number of the component Ba,l within B. Then we define the
partition µ(a) such that the number of length l rows of µ(a) is equal to L
(a)
l . Thus the rigged
configuration (ν, J) is described by the sequence of configurations and the riggings
(ν, J) =
(
(ν, J)(1), . . . , (ν, J)(n−1), (ν, J)(n)
)
together with additional data µ =
(
µ(1), . . . , µ(n−1), µ(n)
)
under certain constraints to be de-
scribed below. We denote by Q
(a)
l (ν) (resp. Q
(a)
l (µ)) the number of boxes in the first l columns
of ν(a) (resp. µ(a));
Q
(a)
l (ν) =
∑
j>0
min{l, j}m(a)
j (ν), Q
(a)
l (µ) =
∑
j>0
min{l, j}L(a)
j .
12 R. Sakamoto
Therefore we have Q
(a)
∞ (ν) = |ν(a)| and Q
(a)
∞ (µ) = |µ(a)| where |ν(a)| and |µ(a)| are the total
number of boxes in ν(a) and µ(a), respectively. From the data µ and ν we define the vacancy
number P
(a)
l (µ, ν) (hereafter we will abbreviate as P
(a)
l (ν)) by the formula
P
(a)
l (ν) = Q
(a)
l (µ) +
∑
b∈I0
AabQ
(b)
l (ν)
= Q
(a)
l (µ)− 2Q
(a)
l (ν) +
∑
b∈I0
b∼a
Q
(b)
l (ν), (3.1)
where Aab is the Cartan matrix of g0 = Dn and a ∼ b means that the vertices a and b are
connected by a single edge on the Dynkin diagram.
Definition 3.1. The data µ and (ν, J) is the highest weight rigged configuration if all the
strings
(
ν
(a)
i , J
(a)
i
)
and the corresponding vacancy numbers satisfy the following condition
0 ≤ J (a)
i ≤ P (a)
ν
(a)
i
(ν).
Remark 3.2. If ν(n−1) = ν(n) = ∅, then we can identify the rigged configuration as type A
(1)
n−2.
In this way we can recover proofs for type A
(1)
n from those for type D
(1)
n .
Example 3.3. The following object is a highest weight rigged configuration corresponding to
the tensor product B3,2 ⊗B3,1 ⊗B2,2 ⊗B1,2 ⊗B1,1 of type D
(1)
5 :
0
0
0
0
1
2
1
1
1
0
0
0
1
0
0
0
0
0
0
0
0
0 2 1
Here we put the vacancy number (resp. rigging) on the left (resp. right) of the corresponding row
of the configuration represented by a Young diagram. By the rigged configuration bijection Φ
(see Section 3.3) the above rigged configuration (ν, J) corresponds to the following highest weight
tensor product:
Φ(ν, J) =
1 3
2 5̄
4 5
⊗
1
4
5
⊗ 2 1
3 1̄
⊗ 1 2 ⊗ 1
The weight λ of the rigged configuration is defined by the relation (sometimes called the
(L, λ)-configuration condition)∑
a∈I0,i>0
im
(a)
i αa =
∑
a∈I0,i>0
iL
(a)
i Λ̄a − λ.
If we expand λ by the basis εi, we can rewrite as
λ =
∑
i
λiεi =
∑
a∈I0
(∣∣µ(a)
∣∣Λ̄a − ∣∣ν(a)
∣∣αa). (3.2)
Then we can use the expressions (2.1) and (2.2) to obtain the explicit expressions for the
weight λi. We write the weight of the rigged configuration by wt(ν, J).
Following [32] we introduce the classical Kashiwara operators on the rigged configurations
and use them to define the general rigged configurations. For the string (l, x) of (ν, J)(i), we
call the quantity P
(i)
l (ν)− x the corigging.
Rigged Configurations and Kashiwara Operators 13
Definition 3.4. The rigged configurations are obtained by all possible applications of the
Kashiwara operators f̃i (i ∈ I0) on the highest weight rigged configurations. Here the Kashiwara
operators on the rigged configurations are defined as follows. Let x` be the smallest rigging
of (ν, J)(i).
(1) Let ` be the minimal length of the strings of (ν, J)(i) with the rigging x`. If x` ≥ 0, define
ẽi(ν, J) = 0. Otherwise ẽi(ν, J) is obtained by replacing the string (`, x`) by (`− 1, x` + 1)
while changing all other riggings to keep coriggings fixed.
(2) Let ` be the maximal length of the strings of (ν, J)(i) with the rigging x`. f̃i(ν, J) is
obtained by the following procedure. If x` > 0, add a string (1,−1) to (ν, J)(i). Otherwise
replace the string (`, x`) by (`+ 1, x`− 1). Change other riggings to keep coriggings fixed.
If the new rigging is strictly larger than the corresponding new vacancy number, define
f̃i(ν, J) = 0.
By definition every rigging is always less than or equal to the corresponding vacancy number.
Proposition 3.5. The above defined Kashiwara operators satisfy the following relations
wt
(
ẽi(ν, J)
)
= wt(ν, J) + αi, wt
(
f̃i(ν, J)
)
= wt(ν, J)− αi.
Moreover f̃i(ν, J) 6= 0 =⇒ ẽif̃i(ν, J) = (ν, J) and ẽi(ν, J) 6= 0 =⇒ f̃iẽi(ν, J) = (ν, J).
Proof. The relations for wt follow from (3.2). In order to show ẽif̃i(ν, J) = (ν, J), suppose
that f̃i creates the string (`+ 1, x` − 1). Let (ν̃, J̃) = f̃i(ν, J). Suppose that 0 < `.
(a) Let (k, xk) be a string of (ν, J)(i) which satisfies k ≤ `. Since f̃i acts on the string with
smallest rigging, we have xk ≥ x`. Recall that we have P
(i)
k (ν̃) = P
(i)
k (ν) since f̃i adds
a box to the (` + 1)-th column. Thus the string (k, xk) remains as it is after f̃i and its
rigging satisfies xk > x` − 1.
(b) Let (k, xk) be a string of (ν, J)(i) which satisfies ` < k. Since f̃i acts on the longest string
with rigging x`, we have x` < xk. Recall that we have P
(i)
k (ν̃) = P
(i)
k (ν) − 2 in this case.
Thus the string (k, xk) becomes (k, xk−2) after f̃i. Then its rigging satisfies xk−2 ≥ x`−1.
Thus the string (` + 1, x` − 1) is the shortest string among the strings of the smallest rigging
of (ν̃, J̃)(i), hence ẽi will act on it. Therefore we obtain ẽif̃i(ν, J) = (ν, J). Let us consider the
case ` = 0. Then we can use the same arguments as before if we put (`, x`) = (0, 0).
The opposite relation f̃iẽi(ν, J) = (ν, J) can be shown similarly. �
Finally, let us define the maps εi, ϕi : (ν, J)→ Z by
εi(ν, J) = max
{
m ≥ 0
∣∣ ẽmi (ν, J) 6= 0
}
, ϕi(ν, J) = max
{
m ≥ 0
∣∣ f̃mi (ν, J) 6= 0
}
.
3.2 Convexity relations of the vacancy numbers and their applications
In this subsection, we introduce a fundamental property of the vacancy numbers called the
convexity relation. As an application, we provide a direct proof of the fact that the crystal
structure on the set of the rigged configurations indeed satisfies the axiom of the Uq(g0)-crystals.
Finally, we prepare refined estimates for the vacancy numbers which are necessary in the later
arguments.
One of the basic properties of the vacancy number is the following convexity relation.
Proposition 3.6 (convexity). Suppose that m
(a)
l (ν) = 0 for some l ≥ 1. Then the vacancy
numbers satisfy the following convexity relation
2P
(a)
l (ν) ≥ P (a)
l−1(ν) + P
(a)
l+1(ν).
14 R. Sakamoto
Proof. Recall that the functions Q
(a)
k (µ) and Q
(b)
k (ν) for b 6= a are all upper convex functions
of k. On the other hand, since m
(a)
l (ν) = 0, the function Q
(a)
k (ν) is a linear function between
l− 1 ≤ k ≤ l+ 1. Thus the combination in (3.1) gives the upper convex relation for the vacancy
numbers. �
By repeated use of the above convexity relation, we obtain several useful criteria.
Corollary 3.7. Suppose that m
(a)
k (ν) = 0 for all l1 < k < l2.
(1) P
(a)
k (ν) ≥ min
{
P
(a)
l1
(ν), P
(a)
l2
(ν)
}
.
(2) For some k satisfying l1 < k < l2 suppose that we have P
(a)
l1
(ν) ≥ P
(a)
k (ν) < P
(a)
l2
(ν) or
P
(a)
l1
(ν) > P
(a)
k (ν) ≤ P (a)
l2
(ν). Then the corresponding rigged configuration is forbidden.
(3) For some k satisfying l1 < k < l2 suppose that we have P
(a)
l1
(ν) ≥ P (a)
k (ν) ≤ P (a)
l2
(ν). Then
the only admissible situation is P
(a)
l1
(ν) = P
(a)
l1+1(ν) = · · · = P
(a)
l2
(ν).
Since we will use these properties so many times in the rest of the paper, we will sometimes
use these relations without giving an explicit reference.
Let us give the first application of the convexity relations.
Theorem 3.8. Let x` be the smallest rigging of (ν, J)(i) and define s = min{0, x`}. If ν(i) = ∅,
set s = 0. Then we have
(1) εi(ν, J) = −s,
(2) ϕi(ν, J) = P
(i)
∞ (ν)− s.
Proof. (1) We proceed by induction on −s. Let (ν̃, J̃) := ẽi(ν, J), x̃˜̀ be the smallest rigging of
(ν̃, J̃)(i) and define s̃ = min{0, x̃˜̀}.
Suppose that −s = 0. This implies that x` ≥ 0. Then by the definition of ẽi we have
ẽi(ν, J) = 0 as requested. Suppose that −s > 0. Then we have s = x` < 0. Suppose that ẽi acts
on the string (`, x`). Let us analyze the behaviors of the riggings by ẽi.
(a) The string (`, x`) becomes (`− 1, x` + 1).
(b) Let (k, xk) be an arbitrary string of (ν, J)(i) satisfying k < `. Since ẽi acts on the shortest
string with rigging x`, we have xk > x`. From k < `, we see that ẽi will not change
corigging of (k, xk), thus the string (k, xk) remains as it is in (ν̃, J̃).
(c) Let (k, xk) be an arbitrary string of (ν, J)(i) satisfying ` ≤ k. Since x` is the minimal
rigging of (ν, J)(i), we have x` ≤ xk. Recall that we have P
(i)
k (ν̃) = P
(i)
k (ν) + 2 since ẽi
removes a box from `th column of ν(i). Thus ẽi makes the string (k, xk) into (k, xk + 2),
in particular, its rigging satisfies x` + 1 < xk + 2.
To summarize, we have x̃˜̀ = x` + 1 and thus s̃ = s+ 1, that is, −s̃ = εi(ν, J)− 1 = εi(ν̃, J̃) as
requested.
(2) This is proved in [32, Lemma 3.6]. However, for the sake of the completeness, we include
a proof here. We prove by induction on P
(i)
∞ (ν) − s. Suppose that f̃i acting on (ν, J) creates
a length `+ 1 string. By definition of s, f̃i creates the string (`+ 1, s− 1) of (ν̃, J̃) := f̃i(ν, J).
Let x̃˜̀ be the smallest rigging of (ν̃, J̃)(i) and define s̃ = min{0, x̃˜̀}.
Step 1. Let us consider the case P
(i)
∞ (ν)− s = 0. Suppose that x` ≤ 0. Then we have ` > 0
and s = x`, thus, P
(i)
∞ (ν) = x`. Let j be the maximal integer such that ` ≤ j and m
(i)
j (ν) > 0.
Suppose if possible that ` < j. Let us consider the string (j, xj). Since f̃i acts on the longest
Rigged Configurations and Kashiwara Operators 15
string with rigging x`, the assumption ` < j implies that x` < xj . Note that we have xj ≤ P (i)
j (ν)
since (ν, J) is the rigged configuration. Thus x` < P
(i)
j (ν). Recall that the definition (3.1)
implies that there exists a large integer L such that P
(i)
L (ν) = P
(i)
L+1(ν) = · · · = P
(i)
∞ (ν) = x`.
Since j is the maximal integer such that m
(i)
j (ν) > 0, we see that P
(i)
k (ν) satisfy the convexity
relation between j ≤ k ≤ ∞. In particular, the relation P
(i)
j (ν) > P
(i)
L (ν) = P
(i)
L+1(ν) for
j < L < L+ 1 is a contradiction. Therefore we have ` = j and thus there is no string of (ν, J)(i)
that is longer than `. Then from the convexity relation of P
(i)
k (ν) between ` ≤ k ≤ ∞, the
only possibility that is compatible with the requirement P
(i)
` (ν) ≥ x` is the relation P
(i)
` (ν) =
P
(i)
`+1(ν) = · · · = x`.
Let us consider the string (`+1, x`−1) of (ν̃, J̃)(i) created by f̃i. Due to the extra box added
by f̃i, we have P
(i)
`+1(ν̃) = P
(i)
`+1(ν)− 2 = x`− 2. Thus the string (`+ 1, x`− 1) of (ν̃, J̃)(i) has the
rigging that is strictly larger than the corresponding vacancy number. Thus we have f̃i(ν, J) = 0
as requested.
Finally let us consider the case x` > 0. Then we have P
(i)
∞ (ν) = 0 by s = 0. Let j be the
maximal integer such that m
(i)
j (ν) > 0 and consider the corresponding string (j, xj). Since x` is
the minimal rigging, we have P
(i)
j (ν) ≥ xj ≥ x` > 0. Then for a sufficiently large integer L, we
have P
(i)
j (ν) > P
(i)
L (ν) = P
(i)
L+1(ν) = · · · = 0. This is a contradiction since P
(i)
k (ν) must satisfy
the convexity relation between j ≤ k ≤ ∞. Therefore we see that there is no string in (ν, J)(i).
This is a contradiction since we assume that x` > 0. Hence this case cannot happen.
Step 2. Let us consider the case P
(i)
∞ (ν) − s > 0. Suppose that x` ≤ 0. Then we have
P
(i)
∞ (ν) > s = x`. Let j be the minimal integer such that ` < j and m
(i)
j (ν) > 0. If there is no
such j, set j = ∞. Then we have x` < P
(i)
j (ν). For, if j < ∞ we have x` < xj ≤ P
(i)
j (ν) as in
the previous step and if j =∞ we have x` < P
(i)
j (ν) by the assumption.
Let us show that
x` < P
(i)
`+1(ν). (3.3)
If j = `+ 1 this relation is already confirmed. Thus suppose that `+ 1 < j. Suppose if possible
that we have x` ≥ P
(i)
`+1(ν). Then we have P
(i)
` (ν) ≥ P
(i)
`+1(ν) < P
(i)
j (ν) by x` ≤ P
(i)
` (ν) and
x` < P
(i)
j (ν). This is a contradiction since P
(i)
k (ν) must satisfy the convexity relation between
` ≤ k ≤ j. In conclusion, we have x` < P
(i)
`+1(ν).
Since f̃i adds a box to the (`+ 1)-th column of ν(i), we have P
(i)
`+1(ν̃) = P
(i)
`+1(ν)− 2 ≥ x` − 1
by (3.3). Therefore the string (` + 1, x` − 1) of (ν̃, J̃)(i) created by f̃i has the rigging which is
smaller than or equal to the corresponding vacancy number. Hence we have f̃i(ν, J) 6= 0.
Let us determine s̃.
(a) f̃i makes the string (`+ 1, x` − 1). Below we consider the remaining strings of (ν, J)(i).
(b) Let (k, xk) be an arbitrary string of (ν, J)(i) satisfying ` < k. Since f̃i acts on the longest
string with rigging x`, ` < k implies that x` < xk. Recall that we have P
(i)
k (ν̃) = P
(i)
k (ν)−2
since f̃i adds a box to the (`+1)th column of ν(i). In order to keep the corigging, the string
(k, xk) becomes (k, xk−2) after f̃i. In particular, the new rigging satisfies x`−1 ≤ xk−2.
(c) Let (k, xk) be an arbitrary string of (ν, J)(i) satisfying k ≤ `. Then we have xk ≥ x`.
Since f̃i does not change the corigging of the string, the string (k, xk) remains as it is
after f̃i.
16 R. Sakamoto
To summarize, we have s̃ = x` − 1 = s − 1. Since we have P
(i)
∞ (ν̃) = P
(i)
∞ (ν) − 2, we have
P
(i)
∞ (ν̃)− s̃ = P
(i)
∞ (ν)− s− 1 = ϕi(ν, J)− 1 = ϕi(ν̃, J̃) as requested.
Finally consider the case x` > 0. In this case we have s = 0. Then f̃i adds the string
(1,−1) to (ν, J)(i). If we show that f̃i(ν, J) 6= 0, then we have s̃ = −1 = s − 1 as requested.
This can be done if we formally setting (`, x`) = (0, 0) in the proof of (3.3) to deduce that
P
(i)
1 (ν̃) = P
(i)
1 (ν)− 2 ≥ −1. �
Corollary 3.9. Let λ = wt(ν, J). Then we have
〈hi, λ〉 = ϕi(ν, J)− εi(ν, J).
Proof. From Theorem 3.8, we have ϕi(ν, J) − εi(ν, J) =
(
P
(i)
∞ (ν) − s
)
− (−s) = P
(i)
∞ (ν). On
the other hand, from (3.2) we have (see also definitions in Section 2.1)
〈hi, λ〉 =
∑
a∈I0
(∣∣µ(a)
∣∣〈hi, Λ̄a〉 − ∣∣ν(a)
∣∣〈hi, αa〉)
=
∣∣µ(i)
∣∣− 2
∣∣ν(i)
∣∣+
∑
a∈I0, a∼i
∣∣ν(a)
∣∣ = P (i)
∞ (ν).
Hence we have 〈hi, λ〉 = ϕi(ν, J)− εi(ν, J). �
Example 3.10. Let us consider the following element of (B1,1)⊗8
b = 1 ⊗ 2 ⊗ 3 ⊗ 1 ⊗ 1 ⊗ 2 ⊗ 1 ⊗ 1
We can see that ϕ1(b) = 3. Then we can apply f̃1 as follows
5
3
1
2 0 0
f̃1−→
3
3
1
−1
−1
0 1 1
f̃1−→
3
−1
−1
−1
−2
−2 1 1
f̃1−→
3
−1
−3
−1
−2
−3 1 1
The leftmost rigged configuration corresponds to the path b. For (ν, J) = Φ−1(b) we have
P
(1)
∞ (ν) = 3 and s in Theorem 3.8 is s = 0 since the smallest rigging is 1. Therefore we have
ϕ1(ν, J) = P
(1)
∞ (ν)− s = 3.
Combining Proposition 3.5 and Corollary 3.9, we obtain the following result.
Theorem 3.11. The maps εi, ϕi, wt and the Kashiwara operators ẽi, f̃i defined for the rigged
configurations satisfy the axioms for the Uq(g0)-crystals for g = A
(1)
n or D
(1)
n given in Defini-
tion 2.1.
For the later purposes, let us prepare some refined estimates related with the convexity
relations. The first one is an intuitive estimate.
Lemma 3.12. Let m
(i)
l (ν) = m. Then P
(i)
l−1(ν), P
(i)
l (ν), P
(i)
l+1(ν) − 2m satisfy the convexity
relation.
Proof. Let (ν ′, J) be the fictitious rigged configuration obtained by replacing all length l strings
of (ν, J)(i) by length l + 1 strings. From the similar argument of the proof of Proposition 3.6,
we can show that P
(i)
l−1(ν ′), P
(i)
l (ν ′), P
(i)
l+1(ν ′) satisfy the convexity relation. Since P
(i)
l−1(ν ′) =
P
(i)
l−1(ν), P
(i)
l (ν ′) = P
(i)
l (ν) and P
(i)
l+1(ν ′) = P
(i)
l+1(ν)− 2m we obtain the assertion. �
For some purposes, we need a more refined estimate as follows.
Rigged Configurations and Kashiwara Operators 17
Lemma 3.13. Suppose that l ≥ 1. For 1 ≤ a < n− 2 we have
−P (a)
l−1(ν) + 2P
(a)
l (ν)− P (a)
l+1(ν) ≥ m(a−1)
l (ν)− 2m
(a)
l (ν) +m
(a+1)
l (ν),
and for n− 2 ≤ a we have
−P (n−2)
l−1 (ν) + 2P
(n−2)
l (ν)− P (n−2)
l+1 (ν) ≥ m(n−3)
l (ν)− 2m
(n−2)
l (ν) +m
(n−1)
l (ν) +m
(n)
l (ν),
−P (n−1)
l−1 (ν) + 2P
(n−1)
l (ν)− P (n−1)
l+1 (ν) ≥ m(n−2)
l (ν)− 2m
(n−1)
l (ν),
−P (n)
l−1(ν) + 2P
(n)
l (ν)− P (n)
l+1(ν) ≥ m(n−2)
l (ν)− 2m
(n)
l (ν).
Proof. Since Q
(a)
l (ν) counts the number of boxes of the left l columns of ν, we have
−Q(a)
l−1(ν) + 2Q
(a)
l (ν)−Q(a)
l+1(ν) =
{
Q
(a)
l (ν)−Q(a)
l−1(ν)
}
−
{
Q
(a)
l+1(ν)−Q(a)
l (ν)
}
=
∑
k≥l
m
(a)
k (ν)−
∑
k≥l+1
m
(a)
k (ν) = m
(a)
l (ν).
Similarly, we obtain
−Q(a)
l−1(µ) + 2Q
(a)
l (µ)−Q(a)
l+1(µ) = L
(a)
l .
Then for a < n− 2 we have
−P (a)
l−1(ν) + 2P
(a)
l (ν)− P (a)
l+1(ν) = L
(a)
l +m
(a−1)
l (ν)− 2m
(a)
l (ν) +m
(a+1)
l (ν).
Since L
(a)
l ≥ 0, we obtain the assertion. Similarly we can show the assertions for n− 2 ≤ a. �
A nice property about these convexity relations is that they do not involve explicitly the
information on L
(a)
l or µ. Indeed, we will not refer to L
(a)
l or µ explicitly during the arguments
of Sections 5, 6 and 7.
3.3 Rigged configuration bijection
Let us define the bijection between the rigged configurations and tensor products of crystals
expressed by the KR tableaux which we call paths. For the map from the rigged configuration
to paths, the basic operation is called δ
δ : (ν, J) 7−→ {(ν ′, J ′), k},
where (ν, J) and (ν ′, J ′) are rigged configurations and k ∈ {1, 2, . . . , n, n̄, . . . , 2̄, 1̄}. In the
description, we call a string (l, xl) of (ν, J)(a) singular if we have xl = P
(a)
l (ν), that is, the
rigging takes the maximal possible value. To begin with we consider the generic case (non-spin
case). Necessary modifications related with the spin cases will be given in Section 4.5.
Definition 3.14. Let us consider a rigged configuration (ν, J) corresponding to the tensor
product of the form Ba,l ⊗ B′ (equivalently, µ(a) has length l row). For 1 ≤ a ≤ n − 2, the
map δ
(a)
l
δ
(a)
l : (ν, J) 7−→ {(ν ′, J ′), k}
is defined by the following procedure. Set `(a−1) = l.
(1) For a ≤ i ≤ n − 2, assume that `(i−1) is already determined. Then we search for the
shortest singular string in (ν, J)(i) that is longer than or equal to `(i−1).
18 R. Sakamoto
(a) If there exists such a string, set `(i) to be the length of the selected string and continue
the process recursively. If there is more than one such string, choose any of them.
(b) If there is no such string, set `(i) =∞, k = i and stop.
(2) Suppose that `(n−2) < ∞. Then we search for the shortest singular string in (ν, J)(n−1)
(resp. (ν, J)(n)) that is longer than or equal to `(n−2) and define `(n−1) (resp. `(n)) similarly.
(a) If `(n−1) =∞ and `(n) =∞, set k = n− 1 and stop.
(b) If `(n−1) <∞ and `(n) =∞, set k = n and stop.
(c) If `(n−1) =∞ and `(n) <∞, set k = n̄ and stop.
(d) If `(n−1) <∞ and `(n) <∞, set `(n−1) = max{`(n−1), `(n)} and continue.
(3) For 1 ≤ i ≤ n− 2, assume that `(i+1) is already defined. Then we search for the shortest
singular string in (ν, J)(i) that is longer than or equal to `(i+1) and has not yet been selected
as `(i). Define `(i) similarly. If `(i) = ∞, set k = i+ 1 and stop. Otherwise continue. If
`(1) <∞, set k = 1̄ and stop.
(4) Once the process has stopped, remove the rightmost box of each selected row specified
by `(i) or `(i). The result gives the output ν ′.
(5) Define the new riggings J ′ as follows. For the rows that are not selected by `(i) or `(i),
take the corresponding riggings from J . In order to define the remaining riggings, replace
one Ba,l in B by Ba−1,1 ⊗ Ba,l−1 (equivalently, replace one of the length l row of µ(a) by
a length (l− 1) row and add a length 1 row to µ(a−1)). Denote the result by B′. Use B′ to
compute all the vacancy numbers for ν ′. Then the remaining riggings are defined so that
all the corresponding rows become singular with respect to the new vacancy number.
We remark that the resulting rigged configuration (ν ′, J ′) is associated with the tensor pro-
duct B′. We write δ2δ1(ν, J) etc. for repeated applications of δ on the rigged configurations.
Definition 3.15. For a given rigged configuration (ν, J) corresponding to the tensor product of
the form B = Br1,s1 ⊗Br2,s2 ⊗· · ·⊗BrL,sL , define the map ΦB (sometimes also just denoted Φ)
ΦB : (ν, J) 7−→ b
as follows.
(1) Suppose that δ
(1)
1 · · · δ
(r1−1)
1 δ
(r1)
s1 (ν, J) = (ν ′, J ′) yields the sequence of letters k(r1), k(r1−1),
. . ., k(1) (k(a) corresponds to δ(a)). Put the transpose of the row
k(1) k(2) . . . k(r1)
as the leftmost column of the rectangle (sr11 ).
(2) Continue the previous step for δ
(1)
1 · · · δ
(r1−1)
1 δ
(r1)
s1−1(ν ′, J ′) = (ν ′′, J ′′) and fill the second
column for the rectangle (sr11 ) with the produced letters. Repeat the process until all
places of (sr11 ) are filled.
(3) Repeat the previous two steps for the remaining rectangles (sr22 ), (sr33 ), . . . , (srLL ).
Remark 3.16. The inverse procedure Φ−1 is obtained by reversing all procedure in Φ step by
step. As an example let us give a sketch of the algorithm for the operation
(
δ
(1)
1
)−1
. Suppose
that we start from an unbarred letter k in a KR tableau. Then we look for the largest singular
string of ν(k). If there is no such string, add a length one row to the bottom of ν(k). Otherwise
add a box to the longest singular row of ν(k). Suppose that we add the box to the `(k)th column
Rigged Configurations and Kashiwara Operators 19
of ν(k). Then we look for the singular string of ν(k−1) which is strictly shorter than `(k). We
continue this process up to ν(1). Finally change the riggings for the modified strings according
to the new vacancy numbers. Here remind that we have to add a length one row to µ(1) for the
computation of the vacancy numbers.
The following fundamental result is conjectured in [24].
Conjecture 3.17. Consider a tensor product of crystals B = Br1,s1 ⊗ Br2,s2 ⊗ · · · ⊗ BrL,sL.
Then the map ΦB gives a well-defined bijection between the rigged configuration of type B and
the elements of B expressed via the KR tableaux representation.
Currently this conjecture is verified in the following cases:
• B = Br1,s1 ⊗Br2,s2 ⊗ · · · ⊗BrL,sL type tensor products of A
(1)
n [19],
• B = Br1,1 ⊗Br2,1 ⊗ · · · ⊗BrL,1 type tensor products of D
(1)
n [31],
• B = B1,s1 ⊗B1,s2 ⊗ · · · ⊗B1,sL type tensor products of D
(1)
n [34],
• B = Br,s (r < n− 1) case of D
(1)
n [24].
A proof for the general case B = Br1,s1 ⊗Br2,s2 ⊗· · ·⊗BrL,sL of type D
(1)
n is the subject of [25].
However, since the paper [25] is yet not available for public, we will not assume this result in
this paper.
If there is no afraid of confusion, we use the abbreviation Φ instead of ΦB.
Example 3.18. Let us consider the following rigged configuration corresponding to the tensor
product B2,2 ⊗B3,2 of type D
(1)
5 :
0 −1
0
1
1
0
1
1
0
−2
−2
−3
0
−2
−2
−3
0
0
0
0
−1
0
−1
0
We start by removing B2,2 part. Then the first operation is δ
(2)
2 which start by searching the
shortest singular string of ν(2) = (4, 3, 1) whose length is equal to or larger than 2. We work out
all the procedure corresponding to B2,2 in the following sequence of diagrams
0 −1
0
1
1
×
0
1
1
0
−2
−2
−3 ×
×
0
−2
−2
−3
0
0 ×
0
0
−1
0 ×
−1
0
δ
(2)
2
?
3̄
0 −1
0
0
1
0
0
1
0
0
−2
−2
0
0
−2
−2
0
0
0
0
−1
−1
−1
−1
δ
(1)
1
?
1
3̄
20 R. Sakamoto
−1 × −1
0
0
1
×
× 0
0
1
0
0
−2
−2
×
× 0
0
−2
−2
0
0
× 0
0
−1
−1
× −1
−1
δ
(2)
1
?
1
3̄ 1̄
−1 × −1
0
1 ×
0
1
0
−2
−2 ×
0
−2
−2 0 × 0
−1
−1
−1
−1
δ
(1)
1
?
1 5
3̄ 1̄
−1 −1
0
1
0
1
0
−2
−2
0
−2
−2 1 1
−1
−1
−1
−1
Here the boxes to be removed by δ
(a)
l indicated below the corresponding rigged configurations
are marked by “×”. The output of each δ
(a)
l is given on the right of the corresponding arrows.
Note that after δ
(2)
2 we recalculate all the vacancy numbers assuming that the resulting rigged
configuration corresponds to the tensor product of type B1,1 ⊗ B2,1 ⊗ B3,2. We remark that
the first δ
(1)
1 cannot remove a box from ν(1) since there is no singular string and, δ
(2)
1 starts by
removing from ν(2) and ends at ν(1). The final output for the given rigged configuration (ν, J)
is as follows:
ΦB2,2⊗B3,2(ν, J) = 1 5
3̄ 1̄
⊗
1 5̄
4 3̄
5 1̄
(3.4)
Remark 3.19. In the above example, we can reverse the order of the tensor product. In this
case, we obtain
ΦB3,2⊗B2,2(ν, J) =
1 4
5 2̄
3̄ 1̄
⊗ 2 5̄
5 3̄
(3.5)
Then the two tensor products in (3.4) and (3.5) are isomorphic under the combinatorial R-matrix
R : 1 5
3̄ 1̄
⊗
1 5̄
4 3̄
5 1̄
7−→
1 4
5 2̄
3̄ 1̄
⊗ 2 5̄
5 3̄
Therefore in this case we have R = ΦB3,2⊗B2,2 ◦ Φ−1
B2,2⊗B3,2 . This relation is valid not only for
two times tensor product but also for multiple applications of pairwise combinatorial R-matrices
for arbitrary tensor products.
Rigged Configurations and Kashiwara Operators 21
4 Rigged configurations and the Kashiwara operators
4.1 Statement of the main result
Let us state the main result of the present paper.
Theorem 4.1. Suppose that the rigged configuration bijection Φ is well-defined. Then the
bijection Φ and the Kashiwara operators ẽi and f̃i (i ∈ I0) of types A
(1)
n and D
(1)
n satisfy the
following commutative diagrams:
RC(B)
Φ−−−−→ B
f̃i
y yf̃i
RC(B) ∪ {0} −−−−→
Φ
B ∪ {0}
RC(B)
Φ−−−−→ B
ẽi
y yẽi
RC(B) ∪ {0} −−−−→
Φ
B ∪ {0}
Here for the tensor product of crystals B = BrL,sL ⊗ BrL−1,sL−1 ⊗ · · · ⊗ Br1,s1 , we denote
by RC(B) the rigged configurations which are mapped to elements of B under ΦB, that is,
ΦB : RC(B)→ B.
Remark 4.2. Note that RC(B) does not fully depend on B but depends only on the shape
of B via the quantities L
(a)
l or µ. Thus it is possible to denote the set of rigged configurations
as RC(L(B)).
Therefore we have the compatibility of the rigged configuration bijection and the classical
Kashiwara operators. We give a few words on the compatibility for the affine Kashiwara ope-
rators ẽ0 and f̃0 of type D
(1)
n . Recall that in [33] the operators ẽ0 and f̃0 are realized via the
bijection between the set of J-highest element of Br,s and the combinatorial objects called the
plus-minus diagrams. Here J = {2, 3, . . . , n}. In [24, Section 4], we introduce an analogue of the
plus-minus diagrams on the set of the rigged configurations and show that this indeed yields
the affine isomorphism with the KR crystal Br,s.
The main points of the arguments in [24] are as follows. In [24, Theorem 5.9] we prove the
combinatorial bijection Φ for the I0-highest element of Br,s directly. Then in [24, Proposition 4.3]
we show that the same sequence of the classical Kashiwara operators f̃i gives the plus-minus
diagram and the corresponding element in the set of rigged configuration. Therefore Theo-
rem 4.1 asserts that the rigged configuration analogue of the plus-minus diagrams introduced
in [24, Section 4] are indeed the image of the plus-diagrams under the rigged configuration bi-
jection Φ. Hence, in view of the construction of [33], we have the compatibility of the affine
Kashiwara operators ẽ0, f̃0 and the rigged configuration bijection for the case Br,s. However,
the compatibility of the affine Kashiwara operators and the combinatorial bijection Φ for more
than two tensor factors is still an open problem.
The rest of the paper is devoted to the proof of Theorem 4.1.
4.2 Preliminary steps
Our main strategy of the proof is to reduce the essential combinatorial aspects of Theorem 4.1 to
the fundamental operation δ = δ
(1)
1 . For this purpose we decompose the original bijection Φ into
several steps according to the operations lh, lb and ls defined on the KR tableau representation
of the crystals. In the following, we define the corresponding operations (summarized in the
following table) on the set of the rigged configurations.
B RC(B)
ls γ
lb β
lh δ
22 R. Sakamoto
First, let us remind that the following symbols bear specific meanings throughout the paper.
• ν: shapes of the configurations,
• J : riggings,
• µ: shape of the tensor product.
Definition 4.3. Let B = BrL,sL ⊗BrL−1,sL−1 ⊗ · · · ⊗Br1,s1 .
(1) Corresponding to BrL,sL = Ba,l, γ replaces a length l row of µ(a) by rows of lengths l − 1
and 1.
(2) Suppose that BrL,sL = Br,1 where r > 1. If r < n − 1, then β removes a length one row
from µ(r), adds a length one row to each of µ(1) and µ(r−1) and adds a length one singular
string to each of (ν, J)(a) for a < r.
(3) Suppose that BrL,sL = B1,1. Then define δ = δ
(1)
1 .
Proposition 4.4. We can decompose the map Φ into the following three steps.
(1) Let B = Br,s ⊗ B̄ with s > 1. Then we reduce the problem to the case of Br,1 by
RC(B)
Φ−−−−→ B
γ
y yls
RC(B′) −−−−→
Φ
B′
where B′ = Br,1 ⊗Br,s−1 ⊗ B̄.
(2) Let B = Br,1 ⊗ B̄ with r > 1. Then we reduce the problem to the case of B1,1 by
RC(B)
Φ−−−−→ B
β
y ylb
RC(B′) −−−−→
Φ
B′
where B′ = B1,1 ⊗Br−1,1 ⊗ B̄.
(3) Let B = B1,1 ⊗ B̄. Then we have to deal with the fundamental operation δ as
RC(B)
Φ−−−−→ B
δ
y ylh
RC(B̄) −−−−→
Φ
B̄
Proof. (1) Suppose that Φ maps (ν, J) ∈ RC(B) to b⊗ b̄ ∈ B. Recall that γ replaces a length s
row of µ(r) by rows of lengths s− 1 and 1. Thus after application of γ there is no singular string
of ν(r) which are strictly shorter than s. Then the length 1 row of µ(r) corresponds to the same
column which appear as the leftmost column of the tableau b.
(2) If we apply δ
(1)
1 for RC(B′), it automatically selects the length 1 singular strings of ν(a)
(a < r) created by β. As for ν(r), since β removes length 1 row from µ(r) and adds a length 1
row to ν(r−1), all the coriggings for ν(r) do not change after β. Thus δ
(1)
1 selects the same string
which is selected by δ
(r)
1 before β.
(3) This follows from the definition. �
Rigged Configurations and Kashiwara Operators 23
4.3 Reduction steps
We show that the operations γ and β commutes with the Kashiwara operators.
Proposition 4.5. Let B = Br,s ⊗ B̄ where s > 1. Then we have the following commutative
diagrams
RC(B)
γ−−−−→ RC(B′)
ẽi
y yẽi
RC(B) −−−−→
γ
RC(B′)
RC(B)
γ−−−−→ RC(B′)
f̃i
y yf̃i
RC(B) −−−−→
γ
RC(B′)
where B′ = Br,1 ⊗Br,s−1 ⊗ B̄ and i ∈ I0.
Proof. For the rigged configuration {µ, (ν, J)}, ẽi and f̃i change (ν, J) part only and γ changes µ
part only. Thus they commute.
We remark that if we have ẽi(ν, J) = 0 or f̃i(ν, J) = 0, we have ẽi◦γ(ν, J) = 0 or f̃i◦γ(ν, J) = 0
by Theorem 3.8 since γ does not alter the vacancy numbers nor riggings. �
Proposition 4.6. Let B = Br,1 ⊗ B̄ where r > 1. Then we have the following commutative
diagrams
RC(B)
β−−−−→ RC(B′)
ẽi
y yẽi
RC(B) −−−−→
β
RC(B′)
RC(B)
β−−−−→ RC(B′)
f̃i
y yf̃i
RC(B) −−−−→
β
RC(B′)
where B′ = B1,1 ⊗Br−1,1 ⊗ B̄ and i ∈ I0.
Proof. Let us prove the f̃i part first. During the proof we denote that f̃i(ν, J) = (ν̃, J̃) and
β(ν, J) = (ν̇, J̇).
Suppose that i > r. Recall that the operation β does nothing on (ν, J)(a) for a ≥ r. Thus β
and f̃i commute since they do not have interaction.
Suppose that i = r. Since the operation β adds a length one string to (ν, J)(r−1), the
interaction between β and f̃r can occur if f̃r adds the string (1,−1) to (ν, J)(r). In this case,
since β does nothing on (ν, J)(r), f̃r adds the same string (1,−1) after β. Let us consider
(ν, J)(r−1). Then we have to check the coincidence of the string added by β:
∅ f̃r7−→ ∅ β7−→
(
1, P
(r−1)
1 (ν̃)
)
,
∅ β7−→
(
1, P
(r−1)
1 (ν)
) f̃r7−→
(
1, P
(r−1)
1 (ν̃)
)
.
Here we use the fact that f̃r changes the riggings so as to keep the coriggings.
Finally, suppose that i < r. Let xk be the smallest rigging of length k strings of (ν, J)(i).
Let x` be the smallest rigging of (ν, J)(i) where ` is the length of the largest string with rigging x`.
Suppose that ν(i) 6= ∅ and x` ≤ 0. Recall that β adds the singular string (1, P
(i)
1 (ν̇))
to (ν, J)(i). If we show
P
(i)
1 (ν̇) ≥ x`,
then we see that f̃i chooses the same string (`, x`) even after β. Let j be the length of the
shortest string of (ν, J)(i). Suppose that j = 1. Then we have P
(i)
1 (ν̇) = P
(i)
1 (ν) by definition
24 R. Sakamoto
of β, P
(i)
1 (ν) ≥ x1 by the definition of the rigged configurations and x1 ≥ x` by the minimality
of x`. Thus we have P
(i)
1 (ν̇) ≥ x` in this case. Next, suppose that j > 1. By the convexity
relation of P
(i)
k (ν) between 0 ≤ k ≤ j we have
P
(i)
1 (ν̇) = P
(i)
1 (ν) ≥ min
{
P
(i)
0 (ν), P
(i)
j (ν)
}
= min
{
0, P
(i)
j (ν)
}
.
If we have P
(i)
j (ν) ≥ 0, we have P
(i)
1 (ν̇) ≥ 0 ≥ x`. On the other hand, suppose that 0 > P
(i)
j (ν).
Then we have P
(i)
1 (ν̇) ≥ P
(i)
j (ν) by the above inequality, P
(i)
j (ν) ≥ xj by the definition of the
rigged configurations and xj ≥ x` by the minimality of x`. Therefore we have P
(i)
1 (ν̇) ≥ x`.
Let us consider the case ν(i) = ∅ or x` > 0. In both cases, if we have
P
(i)
1 (ν̇) > 0,
then we have the following behaviors of the two strings of (ν, J)(i):
∅
∅
β7−→
(
1, P
(i)
1 (ν̇)
)
∅
f̃i7−→
(
1, P
(i)
1 (˜̇ν)
)
(1,−1)
and, on the other hand,
∅
∅
f̃i7−→ ∅
(1,−1)
β7−→ (1, P
(i)
1 ( ˙̃ν))
(1,−1)
.
Here we denote by f̃i◦β(ν, J) = (˜̇ν, ˜̇J) and β◦ f̃i(ν, J) = ( ˙̃ν,
˙̃
J). Since β does not change vacancy
numbers, we have P
(i)
1 (˜̇ν) = P
(i)
1 ( ˙̃ν). Thus f̃i and β commute in this case.
Now let us consider the case P
(i)
1 (ν) = P
(i)
1 (ν̇) ≤ 0.
• Suppose that we have ν(i) = ∅. Then we have P
(i)
∞ (ν) ≥ 0 since we have Q
(i)
∞ (ν) = 0.
Now we invoke the convexity relation of P
(i)
k (ν) between 0 ≤ k ≤ ∞ with 0 = P
(i)
0 (ν) ≥
P
(i)
1 (ν) ≤ P
(i)
∞ (ν) to deduce that P
(i)
0 (ν) = P
(i)
1 (ν) = · · · = P
(i)
∞ (ν) = 0. Since ν(i) = ∅,
this relation also implies that ν(i−1) = ν(i+1) = µ(i) = ∅.
Consider f̃i acting on (ν, J). Suppose that i < r − 1. Then it adds the string (1,−1)
to (ν, J)(i) whereas the corresponding vacancy number is P
(i)
1 (ν̃) = P
(i)
1 (ν) − 2 = −2.
Thus we have f̃i(ν, J) = 0. On the other hand, β acting on (ν, J) adds the length 1
singular string and gives (ν̇, J̇)(i) = {(1, P (i)
1 (ν̇))}. The next f̃i selects this string since we
have P
(i)
1 (ν̇) = P
(i)
1 (ν) = 0. Then we have (˜̇ν, ˜̇J)(i) = {(2,−1)}. In particular we have
Q
(i)
2 (˜̇ν) = 2. Since ν(i−1) = ν(i+1) = ∅, both ˜̇ν(i−1)
and ˜̇ν(i+1)
consist of length 1 string.
Thus Q
(i−1)
2 (˜̇ν) = Q
(i+1)
2 (˜̇ν) = 1. Since we also know µ(i) = ∅, we have Q
(i)
2 (˜̇µ) = 0.
Hence P
(i)
2 (˜̇ν) = Q
(i)
2 (˜̇µ) − 2Q
(i)
2 (˜̇ν) + Q
(i−1)
2 (˜̇ν) + Q
(i+1)
2 (˜̇ν) = −2. Therefore we have
f̃i ◦β(ν, J) = 0. Hence f̃i and β commute in this case. If i = r− 1, we can use the parallel
argument if we replace ν(i+1) by µ(i).
• Since we have completed the proof for the case ν(i) = ∅, we can assume that ν(i) 6= ∅.
Let j be the length of the smallest string of (ν, J)(i). Suppose that j = 1. Since we are
considering the case ν(i) = ∅ or x` > 0, the condition ν(i) 6= ∅ implies that x` > 0. On
the other hand we have 0 ≥ P
(i)
1 (ν) ≥ x1 by the assumption and the definition of the
rigged configurations. This is in contradiction to the relation x1 ≥ x` which follows from
the minimality of x`. Hence this case cannot happen.
Rigged Configurations and Kashiwara Operators 25
• Suppose that j > 1. Again we have x` > 0. By the convexity relation of P
(i)
k (ν) between
0 ≤ k ≤ j and the definition of the rigged configurations we have 0 = P
(i)
0 (ν) ≥ P (i)
1 (ν) ≥
P
(i)
j (ν) ≥ xj . This is in contradiction to the relation xj ≥ x`. Hence this case cannot
occur.
Thus we have checked the commutativity of f̃i and β in all possible cases.
Finally let us consider the proof for ẽi. If ẽi(ν, J) 6= 0, then the commutativity of ẽi and β
follows from the result of f̃i since ẽi and f̃i are mutually inverse in this case. Suppose that
ẽi(ν, J) = 0. Then all the riggings of (ν, J)(i) are non-negative. Recall that β acting on (ν, J)
adds the string (1, P
(i)
1 (ν̇)). Then if we show P
(i)
1 (ν̇) = P
(i)
1 (ν) ≥ 0 we have ẽi ◦ β(ν, J) = 0.
Let j be the length of the shortest string of (ν, J)(i). If ν(i) = ∅, set j = ∞. In both cases
we have P
(i)
j (ν) ≥ 0 since we have P
(i)
j (ν) ≥ xj ≥ 0 if j < ∞ and Q
(i)
j (ν) = 0 if j = ∞.
Now we invoke the convexity relation of P
(i)
k (ν) between 0 ≤ k ≤ j. Then we have P
(i)
1 (ν) ≥
min
{
P
(i)
0 (ν), P
(i)
j (ν)
}
= min
{
0, P
(i)
j (ν)
}
= 0 as desired. �
4.4 Statements to be proved
Due to the arguments in the previous section, the essential points of the proof of Theorem 4.1
are reduced to the following two assertions on the case B = B1,1⊗ B̄. Up to Section 6 from here
we will use the following notation when we consider the case B = B1,1⊗ B̄. Let (ν, J) ∈ RC(B).
Denote the rigged configurations obtained by the actions of f̃i or δ as in the following diagram:
(ν, J)
δ−−−−→ (ν̄, J̄)
f̃i
y yf̃i
(ν̃, J̃) −−−−→
δ
(¯̃ν,
¯̃
J), (˜̄ν, ˜̄J)
Here we denote as δ : (ν̃, J̃) 7→ (¯̃ν,
¯̃
J) and f̃i : (ν̄, J̄) 7→ (˜̄ν, ˜̄J). We denote the smallest rigging
associated with length k rows of ν(i) by xk and express such string as (k, xk). For later purposes,
let us prepare several notation for the specific strings as follows:
• (`, x`) is the string of (ν, J) that is selected by f̃i,
• (¯̀, x¯̀) is the string of (ν̄, J̄) that is selected by f̃i,
• `(a) and `(a) are the lengths of the singular strings of (ν, J)(a) which are selected by δ under
the condition `(a) ≤ `(a),
• ˜̀(a) and ˜̀(a) are the lengths of the singular strings of (ν̃, J̃)(a) which are selected by δ under
the condition ˜̀(a) ≤ ˜̀(a).
Then our main claims to be proved are the following two assertions which are proved in
Sections 5, 6 and 7.
Proposition 4.7. Let B = B1,1 ⊗ B̄ and i ∈ I0. Then we have the following commutative
diagrams
RC(B)
δ−−−−→ RC(B̄)
ẽi
y yẽi
RC(B) −−−−→
δ
RC(B̄)
RC(B)
δ−−−−→ RC(B̄)
f̃i
y yf̃i
RC(B) −−−−→
δ
RC(B̄)
if all ẽi and f̃i in the above diagrams are defined.
26 R. Sakamoto
Proposition 4.8. Let B = B1,1⊗ B̄, i ∈ I0, (ν, J) ∈ RC(B) and b = b′⊗ b̄ ∈ B1,1⊗ B̄. Suppose
that Φ : (ν, J) 7−→ b. Then we have the following properties:
(1) f̃i(ν, J) 6= 0 and f̃i(ν̄, J̄) = 0 =⇒ f̃i(b) 6= 0, f̃i(b̄) = 0 and Φ(f̃i(ν, J)) = f̃i(Φ(ν, J)),
(2) f̃i(b) 6= 0 and f̃i(b̄) = 0 =⇒ f̃i(ν, J) 6= 0, f̃i(ν̄, J̄) = 0 and Φ−1(f̃i(b)) = f̃i(Φ
−1(b)),
(3) f̃i(ν, J) = 0 and f̃i(ν̄, J̄) 6= 0 =⇒ f̃i(b) = 0, f̃i(b̄) 6= 0 and Φ(f̃i(ν, J)) = f̃i(Φ(ν, J)),
(4) f̃i(b) = 0 and f̃i(b̄) 6= 0 =⇒ f̃i(ν, J) = 0, f̃i(ν̄, J̄) 6= 0 and Φ−1(f̃i(b)) = f̃i(Φ
−1(b)),
(5) ẽi(ν, J) 6= 0 and ẽi(ν̄, J̄) = 0 =⇒ ẽi(b) 6= 0, ẽi(b̄) = 0 and Φ(ẽi(ν, J)) = ẽi(Φ(ν, J)),
(6) ẽi(b) 6= 0 and ẽi(b̄) = 0 =⇒ ẽi(ν, J) 6= 0, ẽi(ν̄, J̄) = 0 and Φ−1(ẽi(b)) = ẽi(Φ
−1(b)),
(7) the situation ẽi(ν, J) = 0 and ẽi(ν̄, J̄) 6= 0 cannot happen,
(8) the situation ẽi(b) = 0 and ẽi(b̄) 6= 0 cannot happen.
Here from (1) to (4) we assume the commutativity of f̃i with Φ for B̄ and from (5) to (8) we
assume the commutativity of ẽi and f̃i with Φ for B̄.
Let us show how these properties lead to the proof of Theorem 4.1. In the proof, we will use
the diagram of the following kind as in [19]:
• A //
C
��
��
•
B
��
��
• //
��
•
��
• // •
•
D
//
??
•
i
__
We regard this as a cube with front face given by the large square. Suppose that the square
diagrams given by the faces of the cube except for the front face commute and i is the injective
map. Then the front face also commutes since we have
i ◦B ◦A = i ◦D ◦ C
by a diagram chasing.
Proof of Theorem 4.1. First we prove the f̃i case. Then we can prove the ẽi case by a parallel
argument.
We follow the decomposition of Φ as described in Proposition 4.4. The simplest case B1,1
can be shown directly. We give a list of the explicit rigged configurations in this case:
b ∈ B1,1 (ν, J)(a) of Φ−1(b) values for a
i {(1, 0)} a < i− 1,
{(1,−1)} a = i− 1
∅ i ≤ a.
n̄ {(1, 0)} a ≤ n− 2,
∅ a = n− 1
{(1,−1)} a = n
ī {(1, 0)} a ≤ i− 2,
(i < n) {(1, 1)} a = i− 1
{(1,−1), (1,−1)} a = i
{(1, 0), (1, 0)} i+ 1 ≤ a ≤ n− 2,
{(1, 0)} a = n− 1, n
Rigged Configurations and Kashiwara Operators 27
By using the crystal graph of B1,1 presented in Section 2.2.1 we can verify the commutativity
of f̃i and Φ.3
Let us consider the case B = B1,1⊗ B̄. This case corresponds to an addition of a new tensor
factor. For this we use the following diagram:
RC(B)
Φ //
f̃i
��
δ
%%
B
f̃i
��
lh
zz
RC(B̄)
Φ //
f̃i
��
B̄
f̃i
��
RC(B̄)
Φ // B̄
RC(B)
Φ
//
δ
99
B
lh
dd
Suppose that all f̃i which appear in the above diagram are defined. The top face and the bottom
face commute by Proposition 4.4. The left face commutes by Proposition 4.7. The right face
commutes by definition of lh on tensor products of crystals. The back face commutes by the
assumption. Finally the map lh is injective if we consider the weight of the element of B1,1
subtracted by the operation. Hence the front face commutes. This proves the commutativity
of Φ and f̃i when both f̃i before or after δ (or lh) are defined. On the other hand we know
from Proposition 4.8 that Φ and f̃i also commute even if one of f̃i before or after δ (or lh) is
undefined. By exclusion we have the commutativity for the case when both f̃i before and after δ
(or lh) are undefined.
Let us consider the the case B = Br,1⊗ B̄ with r > 1. Let B′ = B1,1⊗Br−1,1⊗ B̄. Consider
the following diagram:
RC(B)
Φ //
f̃i
��
β
%%
B
f̃i
��
lb
yy
RC(B′)
Φ //
f̃i
��
B′
f̃i
��
RC(B′)
Φ // B′
RC(B)
Φ
//
β
99
B
lb
ee
The top face and the bottom face commute by Proposition 4.4. The left face commutes
by Proposition 4.6. The right face commutes by definition of lb on tensor products of crys-
tals. The back face commutes by the assumption. Since lb is injective, the front face com-
mutes.
3In B1,1 case Φ coincides with δ. In view of the crystal graph of B1,1 and the definition of f̃i on the rigged
configurations we see a connection between the crystal graph and δ in the case B1,1. However this is only the tip
of an iceberg. Indeed, in E
(1)
6 case [26], δ is described by the crystal graph of B1,1 as in the present case and the
same algorithm can be applied for general tensor products of the form (B1,1)L.
28 R. Sakamoto
Finally let us consider the the case B = Br,s ⊗ B̄ with s > 1. Let B′ = Br,1 ⊗ Br,s−1 ⊗ B̄.
Consider the following diagram:
RC(B)
Φ //
f̃i
��
γ
%%
B
f̃i
��
ls
yy
RC(B′)
Φ //
f̃i
��
B′
f̃i
��
RC(B′)
Φ // B′
RC(B)
Φ
//
γ
99
B
ls
ee
The top face and the bottom face commute by Proposition 4.4. The left face commutes by
Proposition 4.5. The right face commutes by the definition of ls on tensor products of crystals.
The back face commutes by the assumption. Since ls is injective, the front face commutes. This
completes the proof of Theorem 4.1. �
4.5 Spin cases
Let us consider the arguments which are needed to treat the spin cases. Here we will concentrate
on the case Bn,s. The other case Bn−1,s can be obtained by replacing the Dynkin nodes n and
n − 1 in the following description. To begin with, following [31], we introduce the three sets
B̂n−1,1, B̂n,1 and ˆ̄Bn,1 which are generated by
ûn−1 =
1
2
...
n− 1
, ûn =
1
2
...
n− 1
n
, ˆ̄un =
1
2
...
n− 1
n̄
Here we regard them as the usual single columns and apply the Kashiwara operators f̃i (i ∈ I0) as
in the case for non-spin KN tableaux to obtain the whole elements of B̂n−1,1, B̂n,1 and ˆ̄Bn,1. On
these sets, we define the operation left-box analogously to Definition 2.4. To be more explicit,
we have lb : B̂n,1 −→ B1,1 ⊗ B̂n−1,1 and lb : B̂n−1,1 −→ B1,1 ⊗Bn−2,1.
Now we describe the algorithm for the rigged configuration bijection for the spin cases Bn−1,l
and Bn,l following [31]. To begin with, we introduce the embedding of the rigged configurations
emb : {µ, (ν, J)} 7−→ {µ′, (ν ′, J ′)},
where the quantities with primes means that the everything is doubled as µ
′(a)
i = 2µ
(a)
i , ν
′(a)
i =
2ν
(a)
i and J
′(a)
i = 2J
(a)
i . Similarly, if the rigged configuration {µ, (ν, J)} is composed of even
integers, we can define emb−1 by dividing all its components by 2. Then the operation to obtain
the leftmost column of an element of Bn,l is
emb−1 ◦ δ(1)
1 · · · δ
(n−2)
1 δ̂
(n−1)
1 δ̂
(n)
2 ◦ emb ◦ γ(ν, J).
Here γ is the same as Definition 4.3; γ replaces one of length l rows of µ(n) by two rows of
length l− 1 and 1. The new operations δ̂
(n)
2 and δ̂
(n−1)
1 are defined as follows. Let us denote by
(ν ′, J ′) = emb ◦ γ(ν, J).
Rigged Configurations and Kashiwara Operators 29
• δ̂(n)
2 : (ν ′, J ′) 7−→ {(ν ′′, J ′′), k} is defined as follows. We look for the shortest singular string
of (ν ′, J ′)(n) which is longer than or equal to 2. If there is no such string, set `(n) = ∞,
k = n and stop. Otherwise we set `(n−1) to be the length of the selected string of (ν ′, J ′)(n)
and do the same procedure as in the usual δ
(a)
l .
The definition of the new rigged configuration (ν ′′, J ′′) is the same as in the usual δ
(a)
l
except for the following point. We replace one of the length 2 rows of µ′(n) by length 1
row and add a length one row to µ′(n−1).
• δ̂(n−1)
1 : (ν ′, J ′) 7−→ {(ν ′′, J ′′), k} is defined as follows. Set `(n−2) = 1. Then, as in the
usual δ
(a)
l we begin to determine `(n−1) and `(n) and proceed similarly to the usual case.
The new rigged configuration (ν ′′, J ′′) is defined similarly as in the usual case except for
the following point. We remove a length one row from each of µ′(n−1) and µ′(n) and add
a length one row to µ′(n−2).
We regard the output of the sequence of δ’s as the element of B̂n,1. Corresponding to the
final emb−1, we halve the width of the column with keeping all its contents. We regard the final
output as the element of Bn,1.
Remark 4.9. In the above algorithm, some readers might feel that the change in µ is peculiar.
To get a better understanding of this point, let us take a look at the following computations
2Λ̄n = ε1 + · · ·+ εn−2 + εn−1 + εn,
Λ̄n + Λ̄n−1 = ε1 + · · ·+ εn−2 + εn−1,
Λ̄n−2 = ε1 + · · ·+ εn−2.
Let us compare with the special case ν(a) = ∅ for all a ∈ I0 with weight 2Λ̄n (see (3.2)). In this
case we should obtain a column filled by 1, 2, . . . , n from top to bottom. Note that each letter i
in this column corresponds to εi.
For the proof of Theorem 4.1, we have to modify β in Definition 4.3 as follows. β̂(n) replaces
a length two row of µ(n) by a length one row, adds a length one row to each of µ(1) and µ(n−1)
and adds a length one singular string to each of (ν, J)(a) for a ≤ n− 1. β̂(n−1) removes a length
one row from each of µ(n−1) and µ(n), adds a length one row to each of µ(1) and µ(n−2) and adds
a length one singular string to each of (ν, J)(a) for a ≤ n − 2. Note that we have the following
property.
Proposition 4.10. On the rigged configurations, we have
emb ◦ ẽi = ẽ2
i ◦ emb, emb ◦ f̃i = f̃2
i ◦ emb.
Proof. We prove the ẽi part. Let us consider the case ẽi(ν, J) 6= 0. Suppose that ẽi acts on the
string (`, x`) of (ν, J)(i). Then ẽi makes the string (2`, 2x`) of emb(ν, J)(i) into (2`− 1, 2x` + 1).
Since x` < 0, we have 2x` + 1 < 0 and thus ẽ2
i ◦ emb(ν, J) 6= 0. We show that the second ẽi acts
on the string (2`− 1, 2x` + 1).
(a) Suppose that there is a string (k, xk) with k < ` in (ν, J)(i). By the minimality of ` we
have xk > x`. Therefore we have 2xk > 2x` + 1.
(b) Suppose that there is a string (k, xk) with ` ≤ k in (ν, J)(i). Then in the ẽi ◦ emb(ν, J)(i),
the string becomes (2k, xk + 2). By the minimality of x`, we have xk + 2 > x` + 1.
Therefore the second ẽi acts on the string (2` − 1, 2x` + 1) to make (2` − 2, 2x` + 2). Thus
emb ◦ ẽi = ẽ2
i ◦ emb in this case. On the other hand, if ẽi(ν, J) = 0, we have ẽi ◦ emb(ν, J) = 0
by Theorem 3.8.
Similarly we can show emb ◦ f̃i = f̃2
i ◦ emb. �
30 R. Sakamoto
Then we can use the same arguments of the proof of Proposition 4.6 to show that β̂(n−1)
and β̂(n) commute with ẽi and f̃i. Thus by using the same arguments we can show Theorem 4.1
in this case.
5 Proof of Proposition 4.7
5.1 Outline
We concentrate on the proof for f̃i part since the other part ẽi follows from the former case. We
divide the proof into the following five cases which exhaust all the possibilities:
Case A: `+ 1 < `(i−1),
Case B: `(i−1) ≤ `+ 1 < `(i) =∞,
Case C: `(i−1) ≤ `+ 1 ≤ `(i) <∞,
Case D: `(i) = `,
Case E: `(i) < `.
The actual proof for each case is based on the following behaviors of the vacancy numbers
induced by δ.
Lemma 5.1. For 1 ≤ i ≤ n we have
P
(i)
`(i)−1
(ν̄) =
P
(i)
`(i)−1
(ν) (if `(i−1) = `(i)), (I)
P
(i)
`(i)−1
(ν)− 1 (if `(i−1) < `(i)). (II)
Here if i = n we understand `(i−1) as `(n−2). For 1 ≤ i ≤ n− 3 we have
P
(i)
`(i)−1(ν̄) =
P
(i)
`(i)−1(ν) (if `(i−1) = `(i) = · · · = `(i)), (III)
P
(i)
`(i)−1(ν)− 1 (if `(i−1) < `(i) = · · · = `(i)), (IV)
P
(i)
`(i)−1(ν) + 1 (if `(i) < `(i+1) = · · · = `(i)), (V)
P
(i)
`(i)−1(ν) (if `(i+1) < `(i+1) = `(i)), (VI)
P
(i)
`(i)−1(ν)− 1 (if `(i+1) < `(i)), (VII)
and for i = n− 2 we have
P
(n−2)
`(n−2)−1(ν̄) =
P
(n−2)
`(n−2)−1(ν) (if `(n−3) = `(n−2) = `(n−1) = `(n) = `(n−2)), (VIII)
P
(n−2)
`(n−2)−1(ν)− 1 (if `(n−3) < `(n−2) = `(n−1) = `(n) = `(n−2)), (IX)
P
(n−2)
`(n−2)−1(ν) + 1 (if `(n−2) < `(n−1) = `(n) = `(n−2)), (X)
P
(n−2)
`(n−2)−1(ν) (if min{`(n−1), `(n)} < `(n−1) = `(n−2)), (XI)
P
(n−2)
`(n−2)−1(ν)− 1 (if `(n−1) < `(n−2)). (XII)
During the proof, the cases (VIII) to (XII) can be treated as the special case of the cases
(III) to (VII). Indeed, we can assume that `(n−1) ≤ `(n) without loss of the generality. Then
we have `(n−1) = `(n) and regard `(n−1) and `(n−1) as lengths of strings of the concatenation(
(ν(n−1), ν(n)), (J (n−1), J (n))
)
of (ν, J)(n−1) and (ν, J)(n).
For the reader’s convenience, we give a summary of all major subcases here.
Rigged Configurations and Kashiwara Operators 31
A `+ 1 < `(i−1).
B `(i−1) ≤ `+ 1 < `(i) =∞,
(1) f̃i creates a non-singular string,
(2) f̃i creates a singular string.
C `(i−1) ≤ `+ 1 ≤ `(i) <∞,
(1) m
(i)
`+1(ν) > 0 and f̃i creates a singular string, or ` + 1 = `(i) and f̃i creates a non-
singular string,
(2) m
(i)
`+1(ν) = 0 and f̃i creates a singular string,
(3) `+ 1 < `(i) and f̃i creates a non-singular string.
D `(i) = `,
(1) ` = `(i) < `(i) and m
(i)
` (ν) > 1,
(2) ` = `(i) = `(i) and m
(i)
` (ν) > 2,
(3) ` = `(i) < `(i) and m
(i)
` (ν) = 1,
(4) ` = `(i) = `(i) and m
(i)
` (ν) = 2.
E `(i) < `,
(1) `(i) < ` and `+ 1 < `(i+1),
(2) `(i) < ` and `(i+1) ≤ `+ 1 < `(i) =∞,
(3) `(i) < ` and `(i+1) ≤ `+ 1 ≤ `(i) <∞,
(4) `(i) < ` and `(i) = `,
(5) `(i) < ` and `(i) < `.
Finally, let us prepare the following lemma which gives an useful criterion.
Lemma 5.2. Suppose that f̃i acts on a string (`, x`) of (ν, J)(i) and creates the new string
(`+ 1, x` − 1).
(1) The string (`+ 1, x` − 1) is singular if and only if P
(i)
`+1(ν) = x` + 1.
(2) Suppose that f̃i creates a singular string. If there is a string (l, xl) of (ν, J)(i) that satisfies
` < l, then we have P
(i)
`+1(ν) ≤ xl.
Proof. (1) Suppose that the string (`+1, x`−1) is singular. Then it satisfies P
(i)
`+1(ν̃) = x`−1.
Since f̃i adds the box to the (`+ 1)-th column of ν(i), we have P
(i)
`+1(ν̃) = P
(i)
`+1(ν)− 2. Thus we
have P
(i)
`+1(ν) = P
(i)
`+1(ν̃) + 2 = x` + 1. We can reverse the arguments to obtain the if part.
(2) If x` ≥ xl, then f̃i will act on (l, xl) instead of (`, x`) since ` < l. Thus we have x` < xl.
Then we have P
(i)
`+1(ν) = x` + 1 ≤ xl. �
32 R. Sakamoto
5.2 Proof for Case A
The proof of this case depends on the following fundamental properties.
Proposition 5.3. If ` < `(i−1) − 1, we have
P
(i)
`(i)−1
(ν̄) > x`, P
(i)
`(i)−1(ν̄) > x`.
Proof. The proof proceeds by case by case analysis according to the classification (I) to (VII) in
Lemma 5.1. We remark that the condition ` < `(i−1)−1 automatically implies that ` < `(i+1)−1.
Case (I). We have to show P
(i)
`(i)−1
(ν) > x` by Lemma 5.1. Suppose if possible that
P
(i)
`(i)−1
(ν) ≤ x`. We have P
(i)
`(i)
(ν) ≥ x`(i) > x` since the string (`(i), x`(i)) is not selected by f̃i
whereas ` < `(i). Let (j, xj) be the longest string of ν(i) that is strictly shorter than `(i). Since
m
(i)
` (ν) > 0 we have ` ≤ j. If j = `(i) − 1, we have ` < j (since ` < `(i−1) − 1 ≤ `(i) − 1)
and xj ≤ P
(i)
j (ν) ≤ x`, which contradict the definition of (`, x`). If j < `(i) − 1, combining
P
(i)
`(i)
(ν) > x` with the assumption P
(i)
`(i)−1
(ν) ≤ x`, we have P
(i)
`(i)−1
(ν) < P
(i)
`(i)
(ν). Then by the
convexity of P
(i)
k (ν) between j ≤ k ≤ `(i), we have xj ≤ P
(i)
j (ν) < P
(i)
`(i)−1
(ν) ≤ x`, which
contradicts the minimality of x`. Thus we have proved P
(i)
`(i)−1
(ν) > x`.
Case (II). We have to show P
(i)
`(i)−1
(ν) > x` + 1 by Lemma 5.1. Suppose if possible that we
have P
(i)
`(i)−1
(ν) ≤ x` + 1. Recall that we have x` < x`(i) ≤ P
(i)
`(i)
(ν) by the minimality of x` and
` < `(i). Thus we have
P
(i)
`(i)−1
(ν) ≤ P (i)
`(i)
(ν). (5.1)
Let j be the length of the longest string of ν(i) under the condition j < `(i). Since m
(i)
` (ν) > 0
we have ` ≤ j.
Let us show that j < `(i−1). Suppose if possible that `(i−1) ≤ j. In particular, we have ` < j
by ` < `(i−1) − 1. Then we shall show that P
(i)
j (ν) ≤ x` + 1.
(i) The case j = `(i) − 1. Then P
(i)
j (ν) ≤ x` + 1 follows from the assumption.
(ii) The case j < `(i) − 1. We can apply the convexity relation of P
(i)
k (ν) between j ≤ k ≤ `(i)
with (5.1) to deduce that P
(i)
j (ν) ≤ P (i)
`(i)−1
(ν) ≤ x` + 1.
On the other hand, since ` < j, we have the condition x` < xj ≤ P
(i)
j (ν). Combining this with
the above result P
(i)
j (ν) ≤ x` + 1 we deduce that P
(i)
j (ν) = x` + 1 and thus xj = P
(i)
j (ν). But
then the string (j, xj) is the singular string whose length satisfies `(i−1) ≤ j < `(i), which is in
contradiction to the definition of `(i). Hence we have proved j < `(i−1).
Then we can apply the convexity relation of P
(i)
k (ν) between `(i−1) ≤ k ≤ `(i) with (5.1)
to deduce that P
(i)
`(i−1)(ν) ≤ P
(i)
`(i)−1
(ν). Since m
(i−1)
`(i−1)(ν) > 0 the convexity relation of P
(i)
k (ν)
between `(i−1) − 1 ≤ k ≤ `(i−1) + 1 is strict. Thus we have
P
(i)
j (ν) < P
(i)
`(i−1)(ν) ≤ P (i)
`(i)−1
(ν) ≤ x` + 1,
in particular, xj ≤ P
(i)
j (ν) ≤ x`. If ` < j, this contradicts the requirement x` < xj . Hence
we are left with the case j = `. Since we are assuming ` < `(i−1) − 1 we can refine the above
relation as
P
(i)
` (ν) < P
(i)
`(i−1)−1
(ν) < P
(i)
`(i−1)(ν) ≤ P (i)
`(i)−1
(ν) ≤ x` + 1,
Rigged Configurations and Kashiwara Operators 33
that is, P
(i)
` (ν) < x`. But this contradicts the requirement that the string (`, x`) must satisfy
x` ≤ P
(i)
` (ν). Thus we have proved P
(i)
`(i)−1
(ν) > x` + 1.
Case (III). We have to show P
(i)
`(i)−1(ν) > x` by Lemma 5.1. Since `(i) = `(i) we can use the
same arguments of (I) to obtain the result.
Case (IV). We have to show P
(i)
`(i)−1(ν) > x` + 1 by Lemma 5.1. Since `(i) = `(i) we can use
the same arguments of (II) to obtain the result.
Case (V). We have to show P
(i)
`(i)−1(ν) ≥ x` by Lemma 5.1. Suppose if possible that
P
(i)
`(i)−1(ν) < x`. Let j be the length of the longest string of ν(i) under the condition j < `(i).
Since m
(i)
`(i)
(ν) > 0 we have `(i) ≤ j. If j = `(i) − 1 we have xj ≤ P
(i)
j (ν) < x` which is in
contradiction to the minimality of x`. Thus assume that j < `(i) − 1. Then by the convexity
relation of P
(i)
k (ν) between j ≤ k ≤ `(i) we have P
(i)
`(i)−1(ν) ≥ min
{
P
(i)
j (ν), P
(i)
`(i)
(ν)
}
. However
this implies that x` > xj or x` > x`(i) both of which are in contradiction to the minimality of x`.
Thus we conclude that P
(i)
`(i)−1(ν) ≥ x`.
Case (VI). We have to show P
(i)
`(i)−1(ν) > x` by Lemma 5.1. Suppose if possible that
P
(i)
`(i)−1(ν) ≤ x`. Let j be the length of the longest string of ν(i) under the condition j < `(i).
Since m
(i)
`(i)
(ν) > 0 we have `(i) ≤ j. In particular, we have ` < j by ` < `(i−1)− 1. If j = `(i)− 1
we have xj ≤ x` which contradicts the relation xj > x` required by ` < j. So suppose that
j < `(i) − 1. Again we have P
(i)
`(i)−1(ν) ≥ min
{
P
(i)
j (ν), P
(i)
`(i)
(ν)
}
. Then we have x` ≥ xj or
x` ≥ x`(i) . This is a contradiction since ` < j and ` < `(i).
Case (VII). We have to show P
(i)
`(i)−1(ν) > x` + 1 by Lemma 5.1. We can prove this
by the same arguments of the case (II) if we replace `(i) and `(i−1) there by `(i) and `(i+1)
respectively. �
Proposition 5.4. If ` < `(i−1) − 1, we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) Step 1. Let us consider the case `(i−1) =∞. Then δ will not choose any string from
(ν, J)(i−1) so that f̃i chooses the same string before and after δ. Next, recall that f̃i does not
change coriggings of untouched strings. Then we have ˜̀(a) = `(a) for all a ≤ i− 1. In particular,
we have ˜̀(i−1) =∞. Thus we have ˜̀(a) = `(a) =∞ for all a ≥ i− 1.
Step 2. Based on the above observation, let us assume that `(i−1) < ∞ in the sequel. Let
us show that δ chooses the same strings before and after f̃i. Again we have ˜̀(a) = `(a) for
all a ≤ i − 1 since f̃i does not change coriggings of untouched strings. Recall that we have
`+ 1 < `(i−1) = ˜̀(i−1) by the assumption. Thus δ cannot choose the string (`+ 1, x`−1) created
by f̃i. Thus we have ˜̀(i) = `(i) which implies that ˜̀(a) = `(a) for all a ≥ i and ˜̀(a) = `(a) for
all a.
Step 3. Let us show that f̃i chooses the same string before and after δ. Recall that δ creates
the strings
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
of (ν̄, J̄)(i). Note that the string (`, x`)
remains as it is in (ν̄, J̄)(i) since δ acting on (ν, J) does not touch the string by the assumption
` < `(i−1). By Proposition 5.3 we have P
(i)
`(i)−1
(ν̄) > x` and P
(i)
`(i)−1(ν̄) > x`. Therefore f̃i acts on
the same string (`, x`) before and after the application of δ.
To summarize, we have ¯̃ν = ˜̄ν.
(2) It is enough to consider the strings
(
`(a), P
(a)
`(a)
(ν)
)
of (ν, J)(a) for a = i − 1, i, i + 1,
since f̃i adds a box to the same place before and after δ and since δ does not change riggings
34 R. Sakamoto
for untouched strings. For example, the string for a = i− 1 case behaves as follows:
(
`(i−1), P
(i−1)
`(i−1)(ν)
) f̃i7−→
(
`(i−1), P
(i−1)
`(i−1)(ν̃)
) δ7−→
(
`(i−1) − 1, P
(i−1)
`(i−1)−1
(¯̃ν)
)
,(
`(i−1), P
(i−1)
`(i−1)(ν)
) δ7−→
(
`(i−1) − 1, P
(i)
`(i−1)−1
(ν̄)
) f̃i7−→
(
`(i−1) − 1, P
(i)
`(i−1)−1
(˜̄ν)
)
.
In all cases, the riggings are determined so that f̃i and δ create the singular strings. Since we
have ¯̃ν = ˜̄ν, we obtain P
(i−1)
`(i−1)−1
(¯̃ν) = P
(i)
`(i−1)−1
(˜̄ν). The other cases are similar. Thus we conclude
that
¯̃
J = ˜̄J . �
5.3 Proof for Case B
Let us show the commutativity of f̃i and δ for the Case B, that is, `(i−1) ≤ ` + 1 < `(i) = ∞.
We divide the proof into the two cases.
5.3.1 Case 1: f̃i creates a non-singular string
Proposition 5.5. Suppose that `(i−1) ≤ ` + 1 < `(i) = ∞. If f̃i acting on (ν, J) creates
a non-singular string, then we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) Since f̃i does not change coriggings, δ chooses the same strings in (ν, J)(a) and
(ν̃, J̃)(a) for all a < i. Also, since f̃i creates a non-singular string, δ does not choose a string
from (ν̃, J̃)(i), which coincides with the action of δ on (ν, J)(i). Next, since δ does not touch
(ν, J)(i), we have J̄ (i) = J (i). Thus f̃i chooses the same string in both (ν, J)(i) and (ν̄, J̄)(i). To
summarize, δ and f̃i choose the same strings in δ ◦ f̃i and f̃i ◦ δ which implies ¯̃ν = ˜̄ν.
(2) We only need to consider J (i−1) and J (i) in this case. If the string is not touched by δ
and f̃i, the riggings after δ ◦ f̃i and f̃i ◦ δ coincide since δ and f̃i choose the same strings in
both cases respectively. The string
(
`(i−1), P
(i−1)
`(i−1)(ν)
)
becomes
(
`(i−1) − 1, P
(i−1)
`(i−1)−1
(¯̃ν)
)
and(
`(i−1)− 1, P
(i−1)
`(i−1)−1
(˜̄ν)
)
after application of δ ◦ f̃i and f̃i ◦ δ respectively. Since ¯̃ν = ˜̄ν we see the
coincidence of the riggings. As for the string (`, x`), we see that both δ ◦ f̃i and f̃i ◦ δ give the
string (`+ 1, x` − 1). Hence we have
¯̃
J = ˜̄J . �
5.3.2 Case 2: f̃i creates a singular string
This situation does not satisfy the condition of Proposition 4.7 due to the following fact.
Proposition 5.6. Suppose that `(i−1) ≤ `+1 < `(i) =∞. If f̃i acting on (ν, J) creates a singular
string, then we have
f̃i ◦ δ(ν, J) = 0.
Proof. Since `(i) = ∞, we have (ν, J)(i) = (ν̄, J̄)(i) and thus ¯̀ = `. Then the string (`, x`) in
(ν, J)(i) will become (`+1, x`−1) in f̃i◦δ(ν, J). Let us compute P
(i)
`+1(˜̄ν). From Lemma 5.2(1), we
have P
(i)
`+1(ν) = x`+1. By the assumption `(i−1)≤`+1 < `(i) we have P
(i)
`+1(ν̄) = P
(i)
`+1(ν)−1 = x`.
Therefore we have P
(i)
`+1(˜̄ν) = P
(i)
`+1(ν̄)−2 = x`−2. But then the string (`+1, x`−1) in f̃i◦δ(ν, J)
satisfies P
(i)
`+1(˜̄ν) < x`−1 which implies that f̃i ◦δ(ν, J) is not a valid rigged configuration. Thus
we have f̃i ◦ δ(ν, J) = 0. �
Rigged Configurations and Kashiwara Operators 35
Example 5.7. Consider the following rigged configuration (ν, J) of type (B1,1)⊗4⊗B1,3⊗B2,1⊗
B2,2 ⊗B3,1 of D
(1)
5
1
1
1
1
0
0
0
0
1
−1
1
1
1
0
0
0
0
1
1
−1
−1
−1
0
0
0
0
1
2
−1
−1
0
0
0
0
0
0
−1
0
0
−1
2
0
2
0
The corresponding tensor product Φ(ν, J) is
2 ⊗ 1̄ ⊗ 1 ⊗ 3 ⊗ 1 1 2 ⊗ 4
4̄
⊗ 1 3
5̄ 1̄
⊗
4
5
4̄
f̃2 acts on the string (4,−1) of (ν, J)(2) and makes it into (5,−2) of (ν̃, J̃)(2). Since we have
P
(2)
5 (ν̃) = −2, the latter string is singular. In particular, we see that f̃2(ν, J) 6= 0. Now (ν̄, J̄) is
0
0
0
0
1
0
0
0
0
−1
1
1
1
−1
−1
−1
0
1
1
−1
−1
−1
0
0
0
0
1
2
−1
−1
0
0
0
0
0
0
−1
0
0
−1
2
0
2
0
Thus we have `(1) = 2 < ` + 1 = 5 < `(2) = ∞ (see Proposition 5.6). Indeed, we have
P
(2)
5 (˜̄ν) = −3 and thus f̃2(ν̄, J̄) = 0.
5.4 Proof for Case C: preliminary steps
5.4.1 Classif ication
Recall that the defining condition of the Case C is `(i−1) ≤ ` + 1 ≤ `(i) < ∞. We divide the
proof into the three cases according to the following extra conditions:
(1) m
(i)
`+1(ν) > 0 and f̃i creates a singular string, or
`+ 1 = `(i) and f̃i creates a non-singular string,
(2) m
(i)
`+1(ν) = 0 and f̃i creates a singular string,
(3) `+ 1 < `(i) and f̃i creates a non-singular string.
We give several remarks on the classification. In case (2), the condition m
(i)
`+1(ν) = 0 automa-
tically implies that ` + 1 < `(i) by the assumption ` + 1 ≤ `(i). On the other hand, in the first
case of case (1), we will show that ` + 1 = `(i). We remark that we may have m
(i)
`+1(ν) > 0 in
case (3).
5.4.2 A common property for Case C
In this case, we can show the following property. Note that the corresponding analysis for the
case `(i) = `(i) is included in the analysis of P
(i)
`(i)−1
(ν̄) which will be given later in this section.
Proposition 5.8. Suppose that `+ 1 ≤ `(i) < `(i) <∞. Then we have
P
(i)
`(i)−1(ν̄) > x`.
36 R. Sakamoto
Proof. We follow the classification in Lemma 5.1. Since `(i) < `(i), we have to consider
Cases (V), (VI) and (VII). Let j be the largest integer satisfying j < `(i) and m
(i)
j (ν) > 0.
From `(i) < `(i), we see that `(i) ≤ j and in particular we have ` < j. In fact we only need the
latter relation ` < j in the following proof4.
Case (V). According to Lemma 5.1 we have to show P
(i)
`(i)−1(ν) ≥ x` under the condition
`(i) < `(i+1) = · · · = `(i). Suppose if possible that P
(i)
`(i)−1(ν) < x`. If j = `(i) − 1, then
we have x`(i)−1 ≤ P
(i)
`(i)−1(ν) < x` which contradicts the minimality of x`. So suppose that
j < `(i) − 1. By the assumption we see that ` < `(i). Then we have P
(i)
`(i)
(ν) ≥ x`(i) > x`.
Thus we have P
(i)
`(i)−1(ν) < P
(i)
`(i)
(ν). By the convexity of P
(i)
k (ν) between j ≤ k ≤ `(i), we have
P
(i)
j (ν) < P
(i)
`(i)−1(ν) < x`. Since xj ≤ P
(i)
j (ν), this contradicts the minimality of x`. Thus we
conclude that P
(i)
`(i)−1(ν) ≥ x`.
Case (VI). According to Lemma 5.1 we have to show P
(i)
`(i)−1(ν) > x` under the condition
`(i+1) < `(i+1) = `(i). Suppose if possible that P
(i)
`(i)−1(ν) ≤ x`. If j = `(i) − 1, then we have
xj ≤ P
(i)
j (ν) ≤ x`. However this relation contradicts the requirement xj > x` which follows
from ` < j. Next, suppose that j < `(i) − 1. Then, as in the Case (V), we have P
(i)
`(i)
(ν) > x` by
` < `(i). Thus we have P
(i)
`(i)−1(ν) < P
(i)
`(i)
(ν). By the similar arguments in the previous case, we
can deduce the contradiction. Therefore we obtain P
(i)
`(i)−1(ν) > x` in this case.
Case (VII). According to Lemma 5.1 we have to show P
(i)
`(i)−1(ν) > x`+1 under the condition
`(i+1) < `(i). Suppose if possible that P
(i)
`(i)−1(ν) ≤ x` + 1. If j = `(i) − 1, then we have
P
(i)
j (ν) ≥ xj > x` by j > `. Then the only possibility that is compatible with the assumption
is P
(i)
j (ν) = xj = x` + 1. In particular, the string (j, xj) is singular whose length satisfies
`(i+1) ≤ j < `(i). However this is in contradiction to the definition of `(i).
If j < `(i) − 1, again we have P
(i)
`(i)
(ν) > x` as in the Case (V). Similarly, by ` < j we have
P
(i)
j (ν) > x`. Then by the convexity of P
(i)
k (ν) between j ≤ k ≤ `(i), the only possibility that is
compatible with P
(i)
`(i)−1(ν) ≤ x` + 1 is P
(i)
j (ν) = · · · = P
(i)
`(i)
(ν) = x` + 1. Since P
(i)
j (ν) ≥ xj > x`
it has to be xj = x` + 1, in particular, the string (j, xj) is singular. If `(i+1) ≤ j, this contradicts
the definition of `(i). On the other hand, if j < `(i+1), m
(i+1)
`(i+1)
(ν) > 0 means that the convexity
of P
(i)
k (ν) between `(i+1) − 1 ≤ k ≤ `(i+1) + 1 is strict. This is in contradiction to the relation
P
(i)
j (ν) = · · · = P
(i)
`(i)
(ν) = x` + 1. �
5.5 Proof for Case C (1)
Proposition 5.9. Suppose that `(i−1) ≤ `+1 ≤ `(i) <∞, m
(i)
`+1(ν) > 0 and f̃i creates a singular
string. Then we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) To begin with let us show that ` + 1 = `(i). By the assumption m
(i)
`+1(ν) > 0 there
exists the string (`+ 1, x`+1) of (ν, J)(i). We show that the string (`+ 1, x`+1) is singular. For
this we derive the following three relations:
4We will use this property in the proof of Proposition 5.24.
Rigged Configurations and Kashiwara Operators 37
• P (i)
`+1(ν) = x` + 1 by Lemma 5.2,
• x`+ 1 ≤ x`+1, that is, x` < x`+1, since the string (`+ 1, x`+1) of (ν, J) is not selected by f̃i
although it is longer than `,
• x`+1 ≤ P
(i)
`+1(ν) by definition of the rigged configurations.
Combining these three inequalities, we obtain P
(i)
`+1(ν) = x`+1 and thus the string (` + 1, x`+1)
is singular. Suppose if possible that `+ 1 < `(i). Then the string (`+ 1, x`+1) is singular whose
length satisfies `(i−1) ≤ ` + 1, contradicting the definition of `(i). Thus ` + 1 = `(i). Then we
can show the commutativity of f̃i and δ as follows.
Let us consider the case δ ◦ f̃i. It is enough to analyze the strings (`, x`) and
(
`+ 1, P
(i)
`+1(ν)
)
in (ν, J)(i):
(`, x`)(
`+ 1, P
(i)
`+1(ν)
) f̃i7−→
(`+ 1, x` − 1)(
`+ 1, P
(i)
`+1(ν̃)
) δ7−→
(`+ 1, x` − 1)(
`, P
(i)
` (¯̃ν)
) .
Recall that f̃i does not change coriggings other than the touched string. In particular, we have˜̀(i−1) = `(i−1). Thus the next δ can act on one of the length `+ 1 singular strings.
Let us consider the case f̃i ◦ δ. Recall that if `(i) < `(i) we have P
(i)
`(i)−1(ν̄) > x` by Proposi-
tion 5.8. Therefore f̃i will not act on the length `(i) − 1 string of (ν̄, J̄)(i) that has been created
by δ. On the other hand, if `(i) = `(i) we do not need to take care of this issue. Thus it is enough
to analyze the behaviors of the lengths ` and `+ 1 strings:
(`, x`)(
`+ 1, P
(i)
`+1(ν)
) δ7−→
(`, x`)(
`, P
(i)
` (ν̄)
) f̃i7−→
(`+ 1, x` − 1)(
`, P
(i)
` (˜̄ν)
) .
Recall that ` + 1 = `(i) so that δ acts on the string
(
` + 1, P
(i)
`+1(ν)
)
. Also, since (ν̄, J̄) is the
valid rigged configuration, we have the requirement P
(i)
` (ν̄) ≥ x`. Thus the string (`, x`) has the
largest length among the strings of the smallest rigging within (ν̄, J̄)(i).
To summarize, we are left with lengths ` and `+ 1 strings in both δ ◦ f̃i and f̃i ◦ δ. Thus we
have ¯̃ν = ˜̄ν.
(2) We have P
(i)
` (¯̃ν) = P
(i)
` (˜̄ν) by ¯̃ν = ˜̄ν. Thus
¯̃
J = ˜̄J . �
Proposition 5.10. Suppose that we have ` + 1 = `(i) and f̃i acting on (ν, J) creates the non-
singular string. Then we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. Since `+ 1 = `(i) there is a singular string
(
`+ 1, P
(i)
`+1(ν)
)
in (ν, J)(i). Then this string
and (`, x`) behave just as in the proof of the previous proposition which confirms the claim. �
5.6 Proof for Case C (2)
5.6.1 Outline
As the second step, we treat the case when m
(i)
`+1(ν) = 0 and f̃i creates a singular string. In this
case, we will show the following assertion by case by case analysis.
Proposition 5.11. Suppose that `(i−1) ≤ ` + 1 < `(i) < ∞ and m
(i)
`+1(ν) = 0. If f̃i acting
on (ν, J) creates a singular string, then the following two conditions are satisfied:
P
(i)
`(i)−1
(ν̄) = x`, (5.2)
m
(i+1)
k (ν) = 0 for ` < k < `(i). (5.3)
38 R. Sakamoto
Note that we have `+1 6= `(i) by m
(i)
`+1(ν) = 0. To begin with, we explain how these properties
provide the proof of ¯̃ν = ˜̄ν.
Proposition 5.12. Suppose that `(i−1) ≤ `+ 1 < `(i) <∞, m
(i)
`+1(ν) = 0 and f̃i acting on (ν, J)
creates a singular string. Then we have ¯̃ν = ˜̄ν.
Proof. Let us analyze f̃i ◦ δ and δ ◦ f̃i respectively.
Step 1. Let us consider the case f̃i ◦ δ. Let us consider the strings
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
of (ν̄, J̄)(i) created by δ. By (5.2) we see that the rigging of the string(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
of (ν̄, J̄) is x`. On the other hand, if `(i) < `(i) we have P
(i)
`(i)−1(ν̄) > x` by
Proposition 5.8. Since we are assuming that ` < `(i) − 1, the string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
has the
largest length among the strings of smallest riggings within (ν̄, J̄)(i). Thus f̃i adds a box to the
length `(i) − 1 string that has been shortened by δ.
Step 2. Let us consider the case δ◦f̃i. Recall that f̃i does not change coriggings of untouched
strings. By the assumption f̃i creates the singular string (` + 1, x` − 1) whose length satisfies
`(i−1) ≤ ` + 1 < `(i). Hence δ choose this string and obtain ˜̀(i) = ` + 1. By definition of δ we
have ˜̀(i) = ` + 1 ≤ ˜̀(i+1) and by (5.3) we conclude that `(i) ≤ ˜̀(i+1). Thus δ choose a string
of length `(i+1) from (ν̃, J̃)(i+1). Therefore we obtain ˜̀(a) = `(a) for all a 6= i and ˜̀(a) = `(a) for
all a.
Step 3. To summarize, we have ¯̃ν
(a)
= ˜̄ν(a)
= ν̄(a) for all a 6= i and ¯̃ν
(i)
= ˜̄ν(i)
is obtained by
removing a box from length `(i) string of ν(i). �
The remaining property
¯̃
J = ˜̄J will be checked from case by case analysis and it will be done
along with the proof of Proposition 5.11.
5.6.2 Classif ication
We divide the proofs of Proposition 5.11 as well as the identity
¯̃
J = ˜̄J into the two cases
according to the following classification.
Lemma 5.13. Suppose that `(i−1) ≤ `+ 1 < `(i) <∞. Then there are only two possibilities as
follows:
(`, x`) is singular and ` = `(i−1) − 1, (5.4)
(`, x`) is non-singular. (5.5)
Proof. Suppose that the string (`, x`) is singular. If its length satisfies `(i−1) ≤ ` < `(i), then
it is in contradiction to the definition of `(i). Therefore we must have ` = `(i−1) − 1. �
5.6.3 Proof for the case (5.4)
Under the present assumptions, let us show the following property.
Proposition 5.14. Suppose that we have `(i−1) = ` + 1 < `(i) < ∞, m
(i)
`+1(ν) = 0, the string
(`, x`) is singular and f̃i creates a singular string. Then we have
P
(i)
` (ν) = x`, P
(i)
`+1(ν) = P
(i)
`+2(ν) = · · · = P
(i)
`(i)
(ν) = x` + 1. (5.6)
Rigged Configurations and Kashiwara Operators 39
Proof. Since f̃i creates a singular string, we have
P
(i)
`+1(ν)
by Lemma 5.2(1)
= x` + 1
since the string (`, x`) is singular
= P
(i)
` (ν) + 1. (5.7)
Due to the existence of the length `(i−1) = `+1 string at ν(i−1) and the assumption m
(i)
`+1(ν) = 0,
the vacancy number P
(i)
k (ν) is strictly convex between the interval ` ≤ k ≤ ` + 2. Then the
relation (5.7) implies P
(i)
`+1(ν) ≥ P (i)
`+2(ν). Let us show that this inequality is in fact equality. For
this let j be the length of the smallest string of ν(i) that is longer than `. Since m
(i)
`+1(ν) = 0 we
have `(i−1) = `+ 1 < j ≤ `(i). By Lemma 5.2(2) we have
P
(i)
`+1(ν) ≤ xj ≤ P (i)
j (ν). (5.8)
Then by the convexity of P
(i)
k (ν) between ` ≤ k ≤ j, the relations P
(i)
`+1(ν) ≥ P
(i)
`+2(ν) and
P
(i)
`+1(ν) ≤ P
(i)
j (ν) imply the relation P
(i)
`+1(ν) = P
(i)
`+2(ν) = · · · = P
(i)
j (ν). Combining this
equality with (5.8) we obtain xj = P
(i)
j (ν), thus the string (j, xj) is singular. By the relation
`(i−1) < j ≤ `(i) we deduce that j = `(i). �
Corollary 5.15. Under the same assumptions of Proposition 5.14, we have
(1) m
(i−1)
`+1 (ν) = 1,
(2) m
(i−1)
k (ν) = 0 for `+ 1 < k < `(i),
(3) m
(i)
k (ν) = 0 for `+ 1 ≤ k < `(i),
(4) m
(i+1)
k (ν) = 0 for `+ 1 ≤ k < `(i).
Proof. The assertions for m
(i)
k (ν) given in (3) are proved at the end of the proof of Proposi-
tion 5.14.
Let us show the assertions for m
(i+1)
k (ν). If m
(i+1)
k (ν) > 0 for some ` + 1 < k < `(i),
the relation (5.6) must be strictly convex around such k, which is a contradiction. Therefore
m
(i+1)
k (ν) = 0 for all ` + 1 < k < `(i). By using (5.6) and the assumption m
(i)
`+1(ν) = 0, the
estimate of Lemma 3.13 with l = `+ 1 gives
−P (i)
` (ν) + 2P
(i)
`+1(ν)− P (i)
`+2(ν) = 1
≥ m(i−1)
`+1 (ν)− 2m
(i)
`+1(ν) +m
(i+1)
`+1 (ν) = m
(i−1)
`+1 (ν) +m
(i+1)
`+1 (ν).
Recall that we have m
(i−1)
`+1 (ν) > 0 by definition of `(i−1)(= ` + 1). Then the only possibility
that is compatible with the above inequality is m
(i−1)
`+1 (ν) = 1 and m
(i+1)
`+1 (ν) = 0. Hence we have
confirmed the relations (4) and (1).
Finally, since m
(i)
k (ν) = 0 and m
(i+1)
k (ν) = 0 for ` + 1 < k < `(i), the relation (5.6) implies
that m
(i−1)
k (ν) = 0 for ` + 1 < k < `(i), since otherwise the relation would be strictly convex.
Thus we have confirmed the relation (2). �
Proof of Proposition 5.11 for the case (5.4). The assertions m
(i+1)
k (ν) = 0 for ` < k < `(i)
are already proved in Corollary 5.15.
From (5.6) we have P
(i)
`(i)−1
(ν) = x`+1 since we have ` < `(i)−1 by the assumption (`(i−1) =)
`+ 1 < `(i). Since `(i−1) ≤ `(i) − 1 we have P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)−1
(ν)− 1 = x`. This completes the
proof of Proposition 5.11 for the case (5.4). �
40 R. Sakamoto
Thus we have confirmed ¯̃ν = ˜̄ν in this case. In order to prove
¯̃
J = ˜̄J , we need additional
relations obtained in Corollary 5.15.
Proposition 5.16. Suppose that we have `(i−1) = ` + 1 < `(i) < ∞, m
(i)
`+1(ν) = 0, the string
(`, x`) is singular and f̃i creates a singular string. Then we have
¯̃
J = ˜̄J .
Proof. From Corollary 5.15 and Proposition 5.12 we see that there are no strings of length
from ` + 1 to `(i) − 1 that are not touched by δ nor f̃i during both δ ◦ f̃i and f̃i ◦ δ. Thus it is
enough to analyze the strings (`, x`) and (`(i), x`(i)) of (ν, J)(i).
Let us analyze the behavior of the string (`, x`). From the proof of Proposition 5.12 we have
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→ (`, x`)
f̃i7−→ (`, x`).
By ` < `(i−1) we have P
(i)
` (¯̃ν) = P
(i)
` (ν). Recall that we have P
(i)
` (ν) = x` by (5.6). Hence we
obtain P
(i)
` (¯̃ν) = x`.
Next let us analyze the behavior of the string (`(i), x`(i)). From the proof of Proposition 5.12
we have
(`(i), x`(i))
f̃i7−→ (`(i), x`(i) − 2)
δ7−→ (`(i), x`(i) − 2),
(`(i), x`(i))
δ7−→
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
) f̃i7−→
(
`(i), P
(i)
`(i)−1
(ν̄)− 1
)
.
Recall that we have ` + 1 < `(i). Thus the first f̃i of δ ◦ f̃i decreases the rigging of the string
(`(i), x`(i)) by 2.
Let us check the coincidence of the final rigging of this case. Recall that we have x`(i) =
P
(i)
`(i)
(ν) since the string (`(i), x`(i)) is singular by definition of `(i). We also have P
(i)
`(i)
(ν) = x`+ 1
by (5.6). Therefore we have x`(i) − 2 = x` − 1.
On the other hand, since `(i−1) ≤ `(i) − 1 we have P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)−1
(ν)− 1. From (5.6) we
have P
(i)
`(i)−1
(ν) = x` + 1. Therefore we have P
(i)
`(i)−1
(ν̄) − 1 = x` − 1. Hence we conclude that
x`(i) − 2 = P
(i)
`(i)−1
(ν̄)− 1. �
Example 5.17. Consider the following rigged configuration (ν, J) of type (B1,1)⊗3 ⊗ B1,3 ⊗
B2,1 ⊗B2,2 ⊗B3,1 of D
(1)
5
1
1
1
0
0
0
1
0
1
−1
−1
0
0
1
−1
−1
0
0
0
0
2
2
2
−1
0
2
2
2
−2
−2
−2
−2
−2
−3
−2
−3
The corresponding tensor product Φ(ν, J) is
1̄ ⊗ 1̄ ⊗ 2̄ ⊗ 5̄ 3̄ 1̄ ⊗ 1
4̄
⊗ 1 3
2 4̄
⊗
1
4
4̄
We see that f̃2 acts on the string (`, x`) = (2,−1) of (ν, J)(2). Then (ν̄, J̄) is
Rigged Configurations and Kashiwara Operators 41
0
0
1
1
0
0
1
1
1
−1
−1
−1
−1
1
−1
−1
−1
−1
0
0
2
2
2
−1
0
2
2
2
−1
−2
−1
−2
−1
−3
−1
−3
and (ν̃, J̃) is
1
1
2
1
0
0
2
1
1
−1
−2
−2
−2
1
−1
−2
−2
−2
0
0
3
3
3
−1
0
3
3
3
−2
−2
−2
−2
−2
−3
−2
−3
Note that we have `(1) = 3 = `+ 1 < `(2) = 4 and m
(2)
3 (ν) = 0. Moreover, we see that the string
(`, x`) is singular since we have P
(2)
2 = −1, and f̃2 makes it into the singular string (3,−2) of
(ν̃, J̃)(2) since we have P
(2)
3 (ν̃) = −2. Hence this is the example for the present case. Then we
have P
(2)
2 (ν) = −1 and P
(2)
3 (ν) = P
(2)
4 (ν) = 0 where `(2) = 4 (see Proposition 5.14). We have
m
(1)
3 (ν) = 1 and m
(2)
3 (ν) = m
(3)
3 (ν) = 0 (see Corollary 5.15). Finally (¯̃ν,
¯̃
J) = (˜̄ν, ˜̄J) is
0
0
1
1
0
0
1
1
1
−1
−1
−1
−2
1
−1
−1
−1
−2
0
0
2
2
3
−1
0
2
2
3
−1
−2
−1
−2
−1
−3
−1
−3
5.6.4 Proof for the case (5.5)
Under the present assumptions, let us show the following property.
Proposition 5.18. Suppose that we have `(i−1) ≤ ` + 1 < `(i) < ∞, m
(i)
`+1(ν) = 0, the string
(`, x`) is non-singular and f̃i creates a singular string. Then we have
P
(i)
` (ν) = P
(i)
`+1(ν) = · · · = P
(i)
`(i)
(ν) = x` + 1. (5.9)
Proof. Let j be the length of the shortest string of ν(i) that satisfies ` < j. By the assumption
m
(i)
`+1(ν) = 0 we have `+ 1 < j ≤ `(i). From the convexity of P
(i)
k (ν) between ` ≤ k ≤ j we have
P
(i)
`+1(ν) ≥ min
{
P
(i)
` (ν), P
(i)
j (ν)
}
. (5.10)
In order to evaluate the minimum in (5.10), let us suppose if possible that P
(i)
` (ν) > P
(i)
j (ν).
Then the inequality (5.10) reads P
(i)
`+1(ν) ≥ P
(i)
j (ν). Recall from Lemma 5.2(2) that we have
P
(i)
`+1(ν) ≤ xj and by definition of the rigged configuration we have xj ≤ P
(i)
j (ν), that is,
P
(i)
`+1(ν) ≤ P
(i)
j (ν). Combining the two inequalities we obtain P
(i)
`+1(ν) = P
(i)
j (ν). By the
assumption, we conclude that P
(i)
` (ν) > P
(i)
`+1(ν) = P
(i)
j (ν) which is in contradiction to the
convexity relation of P
(i)
k (ν) between ` ≤ k ≤ j.
Thus we assume that P
(i)
` (ν) ≤ P (i)
j (ν). Then (5.10) gives P
(i)
` (ν) ≤ P (i)
`+1(ν). We also have
P
(i)
`+1(ν)
by Lemma 5.2(1)
= x` + 1
since the string (`, x`) is non-singular
≤ P
(i)
` (ν).
42 R. Sakamoto
Thus we have P
(i)
` (ν) = P
(i)
`+1(ν). Combining this with the inequality P
(i)
` (ν) ≤ P
(i)
j (ν) and
using the convexity of P
(i)
k (ν) between ` ≤ k ≤ j, we obtain
P
(i)
` (ν) = P
(i)
`+1(ν) = · · · = P
(i)
j (ν) = x` + 1. (5.11)
Suppose if possible that j < `(i). Since ` < j we have P
(i)
`+1(ν) ≤ xj by Lemma 5.2(2). Then
we have P
(i)
`+1(ν) ≤ xj ≤ P
(i)
j (ν). By (5.11) we obtain xj = P
(i)
j (ν), that is, the string (j, xj) is
singular. Since `(i−1) ≤ j < `(i) this contradicts the definition of `(i). Hence we conclude that
j = `(i). �
Proof of Proposition 5.11 for the case (5.5). Since `(i−1) < `(i) we have
P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)−1
(ν)− 1 = x`,
where we have used (5.9). From the result at the end of the proof of the previous lemma, we
see that m
(i)
k (ν) = 0 for ` < k < `(i). Then the relation (5.9) implies that m
(i+1)
k (ν) = 0 for
` < k < `(i) since the existence of such a string would imply that the relation (5.9) have to be
strictly convex. Hence we complete the proof of Proposition 5.11. �
By using the relation ¯̃ν = ˜̄ν we show
¯̃
J = ˜̄J . For this purpose we note the following facts
which are the consequences of the proof of Proposition 5.18.
Corollary 5.19. Under the same assumptions with the previous proposition, we have the fol-
lowing relations.
(1) m
(a)
k (ν) = 0 for all ` < k < `(i) and a = i− 1, i, i+ 1,
(2) `(i−1) ≤ `.
Proof. (1) From m
(i)
k (ν) = m
(i+1)
k (ν) = 0 for ` < k < `(i), we see that if m
(i−1)
k (ν) > 0 for some
` < k < `(i), then P
(i)
k (ν) becomes strictly convex around such the k, which is in contradiction
to the relation (5.9). Thus we obtain the assertion.
(2) If ` < `(i−1), we have m
(i−1)
`(i−1)(ν) > 0 which is in contradiction to the previous assertion
since `(i−1) < `(i). Thus we have `(i−1) ≤ `. �
Proposition 5.20. Suppose that we have `(i−1) ≤ ` + 1 < `(i) < ∞, m
(i)
`+1(ν) = 0, the string
(`, x`) is non-singular and f̃i creates a singular string. Then we have
¯̃
J = ˜̄J .
Proof. By Proposition 5.12 and Corollary 5.19(1) we see that it is enough to check the coinci-
dence of riggings for the strings (`, x`) and (`(i), x`(i)) that are touched by δ and f̃i.
Let us analyze the change of the string (`, x`). Since `+ 1 < `(i) we have
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→ (`, x`)
f̃i7−→ (`, x`).
Since `(i−1) ≤ ` by Corollary 5.19(2) we have P
(i)
` (¯̃ν) = P
(i)
` (ν) − 1. From (5.9), we have
P
(i)
` (ν) = x` + 1. Hence we obtain P
(i)
` (¯̃ν) = x`.
The analysis of the string (`(i), x`(i)) is the same with the corresponding part of the proof of
Proposition 5.16 if we replace (5.6) there by (5.9). �
Rigged Configurations and Kashiwara Operators 43
5.7 Proof for Case C (3)
As the final step of the proof for Case C, let us treat the case where `+ 1 < `(i) and f̃i creates
a non-singular string.
Proposition 5.21. Assume that `(i−1) ≤ ` + 1 < `(i) < ∞. If f̃i acting on (ν, J) creates
a non-singular string, then the following relation is satisfied:
P
(i)
`(i)−1
(ν̄) > x`. (5.12)
Proof. In this case we have
P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)−1
(ν)− 1 (5.13)
since `(i−1) ≤ `(i) − 1.
Throughout the proof of this proposition, let j be the largest integer such that j < `(i) and
m
(i)
j (ν) > 0. Since m
(i)
` (ν) > 0, we have ` ≤ j ≤ `(i) − 1. We divide the proof into four cases.
Case j = `(i) − 1. Let us denote the corresponding string as (`(i)−1, x`(i)−1). Since `(i−1) <
`(i) we have `(i−1) ≤ `(i) − 1 which implies that the string (`(i) − 1, x`(i)−1) is non-singular by
definition of `(i). Thus we have x`(i)−1 < P
(i)
`(i)−1
(ν). On the other hand, from the assumption
`+ 1 < `(i), we have x` < x`(i)−1 since f̃i chooses the string (`, x`) which is strictly shorter than
the string (`(i) − 1, x`(i)−1). Combining both inequalities we obtain x` < P
(i)
`(i)−1
(ν) − 1. Hence
we obtain x` < P
(i)
`(i)−1
(ν̄).
Case ` < j < `(i) − 1. In this case the string (j, xj) have to be non-singular since j satisfies
`(i−1) ≤ `+ 1 ≤ j < `(i). Then we have
P
(i)
j (ν)
since the string (j, xj) is non-singular
> xj
since f̃i acts on the string (`, x`) although we have j > `
> x`.
In particular we have P
(i)
j (ν) > x` + 1. Similarly let us consider the string
(
`(i), P
(i)
`(i)
(ν)
)
whose
existence is assured by the definition of `(i). Then its rigging must satisfy P
(i)
`(i)
(ν) > x` since its
length satisfies `(i) > `. From the convexity relation of P
(i)
k (ν) between j ≤ k ≤ `(i) we have
P
(i)
`(i)−1
(ν) ≥ min
{
P
(i)
j (ν), P
(i)
`(i)
(ν)
}
> x`.
Suppose if possible that P
(i)
`(i)−1
(ν) = x` + 1. Then we would have P
(i)
j (ν) > P
(i)
`(i)−1
(ν) ≤ P (i)
`(i)
(ν)
which is forbidden by the convexity of the vacancy numbers. Thus we have P
(i)
`(i)−1
(ν) > x` + 1.
By (5.13) we have P
(i)
`(i)−1
(ν̄) > x`.
Case j = ` and (`, x`) is non-singular. In this case we have P
(i)
` (ν) > x` since the string
(`, x`) is non-singular. Let us consider the string
(
`(i), P
(i)
`(i)
(ν)
)
. Then its rigging must satisfy
P
(i)
`(i)
(ν) > x` since its length satisfy `(i) > `. Now we invoke the convexity of P
(i)
k (ν) between
` = j ≤ k ≤ `(i) to conclude that P
(i)
`(i)−1
(ν) > x`.
Suppose if possible that P
(i)
`(i)−1
(ν) = x` + 1. Again by the convexity relation the only
possibility is P
(i)
` (ν) = · · · = P
(i)
`(i)
= x` + 1. In particular, we have P
(i)
`+1(ν) = x` + 1 and by
Lemma 5.2 (1) this implies that f̃i makes a singular string. This is in contradiction to the
assumption that f̃i makes a non-singular string. Thus we have P
(i)
`(i)−1
(ν) > x` + 1. By (5.13)
we have P
(i)
`(i)−1
(ν̄) > x`.
44 R. Sakamoto
Case j = ` and (`, x`) is singular. In this case we have P
(i)
` (ν) = x` since the string
(`, x`) is singular. Also the rigging for the string
(
`(i), P
(i)
`(i)
(ν)
)
must satisfy P
(i)
`(i)
(ν) > x` since
`(i) > `. Then from the convexity relation of P
(i)
k (ν) between ` ≤ k ≤ `(i) we have
P
(i)
`(i)−1
(ν) ≥ min
{
P
(i)
` (ν), P
(i)
`(i)
(ν)
}
= x`.
If P
(i)
`(i)−1
(ν) = x` we have P
(i)
` (ν) = P
(i)
`(i)−1
(ν) < P
(i)
`(i)
(ν) which is in contradiction to the
convexity relation since we have ` < `(i) − 1 by the assumption `+ 1 < `(i). Suppose if possible
that P
(i)
`(i)−1
(ν) = x` + 1. Since x` < P
(i)
`(i)
(ν) we have P
(i)
`(i)−1
(ν) ≤ P
(i)
`(i)
(ν). Then the only
possibility that is compatible with the convexity relation is x` = P
(i)
` (ν) < P
(i)
`+1(ν) = · · · =
P
(i)
`(i)
(ν) = x` + 1. In particular, we have P
(i)
`+1(ν) = x` + 1. However, by Lemma 5.2(1), this
implies that f̃i creates a singular string, which is the contradiction. Thus we conclude that
P
(i)
`(i)−1
(ν) > x` + 1. By (5.13) we have P
(i)
`(i)−1
(ν̄) > x`, which completes the whole proof of
Proposition 5.21. �
Proposition 5.22. Assume that `(i−1) ≤ ` + 1 < `(i) < ∞ and f̃i acting on (ν, J) creates
a non-singular string. Then we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) We show that f̃i acts on the same string (`, x`) in both (ν, J) and (ν̄, J̄). For δ creates
the strings
(
`(i)− 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i)− 1, P
(i)
`(i)−1(ν̄)
)
. Then by (5.12) we have P
(i)
`(i)−1
(ν̄) > x`
in this case. Recall also that we have P
(i)
`(i)−1(ν̄) > x` if `(i) < `(i) by Proposition 5.8. Therefore
the string (`, x`) remains as the string with the smallest rigging of the largest length in (ν̄, J̄).
Thus f̃i acts on the string (`, x`) of (ν̄, J̄).
Also, since f̃i creates the non-singular string and does not change other coriggings, δ chooses
the same strings for both (ν, J) and (ν̃, J̃). Thus we have ¯̃ν = ˜̄ν.
(2) Since f̃i acts on the string (`, x`) in both cases, we have to analyze the strings (`, x`),(
`(i), P
(i)
`(i)
(ν)
)
and
(
`(i), P
(i)
`(i)
(ν)
)
. The string (`, x`) behaves as follows:
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→ (`+ 1, x` − 1),
(`, x`)
δ7−→ (`, x`)
f̃i7−→ (`+ 1, x` − 1).
Hence the assertion follows. On the other hand, the string
(
`(i), P
(i)
`(i)
(ν)
)
behaves as follows:
(
`(i), P
(i)
`(i)
(ν)
) f̃i7−→
(
`(i), P
(i)
`(i)
(ν̃)
) δ7−→
(
`(i) − 1, P
(i)
`(i)−1
(¯̃ν)
)
,(
`(i), P
(i)
`(i)
(ν)
) δ7−→
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
) f̃i7−→
(
`(i) − 1, P
(i)
`(i)−1
(˜̄ν)
)
.
By ¯̃ν = ˜̄ν we obtain the desired fact. The analysis of
(
`(i), P
(i)
`(i)
(ν)
)
is similar. Hence we have
¯̃
J = ˜̄J . �
5.8 Proof for Case D: preliminary steps
5.8.1 Classif ication
Recall that the defining condition for Case D is `(i) = `. For the proof, it is convenient to further
divide the case into the following four cases:
Rigged Configurations and Kashiwara Operators 45
(1) ` = `(i) < `(i) and m
(i)
` (ν) > 1,
(2) ` = `(i) = `(i) and m
(i)
` (ν) > 2,
(3) ` = `(i) < `(i) and m
(i)
` (ν) = 1,
(4) ` = `(i) = `(i) and m
(i)
` (ν) = 2.
Note that `(i) = `(i) automatically requires m
(i)
`(i)
(ν) ≥ 2.
5.8.2 A common property for Case D
For the proof of Case D, we will need the following result.
Proposition 5.23. Assume that `(i) = `. Then the following relation holds:
P
(i)
`−1(ν̄) ≥ x`.
Proof. Throughout the proof of this proposition, let j be the largest integer such that j < `
and m
(i)
j (ν) > 0. If there is no such j, set j = 0. We divide the proof into the following four
cases:
(a) j = `− 1,
(b) j < `− 1, `(i−1) = `,
(c) j < `− 1, `(i−1) < ` and `(i−1) ≤ j,
(d) j < `− 1, `(i−1) < ` and j < `(i−1).
Case (a). Denote the corresponding string of (ν, J)(i) as (` − 1, x`−1). Then we have
x` ≤ x`−1 by definition of `. Suppose that `(i−1) < `. Since we have `(i−1) ≤ ` − 1 < `(i) the
string (` − 1, x`−1) cannot be singular. Thus we have P
(i)
`−1(ν) > x` by P
(i)
`−1(ν) > x`−1 ≥ x`.
Thus we obtain P
(i)
`−1(ν̄) ≥ x` since P
(i)
`−1(ν̄) = P
(i)
`−1(ν)− 1 by `(i−1) ≤ `− 1.
Next suppose that `(i−1) = `. In this case the string (`− 1, x`−1) can be singular. Therefore
we have P
(i)
`−1(ν) ≥ x` by P
(i)
`−1(ν) ≥ x`−1 ≥ x`. From `− 1 < `(i−1) we have P
(i)
`−1(ν̄) = P
(i)
`−1(ν).
Thus we obtain P
(i)
`−1(ν̄) ≥ x`.
Case (b). By j < ` we have P
(i)
j (ν) ≥ xj ≥ x`. Note that this relation is valid even if j = 0
since we have P
(i)
0 (ν) = 0 ≥ x` by ` = `(i) > 0. Recall that we also have P
(i)
` (ν) ≥ x`. Thus
by the convexity of P
(i)
k (ν) between j ≤ k ≤ ` we have P
(i)
`−1(ν) ≥ x`. Since ` − 1 < `(i−1) we
conclude that P
(i)
`−1(ν̄) ≥ x`.
Case (c). We can write the defining condition of case (c) as `(i−1) ≤ j < ` − 1. Since
`(i−1) ≤ j < `(i) the string (j, xj) is non-singular. Thus we have P
(i)
j (ν) > x` by P
(i)
j (ν) >
xj ≥ x`. We also have P
(i)
` (ν) ≥ x`. Suppose if possible that P
(i)
`−1(ν) = x`. Then we have
P
(i)
j (ν) > P
(i)
`−1(ν) ≤ P (i)
` (ν) which is in contradiction to the convexity relation of P
(i)
k (ν) between
j ≤ k ≤ `. Thus we have P
(i)
`−1(ν) > x`. By `(i−1) < `− 1 and ` = `(i) we have P
(i)
`−1(ν̄) ≥ x`.
Case (d). We can write the defining condition of case (d) as j < `(i−1) < `. In this
case the string (j, xj) can be singular. Thus we have P
(i)
j (ν) ≥ x` by P
(i)
j (ν) ≥ xj ≥ x`.
We also have P
(i)
` (ν) ≥ x`. Suppose if possible that P
(i)
`−1(ν) = x`. Then by the convexity
relation of P
(i)
k (ν) between j ≤ k ≤ `, the only possibility is P
(i)
j (ν) = · · · = P
(i)
` (ν) = x`.
However this is in contradiction to the fact that P
(i)
k (ν) is a strictly convex function between
`(i−1) − 1 ≤ k ≤ `(i−1) + 1 due to the existence of the length `(i−1) row at ν(i−1). Thus we have
P
(i)
`−1(ν) > x`. By `(i−1) ≤ `− 1 we have P
(i)
`−1(ν̄) ≥ x`. �
46 R. Sakamoto
5.9 Proof for Case D (1)
In this section, let us consider the first case ` = `(i) < `(i) and m
(i)
` (ν) > 1. Throughout this
subsection, let j be the largest integer satisfying j < `(i) and m
(i)
j (ν) > 0. From ` = `(i) < `(i),
we see that ` = `(i) ≤ j. The situation depends on whether ` < j or ` = j.
5.9.1 The case ` < j
Proposition 5.24. Assume that ` = `(i) < `(i), m
(i)
` (ν) > 1 and ` < j. Then we have the
following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) Let us analyze the action of f̃i before and after the application of δ. In this case we
can choose two distinct length ` strings (`, x`) and
(
`, P
(i)
` (ν)
)
of (ν, J)(i) where f̃i will act on the
former one and δ will act on the latter one. Here note that if x` = P
(i)
` (ν) then all the riggings
for the length ` strings are the same since the minimal value of the corresponding riggings is x`
and the maximal one is P
(i)
` (ν).
By Proposition 5.23 we have P
(i)
`(i)−1
(ν̄) ≥ x`. Thus the string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
is shorter
than (`, x`) of (ν̄, J̄)(i) and its rigging is larger than or equal to x`. Also, by ` < j, we can use the
same arguments of Proposition 5.8 to show P
(i)
`(i)−1(ν̄) > x`. Thus the string
(
`(i)− 1, P
(i)
`(i)−1(ν̄)
)
has the rigging that is strictly larger than x`. Therefore f̃i will act on the same string (`, x`) in
both (ν, J) and (ν̄, J̄).
Let us analyze δ. Since f̃i will not change the corigging of the string
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i),
it remains as the shortest possible string in (ν̃, J̃)(i) starting from ˜̀(i−1) = `(i−1). So δ chooses
this string after the application of f̃i too. Let us show that the string
(
`(i), P
(i)
`(i)
(ν)
)
is chosen
by δ in both before and after the application of f̃i. Recall that we have two length ` strings
(`, x`) and
(
`, P
(i)
` (ν)
)
of (ν, J)(i) where f̃i acts on the former one and δ acts on the latter one.
(i) Consider the case P
(i)
` (ν) = x`, that is, the string (`, x`) is singular. Then the assumptions
`(i) < `(i) and m
(i)
`(i)
(ν) ≥ 2 implies that `(i) < `(i+1), as otherwise we would have `(i) = `(i). Then
a subtlety could occur only if `(i) > `(i+1) = `+ 1 and f̃i creates a singular string of length `+ 1.
So suppose if possible that such a situation happens. In this case, we have P
(i)
`+1(ν) = x` + 1 by
Lemma 5.2.
Suppose if possible that m
(i)
`+1(ν) > 0. Then the riggings for such string have to be x` + 1 by
the minimality of x`. Thus the length `+1 strings are singular. However this is in contradiction
to the present assumption `(i) > `+ 1. Thus we have m
(i)
`+1(ν) = 0.
Next we show m
(i+1)
`+1 (ν) = 1. Let j′ be the smallest integer such that ` + 1 ≤ j′ and
m
(i)
j′ (ν) > 0. Since m
(i)
`+1(ν) = 0 and `(i) > `+ 1 we have `+ 1 < j′ ≤ `(i). Also we have j′ ≤ j
by the assumption ` < j. Recall that we have P
(i)
` (ν) = x` and P
(i)
`+1(ν) = x` + 1. By the
minimality of x` we have P
(i)
`+2(ν) ≥ x` + 1 as otherwise we would have P
(i)
j′ (ν) ≤ x`. Let us
write P
(i)
`+2(ν) = x` + 1 + ε with ε ≥ 0. Then the inequality of Lemma 3.13 with l = `+ 1 reads
−P (i)
` (ν) + 2P
(i)
`+1(ν)− P (i)
`+2(ν) = 1− ε
≥ m(i−1)
`+1 (ν)− 2m
(i)
`+1(ν) +m
(i+1)
`+1 (ν) = m
(i−1)
`+1 (ν) +m
(i+1)
`+1 (ν).
Rigged Configurations and Kashiwara Operators 47
Since we have m
(i+1)
`+1 (ν) > 0 by `(i+1) = ` + 1, the only possibility is ε = 0, m
(i−1)
`+1 (ν) = 0 and
m
(i+1)
`+1 (ν) = 1.
Since ` < j′ we have P
(i)
j′ (ν) ≥ xj′ > x`. Then from the convexity relation of P
(i)
k (ν)
between ` ≤ k ≤ j′, the only possibility that is compatible with P
(i)
`+1(ν) = P
(i)
`+2(ν) = x` + 1 is
P
(i)
`+1(ν) = · · · = P
(i)
j′ (ν) = x` + 1. Since xj′ > x`, we see that xj′ = P
(i)
j′ (ν), in particular, the
corresponding string is singular. Since j′ ≥ ` + 1 = `(i+1), we conclude that j′ = `(i). However
this relation contradicts the requirement j′ ≤ j < `(i).
(ii) Consider the case P
(i)
` (ν) > x`. Recall that we have `(i) ≤ `(i+1). Again a subtlety could
occur only if `(i) > `(i+1) = ` + 1 and f̃i creates a singular string of length ` + 1. So suppose
if possible that such a situation happens. By Lemma 5.2, we have P
(i)
`+1(ν) = x` + 1. As in the
previous case, we have m
(i)
`+1(ν) = 0. Let j′ be the smallest integer such that ` + 1 ≤ j′ and
m
(i)
j′ (ν) > 0. Then, as in the previous case, we have `+ 1 < j′ ≤ `(i) and j′ ≤ j. Then from the
convexity relation of P
(i)
k (ν) between ` ≤ k ≤ j′ the only possibility that is compatible with the
relations P
(i)
` (ν) > x`, P
(i)
`+1(ν) = x`+ 1 and P
(i)
j′ (ν) > x` is P
(i)
` (ν) = · · · = P
(i)
j′ (ν) = x`+ 1. On
the other hand, by ` < j′ we have P
(i)
j′ (ν) ≥ xj′ > x`. Therefore we have P
(i)
j′ (ν) = xj′ = x` + 1
and, in particular, the corresponding string is singular. Since ` + 1 = `(i+1) ≤ j′ we conclude
that j′ = `(i). This is a contradiction since we have j′ ≤ j < `(i).
Therefore δ will act on the same strings of lengths `(i) and `(i) in both (ν, J)(i) and (ν̃, J̃)(i).
Thus we conclude that ¯̃ν = ˜̄ν in this case.
(2) It is enough to consider the three strings (`, x`),
(
`(i), P
(i)
`(i)
(ν)
)
and
(
`(i), P
(i)
`(i)
(ν)
)
. As for
(`, x`), we have
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→ (`+ 1, x` − 1),
(`, x`)
δ7−→ (`, x`)
f̃i7−→ (`+ 1, x` − 1),
and as for
(
`(i), P
(i)
`(i)
(ν)
)
we have
(
`(i), P
(i)
`(i)
(ν)
) f̃i7−→
(
`(i), P
(i)
`(i)
(ν̃)
) δ7−→
(
`(i) − 1, P
(i)
`(i)−1
(¯̃ν)
)
,(
`(i), P
(i)
`(i)
(ν)
) δ7−→
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
) f̃i7−→
(
`(i) − 1, P
(i)
`(i)−1
(˜̄ν)
)
.
The situation for
(
`(i), P
(i)
`(i)
(ν)
)
is similar. Thus we have
¯̃
J = ˜̄J . �
5.9.2 The case ` = j
We follow the classification of Lemma 5.1.
Proposition 5.25. Assume that ` = `(i) < `(i+1) = · · · = `(i), m
(i)
` (ν) > 1 and ` = j. Then we
have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) In this case, we can use the same arguments of the proof of Case (V) of Proposi-
tion 5.8 to show P
(i)
`(i)−1(ν̄) > x`. We also have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.23. Thus f̃i acts
on the same string before and after the application of δ. We notice that δ acts on the same string
before and after the application of f̃i even if f̃i creates a singular string since we are assuming
48 R. Sakamoto
that `(i+1) = `(i). Here recall that ˜̀(i) is determined as the length of the minimal possible string
compared with the length ˜̀(i+1) = `(i+1) and the length `(i) singular string remains singular
after the application of f̃i.
Proof of (2) is the same as Proposition 5.24. �
Proposition 5.26. Assume that ` = `(i) ≤ `(i+1) < `(i+1) = `(i), m
(i)
` (ν) > 1 and ` = j. Then
we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) We divide the proof into two cases according to whether ` = `(i) − 1 or ` < `(i) − 1.
Case 1. Let us consider the case ` = `(i) − 1. Then we have at least three strings (`, x`),(
`, P
(i)
` (ν)
)
and
(
`+ 1, P
(i)
`+1(ν)
)
in (ν, J)(i). Let us analyze the action of f̃i before and after the
application of δ. In this case, we show P
(i)
`(i)−1(ν̄) ≥ x`. According to Lemma 5.1 (VI) we have
to show P
(i)
`(i)−1(ν) ≥ x`, which is the consequence of ` = `(i) − 1. Recall also that P
(i)
`−1(ν̄) ≥ x`
by Proposition 5.23.
If P
(i)
` (ν̄) > x` there is no problem since the string (`, x`) of (ν̄, J̄)(i) becomes non-singular.
So suppose that P
(i)
` (ν̄) = x`. Then there are three strings
(
` − 1, P
(i)
`−1(ν̄)
)
, (`, x`), and (`, x`)
in (ν̄, J̄)(i). Thus we can choose the strings such that f̃i acts on the same string before and after
the application of δ. Also, since we are assuming that `(i+1) = `(i), δ acts on the same string
before and after the application of f̃i even if f̃i creates a singular string. Thus we have ¯̃ν = ˜̄ν in
this case.
Case 2. Let us consider the case ` < `(i)− 1. In this case, we can use the same arguments of
Case (VI) of the proof of Proposition 5.8 to show P
(i)
`(i)−1(ν̄) > x`. Thus we see that f̃i acts on
the same string before and after the application of δ. Also, by the assumption `(i+1) = `(i), we
see that δ acts on the same string before and after the application of f̃i. Thus we have ¯̃ν = ˜̄ν in
this case.
Proof of (2) is the same as Proposition 5.24. �
Proposition 5.27. Assume that ` = `(i) ≤ `(i+1) < `(i), m
(i)
` (ν) > 1 and ` = j. Then we have
the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. Step 1. Suppose that we have P
(i)
`(i)−1(ν̄) > x`. Since we have P
(i)
`−1(ν̄) ≥ x` by
Proposition 5.23, we see that f̃i acts on the same string before and after δ. Let us consider the
behavior of δ before and after f̃i. Then a subtlety could occur only if f̃i creates a singular string
which satisfies `(i) > `+ 1 ≥ `(i+1).
Thus suppose if possible that P
(i)
`(i)−1(ν̄) > x`, f̃i creates a singular string of length `+ 1 and
`(i) > `+1 ≥ `(i+1). From Lemma 5.1 (VII) we have P
(i)
`(i)−1(ν) > x`+1 and from Lemma 5.2 we
have P
(i)
`+1(ν) = x`+1. If `(i)−1 = `+1 this is a contradiction. Thus suppose that `(i)−1 > `+1.
Recall that we have P
(i)
` (ν) ≥ x` by definition of the rigged configurations. Suppose if possible
that P
(i)
` (ν) > x`. Then we have P
(i)
` (ν) ≥ P
(i)
`+1(ν) < P
(i)
`(i)−1(ν) which violates the convexity
relation of P
(i)
k (ν) between ` ≤ k ≤ `(i). Thus we have P
(i)
` (ν) = x`. Suppose if possible that
`(i+1) = `. Since P
(i)
` (ν) = x` we see that all length ` strings of (ν, J)(i) are singular. This is
Rigged Configurations and Kashiwara Operators 49
a contradiction since we have m
(i)
` (ν) > 1 and `(i) > `(i+1). Thus we have `(i+1) = `+ 1. From
P
(i)
`+1(ν) = x` + 1 and P
(i)
`(i)−1(ν) > x` + 1 we have
P
(i)
`+2(ν) ≥ min
{
P
(i)
`+1(ν), P
(i)
`(i)−1(ν)
}
= x` + 1.
Let us write P
(i)
`+2(ν) = x` + 1 + ε where ε ≥ 0. From Lemma 3.13 with l = `+ 1, we have
−P (i)
` (ν) + 2P
(i)
`+1(ν)− P (i)
`+2(ν) = 1− ε
≥ m(i−1)
`+1 (ν)− 2m
(i)
`+1(ν) +m
(i+1)
`+1 (ν) = m
(i−1)
`+1 (ν) +m
(i+1)
`+1 (ν).
Since we have m
(i+1)
`+1 (ν) > 0 by `(i+1) = ` + 1, the only possibility is ε = 0, m
(i−1)
`+1 (ν) = 0 and
m
(i+1)
`+1 (ν) = 1. In particular, we have P
(i)
`+2(ν) = x` + 1. However this is a contradiction since
we have P
(i)
`+1(ν) = P
(i)
`+2(ν) < P
(i)
`(i)−1(ν) which violates the convexity relation of P
(i)
k (ν) between
` ≤ k ≤ `(i). Hence this case cannot happen. Therefore we have ¯̃ν = ˜̄ν and
¯̃
J = ˜̄J in this case.
Step 2. Suppose that we have ` = `(i) − 1. By definition of the rigged configuration we
have P
(i)
` (ν) ≥ x`. Then by Lemma 5.1 (VII) we have P
(i)
` (ν̄) ≥ x`− 1. Suppose if possible that
P
(i)
` (ν̄) = x` − 1. Recall that we have the three strings (`, x`),
(
`, P
(i)
` (ν)
)
and
(
`+ 1, P
(i)
`+1(ν)
)
of (ν, J)(i). Then we have the remaining string (`, x`) in (ν̄, J̄)(i). This is a contradiction since
we have P
(i)
` (ν̄) < x`.
Thus suppose that P
(i)
` (ν̄) ≥ x`. As in Case 1 of the proof of the previous proposition we
can show that f̃i chooses the same string before and after δ. Let us consider the behavior of δ
before and after f̃i. In this case, there is no problem even if f̃i creates a singular string since we
have `+ 1 = `(i). Hence we have ¯̃ν = ˜̄ν and
¯̃
J = ˜̄J .
Step 3. We assume that ` < `(i) − 1 and P
(i)
`(i)−1(ν̄) ≤ x` in the following proof. According
to Lemma 5.1 (VII) we have P
(i)
`(i)−1(ν) ≤ x` + 1. Recall that we have P
(i)
`(i)
(ν) > x` by `(i) > `
and P
(i)
` (ν) ≥ x` by definition of the rigged configurations. Since we are assuming that ` =
`(i) = j < `(i) − 1, the convexity relation of the vacancy numbers allows the following three
possibilities:
(1) ` = `(i) − 2, P
(i)
` (ν) = x`, P
(i)
`(i)−1(ν) = x` + 1 and P
(i)
`(i)
(ν) = x` + 2,
(2) P
(i)
` (ν) = x` and P
(i)
`+1(ν) = · · · = P
(i)
`(i)
(ν) = x` + 1,
(3) P
(i)
` (ν) = P
(i)
`+1(ν) = · · · = P
(i)
`(i)
(ν) = x` + 1.
In the following we consider them case by case.
Case (1). According to Lemma 3.13 with l = `(i) − 1, we have
−P (i)
` (ν) + 2P
(i)
`(i)−1(ν)− P (i)
`(i)
(ν) = 0
≥ m(i−1)
`(i)−1(ν)− 2m
(i)
`(i)−1(ν) +m
(i+1)
`(i)−1(ν) = m
(i−1)
`(i)−1(ν) +m
(i+1)
`(i)−1(ν)
where we have used m
(i)
`(i)−1(ν) = 0 by j = `. Thus we have m
(i−1)
`(i)−1(ν) = m
(i+1)
`(i)−1(ν) = 0. Since
we are assuming that `(i+1) < `(i), we conclude that `(i+1) = `(i+1) = `.
Now recall that we are assuming that m
(i)
` (ν) ≥ 2. Then the minimality of x` and the as-
sumption P
(i)
` (ν) = x` imply that all the length ` strings of (ν, J)(i) are
(
`, P
(i)
` (ν)
)
, in particular,
50 R. Sakamoto
they are singular. However this is in contradiction to the assumption `(i+1) < `(i) since we know
that `(i+1) = ` which forces that `(i) = `. Thus this case cannot happen.
Case (2). The assumption P
(i)
`+1(ν) = · · · = P
(i)
`(i)
(ν) implies that m
(i−1)
k (ν) = m
(i+1)
k (ν) = 0
for all ` + 2 ≤ k ≤ `(i) − 1 as otherwise the above relation for the vacancy numbers would be
strictly convex. Next, we apply Lemma 3.13 with l = `+ 1. Then we have
−P (i)
` (ν) + 2P
(i)
`+1(ν)− P (i)
`+2(ν) = 1
≥ m(i−1)
`+1 (ν)− 2m
(i)
`+1(ν) +m
(i+1)
`+1 (ν) = m
(i−1)
`+1 (ν) +m
(i+1)
`+1 (ν),
where we have used m
(i)
`+1(ν) = 0 by j = `.
Suppose if possible that m
(i−1)
`+1 (ν) = 1. Then we have m
(i+1)
`+1 (ν) = 0. Since we are assuming
that `(i+1) < `(i), m
(i+1)
k (ν) = 0 for all `+ 1 ≤ k ≤ `(i) − 1 implies that `(i+1) = `. However, as
in the previous case, the assumption P
(i)
` (ν) = x` implies that all the length ` strings of (ν, J)(i)
are singular. Thus we must have `(i) = ` by `(i+1) = ` and m
(i)
` (ν) > 1. This is a contradiction.
Hence this case cannot happen.
Therefore we assume that m
(i−1)
`+1 (ν) = 0 and m
(i+1)
`+1 (ν) = 1 in the sequel. Then we can show
that `(i+1) = ` and `(i+1) = ` + 1. Since m
(i+1)
k (ν) = 0 for all ` + 2 ≤ k ≤ `(i) − 1, we have
`(i+1) = ` or `+ 1 by ` ≤ `(i+1). Suppose if possible that `(i+1) = `. Since we are assuming that
P
(i)
` (ν) = x`, the length ` = `(i) strings of (ν, J)(i) are always equal to (`, x`) by the minimality
of x`. Then by the assumption m
(i)
` (ν) ≥ 2, we must have `(i) = ` by `(i+1) = `. This is in
contradiction to the other assumption `(i) < `(i). Thus we have `(i+1) = ` + 1. Then from
m
(i+1)
`+1 (ν) = 1, we conclude that `(i+1) = ` since there is the requirement ` ≤ `(i+1) ≤ `(i+1).
Let us analyze f̃i ◦ δ first. Since f̃i acts on the i-th place, we have ˜̄ν(a)
= ν̄(a) for all
a 6= i. Next we consider ˜̄ν(i)
. Recall that we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.23. Also,
since we have P
(i)
`(i)−1(ν) = x` + 1, we have P
(i)
`(i)−1(ν̄) = x` by Lemma 5.1 (VII). Therefore the
string
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
of (ν̄, J̄)(i) has the largest length among the strings with smallest
rigging, hence f̃i acts on it. To summarize, ˜̄ν(i)
is obtained from ν(i) by removing one box from
a length `(i) row.
Next let us analyze δ ◦ f̃i. Since f̃i does not change the coriggings of untouched strings, and
since we are assuming that m
(i)
` (ν) ≥ 2, we have ˜̀(a) = `(a) for all a and ˜̀(a) = `(a) for all
a ≥ i + 1. Let us consider ˜̀(i). Recall that we have ˜̀(i+1) = `(i+1) = ` + 1 in this case. Since
we have P
(i)
`+1(ν) = x` + 1, f̃i creates a singular string of length ` + 1 by Lemma 5.2. Thus the
length ` + 1 string created by f̃i is the shortest possible string starting from ˜̀
(i+1). Therefore
we have ˜̀(i) = `+ 1. Since we have m
(i−1)
k (ν) = 0 for all `+ 1 ≤ k ≤ `(i) − 1 and since f̃i does
not change the coriggings of untouched strings, we have ˜̀(i−1) = `(i−1) and thus ˜̀(a) = `(a) for
all a ≤ i − 1. To summarize, ¯̃ν
(i)
is obtained from ν(i) by removing one box from a length `(i)
row and ¯̃ν
(a)
= ν̄(a) for all a 6= i. Hence we have ¯̃ν = ˜̄ν.
Finally, let us show
¯̃
J = ˜̄J . For this it is enough to check the three strings
(
`(i), P
(i)
`(i)
(ν)
)
,
(`, x`) and
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i). Here note that we have in fact
(
`(i), P
(i)
`(i)
(ν)
)
= (`, x`) under
the present assumption. To begin with, the string
(
`(i), P
(i)
`(i)
(ν)
)
behaves as follows:
(
`(i), P
(i)
`(i)
(ν)
) f̃i7−→
(
`(i), P
(i)
`(i)
(ν̃)
) δ7−→
(
`(i) − 1, P
(i)
`(i)−1
(¯̃ν)
)
,(
`(i), P
(i)
`(i)
(ν)
) δ7−→
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
) f̃i7−→
(
`(i) − 1, P
(i)
`(i)−1
(˜̄ν)
)
.
Rigged Configurations and Kashiwara Operators 51
By ¯̃ν = ˜̄ν we see the coincidence of the riggings. Next, let us consider the string (`, x`):
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→ (`, x`)
f̃i7−→ (`, x`).
Recall that we have ˜̀(i−1) ≤ ˜̀(i) = ˜̀(i+1) = ` and ` + 1 = ˜̀
(i+1) = ˜̀
(i) < ˜̀
(i−1). Therefore we
have P
(i)
` (¯̃ν) = P
(i)
` (ν) = x`. Let us consider the remaining string
(
`(i), P
(i)
`(i)
(ν)
)
. Then we have
(
`(i), P
(i)
`(i)
(ν)
) f̃i7−→
(
`(i), P
(i)
`(i)
(ν̃)
) δ7−→
(
`(i), P
(i)
`(i)
(ν̃)
)
,(
`(i), P
(i)
`(i)
(ν)
) δ7−→
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
) f̃i7−→
(
`(i), P
(i)
`(i)−1(ν̄)− 1
)
.
Since we have ` + 1 ≤ `(i) by the assumption, we have P
(i)
`(i)
(ν̃) = P
(i)
`(i)
(ν) − 2 = x` − 1. Also,
since we are assuming that `(i+1) < `(i), we have P
(i)
`(i)−1(ν̄) = P
(i)
`(i)−1(ν)−1 by Lemma 5.1 (VII).
Thus we have P
(i)
`(i)−1(ν̄) − 1 = P
(i)
`(i)
(ν) − 2 = x` − 1. Therefore we have checked
¯̃
J = ˜̄J in this
case.
Case (3). The assumption P
(i)
` (ν) = · · · = P
(i)
`(i)
(ν) implies that m
(i−1)
k (ν) = m
(i+1)
k (ν) = 0
for all ` + 1 ≤ k ≤ `(i) − 1. Since we are assuming that `(i+1) < `(i), the only possibility is
`(i+1) = `.
Let us analyze f̃i◦δ first. From Lemma 5.1 (VII) we have P
(i)
`(i)−1(ν̄) = x` by P
(i)
`(i)−1(ν) = x`+1.
Recall that we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.23. From the assumption ` ≤ `(i) − 1, we
see that the string
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
has the smallest rigging of the largest length. Thus the
next f̃i acts on this string. To summarize, we have ˜̄ν(a)
= ν̄(a) for all a 6= i and ˜̄ν(i)
is obtained
by removing one box from the length `(i) row.
Next we analyze δ ◦ f̃i. Since f̃i does not change coriggings of untouched strings, and since
we are assuming that m
(i)
` (ν) > 1, we have ˜̀(a) = `(a) for all a and ˜̀(a) = `(a) for all a ≥ i + 1.
In particular, we have ˜̀(i+1) = `(i+1) = ` by the assumption. In order to determine ˜̀(i), recall
that the relation P
(i)
`+1(ν) = x` + 1 implies that f̃i creates the singular string of length `+ 1 by
Lemma 5.2. Thus we have ˜̀(i) = `+ 1. Since we have m
(i−1)
k (ν) = 0 for all `+ 1 ≤ k ≤ `(i) − 1,
we have ˜̀(i−1) = `(i−1) by `(i−1) ≥ `(i), thus ˜̀(a) = `(a) for all a ≤ i − 1. To summarize, we
have ¯̃ν
(a)
= ν̄(a) for all a 6= i and ¯̃ν
(i)
is obtained by removing one box from the length `(i) row.
Hence we have ˜̄ν = ¯̃ν.
Let us show ˜̄J =
¯̃
J in this case. Again it is enough to check the three strings
(
`(i), P
(i)
`(i)
(ν)
)
,
(`, x`) and
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i). The analysis for the string
(
`(i), P
(i)
`(i)
(ν)
)
is the same with
the previous Case (2). Let us analyze the string (`, x`). Then it behaves as
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→ (`, x`)
f̃i7−→ (`, x`).
In this case, we have ˜̀(i−1) ≤ ˜̀(i) = ˜̀(i+1) = ` and ` = ˜̀
(i+1) < ˜̀(i) ≤ ˜̀(i−1). Therefore we have
P
(i)
` (¯̃ν) = P
(i)
` (ν)−1 = x`. The analysis of
(
`(i), P
(i)
`(i)
(ν)
)
is the same with the previous Case (2).
Thus we have ˜̄J =
¯̃
J in this case. This completes the proof of proposition. �
52 R. Sakamoto
Example 5.28. Here we give an example for Case (2) of Step 3 of the proof of Proposition 5.27.
Consider the following rigged configuration (ν, J) of type (B1,1)⊗3 ⊗ B1,3 ⊗ B2,1 ⊗ B2,2 ⊗ B3,1
of D
(1)
5
2
1
1
−1
−1
0
1
−1
0
0
−1
−1
−1
0
0
0
−1
−1
−1
0
0
0
0
2
1
3
0
0
0
2
1
3
0
−1
−3
0
−1
−3
1
−1
1
−2
The corresponding tensor product Φ−1(ν, J) is
1̄ ⊗ 1̄ ⊗ 5 ⊗ 1 4̄ 4̄ ⊗ 5̄
3̄
⊗ 1 5
3 4̄
⊗
2
3
4
Then (ν̄, J̄) is
1
1
1
0
−1
1
0
0
0
0
0
−1
−1
−1
0
0
0
−1
−1
−1
0
0
0
0
1
3
0
0
0
0
1
3
0
0
−3
0
0
−3
2
−1
2
−2
and (ν̃, J̃) is
2
1
1
0
−1
0
1
0
0
0
−1
−1
−2
−2
0
0
−1
−1
−2
−2
0
0
0
2
2
4
0
0
0
2
2
4
0
−1
−3
0
−1
−3
1
−1
1
−2
In this example, we have ` = `(2) = 2 < `(3) = 3 < `(2) = 5, m
(2)
2 (ν) = 3 and m
(2)
k (ν) = 0
for `(2) < k < `(2). Moreover, we have ` < `(2) − 1 and P
(2)
`(2)−1(ν̄) = x` where x` = −1. This
example satisfies the condition for Case (2)
P
(2)
2 (ν) = x` = −1, P
(2)
3 (ν) = P
(2)
4 (ν) = P
(2)
5 (ν) = x` + 1 = 0.
As discussed in the proof, we have m
(1)
4 (ν) = m
(3)
4 (ν) = 0 where `+2 = `(2)−1 = 4, m
(1)
3 (ν) = 0
and m
(3)
3 (ν) = 1. Finally (¯̃ν,
¯̃
J) = (˜̄ν, ˜̄J) is
1
1
1
0
−1
1
0
0
0
0
0
−1
−1
−2
0
0
0
−1
−1
−2
0
0
0
0
1
4
0
0
0
0
1
4
0
0
−3
0
0
−3
2
−1
2
−2
5.10 Proof for Case D (2)
Proposition 5.29. Suppose that ` = `(i) = `(i) and m
(i)
` (ν) > 2. Then we have the following
identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Rigged Configurations and Kashiwara Operators 53
Proof. By the assumption m
(i)
` (ν) ≥ 3 we can choose the three distinct strings(
`(i), P
(i)
`(i)
(ν)
)
=
(
`(i), P
(i)
`(i)
(ν)
)
and (`, x`)
of the i-th configuration. Here we may have P
(i)
`(i)
(ν) = x`.
Let us analyze f̃i ◦ δ. After the application of δ the above three strings become(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
=
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
and (`, x`).
By Proposition 5.23 we have P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)−1(ν̄) ≥ x`. Thus the string (`, x`) remains as
the string with smallest rigging of the largest length. Hence f̃i will act on it. To summarize˜̄ν(a)
= ν̄(a) for all a 6= i and ˜̄ν(i)
is obtained by removing a box from each of the two length
`(i) = `(i) rows and adding a box to the length ` row.
Let us analyze δ ◦ f̃i. Since f̃i does not change coriggings of untouched strings, we have˜̀(a) = `(a) for all a and ˜̀(a) = `(a) for all a ≥ i + 1. Since the string
(
`(i), P
(i)
`(i)
(ν)
)
remains as
the singular string which is longer than ˜̀(i+1), we have ˜̀(i) = `(i) and hence ˜̀(a) = `(a) for all
a ≤ i. Thus we obtain ¯̃ν = ˜̄ν. The remaining statement
¯̃
J = ˜̄J follows from ¯̃ν = ˜̄ν. �
5.11 Proof for Case D (3)
5.11.1 Outline
Let us consider the third case. We prove the following properties in the latter part of this
subsection by case by case analysis.
Proposition 5.30. Assume that `(i) = ` and m
(i)
` (ν) = 1. Then the following relations hold:
P
(i)
`+1(ν) = x` + 1, (5.14)
P
(i)
`−1(ν̄) = x`, (5.15)
`+ 1 ≤ `(i+1). (5.16)
Next let us show the following property:
Proposition 5.31. Assume that ` = `(i) < `(i) and m
(i)
` (ν) = 1. Then the following relation
holds:
P
(i)
`(i)−1(ν̄) > x`.
Proof. We follow the classification of Lemma 5.1. Since we are assuming that `(i) < `(i), we
have to deal with Cases (V), (VI) and (VII). Throughout the proof of this proposition, let j be
the largest integer such that j < `(i) and m
(i)
j (ν) > 0. Since we are assuming that `(i) < `(i), we
have `(i) ≤ j.
Case (V). In this case, we have to show that P
(i)
`(i)−1(ν) ≥ x` under the assumption `(i) <
`(i+1) = · · · = `(i). We can use the same argument of the corresponding part of Proposition 5.8
to show the assertion.
Case (VI). In this case, we have to show that P
(i)
`(i)−1(ν) > x` under the assumption `(i+1) <
`(i+1) = `(i). Suppose if possible that P
(i)
`(i)−1(ν) ≤ x`.
54 R. Sakamoto
Let us consider the case j = `(i) − 1. Note that by (5.16), which is shown independently to
the present proposition, we have ` < `(i+1). Combining with the current assumption `(i+1) < `(i)
we see that ` = `(i) < `(i) − 1. Thus the string (`(i) − 1, x`(i)−1) of (ν, J)(i) is different from the
string (`, x`). Then we have x`(i)−1 ≤ P
(i)
`(i)−1(ν) ≤ x`. However this is in contradiction to the
relation x`(i)−1 > x` which follows from the minimality of x` under the condition `(i) − 1 > `.
Next let us consider the case j < `(i)− 1. Since `(i) < `(i) we have P
(i)
`(i)
(ν) ≥ x`(i) > x` by the
minimality of x`. Combining with the assumption P
(i)
`(i)−1(ν) ≤ x` we have P
(i)
`(i)−1(ν) < P
(i)
`(i)
(ν).
Then from the convexity relation of P
(i)
k (ν) between j ≤ k ≤ `(i), we have xj ≤ P
(i)
j (ν) <
P
(i)
`(i)−1(ν) ≤ x`. This is in contradiction to the relation xj ≥ x` that follows from the minimality
of x`.
Case (VII). In this case, we have to show that P
(i)
`(i)−1(ν) > x` + 1 under the assump-
tion `(i+1) < `(i). Suppose if possible that P
(i)
`(i)−1(ν) ≤ x` + 1. Again recall that we have
` < `(i+1) from (5.16) which is shown independently to the present proposition. Since we have
`(i+1) ≤ `(i+1) by definition, we have the relations ` = `(i) < `(i+1) < `(i).
We divide the proof of Case (VII) into three steps according to the position of j.
Step 1. First, let us consider the case j = `(i) − 1. Then there is the corresponding string
(`(i) − 1, x`(i)−1). Since ` < `(i) − 1, we have x` < x`(i)−1 ≤ P
(i)
`(i)−1(ν) by the minimality of x`.
Combining with the assumption P
(i)
`(i)−1(ν) ≤ x` + 1 we conclude that x`(i)−1 = P
(i)
`(i)−1(ν). Then
the string (`(i) − 1, x`(i)−1) is singular and its length satisfies `(i+1) ≤ `(i) − 1 < `(i). This
contradicts the definition of `(i).
Step 2. Next, let us consider the case ` = `(i) < j < `(i) − 1. Since we are assuming that
`(i) < `(i) we have P
(i)
`(i)
(ν) ≥ x`(i) > x` by the minimality of x`. Similarly, from ` < j, we have
P
(i)
j (ν) > x`. Then, by the convexity of P
(i)
k (ν) between j ≤ k ≤ `(i), the only possibility that
is compatible with the assumption P
(i)
`(i)−1(ν) ≤ x` + 1 is the situation P
(i)
j (ν) = · · · = P
(i)
`(i)
(ν) =
x` + 1.
1. Consider the case `(i+1) ≤ j. Then, combining with the relation P
(i)
j (ν) ≥ xj > x`, we
obtain P
(i)
j (ν) = xj . In particular, the string (j, xj) is singular and its length satisfies that
`(i+1) ≤ j < `(i). This contradicts the definition of `(i).
2. Consider the case j < `(i+1). Since we have m
(i+1)
`(i+1)
(ν) > 0, the relation for P
(i)
k (ν) between
`(i+1) − 1 ≤ k ≤ `(i+1) + 1 must be strictly convex. This contradicts the above relation.
Step 3. Finally, let us consider the case ` = `(i) = j. Since we are assuming that m
(i)
` (ν) = 1,
there is the only one singular string (`, x`). In particular, we have P
(i)
` (ν) = x`. Recall that we
have P
(i)
`+1(ν) = x` + 1 by (5.14) which is shown independently to the present proposition. Then
the convexity relation of P
(i)
k (ν) between j ≤ k ≤ `(i) under the condition P
(i)
`(i)
(ν) > x` admits
the following two possibilities:
1. The case `(i) + 2 = `(i), P
(i)
` (ν) = x`, P
(i)
`+1(ν) = x` + 1 and P
(i)
`(i)
(ν) = x` + 2. Since the
relation for the vacancy numbers is linear, this situation can happen only if m
(i+1)
`+1 (ν) = 0.
Recall that the relation ` = `(i) < `(i+1) < `(i) in this case implies that `(i+1) = ` + 1, in
particular, m
(i+1)
`+1 (ν) > 0. This is a contradiction.
Rigged Configurations and Kashiwara Operators 55
2. The case P
(i)
` (ν) = x` and P
(i)
`+1(ν) = · · · = P
(i)
`(i)
(ν) = x` + 1. Since the relation for P
(i)
k (ν)
between ` + 1 ≤ k ≤ `(i) is linear, we have m
(i+1)
k (ν) = 0 for all ` + 1 < k < `(i). Next,
applying Lemma 3.13 with l = `+ 1 we have
−P (i)
` (ν) + 2P
(i)
`+1(ν)− P (i)
`+2(ν) = 1
≥ m(i−1)
`+1 (ν)− 2m
(i)
`+1(ν) +m
(i+1)
`+1 (ν) = m
(i−1)
`+1 (ν) +m
(i+1)
`+1 (ν),
where we have used the fact that m
(i)
`+1(ν) = 0 which follows from the assumption j = `
and ` + 2 ≤ `(i). Since we know that m
(i+1)
k (ν) = 0 for all ` + 1 < k < `(i), the relation
` = `(i) < `(i+1) < `(i) with m
(i+1)
`(i+1)
(ν) > 0 forces that `(i+1) = ` + 1 and m
(i+1)
`+1 (ν) = 1.
Now recall the relation ` < `(i+1) of (5.16) which is shown independently to the present
proposition. By the assumption `(i+1) < `(i) and the relation `(i+1) ≤ `(i+1) that follows
from the definition, we also have ` = `(i) < `(i+1) < `(i) with m
(i+1)
`(i+1)(ν) > 0. By definition
of the operation δ, the two strings (`(i+1), P
(i)
`(i+1)(ν)) and (`(i+1), P
(i)
`(i+1)
(ν)) of (ν, J)(i+1)
must be distinct. However, as we have seen in the previous paragraph, there is only one
string (k, xk) of (ν, J)(i+1) under the restriction ` < k < `(i). This is a contradiction.
In conclusion, we obtain P
(i)
`(i)−1(ν) > x` + 1 in this Case (VII). �
Proof of “Propositions 5.30 & 5.31 =⇒ ¯̃ν = ˜̄ν”. Step 1. Let us consider the case f̃i ◦ δ.
The first δ creates the strings
(
`− 1, P
(i)
`−1(ν̄)
)
and
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
of (ν̄, J̄)(i). Recall that
the rigging of the latter string satisfies P
(i)
`(i)−1(ν̄) > x` by Proposition 5.31. Then, since the
operation δ does not change the riggings of unchanged strings, the relation (5.15) and m
(i)
` (ν) = 1
imply that the string (`− 1, x`) is the longest strings among the string with rigging x`, hence f̃i
will act on this string.
Step 2. Let us consider the case δ ◦ f̃i. Recall that by definition of f̃i we have ˜̀(a) = `(a) for
all a < i. From Lemma 5.2(1) the relation (5.14) implies that the string (`+1, x`−1) of (ν̃, J̃)(i)
created by f̃i is singular. Since m
(i)
` (ν) = 1 we know that there is no singular string of (ν̃, J̃)(i)
that is shorter than or equal to ` and longer than ˜̀(i−1). Thus we have ˜̀(i) = ` + 1. Then,
since f̃i does not change coriggings of (ν̃, J̃)(i+1), the relation (5.16) implies that ˜̀(a) = `(a) for
all i < a as well as ˜̀(a) = `(a) for all a.
Step 3. To summarize, we have ¯̃ν
(a)
= ˜̄ν(a)
= ν̄(a) for all a 6= i and ¯̃ν
(i)
= ˜̄ν(i)
are obtained
by removing a box from a length `(i) singular string of ν(i). Hence we have ¯̃ν = ˜̄ν. �
Proofs of
¯̃
J = ˜̄J will be given by case by case analysis assuming the relation ¯̃ν = ˜̄ν along with
the proof of Proposition 5.30.
5.11.2 Proof of Proposition 5.30 for the case `(i−1) < `
We divide the proof into two cases depending on whether `(i−1) < ` or `(i−1) = `. In this
subsection, we consider the case `(i−1) < `.
To begin with, let us show the following property.
Proposition 5.32. Suppose that `(i−1) < ` = `(i) and m
(i)
` (ν) = 1. Then we have
P
(i)
`−1(ν) = x` + 1, P
(i)
` (ν) = x`, P
(i)
`+1(ν) = x` + 1. (5.17)
56 R. Sakamoto
Moreover we have
m
(i−1)
` (ν) = m
(i+1)
` (ν) = 0. (5.18)
For the proof, we prepare the following lemma.
Lemma 5.33. Suppose that `(i−1) < ` = `(i) and m
(i)
` (ν) = 1. Then we have the following
relations:
(1) P
(i)
`−1(ν) > P
(i)
` (ν), (2) P
(i)
` (ν) < P
(i)
`+1(ν).
Proof. (1) Note that from the assumptions m
(i)
` (ν) = 1 and ` = `(i), we see that the string
(`, x`) is singular which implies that x` = P
(i)
` (ν). Let j be the largest integer such that j ≤ `−1
and m
(i)
j (ν) > 0. If there is no such j, set j = 0 and xj = 0.
We divide the proof into two cases depending on whether j = `− 1 or j < `− 1.
Step 1. Let us consider the case j = ` − 1. Then the string (` − 1, x`−1) is non-singular
since its length satisfies `(i−1) ≤ ` − 1 < `(i). Thus we have P
(i)
`−1(ν) > x`−1. Recall that we
have x`−1 ≥ x` by the minimality of x`. Then x` = P
(i)
` (ν) implies that P
(i)
`−1(ν) > P
(i)
` (ν) in
this case.
Step 2. Let us consider the case j < `−1. Then we have xj ≥ x` by definition of x` and ` > 0.
Thus we have P
(i)
j (ν) ≥ x` by P
(i)
j (ν) ≥ xj . Since x` = P
(i)
` (ν) we obtain P
(i)
j (ν) ≥ P
(i)
` (ν).
Then by the convexity of P
(i)
k (ν) between j ≤ k ≤ ` we obtain P
(i)
`−1(ν) ≥ min{P (i)
j (ν), P
(i)
` (ν)} =
P
(i)
` (ν). Suppose if possible that P
(i)
`−1(ν) = P
(i)
` (ν). Then by the convexity of the vacancy
numbers the only possibility that is compatible with P
(i)
j (ν) ≥ P
(i)
` (ν) is the case P
(i)
j (ν) =
· · · = P
(i)
` (ν).
(i) Suppose that `(i−1) ≤ j. Recall that the string (j, xj) satisfies xj ≤ P (i)
j (ν) = P
(i)
` (ν) = x`.
On the other hand, by the minimality of x`, we have xj ≥ x`. Thus xj = P
(i)
j (ν). Then
the length of the singular string (j, xj) satisfies `(i−1) ≤ j < `(i) which is in contradiction
to the definition of `(i). Thus we conclude that P
(i)
`−1(ν) > P
(i)
` (ν) if `(i−1) ≤ j.
(ii) Suppose that j < `(i−1). Then the relation of P
(i)
k (ν) between `(i−1) − 1 ≤ k ≤ `(i−1) + 1
must be strictly convex due to the existence of the length `(i−1) string at ν(i−1). This is
in contradiction to the relation P
(i)
j (ν) = · · · = P
(i)
` (ν).
Thus we conclude that P
(i)
`−1(ν) > P
(i)
` (ν).
(2) Since f̃i does not vanish, the rigging for the new string created by f̃i does not exceed the
corresponding vacancy number. Therefore, from the proof of Lemma 5.2(1), we have P
(i)
`+1(ν) ≥
x` + 1. Since the string (`, x`) is singular we have P
(i)
` (ν) = x`. Combining both relations we
obtain P
(i)
` (ν) < P
(i)
`+1(ν). �
Proof of Proposition 5.32. We apply Lemma 3.12 with m
(i)
` (ν) = 1. Then P
(i)
`−1(ν), P
(i)
` (ν)
and P
(i)
`+1(ν)−2 have to satisfy the convex relation. The only possibility that is compatible with
the relation P
(i)
`−1(ν) > P
(i)
` (ν) < P
(i)
`+1(ν) is P
(i)
`−1(ν) = P
(i)
` (ν) + 1 = P
(i)
`+1(ν).
From Lemma 3.13 we have
−P (i)
`−1(ν) + 2P
(i)
` (ν)− P (i)
`+1(ν) = −2
≥ m(i−1)
` (ν)− 2m
(i)
` (ν) +m
(i+1)
` (ν) = m
(i−1)
` (ν)− 2 +m
(i+1)
` (ν),
Rigged Configurations and Kashiwara Operators 57
that is, 0 ≥ m
(i−1)
` (ν) + m
(i+1)
` (ν). Since m
(i−1)
` (ν) ≥ 0 and m
(i+1)
` (ν) ≥ 0 we conclude that
m
(i−1)
` (ν) = m
(i+1)
` (ν) = 0. �
Proof of Proposition 5.30 for the case `(i−1) < `. By (5.17) we have P
(i)
`+1(ν) = x` + 1,
that is, (5.14). By the assumption `(i−1) < ` we see `(i−1) ≤ ` − 1. Thus we obtain P
(i)
`−1(ν̄) =
P
(i)
`−1(ν)− 1 = x`, that is, (5.15). Finally, from (5.18) we have m
(i+1)
` (ν) = 0. Therefore we have
` < `(i+1), that is, (5.16). �
In view of the proof of ¯̃ν = ˜̄ν, we can show
¯̃
J = ˜̄J for this case as follows.
Proposition 5.34. Suppose that `(i−1) < ` = `(i) < `(i) and m
(i)
` (ν) = 1. Then we have
¯̃
J = ˜̄J .
Proof. From (5.18) it is enough to analyze the string (`, x`) of (ν, J)(i) where x` = P
(i)
` (ν).
Then we have
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→
(
`− 1, P
(i)
`−1(ν̄)
) f̃i7−→
(
`, P
(i)
`−1(ν̄)− 1
)
.
Recall that we have `(i−1) < ` = `(i) < `(i+1). Then we have P
(i)
` (¯̃ν) = P
(i)
` (ν)− 1 = x`− 1 since
we have P
(i)
` (ν) = x` by (5.17). Similarly we have P
(i)
`−1(ν̄)− 1 = P
(i)
`−1(ν)− 2 = x` − 1 since we
have P
(i)
`−1(ν) = x` + 1 by (5.17). �
5.11.3 Proof of Proposition 5.30 for the case `(i−1) = `
We begin by showing the following property.
Proposition 5.35. Suppose that `(i−1) = ` = `(i) and m
(i)
` (ν) = 1. Then we have
P
(i)
`−1(ν) = P
(i)
` (ν) = x`, P
(i)
`+1(ν) = x` + 1. (5.19)
Moreover, we have
m
(i−1)
` (ν) = 1, m
(i+1)
` (ν) = 0. (5.20)
For the proof, we prepare the following lemma.
Lemma 5.36. Suppose that `(i−1) = ` = `(i) and m
(i)
` (ν) = 1. Then we have the following
relations:
(1) P
(i)
`−1(ν) ≥ P (i)
` (ν), (2) P
(i)
` (ν) < P
(i)
`+1(ν).
Proof. (1) Let j be the largest integer such that j ≤ `−1 and m
(i)
j (ν) > 0. If there is no such j,
set j = 0. We divide the proof into two cases depending on whether j = `− 1 or j < `− 1.
Step 1. Let us consider the case j = ` − 1. Since the string (`, x`) is singular, we have
P
(i)
` (ν) = x`, and by the minimality of x` we have x` ≤ x`−1. Since x`−1 ≤ P
(i)
`−1, we conclude
that P
(i)
`−1(ν) ≥ P (i)
` (ν).
Step 2. Let us consider the case j < ` − 1. Recall that the assumptions ` = `(i) and
m
(i)
` (ν) = 1 imply that the string (`, x`) is singular and thus P
(i)
` (ν) = x`. If j > 0, then
by the minimality of ` with j < ` we have P
(i)
j (ν) ≥ xj ≥ x` = P
(i)
` (ν). If j = 0 we have
P
(i)
0 (ν) = 0. In this case, we also have P
(i)
j (ν) ≥ P (i)
` (ν) since we have P
(i)
` (ν) = x` ≤ 0 which is
58 R. Sakamoto
the consequence of ` > 0. In both cases, by the convexity of P
(i)
k (ν) between j ≤ k ≤ ` we have
P
(i)
`−1(ν) ≥ min
{
P
(i)
j (ν), P
(i)
` (ν)
}
= P
(i)
` (ν).
Proof of (2) is the same with the previous case. �
Proof of Proposition 5.35. Since `(i−1) = ` we have m
(i−1)
` (ν) > 0. Then Lemma 3.12 with
m
(i)
` (ν) = 1 claims that P
(i)
`−1(ν), P
(i)
` (ν) and P
(i)
`+1(ν)−2 are strictly convex. The only possibility
that is compatible with the relation P
(i)
`−1(ν) ≥ P (i)
` (ν) < P
(i)
`+1(ν) is P
(i)
`−1(ν) + 1 = P
(i)
` (ν) + 1 =
P
(i)
`+1(ν). Then from Lemma 3.13 we have
−P (i)
`−1(ν) + 2P
(i)
` (ν)− P (i)
`+1(ν) = −1
≥ m(i−1)
` (ν)− 2m
(i)
` (ν) +m
(i+1)
` (ν) = m
(i−1)
` (ν)− 2 +m
(i+1)
` (ν),
that is, 1 ≥ m
(i−1)
` (ν) + m
(i+1)
` (ν). Since we know that m
(i−1)
` (ν) > 0 and m
(i+1)
` (ν) ≥ 0, the
only possibility is m
(i−1)
` (ν) = 1 and m
(i+1)
` (ν) = 0. �
Proof of Proposition 5.30 for the case `(i−1) = `. By (5.19) we have P
(i)
`+1(ν) = x` + 1,
that is, (5.14). By the assumption we have ` − 1 < `(i−1). Then from (5.19) we have
P
(i)
`−1(ν̄) = P
(i)
`−1(ν) = x`, which implies (5.15). Finally m
(i+1)
` (ν) = 0 implies that ` < `(i+1),
that is, (5.16). �
This completes the whole proof of Proposition 5.30.
Again, using the proof of ¯̃ν = ˜̄ν, we can show
¯̃
J = ˜̄J for this case as follows.
Proposition 5.37. Suppose that `(i−1) = ` = `(i) < `(i) and m
(i)
` (ν) = 1. Then we have
¯̃
J = ˜̄J .
Proof. By (5.20) it is enough to analyze the string (`, x`) of (ν, J)(i). Then we have
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→
(
`− 1, P
(i)
`−1(ν̄)
) f̃i7−→
(
`, P
(i)
`−1(ν̄)− 1
)
.
From `(i−1) = ` < `(i+1) by Proposition 5.30, P
(i)
` (ν) = x` by (5.19) and the fact that the length
of the string (`, x`) is preserved under both δ◦f̃i and f̃i◦δ, we have P
(i)
` (¯̃ν) = P
(i)
` (ν)−1 = x`−1.
On the other hand, we have P
(i)
`−1(ν̄)− 1 = P
(i)
`−1(ν)− 1 = x`− 1 since we have `− 1 < `(i−1) and
P
(i)
`−1(ν) = x` by (5.19). �
Example 5.38. Here we give an example for Proposition 5.35. Consider the following rigged
configuration (ν, J) of type (B1,1)⊗3 ⊗B1,3 ⊗B2,1 ⊗B2,2 ⊗B3,1 of D
(1)
5
1
1
1
1
−2
0
0
0
1
−2
0
0
0
−1
−1
−1
1
0
0
0
−1
−1
−1
1
0
0
0
1
1
0
0
0
0
0
0
0
0
0
−1
−1
−1
−1
−1
−1
1
−1
−2
0
−1
−2
The corresponding tensor product Φ−1(ν, J) is
1̄ ⊗ 1̄ ⊗ 3̄ ⊗ 5̄ 4̄ 1̄ ⊗ 2
1̄
⊗ 1 2
5̄ 2̄
⊗
1
5
2̄
Then (ν̄, J̄) is
Rigged Configurations and Kashiwara Operators 59
0
0
0
1
−1
0
0
0
1
−1
0
0
0
−1
−1
−1
1
0
0
0
−1
−1
−1
1
0
0
0
1
1
0
0
0
0
0
0
0
0
0
−1
−1
−1
−1
−1
−1
1
−1
−2
0
−1
−2
and (ν̃, J̃) is
1
1
1
1
−1
0
0
0
1
−1
0
0
0
−1
−1
−2
−1
0
0
0
−1
−1
−2
−1
0
0
0
1
1
1
1
0
0
0
0
0
1
1
−1
−1
−1
−1
−1
−1
1
−1
−2
0
−1
−2
In this example, we have `(1) = ` = `(2) = 3, m
(2)
3 (ν) = 1 and x` = −1. As proved in
Proposition 5.35, we have P
(2)
2 (ν) = P
(2)
3 (ν) = −1 = x`, P
(2)
4 (ν) = 0 = x` + 1, m
(1)
3 (ν) = 1 and
m
(3)
3 (ν) = 0. Finally (¯̃ν,
¯̃
J) = (˜̄ν, ˜̄J) is
0
0
0
1
0
0
0
0
1
0
0
0
0
−1
−1
−2
−1
0
0
0
−1
−1
−2
−1
0
0
0
1
1
1
1
0
0
0
0
0
1
1
−1
−1
−1
−1
−1
−1
1
−1
−2
0
−1
−2
5.12 Proof for Case D (4)
5.12.1 Outline
We will prove the following proposition in the latter part of this subsection.
Proposition 5.39. Suppose that ` = `(i) = `(i) and m
(i)
` (ν) = 2. Then the following relations
hold:
P
(i)
`+1(ν) = x` + 1, (5.21)
P
(i)
`−1(ν̄) = x`, (5.22)
`+ 1 ≤ `(i−1). (5.23)
Proof of “Proposition 5.39 =⇒ ¯̃ν = ˜̄ν”. Step 1. Let us consider the case f̃i ◦ δ. The
first δ creates the two strings
(
`− 1, P
(i)
`−1(ν̄)
)
of (ν̄, J̄)(i). Since the operation δ does not change
the riggings of unchanged strings, the relation (5.22) and m
(i)
` (ν) = 2 imply that the strings
(`− 1, x`) are the longest strings among the strings with rigging x`, hence f̃i will act on one of
these strings.
Step 2. Let us consider the case δ ◦ f̃i. Recall that by definition of f̃i we have ˜̀(a) = `(a)
for all a < i. From Lemma 5.2(1) the relation (5.21) implies that the string (` + 1, x` − 1)
of (ν̃, J̃)(i) created by f̃i is singular. Then from the assumption m
(i)
` (ν) = 2 we see that there
is one singular string of length ` and at least one singular string of length `+ 1 within (ν̃, J̃)(i).
Thus we have ˜̀(a) = `(a) for all i ≤ a, ˜̀(a) = `(a) for all i < a and ˜̀(i) = `+ 1. Then from (5.23)
we have ˜̀(i−1) = `(i−1) and hence ˜̀(a) = `(a) for all a < i.
Step 3. To summarize, we have ¯̃ν
(a)
= ˜̄ν(a)
= ν̄(a) for all a 6= i and ¯̃ν
(i)
= ˜̄ν(i)
are obtained
by removing a box from a length ` row of ν(i). Hence we have ¯̃ν = ˜̄ν. �
60 R. Sakamoto
Proofs of
¯̃
J = ˜̄J will be given by case by case analysis along with the proof of Proposition 5.39
assuming the relation ¯̃ν = ˜̄ν we have just proved.
5.12.2 Proof of Proposition 5.39 for the case `(i−1) < `
We divide the proof of Proposition 5.39 into two cases according to whether `(i−1) < ` or
`(i−1) = `. In this subsection, we consider the case `(i−1) < `. Throughout the proof of
Proposition 5.39, let j be the largest integer such that j ≤ `− 1 and m
(i)
j (ν) > 0. If there is no
such j, set j = 0.
Proposition 5.40. Suppose that ` = `(i) = `(i), m
(i)
` (ν) = 2 and `(i−1) < `. Then we have
P
(i)
`−1(ν) = x` + 1, P
(i)
` (ν) = x`, P
(i)
`+1(ν) = x` + 1. (5.24)
Moreover, we have
m
(i−1)
` (ν) = 0, m
(i+1)
` (ν) = 2. (5.25)
For the proof of this proposition, we prepare the following lemma.
Lemma 5.41. Suppose that ` = `(i) = `(i), m
(i)
` (ν) = 2 and `(i−1) < `. Then we have
(1) P
(i)
`−1(ν) > P
(i)
` (ν), (2) P
(i)
` (ν) < P
(i)
`+1(ν).
Proof. Recall that by definition of δ, the strings
(
`(i), P
(i)
`(i)
(ν)
)
and
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i)
must be different. Then the assumptions ` = `(i) = `(i) and m
(i)
` (ν) = 2 imply that the string
(`, x`) must be singular. In particular, we have P
(i)
` (ν) = x`. Then we can use the same
arguments of Lemma 5.33 to prove the assertions. �
Proof of Proposition 5.40. According to the previous lemma, we can write P
(i)
`−1(ν) and
P
(i)
`+1(ν) as follows:
P
(i)
`−1(ν) = x` + ε1, P
(i)
`+1(ν) = x` + ε2,
where ε1, ε2 ≥ 1. Then Lemma 3.13 with l = ` reads
−P (i)
`−1(ν) + 2P
(i)
` (ν)− P (i)
`+1(ν) = −ε1 − ε2
≥ m(i−1)
` (ν)− 2m
(i)
` (ν) +m
(i+1)
` (ν) = m
(i−1)
` (ν)− 4 +m
(i+1)
` (ν),
that is,
4− ε1 − ε2 ≥ m(i−1)
` (ν) +m
(i+1)
` (ν).
Note that the assumption `(i) = `(i) implies that `(i+1) = `(i+1) = ` by definition of δ. In
particular, we have m
(i+1)
` (ν) ≥ 2. Then the only possibility that is compatible with the above
inequality is ε1 = ε2 = 1, m
(i−1)
` (ν) = 0 and m
(i+1)
` (ν) = 2. �
Proof of Proposition 5.39 for the case `(i−1) < `. The relation (5.21) is proved in (5.24).
As for (5.22), since `(i−1) < ` = `(i) we have P
(i)
`−1(ν̄) = P
(i)
`−1(ν) − 1 = x` by (5.24). Finally,
by (5.25) we have m
(i−1)
` (ν) = 0. Recall that we have `(i−1) ≥ `(i) by definition of δ. Then,
since we are assuming that `(i) = `, we have `(i−1) > `, that is, (5.23). �
Rigged Configurations and Kashiwara Operators 61
Assuming that ¯̃ν = ˜̄ν, we can show
¯̃
J = ˜̄J for this case as follows.
Proposition 5.42. Suppose that `(i−1) < ` = `(i) = `(i) and m
(i)
` (ν) = 2. Then we have
¯̃
J = ˜̄J .
Proof. From the proof of ¯̃ν = ˜̄ν after Proposition 5.39, we see that it is enough to look at the
string (`, x`) of (ν, J)(i) which is changed by both f̃i and δ in both f̃i ◦ δ and δ ◦ f̃i. Then we
have
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→
(
`− 1, P
(i)
`−1(ν̄)
) f̃i7−→
(
`, P
(i)
`−1(ν̄)− 1
)
.
Recall that we have˜̀(i−1) < ˜̀(i) = ˜̀(i+1) = ˜̀
(i+1) = ` < ˜̀(i) = `+ 1 ≤ ˜̀(i−1)
by the assumption `(i−1) < ` = `(i) = `(i) and the arguments given in the proof of ¯̃ν = ˜̄ν. Then
we have P
(i)
` (¯̃ν) = P
(i)
` (ν)− 1 = x` − 1 since we have P
(i)
` (ν) = x` by (5.24). Similarly we have
P
(i)
`−1(ν̄)− 1 = P
(i)
`−1(ν)− 2 = x` − 1 since we have P
(i)
`−1(ν) = x` + 1 by (5.24). �
5.12.3 Proof of Proposition 5.39 for the case `(i−1) = `
Proposition 5.43. Suppose that ` = `(i) = `(i), m
(i)
` (ν) = 2 and `(i−1) = `. Then we have
P
(i)
`−1(ν) = P
(i)
` (ν) = x`, P
(i)
`+1(ν) = x` + 1. (5.26)
Moreover, we have
m
(i−1)
` (ν) = 1, m
(i+1)
` (ν) = 2. (5.27)
For the proof of this proposition, we prepare the following lemma.
Lemma 5.44. Suppose that ` = `(i) = `(i), m
(i)
` (ν) = 2 and `(i−1) = `. Then we have
(1) P
(i)
`−1(ν) ≥ P (i)
` (ν), (2) P
(i)
` (ν) < P
(i)
`+1(ν).
Proof. Since we are assuming that `(i) = `(i) and m
(i)
` (ν) = 2, we see that the strings (`, x`) of
(ν, J)(i) are singular. Then we can use the same arguments of the proof of Lemma 5.36. �
Proof of Proposition 5.43. According to the previous lemma, we can express P
(i)
`−1(ν) and
P
(i)
`+1(ν) as follows:
P
(i)
`−1(ν) = x` + ε1, P
(i)
`+1(ν) = x` + ε2,
where ε1 ≥ 0 and ε2 ≥ 1. Then Lemma 3.13 with l = ` reads
−P (i)
`−1(ν) + 2P
(i)
` (ν)− P (i)
`+1(ν) = −ε1 − ε2
≥ m(i−1)
` (ν)− 2m
(i)
` (ν) +m
(i+1)
` (ν) = m
(i−1)
` (ν)− 4 +m
(i+1)
` (ν),
that is,
4− ε1 − ε2 ≥ m(i−1)
` (ν) +m
(i+1)
` (ν).
Since we are assuming that `(i−1) = `, we have m
(i−1)
` (ν) ≥ 1. Also, from the assumption
`(i) = `(i) we see that `(i+1) = `(i+1) by definition of δ. Thus we have m
(i+1)
` (ν) ≥ 2. Then the
only possibility that is compatible with the above inequality is that ε1 = 0, ε2 = 1, m
(i−1)
` (ν) = 1
and m
(i+1)
` (ν) = 2. �
62 R. Sakamoto
Proof of Proposition 5.39 for the case `(i−1) = `. The relation (5.21) is proved in (5.26).
As for (5.22), since `(i−1) = `(i) = ` we have P
(i)
`−1(ν̄) = P
(i)
`−1(ν) = x` by (5.26). Finally,
by (5.27) we have m
(i−1)
` (ν) = 1. Since we are assuming that `(i−1) = `, this string corresponds
to (`(i−1), x`(i−1)) which is different from the string (`(i−1), x`(i−1)
). Since we have `(i−1) ≤ `(i−1)
by the definition of δ, we must have ` < `(i−1), that is, (5.23). �
Assuming that ¯̃ν = ˜̄ν, we can show
¯̃
J = ˜̄J for this case as follows.
Proposition 5.45. Suppose that ` = `(i) = `(i), m
(i)
` (ν) = 2 and `(i−1) = `. Then we have
¯̃
J = ˜̄J .
Proof. Recall that we can assume that f̃i acts on the same string (`, x`) of (ν, J)(i) in both
f̃i ◦ δ and δ ◦ f̃i. Then we have
(`, x`)
f̃i7−→ (`+ 1, x` − 1)
δ7−→
(
`, P
(i)
` (¯̃ν)
)
,
(`, x`)
δ7−→
(
`− 1, P
(i)
`−1(ν̄)
) f̃i7−→
(
`, P
(i)
`−1(ν̄)− 1
)
.
Recall that we have˜̀(i−1) = ˜̀(i) = ˜̀(i+1) = ˜̀
(i+1) = ` < ˜̀(i) = `+ 1 ≤ ˜̀(i−1)
by the assumption `(i−1) = `(i) = `(i) = ` and the arguments given in the proof of ¯̃ν = ˜̄ν. Then
we have P
(i)
` (¯̃ν) = P
(i)
` (ν)− 1 = x` − 1 since we have P
(i)
` (ν) = x` by (5.26). Similarly we have
P
(i)
`−1(ν̄)− 1 = P
(i)
`−1(ν)− 1 = x` − 1 since we have P
(i)
`−1(ν) = x` by (5.26). �
Example 5.46. Consider the following rigged configuration (ν, J) of type (B1,1)⊗3 ⊗ B1,3 ⊗
B2,1 ⊗B2,2 ⊗B3,1 of D
(1)
5
2
1
1
−2
0
0
1
−2
−1
−1
−1
−1
−1
1
−1
−1
−1
−1
−1
1
2
1
1
−1
−2
1
1
1
−1
−2
1
1
1
1
−1
−1
−1
−1
−1
−1
The corresponding tensor product Φ−1(ν, J) is
1̄ ⊗ 1̄ ⊗ 1̄ ⊗ 1 1 4 ⊗ 3̄
2̄
⊗ 2 3̄
1̄ 1̄
⊗
1
3
5̄
Then (ν̄, J̄) is
1
0
0
−2
0
0
0
−2
−1
−1
−1
−1
−1
1
−1
−1
−1
−1
−1
1
2
2
2
−1
−2
1
2
2
−1
−2
1
1
1
1
−1
−1
−1
−1
−1
−1
and (ν̃, J̃) is
2
1
1
−1
0
0
1
−1
−1
−1
−1
−1
−2
−1
−1
−1
−1
−1
−2
−1
2
1
1
0
−1
1
1
1
0
−1
1
1
1
1
−1
−1
−1
−1
−1
−1
Rigged Configurations and Kashiwara Operators 63
In this example, we have ` = `(1) = `(2) = `(2) = 3 and m
(2)
3 (ν) = 2. Then we have x` = −1,
P
(2)
2 (ν) = P
(2)
3 (ν) = −1, P
(2)
4 (ν) = 0, m
(1)
3 (ν) = 1 and m
(3)
3 (ν) = 2 (see Proposition 5.43). We
also have P
(2)
2 (ν̄) = −1 and `(1) = 6 ≥ `+ 1 = 4 (see Proposition 5.39). Finally (¯̃ν,
¯̃
J) = (˜̄ν, ˜̄J)
is
1
0
0
−1
0
0
0
−1
−1
−1
−1
−1
−2
−1
−1
−1
−1
−1
−2
−1
2
2
2
0
−1
1
2
2
0
−1
1
1
1
1
−1
−1
−1
−1
−1
−1
5.13 Proof for Case E: preliminary steps
5.13.1 Classif ication
Recall that the defining condition for Case E is `(i) < `. We further classify this case as follows.
(1) `(i) < ` and `+ 1 < `(i+1),
(2) `(i) < ` and `(i+1) ≤ `+ 1 < `(i) =∞,
(3) `(i) < ` and `(i+1) ≤ `+ 1 ≤ `(i) <∞,
(4) `(i) < ` and `(i) = `,
(5) `(i) < ` and `(i) < `.
5.13.2 A common property for Case E
The fundamental observation for this case is the following property.
Proposition 5.47. If `(i) < `, we have
P
(i)
`(i)−1
(ν̄) ≥ x`.
Proof. We follow the classification of Lemma 5.1.
Case (I). Let us consider the case `(i−1) = `(i). Then by Lemma 5.1, we have to show
P
(i)
`(i)−1
(ν) ≥ x`. Suppose if possible that P
(i)
`(i)−1
(ν) < x`. If `(i) = 1, then we have 0 =
P
(i)
`(i)−1
(ν) < x` which is a contradiction since we have x` ≤ 0 by 0 < `. Therefore we assume
`(i) ≥ 2 in the sequel. From the assumption `(i) < ` we have P
(i)
`(i)
(ν) ≥ x`(i) ≥ x`. Let j be the
length of the longest string of ν(i) that are strictly shorter than `(i). If j = `(i)−1, then we have
xj ≤ P
(i)
j (ν) < x`, which contradicts the minimality of x`. Thus we assume that j < `(i) − 1.
We apply the convexity relation of P
(i)
k for j ≤ k ≤ `(i) with the relation P
(i)
`(i)−1
(ν) < P
(i)
`(i)
(ν)
and obtain xj ≤ P
(i)
j (ν) < P
(i)
`(i)−1
(ν) < x`, which is a contradiction. Therefore we see that
ν(i) does not contain strings that are strictly shorter than `(i). Again we can use the convexity
relation of P
(i)
k for 0 ≤ k ≤ `(i) with the relation P
(i)
`(i)−1
(ν) < P
(i)
`(i)
(ν) and obtain the relation
0 = P
(i)
0 (ν) < P
(i)
`(i)−1
(ν) < x`. This is a contradiction since we have x` ≤ 0. Thus we have
proved P
(i)
`(i)−1
(ν̄) ≥ x` for the case `(i−1) = `(i).
Case (II). Let us consider the case `(i−1) < `(i). Then by Lemma 5.1, we have to show
P
(i)
`(i)−1
(ν) > x`. Suppose if possible that P
(i)
`(i)−1
(ν) ≤ x`. Let j be the length of the longest
row of ν(i) that are strictly shorter than `(i). Suppose that j = `(i) − 1. Then we have xj ≤
P
(i)
j (ν) ≤ x`. By the minimality of x` this relation implies that xj = P
(i)
j (ν), that is, the string
64 R. Sakamoto
(j, xj) is singular in (ν, J)(i). This is in contradiction to the definition of `(i) since we have
`(i−1) ≤ j = `(i) − 1 < `(i).
Therefore we assume j < `(i)−1 in the sequel. If P
(i)
`(i)−1
(ν) < x`, from P
(i)
`(i)
(ν) ≥ x`(i) ≥ x` we
have P
(i)
`(i)−1
(ν) < P
(i)
`(i)
(ν) and by the convexity relation we have xj ≤ P (i)
j (ν) < P
(i)
`(i)−1
(ν), that
is xj < x`. This is a contradiction. Therefore we must have P
(i)
`(i)−1
(ν) = x`. If P
(i)
`(i)
(ν) > x`,
by the similar argument we obtain xj < x`, which is a contradiction. Thus we conclude that
P
(i)
`(i)
(ν) = x`. Suppose that `(i−1) ≤ j. Then from the convexity relation of P
(i)
k (ν) for j ≤ k ≤
`(i), we obtain P
(i)
j (ν) ≤ P
(i)
`(i)−1
(ν) = P
(i)
`(i)
(ν) = x` and from the requirement P
(i)
j (ν) ≥ xj ≥ x`
we obtain P
(i)
j (ν) = xj , that is, the string (j, xj) is singular. But this is a contradiction since
`(i−1) ≤ j < `(i). Thus we have j < `(i−1). Since m
(i−1)
`(i−1)(ν) > 0 the convexity relation for P
(i)
k (ν)
between `(i−1)−1 ≤ k ≤ `(i−1)+1 must be strict. Thus we obtain x` ≥ P
(i)
`(i−1)(ν) > P
(i)
j (ν) ≥ xj ,
which is a contradiction.
Therefore we conclude that (ν, J)(i) does not contain strings that are strictly shorter than `(i).
Then from the convexity relation of P
(i)
k (ν) between 0 ≤ k ≤ `(i), we obtain x` ≥ P
(i)
`(i)−1
(ν) ≥
P
(i)
`(i−1)(ν) > P
(i)
0 (ν) = 0, which is in contradiction to the requirement x` ≤ 0. Thus we have
proved P
(i)
`(i)−1
(ν̄) ≥ x` for the case `(i−1) < `(i). �
5.14 Proof for Case E (1)
To begin with we prepare the following property.
Proposition 5.48. Suppose that `(i) < ` and `+ 1 < `(i+1). Then we have
P
(i)
`(i)−1(ν̄) > x`.
Proof. We follow the classification of Lemma 5.1. From the assumption, we see that `(i) <
`(i+1). Then we only need to consider cases (V), (VI) and (VII). During the proof of the
proposition, let j be the largest integer such that j < `(i) and m
(i)
j (ν) > 0. Since `(i) < ` <
`(i+1) ≤ `(i) we have ` ≤ j. Proofs for cases (V) and (VII) are the same as the corresponding
parts of the proof of Proposition 5.3.
Case (VI). In this case, we have to show P
(i)
`(i)−1(ν) > x`. Suppose if possible that
P
(i)
`(i)−1(ν) ≤ x`. First, let us consider the case j = `(i)−1. Then the corresponding string satisfies
x`(i)−1 ≤ P
(i)
`(i)−1(ν) ≤ x`. Now recall that we have ` < `(i)− 1 since we have `+ 1 < `(i+1) ≤ `(i)
by the assumption. Thus the minimality of x` requires that x` < x`(i)−1. This is a contradiction.
Similarly, if ` < j < `(i) − 1, we have x` ≥ P
(i)
`(i)−1(ν) ≥ min
{
P
(i)
`(i)
(ν), P
(i)
j (ν)
}
which implies
that x` ≥ x`(i) or x` ≥ xj . This is in contradiction to the requirements x` < x`(i) and x` < xj
derived from ` < `(i) and ` < j, respectively.
Finally let us consider the case ` = j. Recall that we have P
(i)
` (ν) ≥ x`, P
(i)
`(i)−1(ν) ≤ x` and
P
(i)
`(i)
(ν) > x`. However these relations are in contradiction to the convexity relation of P
(i)
k (ν)
between ` ≤ k ≤ `(i) which we can use by ` < `(i) − 1. �
Proposition 5.49. Suppose that `(i) < ` and ` + 1 < `(i+1). Then we have the following
identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Rigged Configurations and Kashiwara Operators 65
Proof. Step 1. Let us consider the case δ ◦ f̃i. Since f̃i does not change coriggings of the
untouched strings, and since f̃i creates the string whose length satisfies `(i) < `+ 1 < `(i+1), we
have ˜̀(a) = `(a) and ˜̀(a) = `(a) for all a.
Step 2. Let us consider the case f̃i ◦ δ. Since we are assuming that `(i) < ` < `(i) the string
(`, x`) remains as it is after δ. Recall that we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47 and
P
(i)
`(i)−1(ν̄) > x` by the previous proposition. Thus the string (`, x`) has the largest length among
the strings of (ν̃, J̃)(i) that have the smallest rigging.
Step 3. To summarize, we have ¯̃ν
(a)
= ˜̄ν(a)
= ν̄(a) for all a 6= i and ¯̃ν
(i)
= ˜̄ν(i)
is obtained
by adding a box to the `-th column of ν̄(i). The statement
¯̃
J = ˜̄J is obtained by the equality
P
(i)
k (¯̃ν) = P
(i)
k (˜̄ν) for arbitrary k which is the consequence of ¯̃ν = ˜̄ν. �
5.15 Proof for Case E (2)
Let us consider the case `(i) < ` and `(i+1) ≤ ` + 1 < `(i) = ∞. We divide the proof into two
cases according to whether f̃i creates a singular string or not.
5.15.1 Case 1: f̃i creates a non-singular string
Proposition 5.50. Suppose that we have `(i) < ` and `(i+1) ≤ `+1 < `(i) =∞. Suppose further
that f̃i acting on (ν, J) creates a non-singular string. Then we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. Step 1. Let us consider the case δ ◦ f̃i. Since f̃i does not change coriggings of the
untouched strings, and since f̃i creates a non-singular string, we have ˜̀(a) = `(a) and ˜̀(a) = `(a)
for all a.
Step 2. Let us consider the case f̃i ◦ δ. Since we are assuming that `(i) < ` and `(i) =∞ the
string (`, x`) remains as it is after δ. Recall that we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47.
Thus the string (`, x`) has the largest length among the strings of (ν̃, J̃)(i) that have the smallest
rigging.
Step 3. To summarize, we have ¯̃ν
(a)
= ˜̄ν(a)
= ν̄(a) for all a 6= i and ¯̃ν
(a)
= ˜̄ν(a)
is obtained
by adding a box to the `-th column of ν̄(i). The statement
¯̃
J = ˜̄J follows from ¯̃ν = ˜̄ν. �
5.15.2 Case 2: f̃i creates a singular string
Proposition 5.51. Suppose that we have `(i) < ` and `(i+1) ≤ `+1 < `(i) =∞. Suppose further
that f̃i acting on (ν, J) creates a singular string. Then we have
f̃i ◦ δ(ν, J) = 0.
Proof. Since `(i) = ∞, δ changes the string
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i) only. In particular,
from `(i) < ` the string (`, x`) remains as it is after the application of δ. Since we know that
P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47, the string (`, x`) is the largest string among the strings
with the smallest rigging within (ν̄, J̄)(i). Thus the next f̃i acts on this string and produces
(`+ 1, x` − 1).
Let us compute P
(i)
`+1(˜̄ν). Since f̃i acting on (ν, J) creates a singular string, we have P
(i)
`+1(ν) =
x` + 1 by Lemma 5.2. Then we have P
(i)
`+1(ν̄) = P
(i)
`+1(ν) − 1 = x` since we have `(i−1) ≤ `(i) ≤
66 R. Sakamoto
`(i+1) ≤ ` + 1 and `(i+1) ≤ ` + 1 < `(i) = ∞ by the assumptions. Finally, since f̃i adds a box
to the (` + 1)-th column of (ν̄, J̄)(i), we have P
(i)
`+1(˜̄ν) = P
(i)
`+1(ν̄) − 2 = x` − 2. Then we see
that the string (`+ 1, x` − 1) of (˜̄ν, ˜̄J)(i) satisfies x` − 1 > P
(i)
`+1(˜̄ν). Therefore we conclude that
f̃i ◦ δ(ν, J) = 0. �
Example 5.52. Consider the following rigged configuration (ν, J) of type (B1,1)⊗4 ⊗ B1,3 ⊗
B2,1 ⊗B2,2 ⊗B3,1 of D
(1)
5
1
1
0
0
0
0
0
0
−1
0
1
1
1
1
1
0
0
1
1
1
1
0
1
1
0
0
0
−1
0
0
−1
−1
0
−1
0
0
0
−1
0
0
0
0
−2
0
0
−2
f̃2 acts on the string (`, x`) = (4, 0) of (ν, J)(2) and makes it into (5,−1) of (ν̃, J̃)(2). Since we
have P
(2)
5 (ν̃) = −1, the latter string is singular. In particular, we see that f̃2(ν, J) 6= 0. The
corresponding tensor product Φ−1(ν, J) is
3̄ ⊗ 1̄ ⊗ 5̄ ⊗ 3 ⊗ 1 1 1̄ ⊗ 1
3̄
⊗ 1 5̄
3 1̄
⊗
5
5̄
3̄
Then (ν̄, J̄) is
0
0
0
0
0
0
0
0
−1
0
1
1
1
1
1
0
0
1
1
1
1
0
1
1
0
−1
−1
−1
0
0
0
−1
−1
−1
0
0
1
−1
0
0
0
0
−2
0
0
−2
Thus we have `(2) = 2 < ` = 4 and `(3) = 5 = `+ 1 < `(2) =∞ (see Proposition 5.51). Indeed,
we have P
(2)
5 (˜̄ν) = −2 and thus f̃2(ν̄, J̄) = 0.
5.16 Proof for Case E (3)
Let us consider the case `(i) < ` and `(i+1) ≤ ` + 1 ≤ `(i) < ∞. As in Case C, we use the
following classification:
(1) m
(i)
`+1(ν) > 0 and f̃i creates a singular string, or
`+ 1 = `(i) and f̃i creates a non-singular string,
(2) m
(i)
`+1(ν) = 0 and f̃i creates a singular string,
(3) `+ 1 < `(i) and f̃i creates a non-singular string.
5.16.1 Proof for Case 1
Proposition 5.53. Suppose that `(i) < `, `(i+1) ≤ `+ 1 ≤ `(i) <∞, m
(i)
`+1(ν) > 0 and f̃i creates
a singular string. Then we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Rigged Configurations and Kashiwara Operators 67
Proof. Since we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47, we can use the same arguments of
the proof of Proposition 5.9 if we replace `(i) (resp. `(i−1)) there by `(i) (resp. `(i+1)) and neglect
the arguments concerning the string `(i) there. �
Proposition 5.54. Suppose that `(i) < `, `(i+1) ≤ `+ 1 ≤ `(i) <∞, `+ 1 = `(i) and f̃i creates
a non-singular string. Then we have the following identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. As in Case C we can reduce the proof to the previous proposition. �
5.16.2 Proof for Case 2
As in Case C, we show the following property:
Proposition 5.55. Suppose that `(i) < `, `(i+1) ≤ ` + 1 < `(i) < ∞ and m
(i)
`+1(ν) = 0. If f̃i
acting on (ν, J) creates a singular string, then the following two conditions are satisfied:
P
(i)
`(i)−1(ν̄) = x`, m
(i−1)
k (ν) = 0 for ` < k < `(i).
Proof. As in Case C, the proof depend on the following two cases (see Lemma 5.13);
(`, x`) is singular and ` = `(i+1) − 1, (5.28)
(`, x`) is non-singular. (5.29)
For the case (5.28) (resp. (5.29)) we can use the same arguments of Section 5.6.3 (resp. Sec-
tion 5.6.4) by replacing `(i) (resp. `(i−1)) there by `(i) (resp. `(i+1)) to show the assertions. �
Proposition 5.56. Suppose that `(i) < `, `(i+1) ≤ `+ 1 < `(i) <∞ and m
(i)
`+1(ν) = 0. Then we
have ¯̃ν = ˜̄ν and
¯̃
J = ˜̄J .
Proof. Since we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47, we can use the same arguments of
the proof of Proposition 5.12 if we replace `(i) (resp. `(i−1)) there by `(i) (resp. `(i+1)), neglect the
arguments concerning the string `(i) there and use Proposition 5.55 instead of Proposition 5.11.
Thus we obtain ¯̃ν = ˜̄ν. We can also use the same analysis given for Case C to show
¯̃
J = ˜̄J . �
5.16.3 Proof for Case 3
We can use the same arguments in Section 5.7 if we replace `(i) (resp. `(i−1)) there by `(i)
(resp. `(i+1)) and neglect the arguments concerning the string `(i) there.
5.17 Proof for Case E (4)
5.17.1 Preliminary
In this case, we assume that `(i) < ` and `(i) = `. In view of the classification of Case D we have
to consider the following two cases:
1) `(i) < `, `(i) = ` and m
(i)
` (ν) > 1,
3) `(i) < `, `(i) = ` and m
(i)
` (ν) = 1.
During the proof of Case E (4) we follow the above classification. In both cases, we have the
following common property.
68 R. Sakamoto
Proposition 5.57. Assume that `(i) < ` and `(i) = `. Then the following relation holds:
P
(i)
`−1(ν̄) ≥ x`.
Proof. We can use the same arguments in the proof of Proposition 5.23 if we replace `(i)
(resp. `(i−1)) there by `(i) (resp. `(i+1)). Here in order to evaluate the vacancy numbers in
Case (a) and (b), we note that if we have ` − 1 < `(i+1) we have P
(i)
`−1(ν̄) ≥ P
(i)
`−1(ν) since we
have `(i) ≤ `− 1 by the assumption. �
5.17.2 Proof for Case 1
Proposition 5.58. Assume that `(i) < `, `(i) = ` and m
(i)
` (ν) > 1. Then we have the following
identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Proof. (1) Let us analyze the action of f̃i before and after the application of δ. By the as-
sumption m
(i)
` (ν) > 1 we can choose two distinct length ` strings (`, x`) and
(
`, P
(i)
` (ν)
)
of
(ν, J)(i) where f̃i will act on the former one and δ will act on the latter one. Recall that we
have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47 and P
(i)
`(i)−1(ν̄) ≥ x` by Proposition 5.57. Thus both
strings
(
`(i)− 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i)− 1, P
(i)
`(i)−1(ν̄)
)
of (ν̄, J̄)(i) are shorter than (`, x`) and their
riggings are larger than or equal to x`. Therefore f̃i will act on the same string (`, x`) in both
(ν, J) and (ν̄, J̄).
Let us analyze δ. Recall that f̃i will not change the coriggings of the strings
(
`(i), P
(i)
`(i)
(ν)
)
and
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i). Since `(i) < ` we have ˜̀(a) = `(a) for all a. Also, since `(i) < `+ 1
we have ˜̀(a) = `(a) for all a. To summarize δ chooses the same strings before and after the
application of f̃i. Hence we obtain ¯̃ν = ˜̄ν.
(2) Since each f̃i and δ chooses the same strings in δ ◦ f̃i and f̃i ◦ δ, the coincidence
¯̃
J = ˜̄J is
a consequence of P
(i)
k (¯̃ν) = P
(i)
k (˜̄ν). �
5.17.3 Proof for Case 3
Proposition 5.59. Assume that `(i) < `, `(i) = ` and m
(i)
` (ν) = 1. Then we have the following
identities:
P
(i)
`+1(ν) = x` + 1, (5.30)
P
(i)
`−1(ν̄) = x`, (5.31)
`+ 1 ≤ `(i−1). (5.32)
Proof. The proof is divided into two cases `(i+1) < ` and `(i+1) = ` and we can use the same
arguments in the proof of Proposition 5.30 if we replace `(i−1), `(i) and `(i+1) there by `(i+1), `(i)
and `(i−1), respectively. �
Proposition 5.60. Assume that `(i) < `, `(i) = ` and m
(i)
` (ν) = 1. Then we have the following
identities:
(1) ¯̃ν = ˜̄ν, (2)
¯̃
J = ˜̄J.
Rigged Configurations and Kashiwara Operators 69
Proof. Recall that we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47. Since we have P
(i)
`(i)−1(ν̄) = x`
by (5.31), the string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
of (ν̄, J̄)(i) will not be chosen by the next f̃i. Then
we can use the same arguments in the corresponding parts of Case D (3) if we replace `(i)
(resp. `(i−1)) there by `(i) (resp. `(i+1)) and neglect all the arguments related to `(a). �
5.18 Proof for Case E (5)
In this case, we can show the following property.
Proposition 5.61. Suppose that `(i) < `. Then we have
P
(i)
`(i)−1(ν̄) ≥ x`.
Proof. We can use the arguments of the proof of Proposition 5.47 if we replace `(i) (resp. `(i−1))
there by `(i) (resp. `(i+1)). �
As the consequence of this proposition, we have the following result.
Proposition 5.62. Suppose that `(i) < ` and `(i) < `. Then we have ¯̃ν = ˜̄ν and
¯̃
J = ˜̄J .
Proof. Since we have P
(i)
`(i)−1
(ν̄) ≥ x` by Proposition 5.47 and P
(i)
`(i)−1(ν̄) ≥ x` by Proposi-
tion 5.61, we see that f̃i acts on the same string before and after δ. Also, since `(i) ≤ `(i) < `
and the fact that f̃i does not change coriggings of untouched strings, we see that δ acts on
the same strings before and after f̃i. Thus we have ¯̃ν = ˜̄ν.
¯̃
J = ˜̄J follows from the fact
P
(i)
k (¯̃ν) = P
(i)
k (˜̄ν) which is the consequence of ¯̃ν = ˜̄ν. �
6 Proof of Proposition 4.8: f̃i version
6.1 Proof for (1)
Proposition 6.1. Let us consider the rigged configuration (ν, J) of type B1,1⊗ B̄. Suppose that
we have the commutativity of f̃i and Φ for B̄. Suppose that we have f̃i(ν, J) 6= 0 and f̃i(ν̄, J̄) = 0.
Let b = Φ(ν, J) and b′ = Φ(ν̄, J̄). Then we have one of the following two cases:
(1) b = i⊗ b′, f̃i(b) is defined, f̃i(b
′) is undefined and Φ
(
f̃i(ν, J)
)
= (i+ 1)⊗ b′,
(2) b = i+ 1⊗ b′, f̃i(b) is defined, f̃i(b
′) is undefined and Φ
(
f̃i(ν, J)
)
= i⊗ b′.
For the proof, we start by preparing the following properties.
Proposition 6.2. f̃i(ν, J) 6= 0 and f̃i(ν̄, J̄) = 0 if and only if one the following conditions is
satisfied:
(1) `(i−1) ≤ `+ 1 < `(i) =∞ and f̃i acting on (ν, J) creates a singular string,
(2) `(i) < `, `(i+1) ≤ `+ 1 < `(i) =∞ and f̃i acting on (ν, J) creates a singular string.
Proof. In Section 5, we analyze all possible cases such that f̃i(ν, J) 6= 0. Then only cases
such that f̃i(ν, J) is defined and f̃i(ν̄, J̄) is undefined are the above two cases as described in
Proposition 5.6 and Proposition 5.51. �
70 R. Sakamoto
Lemma 6.3.
(1) Suppose that `(i−1) ≤ ` + 1 < `(i) = ∞ and f̃i acting on (ν, J) creates a singular string.
Then we have ˜̀(i) = `+ 1, ˜̀(i+1) =∞ and P
(i)
` (ν̄) = x`.
(2) Suppose that `(i) < `, `(i+1) ≤ ` + 1 < `(i) = ∞ and f̃i acting on (ν, J) creates a singular
string. Then we have ˜̀(i) = `+ 1, ˜̀(i−1) =∞ and P
(i)
` (ν̄) = x`.
Proof. (1) To begin with we shall show that there is no string of ν̄(i) that is longer than `.
Suppose if possible that such string exist. Let j be the minimal integer satisfying j > ` and
m
(i)
j (ν) > 0. Then there are strings (`, x`) and (j, xj) of (ν, J)(i) that satisfy xj > x` by definition
of `. Since `(i) = ∞ we have (ν̄, J̄)(i) = (ν, J)(i). Thus the string (j, xj) still exists in (ν̄, J̄)(i),
which implies that P
(i)
j (ν̄) ≥ xj . To summarize, we obtain P
(i)
j (ν̄) > x`.
Consider the case j = `+ 1. Then this result is in contradiction to the equality P
(i)
`+1(ν̄) = x`
that appears in the proof of Proposition 5.6. Next consider the general case j > ` + 1. By
the convexity relation between ` and j, we have P
(i)
`+1(ν̄) ≥ min
{
P
(i)
` (ν̄), P
(i)
j (ν̄)
}
. Since we
know that P
(i)
`+1(ν̄) = x` < P
(i)
j (ν̄), we deduce that P
(i)
`+1(ν̄) ≥ P
(i)
` (ν̄). From P
(i)
`+1(ν̄) = x`
and P
(i)
` (ν̄) ≥ x` (since `(i) = ∞), we obtain P
(i)
` (ν̄) = P
(i)
`+1(ν̄) = x`. Then the convexity
relation between ` and j gives x` = P
(i)
` (ν̄) ≥ P
(i)
j (ν̄) which is in contradiction to the relation
P
(i)
j (ν̄) > x`. Hence there is no string of ν̄(i) that is longer than `.
The relations P
(i)
` (ν̄) ≥ x`, P
(i)
`+1(ν̄) = x`, P
(i)
∞ (ν̄) > −∞ and the convexity relation bet-
ween ` and ∞ gives P
(i)
` (ν̄) = P
(i)
`+1(ν̄) = P
(i)
`+2(ν̄) = · · · . This relation forces m
(i+1)
k (ν̄) = 0 for
all k ≥ ` + 1, since otherwise the relation would be strictly convex. Since `(i) = ∞, we have
ν̄(i+1) = ν(i+1) and thus ν̄(i+1) = ν̃(i+1). Therefore we obtain m
(i+1)
k (ν̃) = 0 for all k ≥ `+ 1.
Recall that the definition of f̃i on the rigged configurations implies that ˜̀(a) = `(a) for all
a ≤ i− 1. Then by the assumption we have ˜̀(i−1) ≤ `+ 1. Again from the assumption there is
a length `+ 1 singular string in ν̃(i). Therefore we have ˜̀(i) = `+ 1. Finally from m
(i+1)
k (ν̃) = 0
for all k ≥ `+ 1 we have ˜̀(i+1) =∞.
(2) During the proof of Proposition 5.51 we show P
(i)
`+1(ν̄) = x`. Then we can use the same
arguments of (1) if we replace `(i−1), `(i) and `(i+1) etc. there by `(i+1), `(i) and `(i−1) etc.,
respectively. Note that in this case we have to work under the condition `(i) < ` and `(i) =∞.
Then we can use the fact that all strings of (ν, J)(i−1) that are strictly longer than `(i−1)(< `+1)
do not change after δ. In particular, we can show m
(i−1)
k (ν̄) = m
(i−1)
k (ν̃) = 0 for all k ≥ `+1. �
Proof of Proposition 6.1. According to the previous observations, we have to consider two
distinct cases.
(1) Suppose that `(i−1) ≤ ` + 1 < `(i) = ∞ and f̃i acting on (ν, J) creates a singular string.
From `(i−1) < ∞ and `(i) = ∞ we have Φ(ν, J) = i ⊗ b′. From ˜̀(i) < ∞ and ˜̀(i+1) = ∞ we
have Φ(ν̃, J̃) = Φ ◦ f̃i(ν, J) = (i + 1) ⊗ b′′. Let us show the coincidence b′′ = b′ by proving
(ν̄, J̄) = (¯̃ν,
¯̃
J). Recall that the difference between (ν, J) and (ν̃, J̃) is one box at (` + 1)-th
column of (ν̃, J̃)(i). Then from ˜̀(a) = `(a) for all a ≤ i− 1, ˜̀(i) = `+ 1 and ˜̀(i+1) = `(i) =∞, we
have ν̄ = ¯̃ν. In order to show J̄ =
¯̃
J recall that δ acting on (ν̃, J̃) creates the length ` singular
string of (¯̃ν,
¯̃
J)(i). On the other hand, since P
(i)
` (ν̄) = x` all length ` strings of (ν̄, J̄)(i) are
singular. From P
(i)
` (ν̄) = P
(i)
` (¯̃ν) we see the coincidence of the corresponding riggings and hence
we deduce that J̄ =
¯̃
J . Thus we have b′′ = b′.
As for b′ = Φ(ν̄, J̄), we know that ϕi(b
′) = ϕi(ν̄, J̄) = 0 by the assumption. Thus f̃i◦Φ(ν, J) =
f̃i(i)⊗ b′ = (i+ 1)⊗ b′. Thus Φ and f̃i commutes in this case.
Rigged Configurations and Kashiwara Operators 71
(2) Suppose that `(i) < `, `(i+1) ≤ `+ 1 < `(i) =∞ and f̃i acting on (ν, J) creates a singular
string. From `(i+1) < ∞ and `(i) = ∞ we have Φ(ν, J) = i+ 1 ⊗ b′. From ˜̀
(i) < ∞ and˜̀
(i−1) = ∞ we have Φ(ν̃, J̃) = i ⊗ b′′. The coincidence b′′ = b′ follows from parallel arguments.
Finally, by ϕi(b
′) = ϕi(ν̄, J̄) = 0 we have f̃i ◦ Φ(ν, J) = f̃i(i+ 1) ⊗ b′ = i ⊗ b′. Thus Φ and f̃i
commutes in this case also. �
6.2 Proof for (2)
To begin with, let us show the following property.
Lemma 6.4. Suppose that `(i+1) < ∞, `(i) = ∞, ˜̀(i) < ∞ and ˜̀(i−1) = ∞. Let s (resp. s̄) be
the smallest rigging of (ν, J)(i) (resp. (ν̄, J̄)(i)). Then we have s ≤ s̄.
Proof. Note that if we show that the rigging for the shortened string in (ν̄, J̄)(i) is larger than
or equal to s, then we have s ≤ s̄. In particular, we can assume that `(i) > 1, since otherwise the
corresponding rigging will be erased by δ and will not contribute to s̄, thus s = s̄ by `(i) = ∞.
Since `(i) =∞, it is enough to check
P
(i)
`(i)−1
(ν̄) ≥ s.
Then we have the two possibilities:
(I) `(i−1) = `(i) or (II) `(i−1) < `(i).
During the proof, let j be the largest integer such that j < `(i) and m
(i)
j (ν) > 0. If there is no
such j, we set j = 0.
Case (I). In this case we have to show that P
(i)
`(i)−1
(ν) ≥ s under the assumption `(i−1) = `(i).
Let us consider the case j = `(i) − 1. Then we are done since we have P
(i)
`(i)−1
(ν) ≥ x`(i)−1 ≥ s
by the minimality of s. Next consider the case 0 < j < `(i) − 1. Then by the convexity relation
between j and `(i) we have P
(i)
`(i)−1
(ν) ≥ min
{
P
(i)
j (ν), P
(i)
`(i)
(ν)
}
≥ s. Suppose that j = 0. If
s ≤ 0, we can use the same arguments of the previous case since P
(i)
0 (ν) = 0.
Therefore we have to consider the case j = 0 and s > 0. Since s > 0, f̃i acting on (ν, J) adds
the length 1 string to (ν, J)(i). Recall that f̃i does not change coriggings of untouched strings.
Since we have `(i−1) = `(i) > 1, we see that ˜̀(a) = `(a) and ˜̀(a) = `(a) for all a. However this is
in contradiction to the assumptions `(i) =∞ and ˜̀(i) <∞. Hence this case cannot happen.
Case (II). In this case we have to show that P
(i)
`(i)−1
(ν) > s under the assumption `(i−1) < `(i).
Suppose if possible that P
(i)
`(i)−1
(ν) ≤ s.
Suppose that j = `(i) − 1. Then we have x`(i)−1 ≤ P
(i)
`(i)−1
(ν) ≤ s. Then by the minimality of
s, we have x`(i)−1 = P
(i)
`(i)−1
(ν) = s, in particular, the string (`(i)−1, x`(i)−1) is singular. However
this is a contradiction since its length satisfies `(i−1) ≤ `(i) − 1 < `(i).
Suppose that `(i−1) ≤ j < `(i) − 1. By definition of s we have P
(i)
`(i)
(ν) ≥ s and by the
assumption we have P
(i)
`(i)−1
(ν) ≤ s. Then by the convexity relation of P
(i)
k (ν) between j ≤ k ≤
`(i), we have P
(i)
`(i)
(ν) ≥ P (i)
`(i)−1
(ν) ≥ P (i)
j (ν) ≥ xj . By the minimality of s and s ≥ P (i)
`(i)−1
(ν), we
have P
(i)
j (ν) = xj , that is, the string (j, xj) is singular. This contradicts the definition of `(i).
Suppose that 0 < j < `(i−1). By m
(i−1)
`(i−1)(ν) > 0, the convexity relation of P
(i)
k (ν) between
`(i−1) + 1 ≥ k ≥ `(i−1) − 1 has to be strictly convex. Therefore we have P
(i)
`(i)
(ν) ≥ P
(i)
`(i)−1
(ν) >
P
(i)
j (ν) ≥ xj in this case. Since we have s ≥ P (i)
`(i)−1
(ν), this contradicts the minimality of s.
72 R. Sakamoto
Suppose that j = 0. Then we have s ≥ P
(i)
`(i)−1
(ν) > P
(i)
0 (ν) = 0. If s ≤ 0, this is a contra-
diction. Hence we assume that s > 0. Then f̃i adds the length 1 string (1,−1) to (ν, J)(i). If
`(i−1) > 1, we can use the same arguments in the last part of Case (I) to show that such case
cannot happen. Thus assume that `(i−1) = 1. If P
(i)
1 (ν̃) > −1, then the added string (1,−1)
is non-singular so that we can use the arguments in the last part of Case (I) to show that such
case cannot happen.
To summarize, we have shown that we have s ≤ s̄ except for the case j = 0, s > 0, `(i−1) = 1
and P
(i)
1 (ν̃) = −1. Suppose if possible that this case happen. Since s > 0, f̃i adds the length 1
string (1,−1) to (ν, J)(i). Thus we have s̄ < 0 and P
(i)
1 (ν) = 1. Since m
(i−1)
`(i−1)(ν) = m
(i−1)
1 (ν) > 0
and `(i) > 1, the relation P
(i)
k (ν) for 0 ≤ k ≤ 2 must be strictly convex. In view of P
(i)
0 (ν) = 0,
this implies that 1 = P
(i)
1 (ν) ≥ P
(i)
2 (ν). Since we are assuming that s > 0, the rigging of the
string
(
`(i), P
(i)
`(i)
(ν)
)
must satisfy P
(i)
`(i)
(ν) > 0. Then by the convexity relation of P
(i)
k (ν) between
1 ≤ k ≤ `(i) the only possibility that is compatible with 1 = P
(i)
1 (ν) ≥ P
(i)
2 (ν) is the case
1 = P
(i)
1 (ν) = P
(i)
2 (ν) = · · · = P
(i)
`(i)
(ν). To summarize, we have the following situation:
s > 0, `(i−1) = 1, `(i) > 1 and P
(i)
1 (ν) = P
(i)
2 (ν) = · · · = P
(i)
`(i)
(ν) = 1. (6.1)
Since j = 0 we have m
(i)
1 (ν) = 0. Then from the convexity relation of Lemma 3.13 with l = 1
we have
−P (i)
0 (ν) + 2P
(i)
1 (ν)− P (i)
2 (ν) = 1
≥ m(i−1)
1 (ν)− 2m
(i)
1 (ν) +m
(i+1)
1 (ν) = m
(i−1)
1 (ν) +m
(i+1)
1 (ν).
We have m
(i−1)
1 (ν) ≥ 1 since `(i−1) = 1 and m
(i+1)
1 (ν) ≥ 0 by definition. From the above
inequality we conclude that m
(i−1)
1 (ν) = 1 and m
(i+1)
1 (ν) = 0. We also have m
(i+1)
k (ν) = 0 for
1 < k < `(i) by the relation P
(i)
1 (ν) = P
(i)
2 (ν) = · · · = P
(i)
`(i)
(ν). Since f̃i does not change the
coriggings of (ν, J)(a) for a < i, we have ˜̀(i−1) = `(i−1) = 1. Recall that the string (1,−1) of
(ν̃, J̃)(i) is singular. Thus ˜̀(i) = 1. From ν̃(i+1) = ν(i+1), we have m
(i+1)
k (ν̃) = 0 for 1 ≤ k < `(i).
Since f̃i does not change the coriggings of (ν, J)(i+1), we have ˜̀(i+1) = `(i+1), and thus ˜̀(a) = `(a)
for all i < a and ˜̀(a) = `(a) for all a. In particular we have ˜̀(i) = `(i) = ∞. However this
is a contradiction since we have ˜̀(i) < ∞ by the assumption. Thus the situation (6.1) cannot
happen. �
Proposition 6.5. Let b ∈ B1,1 ⊗ B̄ and suppose that we have shown the commutativity of f̃i
and Φ for the elements of B̄. Suppose that we have f̃i(b) 6= 0 and f̃i(b
′) = 0 where b′ ∈ B̄ is the
corresponding part of b. Then we have f̃i(ν, J) 6= 0, f̃i(ν̄, J̄) = 0 and Φ−1(f̃i(b)) = f̃i(Φ
−1(b)).
Proof. By the assumption, we have the following two possibilities:
(a) b = i⊗ b′ and f̃i(b) = (i+ 1)⊗ b′,
(b) b = i+ 1⊗ b′ and f̃i(b) = i⊗ b′.
Case (a). In this case we have `(i−1) <∞ and `(i) =∞. Then we have P
(i)
∞ (ν) = P
(i)
∞ (ν̄)+1.
By the induction hypothesis, f̃i(b
′) = 0 implies that ϕi(ν̄, J̄) = 0. Then from Theorem 3.8 we
have P
(i)
∞ (ν̄) = s̄ where s̄ is the smallest rigging of (ν̄, J̄)(i). Let s be the smallest rigging of
(ν, J)(i). Since `(i) =∞, we have s = s̄. Again by Theorem 3.8 we have ϕi(ν, J) = P
(i)
∞ (ν)− s =
P
(i)
∞ (ν̄) + 1− s = s̄+ 1− s = 1. Therefore f̃i(ν, J) 6= 0 as desired.
Rigged Configurations and Kashiwara Operators 73
In order to check Φ−1(f̃i(b)) = f̃i(Φ
−1(b)), it is enough to show that f̃i acting on (ν, J)
creates a singular string and it satisfies ˜̀(i−1) ≤ `+1 = ˜̀(i) < ˜̀(i+1) =∞ under the assumptions
f̃i(ν, J) 6= 0 and f̃i(ν̄, J̄) = 0. Then this is the consequence of Proposition 6.2.
Case (b). In this case we have `(i+1) < ∞, `(i) = ∞, ˜̀(i) < ∞ and ˜̀
(i−1) = ∞. By
`(i+1) <∞ and `(i) =∞ we have P
(i)
∞ (ν) = P
(i)
∞ (ν̄)+1. Since we have ϕi(ν̄, J̄) = 0 by f̃i(b
′) = 0,
Theorem 3.8 asserts that P
(i)
∞ (ν̄) = s̄. Now let us invoke the relation s ≤ s̄ of Lemma 6.4. Then
we have ϕi(ν, J) = P
(i)
∞ (ν)− s = P
(i)
∞ (ν̄) + 1− s = s̄− s+ 1 ≥ 1. Thus f̃i(ν, J) 6= 0.
Finally we can check Φ−1(f̃i(b)) = f̃i(Φ
−1(b)) under the assumptions f̃i(ν, J) 6= 0 and
f̃i(ν̄, J̄) = 0 by using Proposition 6.2. �
6.3 Proof for (3)
For the analysis of the present situation, let us prepare several properties of the rigged configu-
rations. Let s (resp. s̄) be the smallest rigging of (ν, J)(i) (resp. (ν̄, J̄)(i)).
Lemma 6.6. Suppose that we have f̃i(ν, J) = 0 and f̃i(ν̄, J̄) 6= 0. Then we have the following
two possibilities:
(1) `(i) <∞ and `(i+1) =∞, or
(2) `(i) <∞ and `(i−1) =∞.
Proof. Let us show that `(i) < ∞. Suppose if possible that `(i) = ∞. From Theorem 3.8 we
have ϕi(ν, J) = P
(i)
∞ (ν) − s = 0. Since `(i) = ∞, we have s = s̄ and P
(i)
∞ (ν̄) ≤ P
(i)
∞ (ν). Thus
ϕi(ν̄, J̄) = P
(i)
∞ (ν̄) − s ≤ 0 contradicting the assumption f̃i(ν̄, J̄) 6= 0. Hence we conclude that
`(i) <∞.
Suppose that `(i+1) <∞. Then we have the following four possibilities:
(a) `(i+1) =∞,
(b) `(i+1) <∞ and `(i) =∞,
(c) `(i) <∞ and `(i−1) =∞,
(d) `(i−1) <∞.
Let us analyze these possibilities case by case. Note that by the assumption f̃i(ν, J) = 0, we
have ϕi(ν, J) = P
(i)
∞ (ν)− s = 0 from Theorem 3.8.
Case (a). By the assumptions `(i+1) <∞ and `(i+1) =∞, we have P
(i)
∞ (ν̄) = P
(i)
∞ (ν). Then
the assumption f̃i(ν̄, J̄) 6= 0 implies that
ϕi(ν̄, J̄) = P (i)
∞ (ν̄)− s̄ = P (i)
∞ (ν)− s̄ = s− s̄ > 0.
By the assumption `(i+1) =∞, we see that s > s̄ can happen only if the string
(
`(i)−1, P
(i)
`(i)−1
(ν̄)
)
which is created by δ has the rigging strictly smaller than s. Thus the present assumptions imply
that P
(i)
`(i)−1
(ν̄) < s. To summarize we are left with the following two possibilities:
(I) `(i−1) = `(i), then P
(i)
`(i)−1
(ν) < s by P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)−1
(ν),
(II) `(i−1) < `(i), then P
(i)
`(i)−1
(ν) ≤ s by P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)−1
(ν)− 1.
Let us show that these cases cannot happen. Let j be the largest integer such that j < `(i) and
m
(i)
j (ν) > 0. If there is no such j, set j = 0.
74 R. Sakamoto
Case (a-I). Suppose that j = `(i) − 1. Then we have xj ≤ P
(i)
j (ν) < s, which contradicts
the minimality of s. Suppose that 0 < j < `(i) − 1. Recall that by the minimality of s we
have P
(i)
`(i)
(ν) ≥ s. Thus we have P
(i)
`(i)
(ν) > P
(i)
`(i)−1
(ν). Then by the convexity relation of P
(i)
k (ν)
between j ≤ k ≤ `(i) we have P
(i)
`(i)
(ν) > P
(i)
`(i)−1
(ν) > P
(i)
j (ν) ≥ xj . Since we are assuming that
s > P
(i)
`(i)−1
(ν), this contradicts the minimality of s.
Finally suppose that j=0. Then the previous inequalities become s>P
(i)
`(i)−1
(ν)>P
(i)
0 (ν)=0.
Note that this relation is always valid since we have `(i) ≥ 1 by the definition. Therefore we have
s > 0 which implies that f̃i acting on (ν, J) adds the length 1 string (1,−1) to (ν, J)(i). Since
we are assuming that f̃i(ν, J) = 0, the rigging of the string (1,−1) of (ν̃, J̃)(i) must be larger
than the corresponding vacancy number. Thus we have P
(i)
1 (ν̃) < −1, that is, P
(i)
1 (ν) ≤ 0 by
P
(i)
1 (ν̃) = P
(i)
1 (ν)− 2. Since P
(i)
0 (ν) = 0, the convexity relation of P
(i)
k (ν) between 0 ≤ k ≤ `(i)
implies that 0 = P
(i)
0 (ν) ≥ P
(i)
1 (ν) ≥ P
(i)
`(i)
(ν). Therefore the rigging of the string
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i) satisfies that P
(i)
`(i)
(ν) ≤ 0. This is a contradiction since we have s > 0.
Case (a-II). Suppose that j = `(i) − 1. Then we have xj ≤ P
(i)
j (ν) ≤ s. By the minimality
of s we have xj = P
(i)
j (ν) = s, in particular, the string (j, xj) of (ν, J)(i) is singular. However
this is in contradiction to the definition of `(i) since we have `(i−1) ≤ j < `(i). Suppose that
`(i−1) ≤ j < `(i) − 1. Then by the convexity relation of P
(i)
k (ν) between j ≤ k ≤ `(i) we have
s ≥ P (i)
`(i)−1
(ν) ≥ P (i)
j (ν) ≥ xj . Then by the minimality of s we have P
(i)
j (ν) = xj , in particular,
the string (j, xj) of (ν, J)(i) is singular. Again this is in contradiction to the definition of `(i)
since `(i−1) ≤ j < `(i). Suppose that 0 < j < `(i−1). Since m
(i−1)
`(i−1)(ν) > 0, the convexity
relation of P
(i)
k (ν) between `(i−1) − 1 ≤ k ≤ `(i−1) + 1 is strictly convex. Therefore we have
s ≥ P (i)
`(i)−1
(ν) > P
(i)
j (ν) ≥ xj . This contradicts the minimality of s.
Finally suppose that j = 0. Then the above relation becomes s ≥ P
(i)
`(i)−1
(ν) > P
(i)
0 (ν) = 0,
that is, s > 0. By the similar arguments of the latter part of Case (a-I), we obtain P
(i)
1 (ν) ≤ 0.
Since P
(i)
0 (ν) = 0 and m
(i−1)
`(i−1)(ν) > 0, the convexity relation of P
(i)
k (ν) between 0 ≤ k ≤ `(i)
becomes 0 = P
(i)
0 (ν) ≥ P
(i)
`(i−1)(ν) > P
(i)
`(i)
(ν). Hence the rigging of the string
(
`(i), P
(i)
`(i)
(ν)
)
of
(ν, J)(i) satisfies that P
(i)
`(i)
(ν) < 0. This is a contradiction since we have s > 0.
Case (b). By the assumptions `(i+1) < ∞ and `(i) = ∞, we have P
(i)
∞ (ν̄) = P
(i)
∞ (ν) − 1.
Then the assumption f̃i(ν̄, J̄) 6= 0 implies that
ϕi(ν̄, J̄) = P (i)
∞ (ν̄)− s̄ = P (i)
∞ (ν)− 1− s̄ = s− s̄− 1 > 0.
By the assumption `(i) =∞, the situation s− 1 > s̄ can happen only if the rigging of the string(
`(i)− 1, P
(i)
`(i)−1
(ν̄)
)
which is created by δ satisfies P
(i)
`(i)−1
(ν̄) < s− 1. Then we can use the same
arguments of the above Case (a) to show that this case cannot happen.
Case (c). By the assumptions `(i) <∞ and `(i−1) =∞, we have P
(i)
∞ (ν̄) = P
(i)
∞ (ν)+1. Then
the assumption f̃i(ν̄, J̄) 6= 0 implies that
ϕi(ν̄, J̄) = P (i)
∞ (ν̄)− s̄ = P (i)
∞ (ν) + 1− s̄ = s− s̄+ 1 > 0 ⇐⇒ s ≥ s̄.
Thus, basically we have the following two cases:
(1) s > s̄, (2) s = s̄.
Case (c-1) To begin with let us consider the case s > s̄. This situation can happen if at
least one of the riggings for the strings
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
or
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
is smaller
Rigged Configurations and Kashiwara Operators 75
than s. If P
(i)
`(i)−1
(ν̄) < s, we can use the same arguments of Case (a) to show that this case
cannot happen. Thus P
(i)
`(i)−1
(ν̄) ≥ s > s̄.
Therefore we suppose that P
(i)
`(i)−1
(ν̄) > s̄ and P
(i)
`(i)−1(ν̄) = s̄ in the sequel. Let us follow the
classification (III) to (VII) of Lemma 5.1. We observe that the Cases (III) and (IV) cannot
happen in this setting, since in these settings we have `(i) = `(i). This is a contradiction since
we are assuming that P
(i)
`(i)−1
(ν̄) > P
(i)
`(i)−1(ν̄). Let us consider the remaining cases under the
assumption `(i) < `(i).
Case (c-1-V). In this case we have P
(i)
`(i)−1(ν) = s̄ − 1. Let j be the largest integer such
that j < `(i) and m
(i)
j (ν) > 0. Since `(i) < `(i) we have `(i) ≤ j. If j = `(i) − 1, we obtain the
immediate contradiction s > s̄− 1 = P
(i)
`(i)−1(ν) ≥ x`(i)−1. Thus suppose that `(i) ≤ j < `(i) − 1.
Since s is the minimal rigging of (ν, J)(i) we see that P
(i)
`(i)
(ν) ≥ s > s̄. Therefore we have
P
(i)
`(i)
(ν) > P
(i)
`(i)−1
(ν). Then by the convexity relation of P
(i)
k (ν) between j ≤ k ≤ `(i) we obtain
P
(i)
`(i)−1
(ν) > P
(i)
j (ν) ≥ xj , in particular, s > xj . This contradicts the minimality of s.
Case (c-1-VI). In this case we have P
(i)
`(i)−1(ν) = s̄. Then we can use the same arguments
of Case (c-1-V) to show that this case cannot happen.
Case (c-1-VII). In this case we have P
(i)
`(i)−1(ν) = s̄+1 under the assumption `(i+1) < `(i). If
s̄+1 < s we can use the same arguments of Case (c-1-V) to show that such case cannot happen.
Thus suppose that s̄ + 1 = s. Let j be the largest integer such that j < `(i) and m
(i)
j (ν) > 0.
Since `(i) < `(i) we have `(i) ≤ j.
Suppose that j = `(i) − 1. Then we see that the string (`(i) − 1, x`(i)−1) is singular by the
minimality of s, since we have s = P
(i)
`(i)−1(ν) ≥ x`(i)−1. However this is in contradiction to the
definition of `(i) since its length satisfies that `(i+1) ≤ `(i) − 1 < `(i).
Suppose that `(i+1) ≤ j < `(i) − 1. Since s is the minimal rigging of (ν, J)(i), we have
P
(i)
`(i)
(ν) ≥ s and P
(i)
j (ν) ≥ s. Since we have P
(i)
`(i)−1(ν) = s, the only possibility that is compatible
with the convexity relation of P
(i)
k (ν) between j ≤ k ≤ `(i) is the case s = P
(i)
`(i)
(ν) = P
(i)
`(i)−1(ν) =
· · · = P
(i)
j (ν) ≥ xj . By the minimality of s we have P
(i)
j (ν) = xj , in particular, the string (j, xj) is
singular. However this is in contradiction to the definition of `(i) since we have `(i+1) ≤ j < `(i).
Suppose that `(i) ≤ j < `(i+1). As in the previous paragraph, we obtain s = P
(i)
`(i)
(ν) =
P
(i)
`(i)−1(ν) = · · · = P
(i)
j (ν) ≥ xj . However, this is a contradiction since we have m
(i+1)
`(i+1)
(ν) > 0,
which implies that the vacancy numbers P
(i)
k (ν) are strictly convex function of k.
To summarize, we have shown that Case (c-1) cannot happen.
Case (c-2). In this case we have s = s̄. This case is indeed possible as the example at the
end of this subsection shows.
Case (d). By the assumption `(i−1) <∞, we have P
(i)
∞ (ν̄) = P
(i)
∞ (ν). Then the assumptions
f̃i(ν̄, J̄) 6= 0 and f̃i(ν, J) = 0 imply that
ϕi(ν̄, J̄) = P (i)
∞ (ν̄)− s̄ = P (i)
∞ (ν)− s̄ = s− s̄ > 0 ⇐⇒ s > s̄.
We can use the same arguments of Case (c-1) to show that this case cannot happen. �
76 R. Sakamoto
Lemma 6.7. Suppose that we have f̃i(ν, J) = 0 and f̃i(ν̄, J̄) 6= 0. Then we have the following
two possibilities:
(1) `(i) <∞, `(i+1) =∞ and s = s̄,
(2) `(i) <∞, `(i−1) =∞ and s = s̄.
Proof. (1) Step 1. Let us show s ≤ s̄. It is enough to show that the new string of (ν̄, J̄)(i)
has the rigging larger than or equal to s, that is, P
(i)
`(i)−1
(ν̄) ≥ s. Suppose that P
(i)
`(i)−1
(ν̄) < s.
Then we can use the same arguments of Case (a) of the proof of Lemma 6.6 to show that this
case cannot happen. Thus s ≤ s̄.
Step 2. Next, let us show s ≥ s̄. Then it is enough to show P
(i)
`(i)
(ν) ≥ s̄ since the only
rigging that is changed by δ−1 is the one associated with the lengthened string (see Remark 3.16
for the explanation of δ−1). By the assumption `(i+1) =∞ we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄)− 1. Thus
the above relation is equivalent to P
(i)
`(i)
(ν̄) > s̄.
Suppose if possible that P
(i)
`(i)
(ν̄) ≤ s̄. Suppose if possible that there are strings of (ν̄, J̄)(i) that
are longer than or equal to `(i). Let j be the length of the shortest such string. By the minimality
of s̄ we have P
(i)
`(i)−1
(ν̄) ≥ s̄. Then we have P
(i)
`(i)−1
(ν̄) ≥ P (i)
`(i)
(ν̄) ≥ P (i)
j (ν̄) ≥ xj by the convexity
relation between `(i)−1 and j (if j = `(i) we can directly show the inequality without the second
term). By the minimality of s̄, we have s̄ ≤ xj . Therefore we obtain s̄ = xj which implies that
the string (j, xj) is singular with length larger than `(i)−1. However, we know that δ−1 will add
a box to the length `(i) − 1 string of (ν̄, J̄)(i). This is a contradiction since `(i+1) = ∞ implies
that δ−1 will add a box to the longest possible string. Therefore we conclude that the longest
string of (ν̄, J̄)(i) has length `(i) − 1. Since we have P
(i)
`(i)−1
(ν̄) ≥ P
(i)
`(i)
(ν̄), the convexity relation
between `(i)−1 and∞ gives P
(i)
`(i)−1
(ν̄) ≥ P (i)
`(i)
(ν̄) ≥ · · · ≥ P (i)
∞ (ν̄). However this is a contradiction
since we already know that P
(i)
`(i)
(ν̄) ≤ s̄ and P
(i)
∞ (ν̄) > s̄ by ϕi(ν̄, J̄) = P
(i)
∞ (ν̄)− s̄ > 0. Thus we
have shown that the relation P
(i)
`(i)
(ν̄) > s̄ is always satisfied. Hence we obtain s ≥ s̄.
By combining the two inequalities, we conclude that s = s̄.
(2) During the proof of the previous lemma, we have shown that the only possible case under
the assumption is s = s̄. �
Example 6.8. Consider the following rigged configuration (ν, J) of type (B1,1)⊗4⊗B1,3⊗B2,1⊗
B2,2 ⊗B3,1 of D
(1)
5
1
1
1
1
1
0
0
1
1
1
1
0
0
0
0
−1
1
0
0
0
0
−1
1
0
−2
−2
−2
−2
1
0
−2
−2
−2
−3
0
1
2
0
1
2
0
−1
−1
0
−1
−1
This is the example for Case (c-2) of the proof of Lemma 6.6. The corresponding image
Φ−1(ν, J) is
2̄ ⊗ 1̄ ⊗ 1 ⊗ 4̄ ⊗ 5 3̄ 1̄ ⊗ 5̄
5
⊗ 1 4
4 5̄
⊗
1
5̄
2̄
We have f̃2(ν, J) = 0 since we have P
(2)
6 (ν̃) = −3. (ν̄, J̄) is given as follows:
Rigged Configurations and Kashiwara Operators 77
0
0
0
1
1
0
0
0
1
1
1
1
0
0
0
−1
1
1
0
0
0
−1
1
1
−2
−2
−2
−2
1
1
−2
−2
−2
−3
0
0
2
0
0
2
0
0
−1
0
0
−1
Note that the minimal rigging −1 of (ν, J)(2) is preserved before and after δ in a nontrivial way.
We have f̃2(ν̄, J̄) 6= 0 since we have P
(2)
5 (˜̄ν) = −2.
Proposition 6.9. Let us consider the rigged configuration (ν, J) of type B1,1⊗ B̄. Suppose that
we have the commutativity of f̃i and Φ for B̄. Suppose that f̃i(ν, J) = 0 and f̃i(ν̄, J̄) 6= 0. Let
b = Φ(ν, J) and b′ = Φ(ν̄, J̄). Then we have f̃i(b) = 0, f̃i(b
′) 6= 0 and Φ(f̃i(ν, J)) = f̃i(Φ(ν, J)).
Proof. According to Lemma 6.6, we see that there are only two possibilities. The first case is
`(i) < ∞ and `(i+1) = ∞. In this case we have b = (i + 1) ⊗ b′ and P
(i)
∞ (ν̄) = P
(i)
∞ (ν) + 1. The
second case is `(i) <∞ and `(i−1) =∞. In this case we have i⊗ b′ and P
(i)
∞ (ν̄) = P
(i)
∞ (ν) + 1.
Recall that we have ϕi(b
′) = ϕi(ν̄, J̄) > 0 by the induction hypothesis. Let us show that in
fact we have ϕi(ν̄, J̄) = 1. From Theorem 3.8, we compute
ϕi(ν̄, J̄) = P (i)
∞ (ν̄)− s̄ = P (i)
∞ (ν)− s+ 1 = 1.
Here we have used the relation s= s̄ from Lemma 6.7 and the assumption ϕ(ν, J)=P
(i)
∞ (ν)−s=0.
Thus ϕi(ν̄, J̄) = 1.
Then by the induction hypothesis we have ϕi(b
′) = ϕi(ν̄, J̄) = 1. On the other hand, recall
that we have εi(i+1) = εi(̄i) = 1. Then from the tensor product rule at Section 2.1, we conclude
that f̃i(b) = 0. Hence we have Φ(f̃i(ν, J)) = f̃i(Φ(ν, J)). �
6.4 Proof for (4)
Proposition 6.10. Let b ∈ B1,1 ⊗ B̄ and suppose that we have shown the commutativity of f̃i
and Φ for the elements of B̄. Suppose that we have f̃i(b) = 0 and f̃i(b
′) 6= 0 where b′ ∈ B̄ is the
corresponding part of b. Then we have f̃i(ν, J) = 0, f̃i(ν̄, J̄) 6= 0 and Φ−1(f̃i(b)) = f̃i(Φ
−1(b)).
Proof. By the assumption, we have b = (i + 1) ⊗ b′ or b = i ⊗ b′ and ϕi(b
′) = 1. In the
first case we have `(i) < ∞ and `(i+1) = ∞ and in the second case we have `(i) < ∞ and
`(i−1) =∞. In both cases we have P
(i)
∞ (ν) = P
(i)
∞ (ν̄)− 1. From the condition ϕi(b
′) = 1 together
with the induction hypothesis we have ϕi(b
′) = ϕi(ν̄, J̄) = 1, which implies P
(i)
∞ (ν̄) = s̄ + 1 by
Theorem 3.8. Here s̄ is the minimal rigging of (ν̄, J̄)(i). Then we have
ϕi(ν, J) = P (i)
∞ (ν)− s = P (i)
∞ (ν̄)− 1− s = s̄− s.
If we show s ≥ s̄ we have ϕi(ν, J) = 0 as desired. Note that this result leads to Φ−1(f̃i(b)) =
f̃i(Φ
−1(b)) also.
Let us show s ≥ s̄. If `(i) < ∞ and `(i+1) = ∞ we can use the same arguments of Step 2 of
the proof of Lemma 6.7(1) since we do not use the assumption ϕi(ν, J) = 0 there. If `(i) < ∞
and `(i−1) = ∞, again we can use the same arguments of Step 2 of the proof of Lemma 6.7(1)
by replacing `(i) and `(i+1) there by `(i) and `(i−1) respectively. �
7 Proof of Proposition 4.8: ẽi version
In this section we use a parallel notation that are introduced in Section 4.4 except for changing
the role of f̃i there by ẽi. For example, ẽi acts on the string (`, x`) of (ν, J)(i).
78 R. Sakamoto
7.1 Proof for (5)
Proposition 7.1. Let us consider the rigged configuration (ν, J) of type B1,1⊗ B̄. Suppose that
we have the commutativity of ẽi and f̃i with Φ for B̄. Suppose that ẽi(ν, J) 6= 0 and ẽi(ν̄, J̄) = 0.
Let b = Φ(ν, J) and b′ = Φ(ν̄, J̄). Then we have one of the following two cases:
(1) b = (i+ 1)⊗ b′, ẽi(b) is defined, ẽi(b
′) is undefined and Φ(ẽi(ν, J)) = i⊗ b′,
(2) b = i⊗ b′, ẽi(b) is defined, ẽi(b
′) is undefined and Φ(ẽi(ν, J)) = i+ 1⊗ b′.
For the proof, we show the following properties.
Proposition 7.2. ẽi(ν, J) is defined and ẽi(ν̄, J̄) is undefined if and only if one of the following
conditions is satisfied:
(1) `(i) <∞, `(i+1) =∞ and all the riggings of (ν, J)(i) are non-negative except for the longest
string whose rigging is −1. Moreover we have
P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)
(ν̄) = · · · = P (i)
∞ (ν̄) = 0.
(2) `(i) <∞, `(i−1) =∞ and all the riggings of (ν, J)(i) are non-negative except for the longest
string(s) whose rigging is −1. Moreover we have
P
(i)
`(i)−1(ν̄) = P
(i)
`(i)
(ν̄) = · · · = P (i)
∞ (ν̄) = 0.
Proof. By the assumptions ẽi(ν, J) 6= 0 and ẽi(ν̄, J̄) = 0, we see that δ changes (ν, J)(i).
Thus we have `(i) < ∞. Also, from ẽi(ν̄, J̄) = 0, we see that all the riggings of (ν̄, J̄)(i) are
non-negative.
(1) Suppose that we have `(i+1) =∞. Suppose if possible that we have m
(i)
j (ν̄) > 0 for some
`(i) ≤ j. Let us choose the minimal such j. Note that by the condition `(i) ≤ j, the string
(j, xj) is different from the string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
created by δ. By the definition of δ−1,
`(i) <∞ and `(i+1) =∞ imply that the string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
is the longest singular string
of (ν̄, J̄)(i). In particular, the string (j, xj) is not singular so that we have P
(i)
j (ν̄) > xj ≥ 0.
On the other hand, we have P
(i)
`(i)−1
(ν̄) ≥ 0 since ẽi(ν̄, J̄) = 0. Let us show P
(i)
`(i)
(ν̄) > 0. If
`(i) = j it is already proved. Thus suppose that `(i) < j. Since we have m
(i)
k (ν̄) = 0 for
all `(i) < k < j, the only possibility that is compatible with P
(i)
`(i)−1
(ν̄) ≥ 0, P
(i)
j (ν̄) > 0 and
the convexity relation of P
(i)
k (ν̄) for `(i) − 1 ≤ k ≤ j is P
(i)
`(i)
(ν̄) > 0. In conclusion, we have
P
(i)
`(i)
(ν) ≥ 0 by P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄)− 1 which follows from the relations `(i) <∞ and `(i+1) =∞.
Since all the riggings of (ν̄, J̄)(i) are non-negative, the relation P
(i)
`(i)
(ν) ≥ 0 implies that all the
riggings of (ν, J)(i) are non-negative. This contradicts the assumption ẽi(ν, J) 6= 0. Hence the
string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
is the longest string of (ν̄, J̄)(i).
Let us show that P
(i)
`(i)
(ν̄) = 0. Suppose if possible that we have P
(i)
`(i)
(ν̄) > 0. Then we have
P
(i)
`(i)
(ν) ≥ 0 by P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) − 1. This contradicts the assumption ẽi(ν, J) 6= 0. Therefore
we have P
(i)
`(i)
(ν̄) = 0. In particular, the string
(
`(i), P
(i)
`(i)
(ν)
)
= (`(i),−1) is the only string of
(ν, J)(i) which has the negative rigging.
Finally let us show that P
(i)
`(i)−1
(ν̄) = · · · = P
(i)
∞ (ν̄) = 0. By the convexity relation of P
(i)
k (ν̄)
between `(i)−1 ≤ k <∞ and the condition P
(i)
∞ (ν̄) > −∞, we have P
(i)
`(i)−1
(ν̄) ≤ P (i)
`(i)
(ν̄) ≤ · · · ≤
Rigged Configurations and Kashiwara Operators 79
P
(i)
∞ (ν̄). Recall that we have P
(i)
`(i)−1
(ν̄) ≥ 0 by ẽi(ν̄, J̄) = 0 and P
(i)
`(i)
(ν̄) = 0 by the previous
paragraph. Then the only possibility is P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)
(ν̄) = · · · = P
(i)
∞ (ν̄) = 0.
(2) In this case we assume that `(i+1) <∞.
Step 1. We shall show that `(i+1) < ∞. Suppose if possible that `(i+1) = ∞. Consider the
case `(i) = `(i+1). Then we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄). This implies that all the riggings of (ν, J)(i)
are non-negative. This contradicts the assumption ẽi(ν, J) 6= 0.
Thus we are left with the case `(i) < `(i+1). Then we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) − 1. In order
to have ẽi(ν, J) 6= 0, we must have P
(i)
`(i)
(ν̄) = 0. Suppose if possible that there are strings of
(ν̄, J̄)(i) which are longer than `(i) − 1. Let j be the minimal integer such that m
(i)
j (ν̄) > 0 and
`(i) ≤ j. Then the string (j, xj) of (ν̄, J̄)(i) must satisfy P
(i)
j (ν̄) ≥ xj ≥ 0. Similarly, the string(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
must satisfy P
(i)
`(i)−1
(ν̄) ≥ 0. Then the only possibility that is compatible
with the convexity relation of P
(i)
k (ν̄) between `(i) − 1 ≤ k ≤ j is the relation
P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)
(ν̄) = · · · = P
(i)
j (ν̄) = 0. (7.1)
From (7.1) we have m
(i+1)
k (ν̄) = 0 for all `(i) ≤ k < j. Recall that by definition of `(i+1) there is
the string (`(i+1)−1, P
(i+1)
`(i+1)−1
(ν̄)) of (ν̄, J̄)(i+1). Then we see that its length satisfies j ≤ `(i+1)−1
by `(i) ≤ `(i+1)−1. Also from (7.1) and the assumption ẽi(ν̄, J̄) = 0, we see that the string (j, xj)
is singular since we have 0 = P
(i)
j (ν̄) ≥ xj ≥ 0. By the relation j ≤ `(i+1)−1, δ−1 will add a box
to the string (j, xj) of (ν̄, J̄)(i) instead of the string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
. This is a contradiction
since we have `(i)−1 < j. In conclusion we see that the string
(
`(i)−1, P
(i)
`(i)−1
(ν̄)
)
is the longest
string of (ν̄, J̄)(i).
We know that P
(i)
`(i)
(ν̄) = 0. Since ẽi(ν̄, J̄) = 0, we have P
(i)
`(i)−1
(ν̄) ≥ 0 and, in particular,
P
(i)
`(i)−1
(ν̄) ≥ P
(i)
`(i)
(ν̄). Since there is no string of (ν̄, J̄)(i) which is longer than `(i) − 1, we can
use the convexity relation of P
(i)
k (ν̄) between `(i) − 1 ≤ k <∞ with P
(i)
∞ (ν̄) > −∞ to conclude
that P
(i)
`(i)−1
(ν̄) ≤ P (i)
`(i)
(ν̄) ≤ · · · ≤ P (i)
∞ (ν̄). By the relation P
(i)
`(i)−1
(ν̄) ≥ P (i)
`(i)
(ν̄) we conclude that
P
(i)
`(i)−1
(ν̄) = P
(i)
`(i)
(ν̄) = · · · = P
(i)
∞ (ν̄) is the only possibility that is compatible with the convexity
relation. Then we have m
(i+1)
k (ν̄) = 0 for all `(i) ≤ k. This is a contradiction since we are
assuming that `(i) < `(i+1). Hence we have `(i+1) <∞.
Step 2. We shall show that `(i) < ∞. Suppose if possible that `(i) = ∞. In order to have
ẽi(ν, J) 6= 0, we must have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) − 1. Under the condition `(i) = ∞, this is only
possible if we have `(i) < `(i+1). Then we can use the same argument of the corresponding part
of Step 1 to deduce a contradiction. Thus we have `(i) <∞.
Step 3. We shall show that `(i−1) = ∞. Suppose if possible that `(i−1) < ∞. Then we can
use the same arguments of Step 1 if we replace `(i) and `(i+1) there by `(i) and `(i−1) respectively
to deduce a contradiction. Thus we have `(i−1) =∞.
Step 4. We shall show that the length `(i) string of (ν, J)(i) created by δ−1 has non-negative
rigging if `(i) < `(i). For this we have to show that P
(i)
`(i)
(ν) ≥ 0. Recall that we have P
(i)
`(i)−1
(ν̄) ≥ 0
by the assumption ẽi(ν̄, J̄) = 0. On the other hand, let j be the smallest integer such that
m
(i)
j (ν̄) > 0 and `(i) ≤ j. By `(i) < `(i) we have j ≤ `(i). Then by the assumption ẽi(ν̄, J̄) = 0
we have P
(i)
j (ν̄) ≥ 0.
If `(i) = `(i+1), we have P
(i)
`(i)
(ν) ≥ P
(i)
`(i)
(ν̄) by `(i) < `(i). Then by the convexity relation
of P
(i)
k (ν̄) between `(i) − 1 ≤ k ≤ j we have P
(i)
`(i)
(ν) ≥ P
(i)
`(i)
(ν̄) ≥ min
{
P
(i)
`(i)−1
(ν̄), P
(i)
j (ν̄)
}
≥ 0.
80 R. Sakamoto
Next let us consider the case `(i) < `(i+1). Then we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) − 1. Then one
can attain P
(i)
`(i)
(ν) = −1 only if P
(i)
`(i)−1
(ν̄) = · · · = P
(i)
j (ν̄) = 0. This relation implies that
m
(i+1)
k (ν̄) = 0 for all `(i)−1 < k < j. Then we have j ≤ `(i+1)−1 since we have `(i)−1 < `(i+1)−1
by the assumption. Let us consider the string (j, xj) of (ν̄, J̄)(i). Then we have P
(i)
j (ν̄) ≥ xj ≥ 0
by the assumption ẽi(ν̄, J̄) = 0. Since P
(i)
j (ν̄) = 0 we see that the string (j, xj) is singular and
its length satisfies `(i)− 1 < j ≤ `(i+1)− 1. Thus δ−1 will add a box to the string (j, xj) instead
of the length `(i) − 1 string. This is a contradiction. To summarize, we have shown that the
length `(i) string of (ν, J)(i) created by δ−1 has non-negative rigging if `(i) < `(i).
Then we can use a parallel arguments of the previous case (1) to deduce that the string(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
is the longest string of (ν̄, J̄)(i).
Step 5. Finally let us show that P
(i)
`(i)−1(ν̄) = · · · = P
(i)
∞ (ν̄) = 0. Since the length `(i) − 1
string is the longest string of (ν̄, J̄)(i), we have P
(i)
`(i)−1(ν̄) ≤ P
(i)
`(i)
(ν̄) ≤ · · · ≤ P
(i)
∞ (ν̄) by the
convexity relation of P
(i)
k (ν̄) between `(i) − 1 ≤ k < ∞ and P
(i)
∞ (ν̄) > −∞. Since we have
P
(i)
`(i)
(ν) = −1, we have P
(i)
`(i)
(ν̄) = 0 by P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) − 1. Since we have P
(i)
`(i)−1(ν̄) ≥ 0 by
ẽi(ν̄, J̄) = 0, we conclude that P
(i)
`(i)−1(ν̄) = P
(i)
`(i)
(ν̄) = · · · = P
(i)
∞ (ν̄) = 0. �
Proof of Proposition 7.1. We divide the proof into two cases following Proposition 7.2.
(1) Let us consider the case `(i) <∞ and `(i+1) =∞. In this case we have b = (i+1)⊗b′. By
the assumption we have εi(b
′) = εi(ν̄, J̄) = 0. From Proposition 7.2 we know that P
(i)
∞ (ν̄) = 0
and the minimal rigging of (ν̄, J̄)(i) is 0. Therefore we have ϕi(ν̄, J̄) = 0 by Theorem 3.8. Then
by the assumption we have ϕi(b
′) = ϕi(ν̄, J̄) = 0. Thus εi(b) = 1, that is, ẽi(b) 6= 0.
Finally let us show Φ(ẽi(ν, J)) = i⊗ b′. Recall that the longest string of (ν̄, J̄)(i) has length
`(i) − 1. Therefore we have m
(i)
`(i)
(ν) = 1. Thus if `(i−1) = `(i) we see that we cannot remove
a box from (ẽi(ν, J))(i) by δ since the length of the longest string of (ẽi(ν, J))(i) is `(i) − 1
and thus all the strings of (ẽi(ν, J))(i) are strictly shorter than `(i−1). In this case we have
Φ(ẽi(ν, J)) = i⊗ b′. Therefore let us suppose that `(i−1) < `(i) in the following. Recall that we
have P
(i)
`(i)−1
(ν̄) = 0. Then we have P
(i)
`(i)−1
(ν) = 1 by `(i−1) < `(i). We know that ẽi changes the
string (`(i),−1) of (ν, J)(i) into (`(i) − 1, 0). In particular, the latter string is non-singular by
P
(i)
`(i)−1
(ν̃) = P
(i)
`(i)−1
(ν) = 1. Thus ˜̀(i) =∞ also. In conclusion we have Φ(ẽi(ν, J)) = i⊗ b′.
(2) The proof for this case is almost identical to the previous case if we replace `(i−1) and `(i)
by `(i) and `(i+1) respectively. �
Example 7.3. Consider the following rigged configuration (ν, J) of type (B1,1)⊗4⊗B1,3⊗B2,1⊗
B2,2 ⊗B3,1 of D
(1)
5
1
0
0
1
−2
0
1
2
2
−1
0
2
2
−1
1
0
0
0
0
−1
0
0
−1
−1
−1
−1
−1
0
−1
0
The corresponding image Φ−1(ν, J) is
2̄ ⊗ 1̄ ⊗ 5 ⊗ 2̄ ⊗ 1 1 2̄ ⊗ 2
3
⊗ 2 2
3 4
⊗
1
4̄
3̄
We have ẽ2(ν, J) 6= 0 since (ν, J)(2) has a negative rigging. Here ẽ2 changes the string (6,−1) of
(ν, J)(2) into (5, 0). Since P
(i)
5 (ν̃) = 1, the latter string is non-singular. Thus the leftmost letter
Rigged Configurations and Kashiwara Operators 81
of Φ(ν̃, J̃) is 3̄ and the rest part is the same as the corresponding part of the above path. (ν̄, J̄)
is given as follows:
2
0
0
2
−2
0
1
1
2
0
0
1
2
0
1
−1
−1
0
0
−1
−1
0
−1
−1
−1
−1
−1
0
−1
0
Since all the riggings of (ν̄, J̄)(2) are non-negative, we have ẽ2(ν̄, J̄) = 0.
7.2 Proof for (6)
Proposition 7.4. Let b ∈ B1,1 ⊗ B̄ and suppose that we have shown the commutativity of ẽi
and f̃i with Φ for the elements of B̄. Suppose that ẽi(b) 6= 0 and ẽi(b
′) = 0 where b′ ∈ B̄ is the
corresponding part of b. Then ẽi(ν, J) 6= 0, ẽi(ν̄, J̄) = 0 and Φ−1(ẽi(b)) = ẽi(Φ
−1(b)).
Proof. By the assumption, we have the following two possibilities:
(a) b = (i+ 1)⊗ b′ and ẽi(b) = i⊗ b′,
(b) b = i⊗ b′ and ẽi(b) = i+ 1⊗ b′.
In both cases we have ϕi(b
′) = 0. Note that we have ẽi(ν̄, J̄) = 0 by the assumption.
Case (a). By the assumption we have εi(ν̄, J̄) = 0 and ϕi(ν̄, J̄) = 0. Then from Theorem 3.8
we have P
(i)
∞ (ν̄) = ϕi(ν̄, J̄) − εi(ν̄, J̄) = 0. Let l := ν̄
(i)
1 be the length of the longest string of
(ν̄, J̄)(i). Then we can use the convexity relation of P
(i)
k (ν̄) between l ≤ k ≤ ∞ to deduce
that P
(i)
l (ν̄) = P
(i)
l+1(ν̄) = · · · = P
(i)
∞ (ν̄) = 0. Consider the string (l, xl) of (ν̄, J̄)(i). Then from
ẽi(ν̄, J̄) = 0 we have P
(i)
l (ν̄) ≥ xl ≥ 0. Thus we have xl = 0 and, in particular, the string
(l, xl) = (l, 0) is singular. Since we have b = (i+ 1)⊗ b′, we see that δ−1 adds a box to the string
(l, 0) of (ν̄, J̄)(i), that is, l+ 1 = `(i). Since `(i+1) =∞, we see that P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄)− 1 = −1.
To summarize, the longest string of (ν, J)(i) is (`(i),−1) and therefore we have ẽi(ν, J) 6= 0.
Since δ−1 does not change the riggings of the strings except for the longest one, we see that
the unique string of (ν, J)(i) which has negative rigging is (`(i),−1). Thus ẽi acts on the string
(`(i),−1) and changes it to (`(i) − 1, 0). If `(i−1) = `(i) we see that δ cannot remove a box
from (ν̃, J̃)(i) since all the strings are strictly shorter than `(i−1). Thus Φ−1(ẽi(b)) = i ⊗ b′. So
suppose that `(i−1) < `(i). Then we have P
(i)
`(i)−1
(ν̃) = P
(i)
`(i)−1
(ν) = P
(i)
`(i)−1
(ν̄) + 1 = 1. Thus
the string (`(i) − 1, 0) created by ẽi is non-singular. Therefore we have ˜̀(i) = ∞ and hence
Φ−1(ẽi(b)) = i⊗ b′.
Case (b). Proof of this case is almost identical to the previous Case (a). �
7.3 Proof for (7)
Proposition 7.5. The situation ẽi(ν, J) = 0 and ẽi(ν̄, J̄) 6= 0 cannot happen.
Proof. Suppose if possible that we have ẽi(ν, J) = 0 and ẽi(ν̄, J̄) 6= 0. Then δ must change
some strings of (ν, J)(i). Thus we have `(i) <∞.
Step 1. Let us consider the case `(i) =∞. Then the string
(
`(i)−1, P
(i)
`(i)−1
(ν̄)
)
is the unique
string of (ν̄, J̄)(i) which has a negative rigging. In particular, we have P
(i)
`(i)−1
(ν̄) < 0. Let j be
the largest integer such that m
(i)
j (ν̄) > 0 and j < `(i) − 1. If there is no such string, set j = 0.
Then we must have P
(i)
j (ν̄) ≥ 0. From the convexity relation of P
(i)
k (ν̄) between j ≤ k ≤ `(i)−1,
we see that P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄).
82 R. Sakamoto
On the other hand, from ẽi(ν, J) = 0 the string
(
`(i), P
(i)
`(i)
(ν)
)
must have a non-negative
rigging so that we have P
(i)
`(i)
(ν) ≥ 0.
1. Let us consider the case `(i) < `(i+1). Then we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) − 1. Therefore we
have P
(i)
`(i)
(ν̄) ≥ 1. Thus
−P (i)
`(i)−2
(ν̄) + 2P
(i)
`(i)−1
(ν̄)− P (i)
`(i)
(ν̄) ≤ −3.
On the other hand, since we cannot have more than one strings of (ν̄, J̄)(i) which have negative
riggings, we have m
(i)
`(i)−1
(ν̄) = 1. Thus
m
(i−1)
`(i)−1
(ν̄)− 2m
(i)
`(i)−1
(ν̄) +m
(i+1)
`(i)−1
(ν̄) ≥ −2.
This contradicts the convexity relation of Lemma 3.13.
2. Next, let us consider the case `(i) = `(i+1) < `(i+1). Then we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) so
that P
(i)
`(i)
(ν̄) ≥ 0. Thus
−P (i)
`(i)−2
(ν̄) + 2P
(i)
`(i)−1
(ν̄)− P (i)
`(i)
(ν̄) ≤ −2.
On the other hand, we have m
(i)
`(i)−1
(ν̄) = 1 and m
(i+1)
`(i)−1
(ν̄) ≥ 1. Thus
m
(i−1)
`(i)−1
(ν̄)− 2m
(i)
`(i)−1
(ν̄) +m
(i+1)
`(i)−1
(ν̄) ≥ −1.
Again this contradicts the convexity relation of Lemma 3.13.
3. Finally let us consider the case `(i) = `(i+1) = `(i+1). Then we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) + 1
so that P
(i)
`(i)
(ν̄) ≥ −1. Thus
−P (i)
`(i)−2
(ν̄) + 2P
(i)
`(i)−1
(ν̄)− P (i)
`(i)
(ν̄) ≤ −1.
In this case, we have m
(i)
`(i)−1
(ν̄) = 1 and m
(i+1)
`(i)−1
(ν̄) ≥ 2. Thus
m
(i−1)
`(i)−1
(ν̄)− 2m
(i)
`(i)−1
(ν̄) +m
(i+1)
`(i)−1
(ν̄) ≥ 0.
This contradicts the convexity relation of Lemma 3.13.
Step 2. Let us consider the case `(i) < ∞. If `(i) = `(i) we can use the previous arguments
to deduce a contradiction. Note that we have `(i+1) = `(i+1) in this case. Thus we assume that
`(i) < `(i). Suppose that we have `(i) = `(i) − 1. Suppose that we also have `(i) − 1 > `(i+1) − 1.
The full condition for this case is as follows
`(i) − 1 = `(i+1) − 1 = `(i+1) − 1 = (`(i) − 1)− 1.
Thus we have P
(i)
`(i)−1(ν) = P
(i)
`(i)−1(ν̄) + 1. Since the string
(
`(i), P
(i)
`(i)
(ν)
)
of (ν, J)(i) has a non-
negative rigging, we have P
(i)
`(i)−1(ν) ≥ 0. To summarize, we have P
(i)
`(i)−1(ν̄) ≥ −1 in this case.
1. Consider the case P
(i)
`(i)−1(ν̄) ≥ 0. Since ẽi(ν̄, J̄) 6= 0 there is a string of (ν̄, J̄)(i) with
negative rigging. Therefore we must have P
(i)
`(i)−1
(ν̄) < 0 since P
(i)
`(i)−1(ν̄) ≥ 0. Let j be the
largest integer such that m
(i)
j (ν̄) > 0 and j < `(i) − 1. If there is no such j, set j = 0. Then
Rigged Configurations and Kashiwara Operators 83
we have P
(i)
j (ν̄) ≥ 0. By the convexity relation of P
(i)
k (ν̄) between j ≤ k ≤ `(i) − 1, we have
P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄). Since `(i) = `(i) − 1, we have
−P (i)
`(i)−2
(ν̄) + 2P
(i)
`(i)−1
(ν̄)− P (i)
`(i)
(ν̄) ≤ −2.
On the other hand, since `(i) − 1 = `(i+1) − 1 = `(i+1) − 1 we have m
(i+1)
`(i)−1
(ν̄) ≥ 2. Recall that
the string
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
is the unique string of (ν̄, J̄)(i) with the negative rigging. Thus
we have m
(i)
`(i)−1
(ν̄) = 1. Therefore we have
m
(i−1)
`(i)−1
(ν̄)− 2m
(i)
`(i)−1
(ν̄) +m
(i+1)
`(i)−1
(ν̄) ≥ 0.
This contradicts the convexity relation of Lemma 3.13.
2. Consider the case P
(i)
`(i)−1(ν̄) = −1 and `(i−1) > `(i). We will show that this case cannot
happen in Lemma 7.6.
3. Consider the case P
(i)
`(i)−1(ν̄) = −1 and `(i−1) = `(i). We will show that this case cannot
happen in Lemma 7.7.
On the other hand, together with the present assumption `(i) = `(i)−1, suppose that we also
have `(i) − 1 = `(i+1) − 1. We will show that this case cannot happen in Lemma 7.8.
Finally let us assume that `(i) < `(i) − 1. Recall that the only strings of (ν̄, J̄)(i) which may
have negative riggings are
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
. Suppose that we have
P
(i)
`(i)−1
(ν̄) < 0. In this case, as in Step 1, we have P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄) and P
(i)
`(i)
(ν) ≥ 0. Then
by the same arguments of Step 1 we can deduce a contradiction. Thus suppose that we have
P
(i)
`(i)−1
(ν̄) ≥ 0 in the following. Then by the assumption ẽi(ν̄, J̄) 6= 0, we have P
(i)
`(i)−1(ν̄) < 0.
Let j be the largest integer such that j < `(i)−1 and m
(i)
j (ν̄) > 0. Then we have P
(i)
j (ν̄) ≥ 0 and
by the convexity relation of P
(i)
k (ν̄) between j ≤ k ≤ `(i) − 1, we have P
(i)
`(i)−2(ν̄) > P
(i)
`(i)−1(ν̄).
On the other hand, from the assumption ẽi(ν, J) = 0 we see that the string
(
`(i), P
(i)
`(i)
(ν)
)
of
(ν, J)(i) must have a non-negative rigging. Thus we have P
(i)
`(i)
(ν) ≥ 0. Then we can use the same
arguments of Step 1 if we replace `(i) and `(i+1) there by `(i) and `(i−1) respectively to deduce
a contradiction. Thus we have confirmed that the final case `(i) < `(i) − 1 cannot happen. �
Lemma 7.6. The following situation cannot happen. ẽi(ν, J) = 0, ẽi(ν̄, J̄) 6= 0,
`(i) − 1 = `(i+1) − 1 = `(i+1) − 1 = (`(i) − 1)− 1,
`(i−1) > `(i) and P
(i)
`(i)−1(ν̄) = −1.
Proof. Since P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄) − 1, we have P
(i)
`(i)
(ν̄) ≥ 1 by ẽi(ν, J) = 0. Let us show that
m
(i)
`(i)−1(ν̄) = 1. Since we know that `(i) − 1 < `(i) − 1, we see that δ−1 will add boxes to the
strings
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
. Thus if we have m
(i)
`(i)−1(ν̄) > 1, there will
be a remaining negative rigging in (ν, J)(i) by P
(i)
`(i)−1(ν̄) < 0, which is in contradiction to the
assumption ẽi(ν, J) = 0. Therefore we have m
(i)
`(i)−1(ν̄) = 1. Then P
(i)
`(i)−2(ν̄), P
(i)
`(i)−1(ν̄) and
(P
(i)
`(i)
(ν̄) − 2) must satisfy the convexity relation by Lemma 3.12. Since P
(i)
`(i)−1(ν̄) = −1 and
P
(i)
`(i)
(ν̄) ≥ 1, we see that
P
(i)
`(i)−2(ν̄) = P
(i)
`(i)−1
(ν̄) ≤ −1. (7.2)
84 R. Sakamoto
From this, we also have m
(i)
`(i)−1
(ν̄) = 1 as otherwise we would have a remaining negative rigging
in (ν, J)(i). Let j be the largest integer such that m
(i)
j (ν̄) > 0 and j < `(i) − 1. If there is no
such j, set j = 0. Then we have P
(i)
j (ν̄) ≥ 0 since the only strings of (ν̄, J̄)(i) with negative
riggings are
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
. Then from the convexity relation of
P
(i)
k (ν̄) between j ≤ k ≤ `(i) − 1 and (7.2), we have
P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄).
Then the left hand side of the convexity relation of Lemma 3.13 is
−P (i)
`(i)−2
(ν̄) + 2P
(i)
`(i)−1
(ν̄)− P (i)
`(i)
(ν̄) ≤ −1
by `(i) − 1 = `(i) and the right hand side of it is
m
(i−1)
`(i)−1
(ν̄)− 2m
(i)
`(i)−1
(ν̄) +m
(i+1)
`(i)−1
(ν̄) ≥ 0.
Here we use the fact m
(i+1)
`(i)−1
(ν̄) ≥ 2 which follows from `(i) − 1 = `(i+1) − 1 = `(i+1) − 1. This is
a contradiction. �
Lemma 7.7. The following situation cannot happen. ẽi(ν, J) = 0, ẽi(ν̄, J̄) 6= 0,
`(i) − 1 = `(i+1) − 1 = `(i+1) − 1 = (`(i) − 1)− 1 = (`(i−1) − 1)− 1
and P
(i)
`(i)−1(ν̄) = −1.
Proof. Since P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄), we have
P
(i)
`(i)
(ν̄) ≥ 0
by ẽi(ν, J) = 0. As in Lemma 7.6, we have m
(i)
`(i)−1(ν̄) = 1 by `(i) − 1 < `(i) − 1. If we use
P
(i)
`(i)−1(ν̄) = −1, the left hand side of the convexity relation of Lemma 3.13 with l = `(i) − 1 is
−P (i)
`(i)−2(ν̄) + 2P
(i)
`(i)−1(ν̄)− P (i)
`(i)
(ν̄) ≤ −P (i)
`(i)−2(ν̄)− 2
and the right hand side of it is
m
(i−1)
`(i)−1(ν̄)− 2m
(i)
`(i)−1(ν̄) +m
(i+1)
`(i)−1(ν̄) ≥ −1 +m
(i+1)
`(i)−1(ν̄) ≥ −1.
Here we have used m
(i−1)
`(i)−1(ν̄) ≥ 1 by `(i−1) − 1 = `(i) − 1. Therefore we have
P
(i)
`(i)−2(ν̄) = P
(i)
`(i)−1
(ν̄) ≤ −1. (7.3)
From this and P
(i)
`(i)−1(ν̄) = −1 we deduce that m
(i)
`(i)−1
(ν̄) = 1 as otherwise we must have
ẽi(ν, J) 6= 0. Let j be the largest integer such that m
(i)
j (ν̄) > 0 and j < `(i) − 1. If there is
no such j, set j = 0. Then we have P
(i)
j (ν̄) ≥ 0 since the only strings of (ν̄, J̄)(i) which may
have negative riggings are
(
`(i) − 1, P
(i)
`(i)−1
(ν̄)
)
and
(
`(i) − 1, P
(i)
`(i)−1(ν̄)
)
. Then by the convexity
relation of P
(i)
k (ν̄) between j ≤ k ≤ `(i) − 1 we have
P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄).
Rigged Configurations and Kashiwara Operators 85
Recall that the left hand side of the convexity relation of Lemma 3.13 with l = `(i) − 1 is
−P (i)
`(i)−2
(ν̄) + 2P
(i)
`(i)−1
(ν̄)− P (i)
`(i)
(ν̄) ≤ P (i)
`(i)−1
(ν̄)
by P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄) and P
(i)
`(i)
(ν̄) = −1 by `(i) = `(i) − 1. On the other hand, the right
hand side of the convexity relation of Lemma 3.13 is
m
(i−1)
`(i)−1
(ν̄)− 2m
(i)
`(i)−1
(ν̄) +m
(i+1)
`(i)−1
(ν̄) = m
(i−1)
`(i)−1
(ν̄)− 2 +m
(i+1)
`(i)−1
(ν̄) ≥ m(i−1)
`(i)−1
(ν̄).
Here we use the fact m
(i+1)
`(i)−1
(ν̄) ≥ 2 which follows from `(i) − 1 = `(i+1) − 1 = `(i+1) − 1. This is
a contradiction since we have (7.3). �
Lemma 7.8. The following situation cannot happen. ẽi(ν, J) = 0, ẽi(ν̄, J̄) 6= 0 and
`(i) − 1 = (`(i+1) − 1)− 1 = (`(i) − 1)− 1.
Proof. In this case we have P
(i)
`(i)−1(ν) ≤ P
(i)
`(i)−1(ν̄). Suppose if possible that we have
P
(i)
`(i)−1(ν̄) < 0. Then the string (`(i) − 1, P
(i)
`(i)−1(ν)) of (ν, J)(i) has a negative rigging. This
is a contradiction since we have ẽi(ν, J) = 0. Thus we can assume that P
(i)
`(i)−1(ν̄) = P
(i)
`(i)
(ν̄) ≥ 0
by `(i) = `(i) − 1. Then we see that we must have
P
(i)
`(i)−1
(ν̄) < 0
since we have ẽi(ν̄, J̄) 6= 0.
Let j be the largest integer such that j < `(i) − 1 and m
(i)
j (ν̄) > 0. If there is no such
string, set j = 0. Then we have P
(i)
j (ν̄) ≥ 0. From the convexity relation of P
(i)
k (ν̄) between
j ≤ k ≤ `(i) − 1, we have
P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄)
since P
(i)
`(i)−1
(ν̄) < 0.
Let us show that m
(i)
`(i)−1
(ν̄) = 1. Suppose if possible that we have m
(i)
`(i)−1
(ν̄) > 1. Since we
have P
(i)
`(i)−1
(ν̄) < 0 and `(i)− 1 < `(i)− 1, we see that there is the remaining negative rigging in
(ν, J)(i). This is a contradiction since we have ẽi(ν, J) = 0. Thus we have m
(i)
`(i)−1
(ν̄) = 1.
Case 1. Suppose that we have `(i)−1 = (`(i+1)−1)−1. Then we have P
(i)
`(i)
(ν) = P
(i)
`(i)
(ν̄)−1.
From ẽi(ν, J) = 0 we have P
(i)
`(i)
(ν̄) ≥ 1. To summarize, we have
P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄) < 0, P
(i)
`(i)
(ν̄) ≥ 1.
Let n
(a)
l be the length of the lth column of ν̄(a)
n
(a)
l :=
∑
l≤k
m
(a)
k (ν̄).
Then the above relations imply that
n
(i−1)
`(i)−1
− 2n
(i)
`(i)−1
+ n
(i+1)
`(i)−1
≤ −1, n
(i−1)
`(i)
− 2n
(i)
`(i)
+ n
(i+1)
`(i)
≥ 2. (7.4)
86 R. Sakamoto
Since we have m
(i)
`(i)−1
(ν̄) = 1, we have n
(i)
`(i)
= n
(i)
`(i)−1
− 1 here. Then the second relation of (7.4)
becomes n
(i−1)
`(i)
−2n
(i)
`(i)−1
+n
(i+1)
`(i)
≥ 0. Combining this with the first relation of (7.4), we obtain
0 >
(
n
(i−1)
`(i)−1
− n(i−1)
`(i)
)
+
(
n
(i+1)
`(i)−1
− n(i+1)
`(i)
)
.
This is a contradiction since we have n
(a)
l ≥ n
(a)
l+1 by definition of n
(a)
l .
Case 2. Suppose that we have `(i) − 1 = `(i+1) − 1. In particular, we have m
(i+1)
`(i)−1
(ν̄) > 0.
As in the previous case, we have
P
(i)
`(i)−2
(ν̄) > P
(i)
`(i)−1
(ν̄) < 0, P
(i)
`(i)
(ν̄) ≥ 0.
Therefore we have
n
(i−1)
`(i)−1
− 2n
(i)
`(i)−1
+ n
(i+1)
`(i)−1
≤ −1, n
(i−1)
`(i)
− 2n
(i)
`(i)
+ n
(i+1)
`(i)
≥ 1. (7.5)
From n
(i)
`(i)
= n
(i)
`(i)−1
− 1, the second relation of (7.5) gives n
(i−1)
`(i)
− 2n
(i)
`(i)−1
+ n
(i+1)
`(i)
≥ −1. Then
from the first relation of (7.5), we obtain
0 ≥
(
n
(i−1)
`(i)−1
− n(i−1)
`(i)
)
+
(
n
(i+1)
`(i)−1
− n(i+1)
`(i)
)
.
Therefore we have n
(i−1)
`(i)−1
= n
(i−1)
`(i)
and n
(i+1)
`(i)−1
= n
(i+1)
`(i)
. In particular, we have m
(i+1)
`(i)−1
(ν̄) = 0,
which is a contradiction. Hence this case cannot happen. �
7.4 Proof for (8)
Proposition 7.9. Let b = b′ ⊗ b̄ ∈ B1,1 ⊗ B̄ and i ∈ I0. Then the situation ẽi(b) = 0 and
ẽi(b̄) 6= 0 cannot happen.
Proof. The proof of this assertion follows from the tensor product rule at Section 2.1. �
8 Comments on Deka–Schilling’s work
In Appendix C of the original preprint version of [4], Deka and Schilling proved the compatibility
of the classical Kashiwara operators and the rigged configuration bijection of type A
(1)
n . I am
pleased to admit that they proved some portion of the entire proof and also I would like to
sincerely admire that they embarked on such a difficult problem. However, due to the following
crucial problems, the author assumes that they did not finish their proof.
(a) In [4], the proofs for ẽi case are completely omitted since they claim at Lemmas C.2 and C.3
(i.e., not in the main proof) that the proofs are “similar” to f̃i case. Actually, as we have
shown in Sections 6 and 7 of the present paper, the logical structure of two proofs are
entirely different. In particular, the proofs for ẽi case require the result of f̃i case which
was proved beforehand. It is natural as the statements of the necessary propositions are
already different.
(b) In [4], the authors give entirely no words about the riggings. As we have shown in the
present paper, the coincidence of the riggings is non-trivial in many cases since the Kashi-
wara operators on the rigged configurations modify riggings. Thus they strongly deserve
individual and careful analysis.
(c) In [4], the authors give no words to Cases A and E of the present paper. They simply claim
that the only non-trivial cases are the Cases B, C and D. However, as we have shown in
the present paper, each proof of Case A or E breaks into subcases and requires somewhat
elaborated arguments even if we restrict to the type A
(1)
n . Thus it is necessary to mention
about these cases.
Rigged Configurations and Kashiwara Operators 87
Acknowledgements
The author would like to thank Professor Masato Okado for valuable discussion throughout the
project. In particular, the project began when both of us worked together to understand the
details of Appendix C of the preprint version of [4]. The author is also grateful to Professor Anne
Schilling for valuable discussion on her results. For both of them, the author is very grateful
for the fruitful collaboration on closely related projects [24, 25] which provides a motivation
and an important application of the present work. I would like to thank anonymous referees
for valuable suggestions which greatly help the author to improve the original manuscript. This
work is partially supported by Grants-in-Aid for Scientific Research No. 21740114 from JSPS.
References
[1] Alba V.A., Fateev V.A., Litvinov A.V., Tarnopolskiy G.M., On combinatorial expansion of the conformal
blocks arising from AGT conjecture, Lett. Math. Phys. 98 (2011), 33–64, arXiv:1012.1312.
[2] Belavin A., Belavin V., AGT conjecture and integrable structure of conformal field theory for c = 1, Nuclear
Phys. B 850 (2011), 199–213, arXiv:1102.0343.
[3] Cai W., Jing N., Applications of a Laplace–Beltrami operator for Jack polynomials, European J. Combin.
33 (2012), 556–571, arXiv:1101.5544.
[4] Deka L., Schilling A., New fermionic formula for unrestricted Kostka polynomials, J. Combin. Theory Ser. A
113 (2006), 1435–1461, math.CO/0509194.
[5] Desrosiers P., Lapointe L., Mathieu P., Superconformal field theory and Jack superpolynomials, J. High
Energy Phys. 2012 (2012), no. 9, 037, 42 pages, arXiv:1205.0784.
[6] Estienne B., Pasquier V., Santachiara R., Serban D., Conformal blocks in Virasoro and W theories: duality
and the Calogero–Sutherland model, Nuclear Phys. B 860 (2012), 377–420, arXiv:1110.1101.
[7] Feigin B.L., Fuchs D.B., Representations of the Virasoro algebra, in Representation of Lie Groups and
Related Topics, Adv. Stud. Contemp. Math., Vol. 7, Gordon and Breach, New York, 1990, 465–554.
[8] Feigin M.V., Hallnäs M.A., Veselov A.P., Baker–Akhiezer functions and generalised Macdonald–Mehta in-
tegrals, J. Math. Phys. 54 (2013), 052106, 22 pages, arXiv:1210.5270.
[9] Fukuda K., Yamada Y., Okado M., Energy functions in box ball systems, Internat. J. Modern Phys. A 15
(2000), 1379–1392, math.QA/9908116.
[10] Hatayama G., Hikami K., Inoue R., Kuniba A., Takagi T., Tokihiro T., The A
(1)
M automata related to
crystals of symmetric tensors, J. Math. Phys. 42 (2001), 274–308, math.QA/9912209.
[11] Hatayama G., Kuniba A., Takagi T., Soliton cellular automata associated with crystal bases, Nuclear Phys. B
577 (2000), 619–645, solv-int/9907020.
[12] Hatayama G., Kuniba A., Takagi T., Simple algorithm for factorized dynamics of the gn-automaton,
J. Phys. A: Math. Gen. 34 (2001), 10697–10705, nlin.CG/0103022.
[13] Kac V.G., Infinite-dimensional Lie algebras, 3rd ed., Cambridge University Press, Cambridge, 1990.
[14] Kashiwara M., On crystal bases of the Q-analogue of universal enveloping algebras, Duke Math. J. 63 (1991),
465–516.
[15] Kashiwara M., Bases cristallines des groupes quantiques, Cours Spécialisés, Vol. 9, Société Mathématique
de France, Paris, 2002.
[16] Kashiwara M., Nakashima T., Crystal graphs for representations of the q-analogue of classical Lie algebras,
J. Algebra 165 (1994), 295–345.
[17] Kerov S.V., Kirillov A.N., Reshetikhin N.Y., Combinatorics, the Bethe ansatz and representations of the
symmetric group, J. Soviet Math. 41 (1988), 916–924.
[18] Kirillov A.N., Reshetikhin N.Y., The Bethe ansatz and the combinatorics of Young tableaux, J. Soviet
Math. 41 (1988), 925–955.
[19] Kirillov A.N., Schilling A., Shimozono M., A bijection between Littlewood–Richardson tableaux and rigged
configurations, Selecta Math. (N.S.) 8 (2002), 67–135, math.CO/9901037.
[20] Kuniba A., Okado M., Sakamoto R., Takagi T., Yamada Y., Crystal interpretation of Kerov–Kirillov–
Reshetikhin bijection, Nuclear Phys. B 740 (2006), 299–327, math.QA/0601630.
http://dx.doi.org/10.1007/s11005-011-0503-z
http://arxiv.org/abs/1012.1312
http://dx.doi.org/10.1016/j.nuclphysb.2011.04.014
http://dx.doi.org/10.1016/j.nuclphysb.2011.04.014
http://arxiv.org/abs/1102.0343
http://dx.doi.org/10.1016/j.ejc.2011.11.003
http://arxiv.org/abs/1101.5544
http://dx.doi.org/10.1016/j.jcta.2006.01.003
http://arxiv.org/abs/math.CO/0509194
http://dx.doi.org/10.1007/JHEP09(2012)037
http://dx.doi.org/10.1007/JHEP09(2012)037
http://arxiv.org/abs/1205.0784
http://dx.doi.org/10.1016/j.nuclphysb.2012.03.007
http://arxiv.org/abs/1110.1101
http://dx.doi.org/10.1063/1.4804615
http://arxiv.org/abs/1210.5270
http://dx.doi.org/10.1142/S0217751X00000616
http://arxiv.org/abs/math.QA/9908116
http://dx.doi.org/10.1063/1.1322077
http://arxiv.org/abs/math.QA/9912209
http://dx.doi.org/10.1016/S0550-3213(00)00105-X
http://arxiv.org/abs/solv-int/9907020
http://dx.doi.org/10.1088/0305-4470/34/48/331
http://arxiv.org/abs/nlin.CG/0103022
http://dx.doi.org/10.1017/CBO9780511626234
http://dx.doi.org/10.1215/S0012-7094-91-06321-0
http://dx.doi.org/10.1006/jabr.1994.1114
http://dx.doi.org/10.1007/BF01247087
http://dx.doi.org/10.1007/BF01247088
http://dx.doi.org/10.1007/BF01247088
http://dx.doi.org/10.1007/s00029-002-8102-6
http://arxiv.org/abs/math.CO/9901037
http://dx.doi.org/10.1016/j.nuclphysb.2006.02.005
http://arxiv.org/abs/math.QA/0601630
88 R. Sakamoto
[21] Kuniba A., Sakamoto R., Yamada Y., Generalized energies and integrable D
(1)
n cellular automaton, in New
Trends in Quantum Integrable Systems, World Sci. Publ., Hackensack, NJ, 2011, 221–242, arXiv:1001.1813.
[22] Mimachi K., Yamada Y., Singular vectors of the Virasoro algebra in terms of Jack symmetric polynomials,
Comm. Math. Phys. 174 (1995), 447–455.
[23] Okado M., Sakamoto R., Stable rigged configurations for quantum affine algebras of nonexceptional types,
Adv. Math. 228 (2011), 1262–1293, arXiv:1008.0460.
[24] Okado M., Sakamoto R., Schilling A., Affine crystal structure on rigged configurations of type D
(1)
n , J. Al-
gebraic Combin. 37 (2013), 571–599, arXiv:1109.3523.
[25] Okado M., Sakamoto R., Schilling A., Kerov–Kirillov–Reshetikhin type bijection for D
(1)
n , in preparation.
[26] Okado M., Sano N., KKR type bijection for the exceptional affine algebra E
(1)
6 , in Algebraic Groups
and Quantum Groups, Contemp. Math., Vol. 565, Amer. Math. Soc., Providence, RI, 2012, 227–242,
arXiv:1105.1636.
[27] Okado M., Schilling A., Shimozono M., A crystal to rigged configuration bijection for nonexceptional affine
algebras, in Algebraic Combinatorics and Quantum Groups, World Sci. Publ., River Edge, NJ, 2003, 85–124,
math.QA/0203163.
[28] Sakamoto R., Kirillov–Schilling–Shimozono bijection as energy functions of crystals, Int. Math. Res. Not.
2009 (2009), 579–614, arXiv:0711.4185.
[29] Sakamoto R., Ultradiscrete soliton systems and combinatorial representation theory, arXiv:1212.2774.
[30] Sakamoto R., Shiraishi J., Arnaudon D., Frappat L., Ragoucy E., Correspondence between conformal field
theory and Calogero–Sutherland model, Nuclear Phys. B 704 (2005), 490–509, hep-th/0407267.
[31] Schilling A., A bijection between type D
(1)
n crystals and rigged configurations, J. Algebra 285 (2005),
292–334, math.QA/0406248.
[32] Schilling A., Crystal structure on rigged configurations, Int. Math. Res. Not. 2006 (2006), Art. ID 97376,
27 pages, math.QA/0508107.
[33] Schilling A., Combinatorial structure of Kirillov–Reshetikhin crystals of type D
(1)
n , B
(1)
n , A
(2)
2n−1, J. Algebra
319 (2008), 2938–2962, arXiv:0704.2046.
[34] Schilling A., Shimozono M., X = M for symmetric powers, J. Algebra 295 (2006), 562–610,
math.QA/0412376.
[35] Sutherland B., Exact results for a quantum many-body problem in one dimension, Phys. Rev. A 4 (1971),
2019–2021.
[36] Sutherland B., Exact results for a quantum many-body problem in one dimension. II, Phys. Rev. A 5 (1972),
1372–1376.
[37] Takahashi D., Satsuma J., A soliton cellular automaton, J. Phys. Soc. Japan 59 (1990), 3514–3519.
http://dx.doi.org/10.1142/9789814324373_0012
http://dx.doi.org/10.1142/9789814324373_0012
http://arxiv.org/abs/1001.1813
http://dx.doi.org/10.1007/BF02099610
http://dx.doi.org/10.1016/j.aim.2011.06.012
http://arxiv.org/abs/1008.0460
http://dx.doi.org/10.1007/s10801-012-0383-z
http://dx.doi.org/10.1007/s10801-012-0383-z
http://arxiv.org/abs/1109.3523
http://dx.doi.org/10.1090/conm/565/11181
http://arxiv.org/abs/1105.1636
http://dx.doi.org/10.1142/9789812775405_0005
http://arxiv.org/abs/math.QA/0203163
http://dx.doi.org/10.1093/imrn/rnn140
http://arxiv.org/abs/0711.4185
http://arxiv.org/abs/1212.2774
http://dx.doi.org/10.1016/j.nuclphysb.2004.10.005
http://arxiv.org/abs/hep-th/0407267
http://dx.doi.org/10.1016/j.jalgebra.2004.12.010
http://arxiv.org/abs/math.QA/0406248
http://dx.doi.org/10.1155/IMRN/2006/97376
http://arxiv.org/abs/math.QA/0508107
http://dx.doi.org/10.1016/j.jalgebra.2007.10.020
http://arxiv.org/abs/0704.2046
http://dx.doi.org/10.1016/j.jalgebra.2005.04.023
http://arxiv.org/abs/math.QA/0412376
http://dx.doi.org/10.1103/PhysRevA.4.2019
http://dx.doi.org/10.1103/PhysRevA.5.1372
http://dx.doi.org/10.1143/JPSJ.59.3514
1 Introduction
2 Background on crystals and tableaux
2.1 Crystal bases
2.2 Kirillov–Reshetikhin tableaux
2.2.1 Reviews on tableau representations
2.2.2 Filling map
3 Rigged configurations and the main bijection
3.1 Definition of the rigged configurations
3.2 Convexity relations of the vacancy numbers and their applications
3.3 Rigged configuration bijection
4 Rigged configurations and the Kashiwara operators
4.1 Statement of the main result
4.2 Preliminary steps
4.3 Reduction steps
4.4 Statements to be proved
4.5 Spin cases
5 Proof of Proposition 4.7
5.1 Outline
5.2 Proof for Case A
5.3 Proof for Case B
5.3.1 Case 1: f"0365fi creates a non-singular string
5.3.2 Case 2: f"0365fi creates a singular string
5.4 Proof for Case C: preliminary steps
5.4.1 Classification
5.4.2 A common property for Case C
5.5 Proof for Case C (1)
5.6 Proof for Case C (2)
5.6.1 Outline
5.6.2 Classification
5.6.3 Proof for the case (5.4)
5.6.4 Proof for the case (5.5)
5.7 Proof for Case C (3)
5.8 Proof for Case D: preliminary steps
5.8.1 Classification
5.8.2 A common property for Case D
5.9 Proof for Case D (1)
5.9.1 The case <j
5.9.2 The case =j
5.10 Proof for Case D (2)
5.11 Proof for Case D (3)
5.11.1 Outline
5.11.2 Proof of Proposition 5.30 for the case (i-1)<
5.11.3 Proof of Proposition 5.30 for the case (i-1)=
5.12 Proof for Case D (4)
5.12.1 Outline
5.12.2 Proof of Proposition 5.39 for the case (i-1)<
5.12.3 Proof of Proposition 5.39 for the case (i-1)=
5.13 Proof for Case E: preliminary steps
5.13.1 Classification
5.13.2 A common property for Case E
5.14 Proof for Case E (1)
5.15 Proof for Case E (2)
5.15.1 Case 1: f"0365fi creates a non-singular string
5.15.2 Case 2: f"0365fi creates a singular string
5.16 Proof for Case E (3)
5.16.1 Proof for Case 1
5.16.2 Proof for Case 2
5.16.3 Proof for Case 3
5.17 Proof for Case E (4)
5.17.1 Preliminary
5.17.2 Proof for Case 1
5.17.3 Proof for Case 3
5.18 Proof for Case E (5)
6 Proof of Proposition 4.8: f"0365fi version
6.1 Proof for (1)
6.2 Proof for (2)
6.3 Proof for (3)
6.4 Proof for (4)
7 Proof of Proposition 4.8: e"0365ei version
7.1 Proof for (5)
7.2 Proof for (6)
7.3 Proof for (7)
7.4 Proof for (8)
8 Comments on Deka–Schilling's work
References
|
| id | nasplib_isofts_kiev_ua-123456789-146830 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 1815-0659 |
| language | English |
| last_indexed | 2025-12-07T15:58:23Z |
| publishDate | 2014 |
| publisher | Інститут математики НАН України |
| record_format | dspace |
| spelling | Sakamoto, R. 2019-02-11T16:42:01Z 2019-02-11T16:42:01Z 2014 Rigged Configurations and Kashiwara Operators / R. Sakamoto // Symmetry, Integrability and Geometry: Methods and Applications. — 2014. — Т. 10. — Бібліогр.: 37 назв. — англ. 1815-0659 2010 Mathematics Subject Classification: 17B37; 05E10; 82B23 DOI:10.3842/SIGMA.2014.028 https://nasplib.isofts.kiev.ua/handle/123456789/146830 For types An⁽¹⁾ and Dn⁽¹⁾ we prove that the rigged configuration bijection intertwines the classical Kashiwara operators on tensor products of the arbitrary Kirillov-Reshetikhin crystals and the set of the rigged configurations. This paper is a contribution to the Special Issue in honor of Anatol Kirillov and Tetsuji Miwa. The full collection is available at http://www.emis.de/journals/SIGMA/InfiniteAnalysis2013.html. The author would like to thank Professor Masato Okado for valuable discussion throughout the project. In particular, the project began when both of us worked together to understand the details of Appendix C of the preprint version of [4]. The author is also grateful to Professor Anne Schilling for valuable discussion on her results. For both of them, the author is very grateful for the fruitful collaboration on closely related projects [24, 25] which provides a motivation and an important application of the present work. I would like to thank anonymous referees for valuable suggestions which greatly help the author to improve the original manuscript. This work is partially supported by Grants-in-Aid for Scientific Research No. 21740114 from JSPS. en Інститут математики НАН України Symmetry, Integrability and Geometry: Methods and Applications Rigged Configurations and Kashiwara Operators Article published earlier |
| spellingShingle | Rigged Configurations and Kashiwara Operators Sakamoto, R. |
| title | Rigged Configurations and Kashiwara Operators |
| title_full | Rigged Configurations and Kashiwara Operators |
| title_fullStr | Rigged Configurations and Kashiwara Operators |
| title_full_unstemmed | Rigged Configurations and Kashiwara Operators |
| title_short | Rigged Configurations and Kashiwara Operators |
| title_sort | rigged configurations and kashiwara operators |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/146830 |
| work_keys_str_mv | AT sakamotor riggedconfigurationsandkashiwaraoperators |