CAPÍTULO 3. LOS CAMINOS RURALES: ANTECEDENTES Y
3.4 ESTABILIZACIÓN QUÍMICA DE SUELOS
3.4.3 Acción de la cal sobre los suelos arcillosos
This section is an informal discussion of the extreme relations in type
Awithout formal proofs.
We relate the combinatorial notions for minuscule posets in this case to established combinatorial
notions for Young diagrams. Interestingly, these combinatorial notions of “content” and “rim hooks”
usually arise from the representation theory of the symmetric group. Use the labelling of typeA
n−1roots and weights from Section 5.2. Fix a minuscule weightλ=ω
j.
Rotate the Hasse diagram for the poset
P
:=a
n−1(j) as pictured in Figure 8.2 clockwise by
45
◦to obtain a rectangular array of dots with
j
rows and
n−j
columns. A rotated filter of the
poset is a left justified subarray of dots such that the number of dots in each of its rows is weakly
decreasing. Fix a filter
J
⊂P. Let
µ1, µ2, . . . , µ
jbe the number of dots in the 1st, 2nd,. . .,
jth
rows of the rotated depiction of
J. Then
µ
:= (µ
1, µ
2, . . . , µ
n−j) is a partition. The dots in the
rotated depiction of
J
is the
Ferrer’s diagram
of
µ. If we replace each dot with a box, then we
obtain the Young diagram ofµ. This rotation process is a bijection from the filters ofP
(and hence
a basis of the
g-moduleV
J(P)) to the Young diagrams fitting inside aj×n−j
rectangle.
In this depiction of filters, the boxes represent elements of
P. It is not difficult to verify the
following description of the coloring functionκ
onP. Each diagonal of a Young diagram is assigned
a distinct color. The diagonal consisting of the location (j,1) at the southwest corner of thej×n−j
rectangle is assigned color 1. The diagonal of locations (j−1,1) and (j,2) is assigned color 2,
and so on. In particular, the main diagonal of locations (1,1),(2,2), . . . ,(j, j) is assigned color
j.
This coloring of the locations is similar to the usual notion of the
content
of a location in a Young
diagram, except that contents on such diagrams range from 1−j
ton−1−j
instead of 1 to
n−1.
We now show that
α-layers for rootsα∈Φ
+are simply “rim hooks” in this formulation.
Definition 9.25.
A
rim hook
is a connected subset of boxes in the
j×n−j
array that does not
have two boxes on the same diagonal.
It is possible to show thatα-layers must be connected. In typeA
n−1, the simple root expansion of a
rootα∈Φ
+isα=α
i+α
i+1+· · ·+α
jfor some 1≤i≤j
≤n−1. Fix such an
α, and suppose that
R⊂P
is anα-layer. Then
R
has a single element of each of the colorsi, i+ 1, . . . , j. Hence when
rotated, the subsetR
does not have two boxes on the same diagonal. Therefore the rotated depiction
ofRis a rim hook. On the other hand, suppose thatH
is a rim hook in thej×n−jarray. It clearly
corresponds to a convex subset of
P. Leti
denote the color of its southwesternmost box, and letj
denote the color of its northeasternmost box. SinceH
is connected, it must have a box in each of
the diagonals of color
i+ 1, . . . , j−1. SinceH
does not have two boxes on the same diagonal, it has
exactly one box of each of these colors. Therefore we have
Pand thatH
is anα-layer.
According to Theorem 9.5, the extreme weight Plücker relations are standard straightening laws
on double-tailed diamond sublattices of
L
λ. As we saw in Section 9.2, in typeA
these double-tailed
diamond lattices are isomorphic to the one that appears in the modelD3
case. Hence there are
only three terms in these straightening laws: the product of the incomparable pair, the product of
their meet and join, and one other standard monomial. It is simple to describe this last standard
monomial in typeA
without using the determinantal Plücker relations from Section 3.1. First we
need some preliminary definitions:
Definition 9.26.
Letα, β∈Φ
+and writeα=α
i+α
i+1+· · ·+α
jandβ
=α
k+α
k+1+· · ·+α
`. If
j < k−1 or
` < i−1, thenα
andβ
are said to beseparated. In that case, without loss of generality
suppose
j < k
and define theirbridge root
γ(α, β) to beγ(α, β) :=α
j+1+α
j+2+· · ·+α
k−1.
Recall that for any root
α∈Φ
+, we defined the actions of root vectors
e
α, f
αof
g
onV
J(P)in
Section 8.2.
Proposition 9.27.
Let
J, K∈J(P)
be incomparable filters such that
J−(J∨K)
and
K−(J∨K)
are rim hooks. Then these rim hooks are root layers for separated rootsα, β
∈Φ
+and we have the
type
A
straightening law:
J K= (J ∨ K)(J ∧ K)−(e
γ(α,β).[J ∨ K])(f
γ(α,β).[J ∧ K]).
The filters of the two vectors in the final standard monomial are obtained from the meet and join by
transferring a rim hook of color
γ(α, β)
from the join filter to the meet filter.
In Section 9.1, we claimed that the Plücker moduleI
is minuscule in typeA
n−1for the minuscule
weights
ω2
andω
n−2. In these cases, the extreme weight Plücker relations given by Theorem 9.5
and described in detail by Proposition 9.27 above form a basis of
I. We now confirm that claim.
First define
ω0
andω
nto be the trivial weight 0. For
λ=ω
jwe can see from the diagrams of the
top incomparable pairZ
], Z
[that the highest weightη
of the submoduleI
isω
j−2