Beruflich Dokumente
Kultur Dokumente
r
X
i
v
:
1
0
0
3
.
0
0
4
6
v
1
[
m
a
t
h
-
p
h
]
1
M
a
r
2
0
1
0
Experimental evidence for the occurrence of E
8
in nature
and the radii of the Gosset circles
Bertram Kostant
Abstract. A recent experimental discovery involving the spin structure of electrons in a cold one-
dimensional magnet points to a validation of a Zamolodchikov model involving the exceptional Lie
group E
8
. The model predicts 8 particles and predicts the ratio of their masses. In more detail,
the vertices of the 8-dimensional Gosset polytope identies with the 240 roots of E
8
. Under the
famous two-dimensional (Peter McMullen) projection of the polytope, the image of the vertices are
arranged in 8 concentric circles, hereafter referred to as the Gosset circles. The Gosset circles are
now understood to correspond to the 8 masses in the model, and in addition it is understood that
the ratio of the their radii is the same as the ratio of the corresponding conjectural masses. A ratio of
the two smallest circles (read 2 smallest masses) is the golden number. Marvelously, the conjectures
have been now validated experimentally, at least for the rst ve masses.
The McMullen projection generalizes to any complex simple Lie algebra whose rank is greater
than 1. The Gosset circles generalize as well, using orbits of the Coxeter element. Using results in
[K-59], I found some time ago a very easily dened operator A whose spectrum is exactly the squares
of the radii r
i
of these generalized Gosset circles. As a conrmation, in the E
8
case, using only the
eigenvalues of a suitable multiple of A, Vogan computed the ratio of the r
i
. Happily these agree
with the corresponding ratio of the Zamolodchikov masses.
The operator A is written as a sum of + 1 rank 1 operators, parameterized by the points in
the extended Dynkin diagram. Involved in this expansion are the coecients n
i
of the highest root.
Suggestively, recalling the McKay correspondence, in the E
8
case the n
i
, together with 1, are the
dimensions of the irreducible representations of the binary icosahedral group.
Key Words: E
8
, Cartan subalgebras in apposition, Gosset circles, Ising chain in E
8
symmetry,
Zamolodchikov theory, 1-dimensional magnet, Coxeter element, Coxeter number, golden number,
conformal eld theory, particle physics.
Mathematics Subject Classication (2010): Primary 20G41, 20G45, 81R05, 81T10.
0. Introduction
0.1. Let g be a complex simple Lie algebra and let (x, y) be the Killing form B on g.
Let = rank g and let h be a Cartan subalgebra of g. Let be the set of roots for (h, g)
and let
+
be a choice of positive roots. For any , let e
be a corresponding
root vector. We assume choices are made so that (e
, e
) = 1.
Let =
1
, . . . ,
+
be the set of simple positive roots. Let h be the Coxeter
number of g. Let w h be the unique element such that
i
, w = 1, i = 1, . . . , . Let G
be a Lie group such that Lie G = g and let H G be the subgroup corresponding to h.
Let c H be given by c = exp 2i w/h. Then c is a regular semisimple element of G and
its centralizer in g is given by
g
c
= h. (0.1)
1
Let
+
be the highest root and let n
i
, i = 1, . . . , , be the coecients (known to
be positive) relative to the simple roots so that =
i=1
n
i
i
. Let
x() = e
i=1
n
i
e
i
. (0.2)
Then results in [K-59] imply x() is a regular semisimple element of g. Let h() be the
Cartan subalgebra of g which contains x(). Furthermore let = e
2i/h
so that is a
primitive h root of unity. The following is proved in [K-59].
Theorem 0.1 The Cartan subalgebra h() is stable under Ad c. Furthermore if
=
Ad c[h(), then
= e
O
i
e
are a basis of h.
0.2. We assume from now on > 1 so that h > 2. Let Vec h() be the real space
of all hyperbolic elements in h() so that Vec h() is a W()-stable real form of h(). If
conjugation in h() is dened with respect to Vec h(), then x() is a regular eigenvector
of
-stable subspace
Y of Vec h() by putting
Y = Vec h() (Cx() +Cx()). (0.4)
Let
Q : Vec h() Y (0.5)
be the orthogonal projection (and
h()
2
dened by B. One denes circles C
i
, i = 1, . . . , , in Y of positive radius r
i
, centered at
the origin, by the condition that
Q(w
) C
i
, O
i
.
In the special case where G = E
8
we will refer to the C
i
as Gosset circles.
Remark. If
. If 0 ,= x()
is a corresponding
(necessarily regular) eigenvector, then one may replace (, x()) by (
, x()
) and replace
(Q, Y ) by a corresponding (Q
, Y
=
h
, (0.6)
where
h
is the Galois group of the cyclotomic eld spanned over Q by the h roots of unity.
In the E
8
case the projection Q of the Gosset polytope appears ubiquitously through-
out the mathematical literature (see e.g., the frontispiece of [CX]). It has been described
by Coxeter as the most symmetric two-dimensional projection of this polyhedron. But
in fact it can be described precisely as the unique such projection, up to isomorphism,
(there are 4 such projections) which commute with the action of the Coxeter element
.
The following Theorem 0.3 below is our main result. It gives the radii r
i
of the circles
C
i
. Its signicance in the E
8
case will be explained in 0.3. If x, y h, let x y be the
rank 1 operator on h, dened so that if z h, then x y(z) = (x, z) y. Also let w
i
h be
the image of
i
in h under the isomorphism h
i=0
n
i
w
i
w
i
, (0.7)
where n
0
= 1 and, we recall, n
i
, i > 0, is the coecient of
i
in the simple root expansion
of the highest root .
Theorem 0.3. The eigenvalues of A are r
2
i
, i = 1, . . . , , and z
i
(see Theorem 0.2) is
an A-eigenvector for r
2
i
.
0.3. I obtained the result Theorem 0.3 sometime (I believe) in the early 1990s. At
that time publication seemed unwarranted since I believed there would be little interest
in a knowledge of the radii r
i
. However in the middle 1990s Peter McMullens image,
by Q, in the E
8
case, of the Gosset polytope was very widely published and became well
known, even to many in the general public, due no doubt, to the very extensive (and well
3
deserved) publicity, given to the determination by a large team of mathematicians, of the
characters of the real forms of E
8
. I showed Theorem 0.3 to David Vogan, one of the
leading members of the aforementioned team. Applying a computer program to a Weyl
group reformulation of a scalar multiple of the operator A, the following list, in increasing
size, of the normalized 8 radii was obtained by Vogan. His normalization was to make the
largest of the 8 radii equal to 1. To avoid decimals we took the liberty of multiplying his
list by 1000. (What will be signicant is the ratio of the radii and not the radii themselves).
209
338
416
502
618
673
813
1000
(0.8)
Remark. The rst 7 numbers in (0.8) are the integral parts of the normalized radii
and to that extent only approximate the normalized radii.
Having been directed by colleagues to the papers [Za] and [Co], we have only recently
become aware of the fact that the ratio of the numbers in (0.8) have physical signicance.
In [Za] Zamolodchikov conjectures the existence of 8 particles in connection with a eld
theory associated with the Ising model. Happily his ratio of the masses of the conjectured
particles agrees with the ratio of the normalized radii in (0.8). (See (1.8), p. 4237 in [Za].)
The use of the quotation marks in agrees, and also below, is because of the statement in
the remark above. Particular emphasis is made in [Za] of the fact that the ratio (m
2
/m
1
)
of the rst two masses should be the golden number. Indeed 338/209 =
1
2
(1 +
5).
Zamolodchikov makes a connection in his paper with E
8
at the bottom of p. 4247 but
attributes the prediction of an E
8
connection to V. Fateev.
The recent nine person authored paper [Co] is an experimental discovery, using a very
cold one-dimensional magnet, validating Zamolodchikovs theory, at least for the rst 5
particles. In particular the equality of m
2
/m
1
(the ratio of the radii of the two inner Gosset
circles) with the golden number is very clearly seen.
0.4. I wish to thank Nolan Wallach for many conversations regarding the subject
matter in this paper. In particular for conversations regarding the two types of Gosset
circles which appear in a number of publications. In addition I thank for him for the
minicourse he gave me on the representations of the Virasoro algebra. Representations of
this algebra make an appearance in [Za]. I also want to thank David Vogan for making
the computations, using Theorem 0.3, which resulted in (0.8). Also David factorized a
relevant characteristic polynomial into a product of two irreducible (over Q) polynomials
of degree 4. The irreducible polynomials relate directly to the two types of Gosset circles
mentioned above.
4
1. Cartan subalgebras in apposition
1.1. Let g be a complex simple Lie algebra and let (x, y) be the Killing form B on g.
Let = rank g and let h be a Cartan subalgebra of g. Let be the set of roots for (h, g)
and let
+
be a choice of positive roots. For any let e
be a corresponding
root vector. We assume choices are made so that
(e
, e
) = 1. (1.1)
Let =
1
, . . . ,
+
be the set of simple positive roots. Let h be the Coxeter
number of g. Let w h be the unique element such that for all i = 1, . . . , ,
i
, w = 1. (1.2)
For any put o() = , w. If
+
is the highest root, then one knows that
o() = h 1. (1.3)
Let G be a Lie group such that Lie G = g and let H G be the subgroup corresponding
to h. If a H and h
is an H-weight, put a
= e
x,
, where a = expx. Let c H be
given by
c = exp2i w/h. (1.4)
Also let
= e
2i/h
.
For g G and z g we will sometimes write g z for Adg(z). Let
g() = x g [ c x = x.
As one readily notes, using e.g., (1.3),
Proposition 1.1. One has dimg() = +1 and in fact the elements e
i
, i = 1, . . . , ,
and e
) C
+1
let x() g() be dened by putting
x() =
i=1
i
e
i
. (1.5)
The following result was established in [K-59].
5
Theorem 1.2. x() is regular semisimple if and only if (C
)
+1
.
If (C
)
+1
, let h() be the Cartan subalgebra which contains x(). It is immediate
that if (C
)
+1
, then h() is stable under Ad c. Let
)
+1
. Then
] where the
J
k
are homogeneous, say of degree d
k
, and algebraically independent. Choose the ordering
so that the d
k
are nonincreasing. In that case d
1
= h and d
k
< h for k > 1. The denition
of a cyclic element x g was introduced in [K-59]. One has that x is cyclic if and only
if J
1
(x) ,= 0 and J
k
(x) = 0 for k > 1. This condition is independent of the choice of the
J
k
. It is established in [K-59] that cyclic elements are regular semisimple. We have also
proved
Theorem 1.4. x() is cyclic for any (C
)
+1
and, up to conjugacy, any cyclic
element in g is of this form.
Let n
i
C, i = 1, . . . , , be dened so that
=
i=1
n
i
i
.
One knows that the n
i
are positive integers. It is immediate from the independence of the
simple roots that if k
i
Z
+
, i = 1, . . . , , and k
Z
+
are such that k
i=1
k
i
= h,
then the monomial
e
k
e
k
1
1
e
k
(1.7)
is in S
h
(g) and
Proposition 1.5. The monomial (1.7) is a zero weight vector if and only if k
= 1
and k
i
= n
i
for i = 1, . . . , .
6
It follows immediately from Theorem 1.4 and Proposition 1.5 that if (C
)
l+1
,
then
Proposition 1.6. With respect to the inner product in S(g) one has
(J
1
, e
e
m
1
1
e
m
) ,= 0
and
J
1
(x()) =
m
1
1
m
1
m
1
! m
!
(J
1
, e
e
m
1
1
e
m
). (1.8)
It is clear that the set of cyclic elelemnts of the form x() for (C
)
l+1
is stable
under conjugation by H. One readily denes an action of H on (C
)
l+1
so that if
(C
)
l+1
and a H, then
a x() = x(a ).
Of course (C
)
l+1
is stable under multiplication by C
on (C
)
l+1
where, for C
and (C
)
l+1
, one lets (C
)
l+1
be given so that
( )
i
=
i
, i = 1, . . . , ,
but
( )
.
Theorem 1.7.
(1) Two cyclic elements v, v
).
(2) Furthermore, if v = x(), v
= x(
) where , (C
)
l+1
, then v and v
are
G-conjugate they are H-conjugate.
(3) Given a cyclic element v g and (C
)
l+1
, there exists a unique C
such
that v and x( ) are G-conjugate.
Proof. (1) follows immediately from the fact that cyclic elements are regular semisim-
ple. To prove (2), assume that x() and x(
)
i
=
i
for i = 1, . . . , . But then (a
by (1.8) since
J
1
(x()) = J
1
(a
). Thus a x(
) is linear
in by (1.8). QED
1.2. Let G
Ad
be the adjoint group. Then we recall from [K-59] there exists a unique
conjugacy class C of regular elements of order h in G
Ad
. Furthermore if a G and
Ad a C, there exists a Cartan subalgebra a which is stable under Ad a and Ada [ a is
a Coxeter element. In such a case we will say that a is Coxeter for a. Conversely, if is
a Coxeter element for a Cartan subalgebra a and Ad a normalizes a and induces , then
Ad a C.
7
Recalling Theorem 1.3, one has that c C. Furthermore since any Cartan subalgebra
which is Coxeter for c necessarily has a regular eigenvector with eigenvalue for Ad c, one
has
Proposition 1.8. h() is Coxeter for c for any (C
)
l+1
. Conversely, any Cartan
which is Coxeter for c is equal to h() for some (C
)
l+1
.
Now note that by Theorem 1.3, if ,
(C
)
l+1
, then
h() = h(
= for some C
. (1.9)
On the other hand, if a H and (C
)
l+1
, then obviously
Ad a (h()) = h(a ). (1.10)
Theorem 1.9. Let c
. In
particular if c = c
)
l+1
.
Proof. Let ,
(C
)
l+1
, We have only to show that there exists a H such that
Ada (h()) = h(
). (1.11)
But now for any C
one has
J
1
(x(
)) =
h
J
1
(x(
)). (1.12)
But then we can choose so that
J
1
(x(
)) = J
1
(x()). (1.13)
But then by Theorem 1.7 there exists a H such that a x() = x(
). But h(
) = h(
). QED
1.3. Let (C
)
l+1
and let () be the set of roots for the pair (h(), g). Then
.
For any (), let e
h()
= e
.
We note, for O
i
, that we may write
e
= z
i
+
d
,
e
,
8
where z
i
h. One further notes that
c e
= z
i
+
o()
d
,
e
= e
= z
i
+
,
e
.
(1.14)
However since c is regular one has
g
c
= h.
But
C
i
e
is an invariant of c and hence lies in h. Thus from (1.14) one must have
C
i
e
= hz
i
. (1.15)
That is, for any ,
C
i
d
,
= 0, (1.16)
and in fact the orbits O
i
consequently dene a distinguished basis of h.
Theorem 1.10. The z
i
, i = 1, . . . , , given by (1.15), are a basis of h.
Proof. Since
, e
] = w
(1.17)
for any , where w
, x) = , x. (1.18)
Now recalling the notation of (1.8) one has by (1.18),
i=1
n
i
w
i
= w
. (1.19)
Now any Cartan subalgebra h
1
is the sum of its vector part Vec h
1
(split real Cartan
subalgebra) and its toroidal part Tor h
1
= i Vec h
1
(Cartan subalgebra of a compact real
form). In particular, for any (C
)
+1
one has
h() = Vec h() +i Vec h(). (1.20)
9
In particular
x() = x() +ix(), (1.21)
where x(), x() Vec h(). But now recalling (1.6) one has
x() = ( +i )(x() +ix())
= ( x() x()) +i ( x() + x()).
(1.22)
Since c
= , x(). (1.25)
One notes that
x() = x()
, x() =
.
(1.26)
Indeed the rst equation in (1.26) is obvious from (1.23). The second follows from the
fact that takes real values on Vec h().
Now let
x
() =
1
i=1
n
i
i
e
i
. (1.27)
It is immediate from (1.3) that
c x
() = x
(). (1.28)
Theorem 1.11. One has x
() h
and
() = x
(). (1.29)
In fact by Colemans uniqueness theorem there exists t
such that
t
x() = x
(). (1.30)
10
Proof. Since x() h() and is regular it clearly suces to prove that x
()
commutes with x(). Here one recalls (1.26) and (1.28). But
[x(), x
()] = [
i=1
i
e
i
,
1
i=1
n
i
i
e
i
]
= [e
, e
] +
i=1
n
i
[e
i
, e
i
]
= w
i=1
n
i
w
i
= 0 by (1.19). QED
Normalize (Weyls normal form) the choice of the e
, so that
(e
) = e
, (1.31)
where is an involution of g such that = 1 on h. In particular there exists a compact
form g
u
of g
e
is contained in g
u
for all . (1.32)
Let
(1)
(C
)
l+1
be dened so that
(1)
= 1 and
(1)
i
=
n
i
, i = 1, . . . , . Then one has
Theorem 1.12. One has x
(
(1)
) = x(
(1)
) so that t
(1) = 1.
Proof. One has
x(
(1)
) = e
i=1
n
i
e
i
(1.33)
and
x
((1)) = e
i=1
n
i
e
i
. (1.34)
But then
x(
(1)
) x
(
(1)
) = e
i=1
n
i
(e
i
e
i
). (1.35)
But then x(
(1)
) x
(
(1)
) g
u
by (1.32) and hence in particular, the corresponding
operator, for the adjoint representation, has pure imaginary spectrum. Thus x(
(1)
)
x
(
(1)
) h(
(1)
). But then
x(
(1)
) x
(
(1)
) h(
(1)
)
11
Hence
(t
(1) 1)(x
(
(1)
)) h(
(1)
)
by (1.30). That is, if s = i(t
(
(1)
)) h(
(1)
)
On the other hand h(
(1)
) is stable under
(
(1)
)) is a eigenvector for
)
l+1
and let
(C
be
dened by deleting the last entry
from . Let x(
) = x()
so that x(
) n
is principal nilpotent. For the opposed direction let b = b and let n = n. Then let
x
) = x
()
1
so that x
,= 0. (1.36)
For k Z let
g(k) = x g [ [w, x] = k x
so that one has the direct sum
g =
k=
g(k), (1.37)
and let
P
k
: g g(k)
be the projection dened by (1.36). We will let e
(k) = P
k
e
so that
e
k=
e
(k), (1.38)
noting that
e
(0) = z
i
. (1.39)
Of course g(k) = 0 for [k[ h so that e
(0) = z
i
.
Theorem 1.13. Let (C
)
l+1
and O
i
. Then for any positive integer k one
has
e
(k) =
1
[x(
), e
(k 1)]
=
1
(ad x(
))
k
z
i
.
(1.40)
12
Proof. By induction we have only to prove the rst line of (1.40). But now from the
root vector property
e
=
1
[x(), e
]. (1.41)
But since e
(k 1) replaces e
. QED
Remark. If one considers the principle TDS dened by
1
x(
) = e
and w, then
note that Theorem 1.12 asserts that e
-module generated by z
i
.
But we can also reverse direction. By (1.26) and (1.30)
, x
() = t
, x()
= t
.
But then e
is an eigenvector for ad (t
)
1
x
(k) = [(t
)
1
x
), e
(k + 1)]
= ((ad (t
)
1
x
))
k
z
i
.
(1.42)
Lemma 1.15. Let z h and (C
)
l+1
. Then the maximum positive integer M
z
such that (ad x(
))
M
z
z ,= 0 is independent of . Moreover M
z
is the maximum integer
such that (ad x
))
M
z
z ,= 0. In fact if a is any principal TDS containing w, then 2M
z
+1
is the dimension of the maximal dimensional irreducible component of the ad a submodule
generated by z.
Proof. Immediate from observation that any two such principal TDS are conjugate
under the action of sAd exp h. QED
Recalling the notation and statements of Theorems 1.12 and 1.14, put M
i
= M
z
i
so
that
e
(h 1) = r
. Also by (1.42)
[x(
), e
(1)] = [x(
), [(t
)
1
x
), z
i
]]
= (t
)
1
[x(
), [x
), z
i
]]
= (t
)
1
j=1
j
, z
i
n
j
w
j
(1.44)
so that
z
i
=
1
([x(
), e
(1)] +
[e
, e
])
=
1
((t
)
1
j=1
j
, z
i
n
j
w
)
=
1
j=1
((t
)
1
j
, z
i
)n
j
w
j
).
(1.45)
Next one recalls that since
[x
(), e
] = t
[x(), e
], (1.46)
computing the component in g
h1
, one has
1
[e
, z
i
] = t
(1.47)
so that
, z
i
= t
. (1.48)
That is,
= (t
)
1
, z
i
so that
[
[
2
t
z
i
=
j=1
j
+, z
i
n
j
w
j
. (1.49)
1.6. We will identify End h with h h, where if x, y h, then x y End h is that
operator such that if z h, then
x y(z) = (x, z) y
Then if A End h is given by
A =
j=1
(w
j
+w
) n
j
w
j
14
, then for z h one has
Az =
=1
j
+, zn
j
w
j
.
Then (1.49) is the statement
Proposition 1.16. For i = 1, . . . , , one has that z
i
is an eigenvector of A with
eigenvalue
[
[
2
/t
. (1.50)
We now want to simplify the expression for A. Indeed
j=1
(w
j
+w
) n
j
w
j
=
j=1
w
j
n
j
w
j
+
j=1
w
n
j
w
j
=
j=1
(w
j
n
j
w
j
) +w
j=1
n
j
w
j
)
=
j=1
n
j
(w
j
w
j
) +w
.
(1.51)
Thus if we consider the extended Dynkin diagram adding another node
0
= and
dene m
0
= 1 as in the McKay correspondence, we have proved
Theorem 1.17. One has
A =
j=0
n
j
w
j
w
j
. (1.52)
1.7. Henceforth we x so that =
(1)
(see Theorem 1.11) so that t
= 1. Also
assume g is not of type A
1
so that is not simple. One then has (see (1.5))
x() = e
i=1
n
i
e
i
(1.53)
and (see (1.30) and (1.27))
x() = e
i=1
n
i
e
i
(1.54)
Recalling (1.21) and (1.24) one notes that then
x() = (x() +x())/2
= (e
+e
)/2 +
i=1
n
i
(e
i
+e
i
)/2,
(1.55)
15
and hence
(x(), x()) = h/2. (1.56)
But by (1.21) and (1.24) one has
x() = i/2((x() x())
= i/2((e
) +
i=1
n
i
(e
i
e
i
)),
(1.57)
and hence
(x(), x()) = h/2. (1.58)
But clearly (1.55) and (1.57) imply
(x(), x()) = 0. (1.59)
Let Y Vec h() be the two real-dimensional plane spanned by the orthogonal vectors
x() and x(), and let
Q : Vec h() Y
be the orthogonal projection. Thus if x Vec h(), then
Qx = 2/h ((x, x()) x() + (x, x()) x()). (1.60)
But this implies
(Qx, Qx) = 2/h ((x, x())
2
+ (x, x())
2
).
But now if z = (x, x()), then
(x, x()) = z
(x, x()) = z
by the top lines in (1.55) and (1.57). Hence we have proved
Proposition 1.18. For any x Vec h() one has
[Qx[
2
= 2/h [(x, x())[
2
. (1.61)
Now for any () (see 1.3) let w
[
2
=
2
h
[
[
2
(1.62)
16
and our main result on the radius of the two-dimensional orbit projections.
Theorem 1.20. Let C
be xed so that =
(1)
is given as in Theorem 1.12.
Let O
i
, i = 1, . . . , , be an orbit of the Coxeter element
i=0
n
i
w
i
w
i
(1.63)
and the corresponding eigenvalue is [Qw
[
2
where is any root in the orbit O
i
.
2. The special case of E
8
2.1. Assume now that g is of type E
8
. Then = 8 and the cardinality of the set
of roots is 240. The Coxeter number h is 30. The group is unique up to isomorphism.
In particular G
= G
ad
. Let (C
)
9
be as in Theorem 1. The Gosset polytope (see
e.g., [Go]) published in 1900 may be taken to be the boundary of the convex hull of the
vectors w
, (), in the 8-dimensional real space Vec h(). The Coxeter element
, he
found that F factors into a product of 2 irreducible (over Q) degree 4 polynomials F
1
and
F
2
, where
F
1
(x) = x
4
15x
3
+ 75x
2
135x + 45
F
2
(x) = x
4
15x
3
+ 60x
2
90x + 45.
(2.1)
Vogan then computed the integral part of the radii of the Gosset circles normalized so that
the maximal integral part is 1000. They are, in increasing size,
209
338
416
502
618
673
813
1000
(2.2)
The use of quotation marks in the following statements is a consequence of the statement
in the Remark of 0.3.
Theorem 2.1. The ratio of the normalized radii in (2.2) agrees with the ratio
of the conjectured 8 masses in [Za]. See (1.8) in [Za].
We later found out that the ratio of the smallest Gosset circles (the larger over the
smaller) should be the Golden number R =
1
2
(1 +