Professional Documents
Culture Documents
Linear Algebraic Groups: Fiona Murnaghan
Linear Algebraic Groups: Fiona Murnaghan
Linear Algebraic Groups: Fiona Murnaghan
Volume 4, 2005
Linear Algebraic Groups
Fiona Murnaghan
Abstract. We give a summary, without proofs, of basic properties of linear
algebraic groups, with particular emphasis on reductive algebraic groups.
1. Algebraic groups
Let K be an algebraically closed eld. An algebraic K-group G is an algebraic
variety over K, and a group, such that the maps : G G G, (x, y) = xy,
and : G G, (x) = x
1
, are morphisms of algebraic varieties. For convenience,
in these notes, we will x K and refer to an algebraic K-group as an algebraic
group. If the variety G is ane, that is, G is an algebraic set (a Zariski-closed set)
in K
n
for some natural number n, we say that G is a linear algebraic group. If G
and G
is a homomorphism of algebraic
groups if is a morphism of varieties and a group homomorphism. Similarly, is
an isomorphism of algebraic groups if is an isomorphism of varieties and a group
isomorphism.
A closed subgroup of an algebraic group is an algebraic group. If H is a closed
subgroup of a linear algebraic group G, then G/H can be made into a quasi-
projective variety (a variety which is a locally closed subset of some projective
space). If H is normal in G, then G/H (with the usual group structure) is a linear
algebraic group.
Let : G G
are
algebraic varieties dened over k, a morphism : X X
k
: k[X
: K[X
] K[X] dening is
k
id. Equivalently,
the coordinate functions of all have coecients in k. The set X(k) := X k
n
is
called the K-rational points of X.
2000 Mathematics Subject Classication. 20G15.
c 2005 Clay Mathematics Institute
379
380 FIONA MURNAGHAN
If k is a subeld of K, we say that a linear algebraic group G is dened over k
(or is a k-group) if the variety G is dened over k and the homomorphisms and
are dened over k. Let : G G
is a ho-
momorphism of linear algebraic groups. Let g G. Then (g)
s
= (g
s
) and
(g)
u
= (g
u
).
3. Lie algebras
Let G be a linear algebraic group. The tangent bundle T(G) of G is the set
Hom
Kalg
(K[G], K[t]/(t
2
)) of K-algebra homomorphisms from the ane algebra
K[G] of G to the algebra K[t]/(t
2
). If g G, the evaluation map f f(g) from
K[G] to K is a K-algebra isomorphism. This results in a bijection betweeen G and
Hom
Kalg
(K[G], K). Composing elements of T(G) with the map a+bt +(t
2
) a
from K[t]/(t
2
) to K results in a map from T(G) to G = Hom
Kalg
(K[G], K).
The tangent space T
1
(G) of G at the identity element 1 of G is the bre of T(G)
over 1. If X T
1
(G) and f K[G], then X(f) = f(1) + t d
X
(f) + (t
2
) for some
d
X
(f) K. This denes a map d
X
: K[G] K which satises:
d
X
(f
1
f
2
) = d
X
(f
1
)f
2
(1) +f
1
(1)d
X
(f
2
), f
1
, f
2
K[G].
Let
: K[G] K[G]
K
K[G] be the K-algebra homomorphism which
corresponds to the multiplication map : G G G. Set
X
= (1 d
X
)
.
The map
X
: K[G] K[G] is a K-linear map and a derivation:
X
(f
1
f
2
) =
X
(f
1
)f
2
+f
1
X
(f
2
), f
1
, f
2
K[G].
Furthermore,
X
is left-invariant:
g
X
=
X
g
for all g G, where (
g
f)(g
) =
f(g
1
g
: K[G
, and satises
d([X, Y ]) = [d(X), d(Y )], X, Y g.
That is, d is a homomorphism of Lie algebras. If is bijective, then is an isomor-
phism if and only if d is an isomorphism of Lie algebras. If K has characteristic
zero, any bijective homomorphism of linear algebraic groups is an isomorphism.
If H is a closed subgroup of a linear algebraic group G, then (via the dierential
of inclusion) the Lie algebra h of H is isomorphic to a Lie subalgebra of g. And H
is a normal subgroup of G if and only if h is an ideal in g ([X, Y ] h whenever
X g and Y h).
If g G, then Int
g
: G G, Int
g
= gg
0
g
1
, g
0
G, is an isomorphism of
algebraic groups, so Ad g := d(Int
g
) : g g is an isomorphism of Lie algebras.
Note that (Ad g)
1
= Ad g
1
, g G, and Ad (g
1
g
2
) = Ad g
1
Ad g
2
, g
1
, g
2
G.
The map Ad : G GL(g) is a homomorphism of algebraic groups, called the
adjoint representation of G.
If G is a k-group, then its Lie algebra g has a natural k-structure g(k), with
g K
k
g(k). Also, Ad is dened over k.
Jordan decomposition in the Lie algebra. We can dene semisimple and nilpo-
tent elements in g in manner analogous to denitions of semisimple and unipotent
elements in G (as g is isomorphic to a Lie subalgebra of gl
n
(K) for some n). If
X g, there exist unique elements X
s
and X
n
g such that X = X
s
+ X
n
,
[X
s
, X
n
] = 0, X
s
is semisimple, and X
n
is nilpotent. If : G G
is a homomor-
phism of algebraic groups, then d(X)
s
= d(X
s
) and d(X)
n
= d(X
n
) for all
X g.
4. Tori
A torus is a linear algebraic group which is isomorphic to the direct product
G
d
m
= G
m
G
m
(d times), where d is a positive integer. A linear algebraic
group G is a torus if and only if G is connected and abelian, and every element of
G is semisimple.
A character of a torus T is a homomorphism of algebraic groups from T to G
m
.
The product of two characters of T is a character of T, the inverse of a character
of T is a character of T, and characters of T commute with each other, so the set
X(T) of characters of T is an abelian group. A one-parameter subgroup of T is a
homomorphism of algebraic groups from G
m
to T. The set Y (T) of one-parameter
subgroups is an abelian group. If T G
m
, then X(T) = Y (T) is just the set of
maps x x
r
, as r varies over Z. In general, T G
d
m
for some positive integer d,
so X(T) X(G
m
)
d
Z
d
Y (T). We have a pairing
, ) : X(T) Y (T) Z
, ) r where (x) = x
r
, x G
m
.
LINEAR ALGEBRAIC GROUPS 383
Let k be a subeld of K. A torus T is a k-torus if T is dened over k. Let T
be a k-torus. Let X(T)
k
be the subgroup of X(T) made up of those characters of
T which are dened over k. We say that T is k-split (or splits over k) whenever
X(T)
k
spans k[T], or, equivalently, whenever T is k-isomorphic to G
m
G
m
(d times, d = dimT). In this case, T(k) k
. If X(T)
k
= 0, then we
say that T is k-anisotropic. There exists a nite Galois extension of k over which
T splits. There exist unique tori T
spl
and T
an
of T, both dened over k, such that
T = T
spl
T
an
, T
spl
is k-split and T
an
is k-anisotropic. Also, T
an
is the identity
component of
X(T)
k
ker .
Example 4.1. Let T be the subgroup of GL
n
(K) consisting of diagonal ma-
trices in GL
n
(K). Then T is a k-split k-torus for any subeld k of K.
Example 4.2. Let T be the closed subgroup of GL
2
(C) dened by
T =
__
a b
b a
_
[ a, b C, a
2
+b
2
,= 0
_
.
Then T is an R-torus and is R-anisotropic.
5. Reductive groups, root systems and root datathe absolute case
Let G be a linear algebraic group which contains at least one torus. Then the
set of tori in G has maximal elements, relative to inclusion. Such maximal elements
are called maximal tori of G. All of the maximal tori in G are conjugate. The rank
of G is dened to be the dimension of a maximal torus in G.
Now suppose that G is a linear algebraic group and T is a torus in G. Recall
that the adjoint representation Ad : G GL(g) is a homomorphism of algebraic
groups. Therefore Ad (T) consists of commuting semisimple elements, and so is
diagonalizable. Given X(T), let g
= X g [ Ad (t)X = (t)X, t T.
The nonzero X(T) such that g
and dimg
= 1 for all .
(2) If , let T
= (Ker )
0
. Then T
= Z
G
(T
). Then Z
of Z
satises z
= t g
. The
group G is generated by the subgroups Z
, .
(4) The centre Z(G) of G is equal to
.
(5) If , there exists a unique connected T-stable (relative to conjugation
by T) subgroup U
. Also, U
.
(6) Let n N
G
(T), and let w be the corresponding element of W = W(G, T).
Then nU
n
1
= U
w()
for all .
(7) Let . Then there exists an isomorphism
: G
a
U
such that
t
(x)t
1
=
((t)x), t T, x G
a
.
(8) The groups U
() .
(A reection relative to is a linear transformation sending to that
restricts to the identity map on a subspace of codimension one).
(R3): If , , then s
() is an integer multiple of .
A root system is reduced if it has the property that if , then are the
only multiples of which belong to .
The rank of is dened to be dimV . The abstract Weyl group W() is the
subgroup of GL(V ) generated by the set s
[ .
If T is a maximal torus in G, then = (G, T) is a root system in V =
)
Z
R, and it is reduced. The rank of is equal to the semisimple rank of G,
and the abstract Weyl group W() is isomorphic to W = W(G, T).
A base of is a subset =
1
, . . . ,
i=1
c
i
i
, where the
c
i
s are all integers, no two of which have dierent signs. The elements of are
called simple roots. The set of positive roots
+
is the set of such that the
coecients of the simple roots in the expression for , as a linear combination of
simple roots, are all nonnegative. Similarly,
.
Given , there exists a base containing . Given a base , the set s
[
generates W = W(). The subgroups Z
, , and T, generate G.
There is an inner product (, ) on V with respect to which each w W is an
orthogonal linear transformation. If , , then s
) = 2(, )/(, )
for all . The set
of elements
)
Z
R, called the dual of . The Weyl group W(
) is isomorphic to W(),
via the map s
.
A root system is said to be irreducible if cannot be expressed as the
union of two mutually orthogonal proper subsets. In general, can be partitioned
uniquely into a union of irreducible root systems in subspaces of V . The group G
is simple (or almost simple) if G contains no proper nontrivial closed connected
normal subgroup. When G is semisimple and connected, then G is simple if and
only if is irreducible.
The reduced irreducible root systems are those of type A
n
, n 1, B
n
, n 1,
C
n
, n 3, D
n
, n 4, E
6
, E
7
, E
8
, F
4
, and G
2
. For each n 1 there is one
irreducible nonreduced root system, BC
n
. (These root systems are described in
many of the references). If n 2, the root system of GL
n
(K) (relative to any
maximal torus) is of type A
n1
. The root system of Sp
2n
(K) is of type C
n
, if
n 3, and of type A
1
and B
2
for n = 1 and 2, respectively.
The quadruple (G, T) = (X, Y, ,
(G, T))
is a root datum. An abstract root datum is a quadruple = (X, Y, ,
), where
X and Y are free abelian groups such that there exists a bilinear mapping , ) :
X Y Z inducing isomorphisms X Hom(Y, Z) and Y Hom(X, Z), and
X and
of
onto
) = 2
(RD2): If s
: X X and s
: Y Y are dened by s
(x) = xx,
)
and s
(y) = y , y)
, then s
() and s
(for all
).
The axiom (RD2) may be replaced by the equivalent axiom:
(RD2): If , then s
() , and the s
, , generate a nite
group.
If ,= , then is a root system in V := )
Z
R, where ) is the subgroup
of X generated by . The set
= (Y, X,
=
(X
, Y
) is a group isomorphism f : X X
onto
. If G
is a linear
algebraic group which is isomorphic to G, and T
is a maximal torus in G
, then
the root data (G, T) and (G
, T
) are isomorphic.
If is a reduced root datum, there exists a connected reductive K-group G
and a maximal torus T in G such that = (G, T). The pair (G, T) is unique
up to isomorphism.
386 FIONA MURNAGHAN
6. Parabolic subgroups
Let G be a connected linear algebraic group. The set of connected closed
solvable subgroups of G, ordered by inclusion, contains maximal elements. Such a
maximal element is called a Borel subgroup of G. If B is a Borel subgroup, then
G/B is a projective variety and any other Borel subgroup is conjugate to B. If P is
a closed subgroup of G, then G/P is a projective variety if and only if P contains a
Borel subgroup. Such a subgroup is called a parabolic subgroup. If P is a parabolic
subgroup, then P is connected and the normalizer N
G
(P) of P in G is P. If P
and P
are
conjugate, then P = P
.
Now assume that G is a connected reductive linear algebraic group. Let T
be a maximal torus in G. Then T lies inside some Borel subgroup B of G. Let
U = R
u
(B) be the unipotent radical of B. There exists a unique base of
= (G, T) such that U is generated by the groups U
,
+
, and B = TU.
Conversely if is a base of , then the group generated by T and by the groups
U
,
+
, is a Borel subgroup of G. Hence the set of Borel subgroups of G which
contain T is in one to one correspondence with the set of bases of . The Weyl
group W permutes the set of Borel subgroups containing T simply transitively.
The set of Borel subgroups containing T generates G.
The Bruhat decomposition. Let B be a Borel subgroup of G, and let T be a
maximal torus of G contained in B. Then G is the disjoint union of the double
cosets BwB, as w ranges over a set of representatives in N
G
(T) of the Weyl group
W (BwB = Bw
B if and only if w = w
in W).
Let G, B and T be as above. Let be the base of (G, T) corresponding to
B. If I is a subset of , let W
I
be the subgroup of W generated by the subset
S
I
= s
[ I of I. Let P
I
= BW
I
B (note that P
= B). Then P
I
is a
parabolic subgroup of G (containing B). A subgroup of G containing B is equal
to P
I
for some subset I of . If I and J are subsets of then W
I
W
J
implies
I J and P
I
P
J
implies I J. Also, P
I
is conjugate to P
J
if and only if
I = J. A parabolic subgroup is called standard if it contains B. Any parabolic
subgroup P is conjugate to some standard parabolic subgroup.
Let I . The set
I
of such that is an integral linear combination of
elements of I forms a root system, with Weyl group W
I
. The set of roots (P
I
, T)
of P
I
relative to T is equal to
+
(
I
). Let N
I
= R
u
(P
I
). Then N
I
is a
T-stable subgroup of U = B
u
, and is generated by those U
such that
+
and /
I
. Let T
I
= (
I
Ker )
0
,
and let M
I
= Z
G
(T
I
). The set
I
coincides with the set of roots in which are
trivial on T
I
. The group M
I
is reductive and is generated by T and by the set of
U
,
I
, T
I
is the identity component of the centre of M
I
, and (M
I
, T) =
I
.
The Lie algebra of M
I
is equal to t
I
g
: G
a
U
are k-isomorphic.
The centralizer Z
G
(T) of a k-torus T in G is reductive and dened over k, and
if T is k-split, Z
G
(T) is the Levi factor of a parabolic k-subgroup of G. (Here, we
say a closed subgroup H of G is a k-subgroup of G if H is a k-group). Any k-torus
in G is contained in some maximal torus which is dened over k. If k is innite,
then G(k) is Zariski dense in G.
The maximal k-split tori of G are all conjugate under G(k). Let S be a maximal
k-split torus in G. The k-rank of G is the dimension of S. The semisimple k-rank
of G is the k-rank of G/R(G). The nite group
k
W = N
G
(S)/Z
G
(S) is called the
k-Weyl group. The set
k
= (G, S) of roots of G relative to S is called the k-roots
of G. The k-roots form an abstract root system, which is not necessarily reduced,
with Weyl group isomorphic to
k
W. The rank of
k
is equal to the semisimple
k-rank of G.
A Borel subgroup B of G might not be dened over k. We say that G is
k-quasisplit if G has a Borel subgroup that is dened over k. If P is a parabolic
k-subgroup of G, then R
u
(P) is dened over k. A Levi factor M of a parabolic
k-subgroup is called a Levi k-factor of P if M is a k-group. Any two Levi k-
factors of P are conjugate by a unique element of R
u
(P)(k). If two parabolic
k-subgroups of G are conjugate by an element of G then they are conjugate by an
element of G(k). The group G contains a proper parabolic k-subgroup if and only
if G contains a noncentral k-split torus, that is, if the semisimple k-rank of G is
positive. The results described in this section give no information in the case where
G has semisimple k-rank zero.
Let P
0
be a minimal element of the set of parabolic k-subgroups of G (such
an element exists, since the set is nonempty, as it contains G). Any minimal
parabolic k-subgroup of G is conjugate to P
0
by an element of G(k). The group
P
0
contains a maximal k-split torus S of G, and Z
G
(S) is a k-Levi factor of P
0
.
The semisimple k-rank of Z
G
(S) is zero. Because N
G
(S) = N
G
(S)(k) Z
G
(S),
G(k) contains representatives for all elements of
k
W. The group
k
W acts simply
transitively on the set of minimal parabolic k-subgroups containing Z
G
(S).
Let Lie(Z
G
(S)) be the Lie algebra of Z
G
(S). Then
g = Lie(Z
G
(S))
.
388 FIONA MURNAGHAN
If
k
and 2 /
k
, then g
is a subalgebra of g. If and 2
k
, then
g
+ g
2
is a subalgebra of g. For each
k
, set
g
()
=
_
g
, if 2 /
k
g
2
, if 2
k
.
There exists a unique closed connected unipotent k-subgroup U
()
of G which
is normalized by Z
G
(S) and has Lie algebra g
()
.
Let P
0
be as above. Then there exists a unique base
k
of
k
such that
R
u
(P
0
) is generated by the groups U
()
,
k
+
. The set of standard parabolic
k-subgroups of G corresponds bijectively with the set of subsets of
k
. Fix I
k
.
Let S
I
= (
I
Ker )
0
and let
k
I
be the set of
k
which are integral linear
combinations of the roots in I. Let
k
W
I
be the subgroup of
k
W generated by
the reections s
k
W
I
P
0
. The unipotent radical of P
I
is equal to N
I
, the subgroup of G
generated by the groups U
()
, as ranges over the elements of
k
+
which are not
in
k
I
. The k-subgroup M
I
:= Z
G
(S
I
) is a Levi k-factor of P
I
, (M
I
, S) =
k
I
,
and
k
W
I
=
k
W(M
I
, S).
A parabolic k-subgroup of G is conjugate to exactly one P
I
, and it is conjugate
to P
I
by an element of G(k).
Relative Bruhat decomposition. Let U
0
= R
u
(P
0
). Then G(k) = U
0
(k)
N
G
(S)(k) U
0
(k), and G(k) is the disjoint union of the sets P
0
(k)wP
0
(k), as w
ranges over a set of representatives for elements of
k
W in N
G
(S)(k).
A parabolic subgroup of G(k) is a subgroup of the form P(k), where P is a
parabolic k-subgroup of G. A subgroup of G(k) which contains P
0
(k) is equal to
P
I
(k) for some I
k
. If I
k
, choosing representatives for
k
W
I
in N
G
(S)(k),
we have P
I
(k) = P
0
(k)
k
W
I
P
0
(k). The group P
I
(k) is equal to its own normalizer
in G(k). The Levi decomposition P
I
= M
I
N
I
carries over to the k-rational
points: P
I
(k) = M
I
(k) N
I
(k). If I, J
k
and g G(k), then gP
J
(k)g
1
P
I
(k) if and only if J I and g P
I
(k).
8. Examples
Example 8.1. G = GL
n
(K), n 2.
The group T = diag (t
1
, t
2
, . . . , t
n
) [ t
i
K
i=1
k
i
i
n
P
i=1
kii
, where
n
P
i=1
kii
(diag (t
1
, , t
n
)) = t
k1
1
t
kn
n
,
is an isomorphism from Z
n
to X(T). If n
P
i=1
kii
(t) = diag (t
k1
, , t
kn
), t
K
i=1
k
i
i
n
P
i=1
kii
is an isomorphism from Z
n
to Y (T). Also,
kii
,
iei
) =
n
i=1
k
i
i
. The root system = (G, T) =
ij
[ 1 i ,= j
n.
For 1 i ,= j n, let E
ij
M
n
(K) = g be the matrix having a 1 in the ij
th
entry, and zeros elsewhere. If =
ij
, i ,= j, then g
is spanned by E
ij
, and
LINEAR ALGEBRAIC GROUPS 389
U
= I
n
+ tE
ij
[ t K. The reection s
permutes
i
and
j
, and xes all
k
with k / i, j. The co-root
is
ij
The Weyl group W is isomorphic to the
symmetric group S
n
. The root system
is of type A
n1
.
The set :=
ii+1
[ 1 i n 1 is a base of . The corresponding
Borel subgroup B is the subgroup of G consisting of upper triangular matrices.
If I , there exists a partition (n
1
, n
2
, , n
r
) of n (n
i
a positive integer,
1 i r, n
1
+n
2
+ +n
r
= n), such that
T
I
= diag (a
1
, . . . , a
1
. .
n1 times
, a
2
, . . . , a
2
. .
n2 times
, . . . , a
r
, . . . , a
r
. .
nr times
) [ a
1
, a
2
, . . . , a
r
K
The group M
I
:= Z
G
(T
I
) is isomorphic to GL
n1
(K)GL
n2
(K) GL
nr
(K),
N
I
consists of matrices of the form
_
_
I
n1
I
n2
.
.
.
0
.
.
.
I
nr
_
_
,
and P
I
= M
I
N
I
.
Example 8.2. G = Sp
4
(K) (the 4 4 symplectic group). Let
J =
_
_
0 0 0 1
0 0 1 0
0 1 0 0
1 0 0 0
_
_
Then G = g GL
4
(K) [
t
gJg = J and g = X M
4
(K) [
t
XJ +JX = 0.
The group T := diag (a, b, b
1
, a
1
) [ a, b K
is a maximal torus in G
and X(T) Z Z, via
(i,j)
(i, j), where
(i,j)
(diag (a, b, b
1
, a
1
)) = a
i
b
j
.
And Y (T) Z Z, via
(i,j)
(i, j), where
(i,j)
(t) = (diag (t
i
, t
j
, t
j
, t
i
).
Note that
(i,j)
,
(k,)
) = ki +j.
Let =
(1,1)
and =
(0,2)
. Then
= , , ( +), (2 + ,
:= , is a base of = (G, T), and
g
= Span
K
(E
12
E
34
), g
= Span
K
(E
21
E
43
) g
= Span
K
E
23
g
+
= Span
K
(E
13
+E
24
), g
2+
= Span
K
E
14
, etc.
Identifying and with (1, 1) and (0, 2) Z Z, respectively, we have
s
_
0 1 0 0
1 0 0 0
0 0 0 1
0 0 1 0
_
_
.
390 FIONA MURNAGHAN
We also have s
_
1 0 0 0
0 0 1 0
0 1 0 0
0 0 0 1
_
_
.
The Weyl group W = W() is equal to 1, s
, s
, s
, s
, s
, s
, (s
)
2
is described by
, ( +)
, (2 +)
= (1, 1) ( +)
= (1, 1)
= (0, 1) (2 +)
= (1, 0)
The root system is of type C
2
and
is of type B
2
, isomorphic to C
2
.
Remark 8.3. If n > 2 the root system of Sp
2n
(K) is of type C
n
, and the dual
is of type B
n
, and B
n
and C
n
are not isomorphic.
The Borel subgroup of G which corresponds to is the subgroup B of upper
triangular matrices in G. Apart from G and B, there are two standard parabolic
subgroups, P
and P
= {diag (a, a, a
1
a
1
) | a K
}
M = Z
G
(T) =
8
<
:
2
4
A 0
0
0 1
1 0
t
A
1
0 1
1 0
3
5
A GL
2
(K)
9
=
;
N =
I
2
B
0 I
2
B M
2
(K),
t
B = B
= (Ker )
= { diag (a, 1, 1, a
1
) | a K
}
M
= Z
G
(T
) =
8
>
>
<
>
>
:
2
6
6
4
d 0 0 0
0 c
11
c
12
0
0 c
21
c
22
0
0 0 0 d
1
3
7
7
5
d K
, c
11
c
22
c
12
c
21
= 1
9
>
>
=
>
>
;
SL
2
(K) K
=
8
>
>
<
>
>
:
2
6
6
4
1 x y z
0 1 0 y
0 0 1 x
0 0 0 1
3
7
7
5
x, y, z K
9
>
>
=
>
>
;
9. Comments on references
For the basic theory of linear algebraic groups, see [B1], [H] and [Sp2], as
well as the survey article [B2]. For information on reductive groups dened over
non algebraically closed elds, the main reference is [BoT1] and [BoT2]. Some
material appears in [B1], and there is a survey of rationality properties at the end
of [Sp2]. See also the survey article [Sp1]. For the classication of semisimple
algebraic groups, see [Sa] and [T2]. For information on reductive groups over
local nonarchimedean elds, see [BrT1], [BrT2], and the article [T1]. Adeles and
algebraic groups are discussed in [W].
LINEAR ALGEBRAIC GROUPS 391
References
[B1] A. Borel, Linear algebraic groups, Second Enlarged Edition, Graduate Texts in Mathe-
matics, 126, Springer-Verlag, New York 1991.
[B2] A. Borel, Linear algebraic groups, in Algebraic Groups and Discontinuous Subgroups,
Proc. Symp. Pure Math. 9, (1966), 319.
[BoT1] A. Borel and J. Tits, Groupes reductifs, Publ. Math. I.H.E.S. 27 (1965), 55-151.
[BoT2] A. Borel and J. Tits, Complements `a larticle: Groupes reductifs, Publ. Math. I.H.E.S. 41
(1972), 253276.
[B] N. Bourbaki, Groupes et alg`ebres de Lie, Hermann, Paris, 1968, Chapters IV, V, VI.
[BrT1] F. Bruhat and J. Tits, Groupes reductifs sur un corps local I, Donnees radicielles valuees,
Publ. Math. IHES 41, (1972), 5251.
[BrT2] F. Bruhat and J. Tits, Groupes reductifs sur un corps local II, Schemas en groupes.
Existence dune donnee radicielle valuee, Publ. Math. IHES 60 (1984), 197376.
[H] J.E. Humphreys, Linear algebraic groups, Graduate Texts in Mathematics, Springer-
Verlag, New York, 1975.
[Sa] Satake, Classication theory of semisimple algebraic groups, Lecture Notes in Pure and
Applied Mathematics, Marcel Dekker, New York 1971.
[Sp1] T.A. Springer, Reductive groups, in Automorphic Forms, Representations, and L-
functions, Proc. Symp. Pure Math. 33, (1979), part 1, 327.
[Sp2] T.A. Springer, Linear algebraic groups, Birkhauser, Boston, 1981.
[T1] J. Tits, Reductive groups over local elds, Proc. Symp. Pure Math. 33 (1979), part 1,
2969.
[T2] J. Tits, Classication of algebraic semisimple groups, Proc. Symp. Pure Math. 9 (1966),
33-62.
[W] A. Weil, Adeles and algebraic groups, Insitute for Advanced Study, Princeton, N.J., 1961.
Department of Mathematics, University of Toronto, Toronto, Canada M5S 3G3
E-mail address: fiona@math.toronto.edu