# C HAPTER

I

**Basic Theory of Algebraic Groups
**

The emphasis in this chapter is on afﬁne algebraic groups over a base ﬁeld, but, when it requires no extra effort, we often study more general objects: afﬁne groups (not of ﬁnite type); base rings rather than ﬁelds; afﬁne algebraic monoids rather than groups; afﬁne algebraic supergroups (very brieﬂy); quantum groups (even more brieﬂy). The base ﬁeld (or ring) is alway denoted k, and R is always a commutative k-algebra.

I

Basic Theory of Algebraic Groups 1 Introductory overview . . . . . . . . . . . . . . . . . . . 1a The building blocks . . . . . . . . . . . . . . . . 1b Extensions . . . . . . . . . . . . . . . . . . . . 1c Summary . . . . . . . . . . . . . . . . . . . . . 1d Exercises . . . . . . . . . . . . . . . . . . . . . 2 Deﬁnitions; examples . . . . . . . . . . . . . . . . . . . 2a Deﬁnition . . . . . . . . . . . . . . . . . . . . . 2b Examples . . . . . . . . . . . . . . . . . . . . . 2c Afﬁne monoids . . . . . . . . . . . . . . . . . . 2d Afﬁne supergroups . . . . . . . . . . . . . . . . 2e Relaxing the hypothesis on k . . . . . . . . . . . 2f Terminology . . . . . . . . . . . . . . . . . . . 3 Some basic constructions . . . . . . . . . . . . . . . . . 3a Products of afﬁne groups . . . . . . . . . . . . . 3b Fibred products of afﬁne groups . . . . . . . . . 3c Extension of the base ring (extension of scalars) . 3d Restriction of the base ring (restriction of scalars) 3e The Greenberg functor . . . . . . . . . . . . . . 4 Afﬁne groups and Hopf algebras . . . . . . . . . . . . . 4a Algebras . . . . . . . . . . . . . . . . . . . . . 4b Co-algebras . . . . . . . . . . . . . . . . . . . . 4c The duality of algebras and co-algebras . . . . . 4d Bi-algebras . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . .

1 5 5 7 8 9 10 10 14 18 19 20 22 23 23 23 24 24 27 28 28 28 29 30

This is Chapter 1 of Algebraic Groups, Lie Groups, and their Arithmetic Subgroups, available at www.jmilne. org/math/. Version 2.21, April 27, 2010. Copyright c 2005, 2006, 2009, 2010 J.S. Milne.

1

5

6

7

4e Afﬁne monoids and bi-algebras . . . . . . . . . . . . . . . . . . . . . 4f Afﬁne groups and Hopf algebras . . . . . . . . . . . . . . . . . . . . . 4g Abstract restatement . . . . . . . . . . . . . . . . . . . . . . . . . . . 4h Explicit description of , , and S . . . . . . . . . . . . . . . . . . . . 4i Commutative afﬁne groups . . . . . . . . . . . . . . . . . . . . . . . . 4j Finite ﬂat algebraic groups; Cartier duality . . . . . . . . . . . . . . . 4k Galois descent of afﬁne groups . . . . . . . . . . . . . . . . . . . . . . 4l Quantum groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4m Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4n Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Algebraic groups and afﬁne algebraic schemes . . . . . . . . . . . . . . . . . . 5a Afﬁne k-algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5b The max spectrum of a ring . . . . . . . . . . . . . . . . . . . . . . . 5c Afﬁne algebraic schemes . . . . . . . . . . . . . . . . . . . . . . . . . 5d Properties of afﬁne algebraic schemes . . . . . . . . . . . . . . . . . . 5e Algebraic groups as groups in the category of afﬁne algebraic schemes 5f Homogeneity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5g Reduced algebraic groups . . . . . . . . . . . . . . . . . . . . . . . . 5h Smooth algebraic schemes . . . . . . . . . . . . . . . . . . . . . . . . 5i Smooth algebraic groups . . . . . . . . . . . . . . . . . . . . . . . . . 5j Algebraic groups in characteristic zero are smooth (Cartier’s theorem) . 5k Transporters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5l Relaxing the conditions on k and G . . . . . . . . . . . . . . . . . . . 5m Appendix: The faithful ﬂatness of Hopf algebras . . . . . . . . . . . . 5n Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Group theory: subgroups and quotient groups. . . . . . . . . . . . . . . . . . . 6a A criterion to be an isomorphism . . . . . . . . . . . . . . . . . . . . . 6b Subgroups; injective homomorphisms . . . . . . . . . . . . . . . . . . 6c Kernels of homomorphisms . . . . . . . . . . . . . . . . . . . . . . . 6d Dense subgroups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6e Normalizers; centralizers; centres . . . . . . . . . . . . . . . . . . . . 6f Quotient groups; surjective homomorphisms . . . . . . . . . . . . . . 6g Existence of quotients . . . . . . . . . . . . . . . . . . . . . . . . . . 6h Semidirect products . . . . . . . . . . . . . . . . . . . . . . . . . . . 6i Algebraic groups as sheaves . . . . . . . . . . . . . . . . . . . . . . . 6j Limits of afﬁne groups . . . . . . . . . . . . . . . . . . . . . . . . . . 6k Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Representations of afﬁne groups . . . . . . . . . . . . . . . . . . . . . . . . . 7a Deﬁnition of a representation . . . . . . . . . . . . . . . . . . . . . . . 7b Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7c Comodules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7d The category of comodules over C . . . . . . . . . . . . . . . . . . . . 7e Representations and comodules . . . . . . . . . . . . . . . . . . . . . 7f The category of representations of G . . . . . . . . . . . . . . . . . . 7g Afﬁne groups are inverse limits of algebraic groups . . . . . . . . . . . 7h Algebraic groups admit ﬁnite-dimensional faithful representations . . . 7i The regular representation contains all . . . . . . . . . . . . . . . . . . 7j Every faithful representation contains all . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31 32 33 34 35 36 36 37 37 38 40 40 40 41 43 43 45 46 47 47 48 50 52 52 55 56 56 56 58 60 62 65 68 68 69 72 72 74 74 76 76 79 79 82 83 84 85 87

2

8

9

10

11

12

13

14

7k Stabilizers of subspaces . . . . . . . . . . . . . . . . . . . . . 7l Sub-coalgebras and subcategories . . . . . . . . . . . . . . . . 7m Quotient groups and subcategories . . . . . . . . . . . . . . . . 7n Normal subgroups and subcategories . . . . . . . . . . . . . . 7o Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Group theory: the isomorphism theorems . . . . . . . . . . . . . . . . 8a Review of abstract group theory . . . . . . . . . . . . . . . . . 8b The existence of quotients . . . . . . . . . . . . . . . . . . . . 8c The homomorphism theorem . . . . . . . . . . . . . . . . . . . 8d The isomorphism theorem . . . . . . . . . . . . . . . . . . . . 8e The correspondence theorem . . . . . . . . . . . . . . . . . . . 8f The Schreier reﬁnement theorem . . . . . . . . . . . . . . . . . 8g The category of commutative algebraic groups . . . . . . . . . 8h Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recovering a group from its representations; Jordan decompositions . . 9a Recovering a group from its representations . . . . . . . . . . . 9b Application to Jordan decompositions . . . . . . . . . . . . . . Characterizations of categories of representations . . . . . . . . . . . . 10a Characterization of categories of comodules . . . . . . . . . . . 10b Characterization of categories of representations of afﬁne groups Finite algebraic groups . . . . . . . . . . . . . . . . . . . . . . . . . . 11a Etale groups . . . . . . . . . . . . . . . . . . . . . . . . . . . 11b Finite group schemes in characteristic p ¤ 0 . . . . . . . . . . 11c Cartier duality revisited . . . . . . . . . . . . . . . . . . . . . . 11d Finite ﬂat group schemes over commutative rings . . . . . . . . 11e Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The connected components of an algebraic group . . . . . . . . . . . . 12a Some algebraic geometry . . . . . . . . . . . . . . . . . . . . . 12b Separable subalgebras . . . . . . . . . . . . . . . . . . . . . . 12c The group of connected components of an algebraic group . . . 12d Connected algebraic groups . . . . . . . . . . . . . . . . . . . 12e Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12f Afﬁne groups . . . . . . . . . . . . . . . . . . . . . . . . . . . 12g Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12h Where we are . . . . . . . . . . . . . . . . . . . . . . . . . . . Groups of multiplicative type; tori . . . . . . . . . . . . . . . . . . . . 13a Group-like elements . . . . . . . . . . . . . . . . . . . . . . . 13b The characters of an algebraic group . . . . . . . . . . . . . . . 13c The algebraic group D.M / . . . . . . . . . . . . . . . . . . . . 13d Characterizing the groups D.M / . . . . . . . . . . . . . . . . 13e Diagonalizable groups . . . . . . . . . . . . . . . . . . . . . . 13f Diagonalizable groups are diagonalizable . . . . . . . . . . . . 13g Split tori and their representations . . . . . . . . . . . . . . . . 13h Groups of multiplicative type . . . . . . . . . . . . . . . . . . 13i Rigidity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13j Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solvable algebraic groups . . . . . . . . . . . . . . . . . . . . . . . . . 14a Commutative groups are triangulizable . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

88 90 91 91 94 95 95 95 96 98 99 99 101 101 102 102 105 110 110 113 115 115 118 119 120 120 121 121 122 123 125 126 127 127 128 129 129 129 130 131 132 133 134 135 136 137 138 138 3

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . The exceptional almost-simple groups . . . 15g The Spin group . . . . . . . . . . . . 18d Gradations on tensor categories . . . . . 17d Structure of a general algebraic group . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
139 140 142 143 144 145 148 149 149 150 150 151 154 154 156 156 160 161 163 163 164 164 165 165 166 167 168 169 169 171 172 174 177 179
Bibliography Index of deﬁnitions
4
. . . . . . . . . .
. . . .
. . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 14j Exercises . . . 17b Semisimple and reductive groups . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14e Independence of characters . . . . . . . . . . . . . . . . . . . . . 14i Tori in solvable groups . . . . . . 15i Action of O. . . . . . . . . . . . . . . Tannakian categories . . . . . . . . . . . . . . . . . . . . . 15b Theorems of Witt and Cartan-Dieudonn´ .
. . . . . . . . . . . . . . . . . . . . . . 14h Structure of solvable groups . . . . . . . . . General algebraic groups . . . . . . . . . . . . . . . . . . . . . . . . . . 14g Unipotent groups . . .
. . . . 15h The Clifford group . . . . . . 14d Solvable algebraic groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18a Properties of G versus those of Repk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .q/ on Spin. . .q/ . . . . . . . . 15d Super algebras .
. . . . . . . .
. . . 15f The Clifford algebra . . . .15
16 17
18
14b Decomposition of a commutative algebraic group 14c The derived group of algebraic group . . . . . . . . . . . . . 15j Restatement in terms of algebraic groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . The classical almost-simple groups . . . . . . . . . . . . . . . . . . . . . . . 15e Brief review of the tensor algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .G/ . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17a The radical of an algebraic group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18c Neutral tannakian categories . . . . . . 14f The Lie-Kolchin theorem . . . 16a The group G2 . . . . . . . . . . . . . . . . . . . . . . .
. . . e 15c The orthogonal group . . . . 18b Tensor categories . . . . . . 17c Reductive groups . . . . . . . . . . . . . . . . . . . . . 15a Quadratic spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .

1

Introductory overview

Loosely speaking, an algebraic group over a ﬁeld k is a group deﬁned by polynomials. Before giving the precise deﬁnition in the next section, we look at some examples of algebraic groups. Consider the group SLn .k/ of n n matrices with entries in a ﬁeld k and with determinant 1. The determinant of a matrix .aij / is a polynomial in the entries aij of the matrix, namely, X det.aij / D sign. / a1 .1/ an .n/ (Sn D symmetric group),

2Sn

and so SLn .k/ is the subset of Mn .k/ D k n deﬁned by the polynomial condition det.aij / D 1. The entries of the product of two matrices are polynomials in the entries of the two matrices, namely, .aij /.bij / D .cij / with cij D ai1 b1j C C ai n bnj ;

2

and Cramer’s rule realizes the entries of the inverse of a matrix with determinant 1 as polynomials in the entries of the matrix, and so SLn .k/ is an algebraic group (called the special linear group). The group GLn .k/ of n n matrices with nonzero determinant is also an algebraic group (called the general linear group) because its elements can be identiﬁed with the n2 C1-tuples ..aij /1Äi;j Än ; d / such that det.aij / d D 1. More generally, for a ﬁnite-dimensional vector space V , we deﬁne GL.V / (resp. SL.V /) to be the group of automorphisms of V (resp. automorphisms with determinant 1). These are again algebraic groups. To simplify the statements, for the remainder of this section, we assume that the base ﬁeld k has characteristic zero.

1a

The building blocks

We now list the ﬁve types of groups from which all others can be constructed by successive extensions: the ﬁnite groups, the abelian varieties, the semisimple groups, the tori, and the unipotent groups.

F INITE GROUPS

Every ﬁnite group can be realized as an algebraic group, and even as an algebraic subgroup of GLn .k/. Let be a permutation of f1; : : : ; ng and let I. / be the matrix obtained from the identity matrix by using to permute the rows. For any n n matrix A, the matrix I. /A is obtained from A by using to permute the rows. In particular, if and 0 are two permutations, then I. /I. 0 / D I. 0 /. Thus, the matrices I. / realize Sn as a subgroup of GLn . Since every ﬁnite group is a subgroup of some Sn , this shows that every ﬁnite group can be realized as a subgroup of GLn , which is automatically deﬁned by polynomial conditions. Therefore the theory of algebraic groups includes the theory of ﬁnite groups. The algebraic groups deﬁned in this way by ﬁnite groups are called constant ﬁnite algebraic groups. An algebraic group is connected if it has no constant ﬁnite group as a quotient, even over the algebraic closure of k.

A BELIAN VARIETIES

Abelian varieties are connected algebraic groups that are projective when considered as algebraic varieties. An abelian variety of dimension 1 is an elliptic curve, which can be described by a homogeneous equation Y 2 Z D X 3 C bXZ 2 C cZ 3 : 5

6

BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)

Therefore, the theory of algebraic groups includes the theory of abelian varieties. We shall ignore this aspect of the theory. In fact, we shall study only algebraic groups that are afﬁne when considered as algebraic varieties.

**S IMPLE GROUPS ; SEMISIMPLE GROUPS
**

A connected algebraic group G is simple if it is not commutative and has no normal algebraic subgroups, and it is almost-simple1 if its centre Z is ﬁnite and G=Z is simple. For example, SLn is almost-simple for n > 1 because its centre ( ZD

0

! ˇ ˇ ˇ :: ˇ : ˇ

0

)

n

D1

is ﬁnite and the quotient PSLn D SLn =Z is simple. An isogeny of algebraic groups is a surjective homomorphism G ! H with ﬁnite kernel. Two algebraic groups H1 and H2 are isogenous if there exist isogenies H1 G ! H2 :

This is an equivalence relations. When k is algebraically closed, every almost-simple algebraic group is isogenous to exactly one algebraic group on the following list: An .n Bn .n 1/; the special linear group SLnC1 I 2/; the special orthogonal group SO2nC1 consisting of all 2n C 1 2n C 1 matrices A such that At A D I and det.A/ D 1; Cn .n 3/; the symplectic group Sp2n consisting of all invertible 2n 2n matrices A such that Â Ã 0 I t J A D J where J D A ; I 0 Dn .n 4/; the special orthogonal group SO2n ; E6 ; E7 ; E8 ; F4 ; G2 the ﬁve exceptional groups. We say that an algebraic group G is an almost-direct product of its algebraic subgroups G1 ; : : : ; Gr if the map .g1 ; : : : ; gr / 7! g1 gr W G1 Gr ! G

is an isogeny. In particular, this means that each Gi is a normal subgroup of G and that the Gi commute with each other. For example, G D SL2 SL2 =N; N D f.I; I /; . I; I /g (1)

is the almost-direct product of SL2 and SL2 , but it is not a direct product of two almost-simple algebraic groups. A connected algebraic group is semisimple if it is an almost-direct product of almost-simple subgroups. For example, the group G in (1) is semisimple. Semisimple algebraic groups will be our main interest.

1 Other

authors say “quasi-simple” or “simple”.

1. Introductory overview

7

**G ROUPS OF MULTIPLICATIVE TYPE ; ALGEBRAIC TORI
**

An algebraic subgroup T of GL.V / is said to be of multiplicative type if, over k al , there exists a basis of V relative to which T is contained in the group Dn of all diagonal matrices 0 1 0 0 0 B0 0 0C B C B : : :: : :C B: : : : : C: : :C B: : @0 0 0A 0 0 0 In particular, the elements of an algebraic torus are semisimple endomorphisms of V . A connected algebraic group of multiplicative type is a torus.

U NIPOTENT GROUPS

An algebraic subgroup G of GL.V / is unipotent if there exists a basis of V relative to which G is contained in the group Un of all n n matrices of the form 0 1 1 C B0 1 B C B: : :: : :C (2) B: : : : : C: : :C B: : @0 0 A 1 0 0 0 1 In particular, the elements of a unipotent group are unipotent endomorphisms of V .

1b

Extensions

We now look at some algebraic groups that are nontrivial extensions of groups of the above types.

S OLVABLE GROUPS

An algebraic group G is solvable if it there exists a sequence of algebraic subgroups G D G0 Gi Gn D 1

such that each Gi C1 is normal in Gi and Gi =Gi C1 is commutative. For example, the group Un is solvable, and the group Tn of upper triangular n n matrices is solvable because it contains Un as a normal subgroup with quotient isomorphic to Dn . When k is algebraically closed, a connected subgroup G of GL.V / is solvable if and only if there exists a basis of V relative to which G is contained in Tn (Lie-Kolchin theorem 14.21).

R EDUCTIVE GROUPS

A connected algebraic group is reductive if it has no nontrivial connected normal unipotent subgroups. According to the table below, they are extensions of semisimple groups by tori. For example, GLn is reductive. It is an extension of the simple group PGLn by the torus Gm , 1 ! Gm ! GLn ! PGLn ! 1: Here Gm D GL1 , and the map sends it onto the subgroup of nonzero scalar matrices.

) independently gave different proofs.
2 Rosenlicht (1956) notes that this was proved by Chevalley in 1953. I. let O. In the following tables.Dn / sends
Â
Sn
to the automorphism in (3).
Barsotti (1955) and Rosenlicht (ibid. Let M be the group of monomial matrices.n/ is not connected: it contains Â Ã SO.
def
det
The monomial matrices. i.n/ D Ker O. (d) A connected afﬁne group G contains a largest connected normal unipotent subgroup N . A modern proof can be found in Conrad 2002. : : : . those with exactly one nonzero element in each row and each column. /
1
D diag. and a unipotent group. Rosenlicht).8
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
N ONCONNECTED GROUPS
We give some examples of naturally occurring nonconnected algebraic groups. : : : . this shows that Dn is normal in M . the quotient G=N semisimple..e. The matrix diag. when G is solvable. the group at left has a composition series whose quotients are the groups at right. This group contains both the algebraic subgroup Dn and the algebraic subgroup Sn of permutation matrices.
(3)
As M D Dn Sn .At / det.n/ denote the group of n n matrices A such that At A D I . Clearly D \ Sn D 1.2 (c) A connected afﬁne algebraic group G contains a largest connected normal solvable subgroup N (see 17a).a1 . Barsotti.35). More precisely: (a) An algebraic group G contains a unique normal connected subgroup G ı such that G=G ı is ﬁnite (see 12.1/ . an /.A/ D 1.
. a semisimple group.A/2 D det.n/ ! f˙1g as a normal algebraic subgroup of index 2 with quotient the constant ﬁnite group f˙1g.a
. and so det. an / I. the quotient G=N is an abelian variety (Chevalley. Every algebraic group has a composition series whose quotients are respectively a ﬁnite group.n/ /. G=N is of multiplicative type (see 17.9). 1. : : : .
The orthogonal group. (b) A connected algebraic group G contains a largest normal afﬁne algebraic subgroup N .
1c
Summary
Recall that we are assuming that the base ﬁeld k has characteristic zero. For an integer n 1.a1 .A/ 2 f˙1g. a . : : :/ lies in O. a torus. Moreover. who only published his proof in Chevalley 1960. and so O.2. / diag.n/ and has determinant 1. 1. and so M is the semi-direct product M D Dn where ÂW Sn ! Aut. Then det. an abelian variety. for any diagonal matrix diag. 14.

y/ for all x. / is not diagonalizable because it is not commutative). Y . Show that every element of GLn .x. / be the algebraic subgroup of SL.
. Hence the subset f. E XERCISE 1-3 Let be a positive deﬁnite bilinear form on a real vector space V . / is semisimple (but SO. but it is not diagonalizable because it is not commutative).x. Show that there exists a basis for V relative to which T Dn . Y / 2 RŒX. E XERCISE 1-4 Let k be a ﬁeld of characteristic zero.V / of ˛ such that .x. E XERCISE 1-2 Let T be a commutative subgroup of GL.1. ˛y/ D . and let SO.˛x. Show that if f .k/ of ﬁnite order is semisimple. (Hence the group of permutation matrices in GLn .k/ consists of semisimple elements. e x / D 0 for all x 2 R.X. Introductory overview
9
General algebraic group
general
Afﬁne algebraic group
afﬁne
Reductive algebraic groups
j
connected
´ ﬁnite etale abelian variety semisimple torus unipotent
j j
connected
´ ﬁnite etale reductive semisimple torus torus f1g unipotent
connected afﬁne
j j
solvable
j j
semisimple torus
solvable
j j
unipotent
unipotent
j j
f1g
f1g
1d
Exercises
E XERCISE 1-1 Let f .V / consisting of diagonalizable endomorphisms. Y ). then f is zero (as an element of RŒX. y 2 V . Show that every element of SO. e x / j x 2 Rg of R2 is not the zero-set of a family of polynomials.

: : : . According to the Hilbert basis theorem (CA 3. R/: Since the k-algebras that can be expressed in the form kŒX1 .R/ D f1g otherwise. Xn =a. We adopt this as our deﬁnition of an afﬁne algebraic group together with the requirement that the functor be “deﬁned by polynomials”. : : : .a1 .R/ is for any commutative ring R. : : : . Xn =a are exactly the ﬁnitely generated k-algebras.R0 /. For any k-algebra R. we know that a homomorphism R ! R0 of rings deﬁnes a homomorphism of groups SLn . an / 2 Rn j f . in 2a–2d we require k to be a ﬁeld. an / D 0 for all f 2 S g: A homomorphism of k-algebras R ! R0 deﬁnes a map S. We have just observed that an afﬁne algebraic group G is isomorphic to the functor deﬁned by an ideal a of polynomials in some polynomial ring kŒX1 . This suggest deﬁning an afﬁne algebraic group to be a functor from the category of k-algebras to groups that is is isomorphic (as a functor to sets) to the functor deﬁned by a set of polynomials in a ﬁnite number of symbols.Xij / 1 where X det. fi 2 S: Clearly S and a have the same zero-sets for any k-algebra R.a1 . Moreover. X12 . For any k-algebra R. For simplicity.6).n/ 2 kŒX11 . / X1 . the functor with ( Z=2Z if R D k G. examples
What is an algebraic group? For example.R0 /. Let A D kŒX1 . Xn can be generated by a ﬁnite set of polynomials. gi 2 kŒX1 . For example. The ideal a generated by S consists of the ﬁnite sums X gi fi . and the ntuples .A. for example. we conclude that the functors Algk ! Set deﬁned by a set of polynomials 10
. and these maps make R S. Xn . the zero-set of S in Rn is S. Therefore the functor R a.R/ D f. a homomorphism A ! R is determined by the images ai of the Xi . Xnn : (4)
2Sn
The condition that G can be deﬁned by polynomials is very strong: it excludes. So what is SLn without the “.2
Deﬁnitions. it is the group of n n matrices with entries in R and determinant 1. an / that arise from a homomorphism are exactly those in the zero-set of a. Xn . and so an afﬁne algebraic group is isomorphic (as a functor to sets) to the functor deﬁned by a ﬁnite set of polynomials.R/ satisﬁes this condition because it is isomorphic to the functor deﬁned by the polynomial det. : : : .R/ ! SLn . namely.R/”? Obviously. what is SLn ? We know what SLn .R/ into a functor from the category of k-algebras to the category of sets.R/ sending a k-algebra R to the zero-set of a in Rn is canonically isomorphic to the functor R Homk-alg . : : : . : : : .R/ ! S. : : : . : : : .
2a
Deﬁnition
M OTIVATING DISCUSSION
We ﬁrst explain how a set of polynomials deﬁnes a functor. Xn . it is a functor from the category of rings to groups. the functor R SLn .Xij / D sign. : : : . Let S be a subset of kŒX1 . : : : . : : : . Xn .1/ Xn . every ideal in kŒX1 . Let S be a subset of kŒX1 .a1 .

it is a natural transformation hA0 ! hA ): Let F W A ! Set be a functor from A to the category of sets.A/.f /.A/.
.f /.f /
(5)
hA .R/ D Hom. For any morphism f W A ! R.g/ D f ı g. hA . g 2 hA . the map T 7! aT is surjective. the commutative diagram hA . aT / represents F — in fact. R/ deﬁned by a ﬁnitely generated k-algebra A.A/ shows that
F .A.aT / D TR .2 A functor F W A ! Set is said to be representable if it is isomorphic to hA for some object A.f / D F .1 (YONEDA LEMMA ) For any functor F W A ! Set and object A of A. is said to represent F if Ta W hA ! F is an isomorphism.3 Let F1 and F2 be functors A ! Set. a 2 F .A. but the Yoneda lemma shows that Hom. and so the map T 7! aT is injective.Ta /R .a/ D a.hA .R/ ! hA . if F is representable. and let A be an object of A. 2.R/
aT
F .A/ such that . The proof of the naturality of (5) is left as an (easy) exercise for the reader.R/ D Hom.aT /:
(6)
Therefore T is determined by aT .a/.
T HE YONEDA LEMMA
An object A of a category A deﬁnes a functor hA W A ! Set by hA . examples
11
in a ﬁnite number of symbols are exactly the functors R Homk-alg . In general. hB / ' Hom.
2. Before continuing.e.Ta /A .f / D F .R/
TR
idA
f
F . it is convenient to review some category theory. the map T 7! aT is a bijection Hom.A. f W R ! R0 .R/ which is natural in R (i.A. F / ' F .f /
F . T / represents F .idA / D F .idA /.A/. say F hA . F2 / is a set if F1 is representable (because it is isomorphic to a set).A.R/ D Hom.f /. T D TaT — and so we sometimes say that . Let T be a natural transformation hA ! F . when F D hB .f /
TR . On the other hand.A.A/ which is natural in both A and F .A/
TA hA . then the choice of an isomorphism T W hA ! F determines an element aT 2 F .B.hA . A pair . R/. In particular.. an element a of F . 2. f 2 hA . R 2 ob.F1 . a/.A/ deﬁnes a natural transformation Ta W hA ! F by the rule . R/:
A morphism ˛W A ! A0 of objects deﬁnes a map f 7! f ı ˛W hA0 . Hom. A/: P ROOF.idA / of F . R/:
2
As .2.f /.A. Note that. Deﬁnitions. A natural transformation T W hA ! F deﬁnes an element aT D TA . the natural transformations F1 ! F2 will form a proper class (not a set).

a/ representing .R/ ? ? yH. fn g of representatives for the images of Y1 .A. the afﬁne groups over k form a category with the afﬁne algebraic groups as a full subcategory.R/ G. H / is a set.A. The map g 7! g 1 W G. Later (7. ! H. G.5 The coordinate ring of an afﬁne group is uniquely determined up to a unique isomorphism. Yn =b.21) we shall see that the afﬁne groups are exactly the functors that can be expressed as inverse limits of afﬁne algebraic groups. Therefore.4 An afﬁne group (resp. Yn .R/ is represented by O. A homomorphism G ! H of afﬁne groups over k is a natural transformation. the diagram G.R/ is natural in R. we make the following deﬁnition. an afﬁne group can be regarded as a functor from k-algebras to groups that is isomorphic (as a functor to sets) to the functor deﬁned by a (possibly inﬁnite) set of polynomials in a (possibly inﬁnite) set of symbols. We discuss the maps and S in detail in Section 4.R0 /
. 2.e. a family of homomorphisms ˛.G/ (CA (2). : : : . then Hom.R/ arises from a unique homomorphism W A ! A ˝k A. which can be described by polynomials. When we write A and B as quotients of polynomial rings. and so arises from a unique homomorphism S W O.
˛. 2.R/
/
! H. The Yoneda lemma shows that every natural transformation hA ! hB arises from a unique homomorphism of k-algebras B ! A. : : : .12
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
D EFINITIONS
After these preliminaries. afﬁne algebraic group) over k is a functor GW Algk ! Grp such that the composite
Algk ! Grp ! Set
G forget
is representable by a k-algebra (resp. say. namely.G/ or kŒG. i. It is often convenient to regard the coordinate ring .R0 / commutes.3 shows that if G and H are afﬁne groups over k. and so the multiplication map G. p. The functor R G.27).G.R/ ! G. D EFINITION 2.G/ ! O. by a set ff1 .forget/ıG is called the coordinate ring (or coordinate algebra) of G. and A is denoted O. Let G be an algebraic group.G/. a/ of an afﬁne group G as a k-algebra A together with an isomorphism aW hA ! G of functors to sets (see 2.R/ G.R/ of groups such that.. which again can be described by polynomials. Let A and B be ﬁnitely generated k-algebras.R/ ! G.R/ ! H. for every homomorphism of k-algebras W R ! R0 . A D kŒX1 . : : : .R0 /
/
˛.2). Below we shall see that there is a canonical choice for it. : : : . Less formally.R/W G.R/ ? ? yG. we see that a homomorphism B ! A can be described by polynomials. 2.7 The group structure of an afﬁne algebraic group is also described by polynomials. a ﬁnitely generated k-algebra). We explain brieﬂy how this follows from the Yoneda lemma.G/ ˝k O. Xm =a and B D kŒY1 .6 Statement 2. A pair .

R/.g/W A ! R of k-algebras. f2 7! f1 f2 D . the diagram G. cR . R a k-algebra. .R/ ! R.R/ into a group for all k-algebras R.R.R0 / commutes.
.
and let G be a functor from the category of k-algebras to groups. To give a functor GW Algk ! Grp such that G0 D .8 Let G0 be a functor Algk ! Set.R/ deﬁnes a homomorphism f 7! fR . deﬁne
0 . then they make hA into an afﬁne group over k (by 2. f2 / be the unique homomorphism A ˝k A ! R such that . to give an afﬁne group over k is the same as giving a representable functor G0 W Algk ! Set and a natural transformation G0 G0 ! G0 that makes G.g/: fR
! R ? ? y ! R0
fR 0
With these operations. . . we get a natural transformation ˛W G0 ! hA of functors to sets. Given homomorphisms f1 . examples
13
2.R/:
An element g 2 G. f2 / ı W hA .2.ff 0 /R . and even a k-algebra because each c 2 k deﬁnes a constant natural transformation cR W G.a/ and . Deﬁnitions.)
def
T HE CANONICAL COORDINATE RING
Let A1 be the functor sending a k-algebra R to its underlying set. We saw in (2.g/ D c for all g 2 G. which is natural in R.R/.a ˝ 1/ D f1 . there is an equivalence of categories .R/ into a group for all R. for every homomorphism of k-algebras R ! R0 .g/fR . .forget/ıG be the underlying functor to Set.G. A1 /: Thus an element f of A is a family of maps of sets fR W G.g/ D fR . f2 /. Thus.a/ for all a 2 A. and let W A ! A ˝k A be a homomorphism. 1/ R.R/ ? ? y G. A becomes a commutative ring.R/ into a group for all R. A1 W Algk ! Set.8).R/ ! hA . f2 W A ! R. Now deﬁnes a map f1 . Let G0 D .f1 .R/ ! R. f 0 2 A and g 2 G.f1 . In this way.7) that all afﬁne groups arise in this way. (More precisely. let . 2.R/ hA . GW Algk ! Grp. f2 /. A D Hom.
such that. . C.f1 .f ˙ f 0 /R .f1 . For f.g/ D fR .R/ into a group for all k-algebras R. If these maps make hA .9 Let A be a k-algebra.forget/ ı G is the same as giving a natural transformation G0 G0 ! G0 that makes G0 . to give an afﬁne group over k is essentially the same as giving a k-algebra A and a homomorphism W A ! A ˝k A such that makes hA .g/ 0 .g/ ˙ fR . and so.1 ˝ a/ D f2 . and let A be the set of natural transformations from G to A1 .

g/ D g.13 Let Gm be the functor R inverse. Y 7! X 1 deﬁnes an isomorphism kŒX. called the multiplicative group. A1 / is the coordinate ring of G.R.a. Then. Y ! kŒX. E XAMPLE 2.X Y
Therefore Gm is an afﬁne algebraic group. 2 Thus.
1
fR . and let . X 1 . g 2 G. and one checks that ˛W hB ! hA is the natural transformation deﬁned by this isomorphism.a/ D f . then G is an afﬁne group (by deﬁnition). A D Hom.a.kŒX.f /.B. A1 /. an element g 2 G.11 Let A be a coordinate ring for G. O. it is called the additive group. A1 / ' A1 .kŒX. Y =. The homomorphism kŒX.G/.G. On the other hand.G0 . g 2 G. and so Ga .
. X
1 .R/ ' Homk-alg . Conversely.X / be the ﬁeld of fractions of kŒX . If ˛ is an isomorphism.12 Let Ga be the functor sending a k-algebra R to itself considered as an additive group.B/ ' B. A1 / ' Hom. an element f 2 A is a family of maps fR W G. let G be an afﬁne group. for an afﬁne group G. X
1 . and so R (multiplicative group).R/ D . R/:
Gm . X 1 be the subring of polynomials in X and X 1 . Ga . f 2 O.G0 . therefore ˛ is an isomorphism. Let k. aW hB ! G0 / be a coordinate ring for G. Thus A ' B.H /.
and so
1
Gm .Gm / D kŒX.e. N OTATION 2. because A1 D hkŒX . Each a 2 R has a unique 1/.X Y 1/ ' kŒX.Ga / D kŒX . a homomorphism ˛W G ! H deﬁnes a homomorphism of rings ˛ W O. when we regard A as a k-algebra representing G.g/ D fR .R/: (7)
def
def
a
Yoneda
According to the Yoneda lemma. X
1
and a 2 Gm .14
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROPOSITION 2.G/ D Hom.R/ ' Homk-alg . C/.hB . Y =.˛ f /R .R/ is a homomorphism of k-algebras gW A ! R. This proves the ﬁrst part of the statement (and the parenthetical statement is obvious). X 7! X. R/: Therefore Ga is an afﬁne algebraic group with O. f 2 A.R/ ! R (of sets) indexed by the k-algebras R.˛R g/. When we regard A as Hom. i. b/ 2 R2 j ab D 1g ' Homk-alg . To give an element of R is the same as giving a k-algebra homomorphism kŒX ! R. a
/:
. R/:
for f 2 kŒX.H / ! O.. .R/ ' f. The two points of views are related by the equation fR . and let kŒX.kŒX .10 The functor G is an afﬁne group if and only if ˛ is an isomorphism (in which case it is an algebraic group if and only if A is ﬁnitely generated). Explicitly.R/: (8)
2b
Examples
E XAMPLE 2. X Thus O. P ROOF.R/ D R .

kŒT =. Then R Homk-alg . examples E XAMPLE 2. Such an afﬁne algebraic group is called a constant ﬁnite algebraic group.14 Let G be the functor such that G.R/ D fr 2 R j r p D 0g is an additive group. 1/. Therefore.V / of V : Homk-alg . Moreover.
suggested by those in DG II. R/. 1. E XAMPLE 2.V / is an afﬁne group over k (and even an afﬁne algebraic group when V is ﬁnite dimensional).T p /.T p /.
.V /W R Homk-lin .a C b/p D ap C b p . Deﬁnitions.O.kŒX =.V /. Then G. for any Q ﬁnite group G.
3 Notations
R ˝k V (additive group) is an afﬁne algebraic group with coordinate ring Sym. Moreover. the functor of k-algebras3 D a .R/
D fr 2 R j r n D 1g
is a multiplicative group.15 For an integer n 1.R/
n .R/ ' Homk-alg . R/. R/ ' R ˝k V .
' Homk-alg . and R
n . for any k-algebra R over a ﬁeld k of characteristic p ¤ 0. ˛p .2.k. R/.V _ /.Sym.
D kŒX =.R ˝k V / ' R ˝k Endk-lin . More generally.G/.V.X n
n/
and so
n
is an afﬁne algebraic group with O. ˛p . R/. 1/.G/k . and so R Moreover.G/k over k such that .
(see CA 8). E XAMPLE 2. and so ˛p is an afﬁne algebraic group with O. Now assume that V is ﬁnite dimensional. R/ ' Homk-lin . called the trivial algebraic group.X n
E XAMPLE 2. the binomial theorem takes the form . Then Homk-lin .V _ ˝k V /.R/ D G for any k-algebra R with no nontrivial idempotents (see 4.16 In characteristic p ¤ 0. 2. R/ (additive group) (9)
is represented by the symmetric algebra Sym.R/
is a functor.V _ . let O.
15
and so G is an afﬁne algebraic group.1.
n . Therefore D a .13 below).R/ ' Homk-alg .R/ D f1g for all k-algebras R. R/ is an afﬁne algebraic group .˛p / D kŒT =.R/ is a functor.G/ D g2G kg (product of copies of k indexed by the elements of G).17 For any k-vector space V . EndR-lin .V. and R ˛p . R a k-algebra.V / ' R ˝k .

: : : . en for V deﬁnes isomorphisms EndR-lin . : : : .R ˝k V / ' Mn . For f 2 O. . : : : .N / and adj. It is called the ij /Y 1/ general linear group.16 and so the functor R
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
EndR-lin .Xij /1Äi.R/ ' Homk-alg . 12.X12 .Y .Xij / is the polynomial (4).:::. ann /. there is a functor SLn sending a k-algebra R to the group of n n matrices with entries in R and with determinant 1. For f 2 O. Xnn det. Xnn .R . X12 .R/. These formulas can be proved by the same argument as for R a ﬁeld.A.det.3).M / D .Xij / 1/ where det.M / M D det.V ˝k V _ / ' kŒX11 .R/.R/ and Sym. B 2 Mn . ann /: E XAMPLE 2.det.M / where I denotes the identity matrix and Á adj. GLn .a/ D f .X .19 Similar arguments show that the n n matrices with entries in a k-algebra R and with determinant a unit in R form a group GLn .det. : : : .GLn /. : : : .Xij /.V ˝k V _ /.SLn / D kŒX11 .X / 1/
ij
(10)
(Cramer’s rule)
(11)
fR .a11 . X12 . Therefore.GLn / and a D .aij / D f . O. : : : .M / det. X12 .MN / D det.R/ with Mij the matrix obtained from M by deleting the i th row and the j th column.Xnn .
other words. Moreover.aij /
1
/:
Alternatively.R/ j A is invertibleg (because a right inverse of a square matrix is unique if it exists. R/ D f. B/ 7! A projects this bijectively onto fA 2 Mn .R/ ' Homk-alg . Xnn for the multiplicative subset generated by det. det.GLn / is the ring of fractions of kŒX11 . det.det.j Än ). Xnn SLn .aij / 2 SLn . Moreover. : : : . : : : . B/ j A. AB D I g:
The map . or by applying the principle of permanence of identities (Artin 1991.SLn / and a D . Therefore A ' O. 1/i Cj det Mj i 2 Mn .a/ D f .GLn / D kŒX11 . It is called the special linear group. Â Ã kŒX11 . ann . Ynn modulo the ideal generated by the n2 entries of the matrix . let A be the k-algebra in 2n2 symbols.R/ is a functor. X12 . and so SLn is an afﬁne algebraic group with O.2.Yij / I .18 For n n matrices M and N with entries in a k-algebra R.a11 . Example 6. The choice of a basis e1 . For f 2 kŒX11 . : : : .
fR .A.Xnn . X12 .aij / 2 Mn . .:::.R/.aij / 2 GLn .
4 In
.R/.Xij /. : : : . Then Homk-alg . Y .R/.M / I D M adj. Â Ã kŒX11 . Xnn (polynomial algebra in the n2 symbols . O. Y11 .Xij / : See CA.X12 and so GLn is an afﬁne algebraic group with coordinate ring4 kŒX.a11 .R ˝k V / (additive group)
is an algebraic group Va with coordinate ring Sym.R . X11 . Xnn . and that R GLn .A. X12 . E XAMPLE 2. Xnn and a D . a12 . and is also a left inverse). fR .Xij /Y 1/
11 . : : : . : : : . X12 .

20 Let C be an invertible n n matrix with entries in k.R/ j . b/ D .R/ is a group.aij / with entries in K. ˛w/ D . Sp.
A choice of a basis for V deﬁnes an isomorphism of each of these functors with one of those in (2. with the obvious N notation. which shows that they are afﬁne algebraic groups. and when 0 C D I I .21 There are abstract versions of the last groups. l D 1. it is the symplectic group Spn .v. Xnn .R/ D f˛ 2 GLV .
j. There are only two possibilities: N N (a) K is a separable ﬁeld extension of k of degree 2 and a 7! a is the nontrivial element of the N Galois group. G. n. 0 E XAMPLE 2.R/ D f˛ 2 GLV . n:
Therefore G is an afﬁne algebraic group.A/ is a unit. Then there is an algebraic group G over k such that G.A/ D 1. i. : : : . deﬁne a ˝ r D a ˝ r for a ˝ r 2 K ˝k R.a.22 Let K be a separable k-algebra of degree 2.k Xj i cj k Xkl ci l . When C D I .R/ j . If C D . /. let G. Y by the ideal generated by the polynomials det. Let V be a ﬁnite-dimensional vector space over k.aij / and A to be the transpose N of A.Xij /Y 1 P j.R/ D f. R/ with A equal to the quotient of kŒX11 . i.R/ D fT 2 GLn . In particular. B/ 2 Mn . : : : . it is the orthogonal group On . and let G.18).k
17
and so G. det. and.R/ consists of the invertible matrices . deﬁne A to be .R/ D fR-linear automorphisms of R ˝k V g.R/ j T t C T D C g.A/ det.R/ j AB D I g
.b.k/ D fA 2 Mn . There is a unique k-automorphism a 7! a of K such that a D a if and only if a 2 k.A. l D 1. Deﬁnitions. then G.˛v. or (2. for a k-algebra R.19). w/ for all v.K ˝k R/ j A A D I g: Note that A A D I implies det. E XAMPLE 2.cij /. w 2 R ˝k V g. a/: N For an n n matrix A D . X12 . ˛w/ D .2. let be a nondegenerate symmetric bilinear form V V ! k. or (b) K D k k and .R/ D fR-linear automorphisms of R ˝k V with determinant 1g. (2. GLV . : : : .v. In case (b).A. O. and let be a nondegenerate alternating form V V ! k.R/ ' Homk-alg . and so G.tij / such that X tj i cj k tkl D ci l .20). w/ for all v. w 2 R ˝k V g.K/ j A A D I g: More precisely. Then there are afﬁne algebraic groups with SLV . /.˛v.R/ D fA 2 Mn . examples E XAMPLE 2.

Yij .n/ Y
1/.18
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
and so . it is a “group without inverses”. B/ 7! A is an isomorphism of G with GLn . Then e satisﬁes a quadratic polynomial with coefﬁcients in k.G/. we can “complete the square” and choose e so that e 2 2 k and e D e.1/ .y/ for all x.
. : : :. when R has no nontrivial idempotents. R/:
.R/ if and only if . Xn .xy/ D '. y 2 M .
E XAMPLE 2.eM / D eM 0 .GLn /=. When M 0 is a group.24 The determinant deﬁnes a homomorphism of algebraic groups detW GLn ! Gm : E XAMPLE 2. An afﬁne monoid (resp. G is represented by a quotient of kŒ: : : . and (b) '. Â Ã 1 a a 7! .k/ ¤ 2. the set M of elements in M with inverses is a group (the largest subgroup of M ).Xij j j ¤ . or e). A matrix N with entries in K ˝k R can be written in the form A C eB with A. called the group of monomial matrices. B 2 Mn .23 There exists an afﬁne algebraic group G. For each 2 Sn (symmetric group).25 The homomorphisms R ! SL2 . a ﬁnitely generated k-algebra). G.O. Then A ' kŒX1 and so G. Xij . E XAMPLE 2. : : : . Y =.
2c
Afﬁne monoids
Recall that a monoid is a set M together with an associative binary operation M M ! M and an identity element (usually denoted 0. A homomorphism of monoids is a map 'W M ! M 0 such that (a) '. let e 2 K k.R/ ' G Homk-alg .n/ . (a) holds automatically because a group has only one element such that ee D e. For any monoid M . In case (a). Assuming char.1/
Xn
.sign..A C eB/ D I i. B A D 0:
t
and
Evidently.A . In other words. 0 1
deﬁne a homomorphism of algebraic groups Ga ! SL2 . R/ ' Homk-alg . 1. : : : ˝k kŒ: : : .R/ having exactly one nonzero element in each row and column. It lies in G. such that. afﬁne algebraic monoid) over k is a functor from the category of k-algebras to monoids that is representable (as a functor to sets) by a k-algebra (resp.x/'.R/. let A D O.G/ D 2Sn A .i // Q and let O.
/ X1
.At eB t /. if and only if At A AB
t
e 2 BB t D I.R/ is the group of invertible matrices in Mn .R/.e.A.

P ROOF. : : : . Xdimk B . P ROPOSITION 2. we saw in (2.a0 . Xnn . X12 .b/7!a
. : : : . it is odd.a/ D 0.EndR-lin . R is an associative k-algebra equipped with a decomposition R D R0 ˚ R1 (as a k-vector space) such that k R0 and Ri Rj Ri Cj (i. : : : .a. if B D Endk-lin . We explain brieﬂy how it leads to the notion of an afﬁne “supergroup”. and has parity p.R ˝k B. the functor R M.28 An associative k-algebra B with identity (not necessarily commutative) deﬁnes a functor R 7! .R/ can be obtained from M by forming ﬁbre products.R ˝k V /. b/. .26 For a k-vector space V . 2 E XAMPLE 2. We let GB m denote the corresponding afﬁne algebraic group R 7! . Deﬁnitions. then GB D GLV . as a functor to sets. b/ 2 M M ' f.27 For any afﬁne monoid M over k.a. and have parity p. it is the study of Z=2Z-graded versions of some of the usual objects of mathematics.2. / (multiplicative monoid of n n matrices).b/7!ab
! M
! M:
It follows that.V /.a.b/7!b
.. If B is a ﬁnite k-algebra. the functor R M over k.R/. On choosing a basis e1 . Roughly speaking. which is represented by kŒX1 .R ˝ B/ : For example. b 0 // 2 M1 M.Mn . and so it is an algebraic monoid. if it lies in R0 . then M1 j a D b 0 g:
This shows that M can be constructed from M by using only ﬁbred products: M1 ? ? y M M ! f1g ? ? y M ? ? y M1 ! M1 ? ? y. j 2 Z=2Z). An element a of R is said to be even. let EndV be the functor R . For an abstract monoid M .
which is represented by the polynomial ring kŒX11 . The homogeneous
. we obtain an isomorphism of EndV with the functor R . m
2d
Afﬁne supergroups
The subject of supersymmetry was introduced by the physicists in the 1970s as part of their search for a uniﬁed theory of physics consistent with quantum theory and general relativity.V ˝k V _ /.R/ is an afﬁne group
M j ab D 1g. A superalgebra over a ﬁeld k is a Z=2Z-graded associative algebra R over k. which shows that it is representable (see 3b below). en for V .a/ D 1. for an afﬁne monoid M . let M1 D f. ı/:
19
When V is ﬁnite dimensional. examples E XAMPLE 2.17) that. / from the category of k-algebras to monoids. if it lies in R1 . then (as a functor to sets) it becomes R 7! Rdimk B . EndV is represented by Sym. In other words. so also is M . then this is an afﬁne algebraic monoid: when we choose a basis for B as a k-vector space.a. when M is algebraic.a.

Xn by a ﬁnitely generated ideal. 1/p. Y1 . Yn in the even symbols Xi and the odd symbols Yi is deﬁned to be the quotient of the k-algebra of noncommuting polynomials in X1 .m . ab C ba D 0. Xm .1). an afﬁne algebraic group G is described by a ﬁnite set of polynomials in a ﬁnite number of variables with coefﬁcients in k.Xij /mC1Äi:j ÄmCn D 1: Much of the theory of afﬁne groups extends to afﬁne supergroups (see. Recall that a k-algebra is ﬁnitely presented if it is isomorphic to the quotient of a polynomial ring kŒX1 . Thus even elements commute with all elements.det.b/ ab for all a. for m.R0 /.Xij /1Äi. let GLmjn be the functor ˚ AB ˇ « ˇ R B 2 Mm. 1 Ä i. A super k-algebra R is said to be commutative if ba D . this is the polynomial ring in the commuting symbols X1 . by a ﬁnitely presented k-algebra). : : : .1 Ä i. . It is an afﬁne supergroup because it is represented by the commutative super k-algebra obtained from the commutative super k-algebra kŒX11 . n 2 N. 1 Ä j. For example. The only other changes are that “vector space over k” must be replaced by “module over k” and “ﬁnite-dimensional vector space over k” by “ﬁnitely generated projective module over k”. : : : . but for odd elements a. Xij . : : : . j 0 Ä n:
When n D 0.j Äm D 1. : : : . A homomorphism of super k-algebras is a homomorphism of k-algebras preserving the parity of homogeneous elements. b. C 2 Mn.n . Y. as a functor to sets. j /)
2e
Relaxing the hypothesis on k
Everything in this section holds with only minor changes when k is not a ﬁeld. A functor from the category of commutative super k-algebras to groups is an afﬁne supergroup if it is representable (as a functor to sets) by a commutative super k-algebra. The commutative super k-algebra kŒX1 . XmCn.i. : : : .remaining pairs . Fioresi and Gavarini 2008).6. D 2 GLn . j Ä m C n/
and the odd symbols Xij by setting Y . Y j Yj 0 D Yj 0 Yj . m C 1 Ä i. i 0 Ä m.30) and (2. j Ä m.31) below.a/p.R1 /. Thus. Yn by the relations Xi Xi 0 D Xi 0 Xi . The explanation for the second change is provided by the statements (2. : : : . for example. Z det. and when m D 0. X12 . Let k be a commutative ring. Z in the even symbols Y.R1 /. 3.20
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
elements of R are those that are either even or odd.
. Yn g provided 2 ¤ 0 in k. X i Yj D Yj X i .mCn . and so GLmjn is a functor to groups. Z.R0 / : C D A 2 GLm . A functor from commutative k-algebras to groups is an afﬁne group (resp. : : : . b 2 R. it is the exterior algebra of the vector space with basis fY1 . an afﬁne algebraic group) if it is representable (as a functor to sets) by a k-algebra (resp. Xm .
A B It is known that such a matrix C D is invertible (Varadarajan 2004.

30 Let A be a commutative ring.26. examples
21
In more detail. the canonical maps HomA-lin . For any k-algebra R.T r /_ .
5 In
.17. P ROPOSITION 2.3.
F INITELY GENERATED PROJECTIVE MODULES OVER COMMUTATIVE RINGS
In many situations.S ˝A M _ . no changes to the text through Example 2. and the proof that functors are afﬁne algebraic groups is a little more complicated. for all i 2 I . the general case follows.Ri /
i
for all k-algebras R D R1
Rn .R/ D Gi . T / M ˝T
_ _
(12) (13) (14)
! .2. the Afi -module Mfi is free of ﬁnite rank.16 are needed. the functor R R ˝k V is representable only if V is ﬁnitely generated and projective (cf.29 Let k D k1 kn . 2
fact. See CA 10.10. (c) there exists a ﬁnite family .Gi /1Äi Än .M ˝ T /
__
_
M !M are isomorphisms. If G $ . when M D Ar and (13) becomes the isomorphism . To give an afﬁne group G over k is the same as giving an afﬁne group Gi over each ki . p. : : : . then Y G.5 and the ﬁnal two sentences make sense only for free k-modules. T with M ﬁnitely generated and projective.S. R D R1 Rn C en . T ˝A M / ! HomA-lin . 1. and write 1 D e1 C set of orthogonal idempotents in k. In Example 2. Then fe1 . When M is free of ﬁnite rank the statements are easy to prove. In the second paragraph of Example 2.21 one should take V to be a ﬁnitely generated projective k-module. Since the ﬁnitely generated projective modules are exactly the direct summands of free modules of ﬁnite rank. Deﬁnitions. S . m one should require B to be ﬁnitely generated and projective as a k-module (DG II.fi /i 2I of elements of A generating the ideal A and such that. (b) M is ﬁnitely presented and Mm is a free Am -module for every maximal ideal m.28). EGA I (1971) 9. where this is stated without proof).T _ /r ! . P ROOF. and the ﬁnal sentence makes sense only when V is free.
2
P ROPOSITION 2. en g is a complete
where Ri is the k-algebra Rei . one should take V to be a ﬁnitely generated projective k-module.149).4. For example.4. (d) M is ﬁnitely presented and ﬂat.
P ROOF. one should take V to be a ﬁnitely generated projective k-module. E XAMPLE 2. the correct generalization of “ﬁnite-dimensional vector space” is not “ﬁnitely generated module” but “ﬁnitely generated projective module”. The following conditions on an A-module M are equivalent: (a) M is ﬁnitely generated and projective. 2.31 For any A-modules M . In Example 2. In order for GB to be an afﬁne algebraic group in (2.

22
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
2f
Terminology
From now on “algebraic group” will mean “afﬁne algebraic group” and “algebraic monoid” will mean “afﬁne algebraic monoid”.
.

.A1 .B. R/ 23
Homk-alg . R/ Homk-alg . We deﬁne the ﬁbred product functor G1 H G2 to be the functor sending a k-algebra R to the set of pairs . Moreover. p.
3a
Products of afﬁne groups
Let G and H be afﬁne groups over k. then i 2I Gi is algebraic. for any k-algebras A. The k-algebra A1 ˝B A2 is the largest quotient of A1 ˝k A2 such that the diagram B ? ? y A1 ! A2 ? ? ya7!1˝a
a7!a˝1
! A1 ˝B A2
commutes. More precisely. Homk-alg . If each Gi is an algebraic group and I is ﬁnite. let .A.Gi / (in the inﬁnite case.R/ G2 . More generally.R/ H.R/. and let A1 and A2 be B-algebras.R/
H over k with coordinate ring O. R/ (16)
(CA (2).H /. B. g2 / 2 G1 .R/:
Obviously G1 H G2 is the ﬁbred product of (17) in the category of functors from Algk to Set. R/ ' Homk-alg .R/
H.R/: i2I N Then G is an afﬁne group with coordinate ring i 2I O. R. and H be functors from the category of k-algebras to sets.G/ ˝k O.R/ such that g1 and g2 have the same image in H. G together with the projection maps is the product of the Gi in the Q category of afﬁne groups. and let G be the functor Y R Gi .
3b
Fibred products of afﬁne groups
Let G1 .R/ Homk-alg . R/:
(18)
.A ˝k B. k is a commutative ring. for any k-algebra R. R/ ' Homk-alg . 8).g1 . i.B.A1 ˝B A2 ..27). Prop. III 5. apply Bourbaki A. Homk-alg .G H / D O.G1
H
G2 / . (15)
because. Let B be a k-algebra.e.Gi /i 2I be a (possibly inﬁnite) family of afﬁne groups over k. and let G1 ! H G2 (17)
be natural transformations. G2 .3
Some basic constructions
Throughout this section.R/ D G1 .A2 . The functor R is an afﬁne group G G.R/ G2 .

Homk 0 -alg .F 0 /k 0 =k .G ! H // D O.G1 / ˝O.H / k: (20)
3c
Extension of the base ring (extension of scalars)
Let G be an afﬁne group over k. A natural transformation 'W Fk 0 ! F 0 deﬁnes homomorphisms F . a functor F 0 W Algk 0 ! Set deﬁnes a functor . .24
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
Therefore. A2 . E XAMPLE 3. Gk 0 is the unitary group deﬁned by the k 0 -algebra K ˝k k 0 . R/ consisting of the homomorphisms that restrict to the given homomorphism k 0 ! R on k 0 .G/.R/ is deﬁned.F 0 /k 0 =k .Ker. As above. We shall show that there is a right adjoint to the functor G Gk 0 . R/.
3d
Restriction of the base ring (restriction of scalars)
Throughout this subsection.r7!1˝r/
!
F .k 0 ˝k R/
'.R/.R//:
/ is called the kernel of ˛.G2 /.G The afﬁne group . let H be an afﬁne group and let ! H be the unique homomorphism from the trivial group to H .k 0 ˝k A.G/.
and so.Vk 0 /a . E XAMPLE 3.F 0 /k 0 =k W Algk ! Set with .R/
F . and so. For any ﬁeld extension k ! k 0 . G1 H G2 is a functor to Grp.R/ D F . we obtain an afﬁne group Gk 0 over k 0 with coordinate ring O. If G is an algebraic group. The afﬁne group Gk 0 is said to have been obtained from G by extension of the base ring or by extension of scalars. Homk 0 -alg . We ﬁrst explain this for functors to sets.R/ D Ker. and is denoted Ker. G2 . and the above remark shows that it is an afﬁne group with coordinate ring O.R// . R/ ' Homk-alg . When the natural transformations G1 ! H G2 are homomorphisms of afﬁne groups.R/ ! H. In particular.R/ D F 0 . then G1 H G2 is represented by the k-algebra A1 ˝B A2 . if the functors G1 . R/ ' Homk-alg . and so G. and B.˛.G1
H
G2 / D O. Note that O.1 Let V be a ﬁnitely generated projective k-module. For example.H / O. R/ .R/W G. For any homomorphism ˛W G ! H.G
H H
/ . and let Vk 0 D k 0 ˝k V .G/.G/ ˝O.Va /k 0 ' . and H in (17) are represented by k-algebras A1 . so also is Gk 0 .A. on restricting the functor G to k 0 -algebras. for example.k 0 ˝k R/.' G. so also is G1 H G2 .k 0 ˝k R/ D . Homomorphisms of afﬁne groups over k give homomorphisms of afﬁne groups over k 0 .k 0 ˝k A. then .R/
def
.2 Let G be the unitary group deﬁned by a separable k-algebra K of degree 2 (see 2. On the other hand. Gk al ' GLn . k 0 is a k-algebra that is ﬁnitely generated and projective as a k-module.
(19)
If G1 and G2 are algebraic groups. and let k 0 be a k-algebra. a functor F W Algk ! Set deﬁnes a functor Fk 0 W Algk 0 ! Set with Fk 0 . A k 0 -algebra R can be regarded as a kalgebra through k ! k 0 ! R. and so G Gk 0 is a functor.k 0 ˝k A. For any k-algebra A. R/ is the subset of Homk-alg .22).O.Gk 0 / D k 0 ˝k O.k 0 ˝k R/
!
F 0 .˛/. Therefore (18) shows that Homk 0 -alg .k 0 ˝k O.

: : : .R0 / D Fk 0 . : : : .3 If F W Algk 0 ! Set is represented by a (ﬁnitely-generated) k-algebra.bij / 1Äi Än W R0n ! Rnd
1Äj Äd
def
˚ ked .F 0 /k 0 =k /: (21)
Because it is a right adjoint. Ynd / with the property that f . Hence we have F . Xn .F. Now suppose that F is the subfunctor of An deﬁned by a polynomial f . : : : D kŒY11 .bij / deﬁned by the equations ai D Pd
j D1 bij ej .F 0 /k 0 =k .X1 . .R/ D Rn for all k 0 -algebras R.
i D 1. : : : . b12 . The polynomial g has coefﬁcients in k 0 .
C gd ed . say. In particular. Thus. and so it is a bijection.ai /1Äi Än 7! . bnd / D 0 when the a’s and b’s are related by (22). but k 0 ŒY11 . so that F . : : : . : : : . : : :e1 ˚ and so we can write it as a sum g D g1 e1 C
6 Given
˚ kŒY11 . The given k-algebra map k 0 ! R and the identity map R0 ! R deﬁne a map k 0 ˝k R0 ! R (of k 0 -algebras).3. it takes (ﬁbred) products to (ﬁbred) products. and let R0 be R regarded as a k-algebra.
which sends . and F . Ynd :
F ! . and so form a natural transformation F ! .F 0 /k 0 =k : Hom. then so also is .a1 .F /k 0 =k And (the isomorphism depends only on the choice of the basis e1 . Some basic constructions
25
which are natural in the k-algebra R.F 0 /k 0 =k /: This has an obvious inverse6 .k 0 ˝k R0 / ! F 0 . We prove this only in the case that k 0 is free as a k-module.F. : : : . On substituting P Xi D dD1 Yij ej j into f .Fk 0 .R/:
.F /k 0 =k . F 0 / ' Hom. ed ). .F 0 /k 0 =k preserves inverse limits. : : : . R0 D k 0 ˝k R ' Re1 ˚ and so there is a bijection . : : :ed . Y12 . L EMMA 3. Y12 .R/.Y11 . an / D 0 ” g. k 0 D ke1 ˚ R.
ei 2 k 0 :
Consider ﬁrst the case that F D An .
gi 2 kŒY11 . we need Fk 0 ! F 0 . we obtain a polynomial g. F 0 . we have a morphism Hom. Xn / 2 k 0 ŒX1 . and shows that . We have shown that the extension of scalars functor F Fk 0 has a right adjoint F 0 .ai / to the family .F 0 /k 0 =k . For any k-algebra ˚ Red . : : : .
(22)
The bijection is natural in R.b11 .Fk 0 . F 0 / ! Hom. Let R be a k 0 -algebra. P ROOF. This can also be checked directly.R0 / ! F 0 . n.

k 0 ˝k R/ ' G. : : : . b12 . k 0 ). This argument extends in an obvious way to the case that F is the subfunctor of An deﬁned by a ﬁnite set of polynomials. then . and even to the case that it is a subfunctor of an inﬁnite dimensional afﬁne space deﬁned by inﬁnitely many polynomials.
i
We sketch a proof of this. Indeed.R/ D G.26 Clearly.F /k 0 =k is isomorphic to the subfunctor of And deﬁned by the polynomials g1 . for a K-algebra R.G.G 0 /k 0 =k is right adjoint to the functor “extension of scalars”: Hom. 3.
. fr 2 k be such that . (Add details.f1 . This is obvious in terms of sheaves. : : : . k 00 ) is ﬁnitely generated and projective over k (resp.k 00 ˝k R/ D ˘k 00 =k G . and let F be a functor Algk ! Set.GK /I
(23)
in other words.5 For any homomorphisms k ! k 0 ! k 00 of rings such that k 0 (resp. ˘k 0 =k G
K
' ˘k 0 ˝k K=K . Weil restriction commutes with base extension. 3. bnd / D 0 for i D 1.GK /. bnd / D 0 ” gi .R/ because k 00 ˝k 0 k 0 ˝k R ' k 00 ˝k R. ˘k 0 =k ı ˘k 00 =k 0 ' ˘k 00 =k . : : : . then F itself is representable. ˘k 0 =k G
K
. for any afﬁne group G over k 00 and k-algebra R.R/ D ˘k 0 =k .R/ D G.
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
g.b11 . G 0 /: The afﬁne group . and so . 2
A SIDE 3.G/ .4 In order to extend the proof to the general case it sufﬁces to show: Let f1 .6 For any k-algebra K and any afﬁne group G over k 0 .G/k 0 =k . and (21) shows that the functor G 0 . Then kfi fj ˝kf Ai ' kfi fj ˝kf Aj for all i.k 0 ˝k K ˝K R/ D ˘k 0 ˝k K=K .b11 . : : : . j i j i because they both represent the restriction of F to Algkf f . and called the Weil restriction of G. . gd .˘k 00 =k 0 G/ .G/k 0 =k is a functor Algk ! Grp. : : : . Let Ai represent F jAlgkf .G 0 /k 0 =k / ' Hom. The lemma shows that if G is an afﬁne (algebraic) group.G/k 0 =k is said to have been obtained from G by (Weil) restriction of scalars (or by restriction of the base ring). The Ai patch together (on Spec k) to give a i j k-algebra A such that Afi ' Ai for all i.R/
def
def
because k 0 ˝k R ' k 0 ˝k K ˝K R. : : : .Gk 0 . b12 .)
If G is a functor Algk 0 ! Grp. then so also is . If the restriction of F to Algkf is representable for all i . It is sometimes denoted Resk 0 =k G or ˘k 0 =k G. ˘k 0 =k ı ˘k 00 =k 0 . d . and that A represents F . Indeed. fr / D k.k 00 ˝k 0 k 0 ˝k R/ ' G.

For an afﬁne group G over A. (To be extended. A could be the ring Wm . and let K be a ﬁeld containing all k-conjugates of k 0 .R//: Then G. Then
0 .10 (Add a discussion of G .kn ˝ R/ Á 0 .
(24)
Indeed.G1 /k1 =k
def
def
0 Gn .k 0 .R/ ! ˘k 0 =k G .R/
0 because k 0 ˝ R ' k1 ˝ R
0 kn ˝ R.G/k 0 =k .Gn /kn =k . then any homomorphism G ! Hk 0 (over k 0 ) factors uniquely through i .7 Let k 0 D k1 afﬁne groups over the ki0 (see 2. See Greenberg 1961.. For example. Let G be the afﬁne group over k 0 corresponding to a family .A ˝Wm .G/ is an afﬁne group over k. consider the functor G. i.˘k 0 =k G/k 0 of afﬁne groups over k 0 such that.29). K/j D Œk 0 W k.G/k 0 =k when k 0 is purely inseparable over k. A is a Wm .k/-module for some m.3. Then Y G˛ ˘k 0 =k G K ' 0
˛Wk !K
def
where G˛ is the afﬁne group over K obtained by extension of scalars ˛W k 0 ! K. Greenberg 1963.
3.)
.K/ . for all k 0 algebras R.k1 ˝ R/
0 D . In general. and has the following universal property: let H be an afﬁne group over k.k/ Wm . 3. for any k-algebra R.k .Gn /kn =k .R/ D G. Indeed ˘k 0 =k G because k 0 ˝ K ' K Homk .G/k 0 =k ' .8 There is a homomorphism iW G ! . such that jHomk . i.R/ D G.k 0 ˝ R/ deﬁned by a 7! 1 ˝ aW R ! k 0 ˝ R.9 Let k 0 be a ﬁnite separable ﬁeld extension of a ﬁeld k.k/ of Witt vectors of length m.Gi /i of 3. Then i is injective (obviously). Some basic constructions
27
0 0 kn .k 0 ˝ R/
0 D G1 .)
0
(23) K
' ˘k 0 ˝K=K GK '
(24)
Y
˛Wk 0 !K
G˛
3e
The Greenberg functor
Let A be a local artinian ring with residue ﬁeld k.G/: R G. .R/ is the map G.e. 3.G1 /k1 =k 0 .

. a co-algebra over k) is a module C over k together with a pair of k-linear maps WC ! C ˝C such that the diagrams C ˝C ˝C o O
˝C C˝
WC ! k
C ˝C O
k ˝ CfMo
˝C
C ˝C o
C co-associativity
MMM MMM MM ' MMM MM
C ˝C O
C co-identity
/ C ˝k oo7 oo ooo ooo oo ' ooo
C˝
(26)
commute. p39. such that .e.
4a
Algebras
Recall that an associative algebra over k with identity is a module A over k together with a pair of k-linear maps7 mW A ˝ A ! A eW k ! A such that the following diagrams commute:
A˝m
A˝A˝A
m˝A
/ A˝A
m
k ˝ AM
A˝A
m
/A
e˝A A˝e / A˝A o A˝k MMM oo MMM ooo o MM m ooo ' MMM ooo ' MM& oo wo
(25)
associativity
A identity
On reversing the directions of the arrows..4
Afﬁne groups and Hopf algebras
Un principe g´ n´ ral: tout calcul relatif aux e e cog` bres est trivial et incompr´ hensible.G/ that makes O.
4b
Co-algebras
D EFINITION 4.G/ into a k-algebra. Throughout k is a commutative ring.C ˝ / ı . Unadorned tensor products are over k.k/ and also for the homomorphism k ! O. ˝C/ı D idC D .
28
.C ˝ / ı
7 Warning:
D . ˝ C / ı :
(27)
I sometimes use “e” for the neutral element of G. we obtain the notion of a co-algebra. we examine the extra structure that the coordinate ring of an afﬁne group G acquires from the group structure on G. i.1 A co-associative co-algebra over k with co-identity (henceforth. e e Serre 1993.
In this section.

x/ D 1. then the formula
deﬁnes a linear map V _ ˝ W _ ! . E XAMPLE 4.! C is a co-algebra homomorphism.V ˝ W /_ which is always injective.A.D.x/ D cx:
. C . . g 2 W _ . If . i.e.f ˝ g/.
C ˝D /
C ˝D
! k ˝ k ' k. Similarly. x 2 X.A ˝ A/_ ' A_ ˝ A_ and the co-identity k ' k _ ! A_ .x/ D f . Then . co-algebra over k.! . On taking D D C . w 2 W.C.c/. These statements are proved by applying the functor _ to one of the diagrams (25) or (26). If V and W are k-modules. m. then C _ becomes an associative algebra over k with the multiplication C _ ˝ C _ .f. Such a D is called a sub-co-algebra of C . v 2 V .2 Let X be a set. C ˝D .x/ D x ˝ x.D/ D ˝ D. D / be co-algebras over k. and the inclusion D .
is a co-algebra over k. and is an isomorphism when at least one of V or W is ﬁnitely generated and projective (2. and let C be the free k-module with basis X .
C ? ? y k
f
C
! D ? ? yD k
f
D ıf
Let D be a subspace of a co-algebra C such that . and deﬁne C ˝D then . deﬁne C ˝D to be the composite C ˝D
C˝
! C ˝ C ˝k D ˝ D
C˝ D
D
C ˝t ˝D
'
C ˝ D ˝k C ˝ D to be the composite
where t is the transposition map c ˝ d 7! d ˝ c. .v/ ˝ g. such that ( . e/ is an associative algebra over k with an identity and A is ﬁnitely generated and projective as a k-module. if .w/.C ˝ C /_ ! C _ and the identity k ' k _ ! C _ .D.4. C / and . The dual algebra C _ can be identiﬁed with the k-module of maps X ! k endowed with the k-algebra structure m.x/ e. (28)
Recall that V _ denotes the dual of a k-module V ..
m_
_ _ _
endow C with the structure of co-algebra over k. x 2 X. jD. Afﬁne groups and Hopf algebras A homomorphism of co-algebras over k is a k-linear map f W C ! D such that the diagrams C ˝C x ? ? C C commute. The k-linear maps W C ! C ˝ C.v ˝ w/ D f . / is a co-associative co-algebra over k with a co-identity. g/. D .f ˝ f / ı
C f ˝f
29
! D ˝D x ? ? D ! D D D
D ıf C. then A_ becomes a co-associative co-algebra over k with the co-multiplication A_ ! .x/g.31).C ˝ D. Let . f 2 V _ . we see that C ˝ C is a
4c
The duality of algebras and co-algebras
.C. jD/ is a co-algebra (obvious). W C ! k. .

m. The next proposition shows that the deﬁnition is self dual.a ˝ a0 / ˝ mA . each of (a) and (b) is equivalent to the commutativity of all four diagrams. m. and the second and fourth commute if and only if e is a co-algebra homomorphism.b ˝ b 0 / eA˝B .c/ D eA . a bi-algebra over k is a quintuple .a ˝ b/ ˝ . .A.c/ ˝ 1 D 1 ˝ eB .A. ﬁnitely presented) if its underlying k-algebra is commutative (resp. / is a co-associative co-algebra over k with co-identity .A. the ﬁrst and third diagrams commute if and only if m is a co-algebra homomorphism. 2 A bi-algebra is said to be commutative (resp. In detail. On the other hand. (b) .4 For a quintuple . (d) W A ! k is a homomorphism of algebras.
. e.30
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
4d
Bi-algebras
For k-algebras A and B. Therefore. / where (a) . Note that these notions are not self dual. m0 . A homomorphism of bi-algebras . (b) m and e are co-algebra homomorphisms. m. and the third and fourth diagrams commute if and only if is an algebra homomorphism. . : : :/ is a k-linear map A ! A0 that is both a homomorphism of k-algebras and a homomorphism of k-co-algebras. : : :/ ! . e. .A. A ˝ B becomes a k-algebra with the maps mA˝B . e/ is an associative algebra over k with identity e.a0 ˝ b 0 // D mA . m. the following conditions are equivalent: (a) and are algebra homomorphisms.. P ROPOSITION 4.A. ﬁnitely presented).A0 .c/: D EFINITION 4. / satisfying (a) and (b) of (4.3). (c) W A ! A ˝ A is a homomorphism of algebras. P ROOF Consider the diagrams: A˝A
˝ A˝t ˝A m
/A
/ A˝A O
m˝m
(29)
A˝A˝A˝A
/ A˝A˝A˝A /A
A˝A o O
e˝e
A O
e '
A˝A
˝
m
k ˝k o
k
k ˝k
'
/k
A3 E 333 e 33 33 id /k k
(30)
The ﬁrst and second diagrams commute if and only if is an algebra homomorphism.3 A bi-algebra over k is a k-module with compatible structures of an associative algebra with identity and of a co-associative co-algebra with co-identity.

eW k ! A.e.
endow A with the structure of a k-algebra. The k-linear maps mW A ˝ A ! A. A homomorphism of Hopf algebras . such that m ı . m0 .id ˝S /
! A x ?e ? k
!
(31)
D EFINITION 4.S ˝ id/ ı D e ı D m ı .4. The reader encountering bi-algebras for the ﬁrst time should do Exercise 4-1 below before continuing. i. : : :/ is a homomorphism f of bi-algebras such that f ı S D S 0 ı f . this makes A into a bi-algebra over k.id ˝S / ı . iW B A!k
. and let A be the k-module with basis X.x ˝ x 0 / D xx 0 . e. the map SW A ! A. When X is a group. : : :/ ! .A0 .5 A Hopf algebra over k is a bi-algebra over k together with an inversion.
A SIDE 4.
4e
Afﬁne monoids and bi-algebras
An afﬁne monoid over k can be regarded as a representable functor from the category of commutative k-algebras to sets together with natural transformations mW M M !M eW ! M (32)
such that the following diagrams commute: M M
M
idM
m
/M
M
m
MM M
e idM
m idM
M
M
m
/M
MMM ooo MMM ooo o m M ooo ' MMM MM& ooooo ' wo
/M
M o
idM
e
M (33)
M
.7 To give a k-bialgebra that is ﬁnitely generated and projective as a k-module is the same as giving a pair of ﬁnitely generated projective k-algebras A and B together with a nondegenerate k-bilinear pairing h .6 Let X be a monoid..x/ D f .A.
mı.c/ D c1X . When combined with co-algebra structure in (4. Afﬁne groups and Hopf algebras An inversion (or antipode) for a bi-algebra A is a k-linear map S W A ! A such that A x ?e ? k commutes.x
1
/
satisfying compatibility conditions that we leave to the reader to explicate. x. m. m. is an inversion.Sf /. E XAMPLE 4.2). x 0 2 X.S˝id/
31
A˝A x ? ? A
mı. c 2 k.

which is represented by k. there exists at most one pair . More precisely.A.M / is ﬁnitely presented. If A represents M . M O. e/ of a monoid on the set-valued functor M deﬁnes the structure of a bi-algebra on the k-algebra A representing it. The afﬁne group M is algebraic if and only if the k-bi-algebra O. Here is the trivial monoid and . e 0 / is a homomorphism of Hopf algebras if .e.id. the product of two elements g. : : :/ and .A0 .id.32
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
(cf.A0 .R/ is the composite of A ! k ! R: A homomorphism M ! N of afﬁne monoids deﬁnes a homomorphism O. for a commutative k-algebra A.. S / such that . A homomorphism f W .R/ D Homk-alg .inv. and so the maps in (32) can be rewritten as mW hA˝A ! hA eW hk ! hA :
The Yoneda lemma shows that these maps arise from unique homomorphisms of k-algebras WA ! A˝A W A ! k.
and that the commutativity of the diagrams (33) is equivalent to that of the diagrams (26) (with A for C ).N / ! O. m. e. m.M / of kbi-algebras (by the Yoneda lemma again). .A. e/ and homomorphism W A ! A ˝ A.M / is a contravariant equivalence from the category of afﬁne groups over k to the category of commutative k-bi-algebras with inversion. i.8 (a) For any commutative k-algebra . Explicitly.id/
! M
M ? ?m y
.A. the structure of a bi-algebra on A deﬁnes the structure of an afﬁne monoid on hA . m. S / is a Hopf algebra. R/ is the composite of A ! A ˝ A ! R. the structure . to an inversion for A. Such a natural transformation corresponds to a k-linear map SW A ! A satisfying (31). the functor M O. P ROPOSITION 4. More precisely.h/
(34)
4f
Afﬁne groups and Hopf algebras
An afﬁne monoid M is an afﬁne group if and only if there exists a natural transformation invW M ! M such that M ? ? y
. then A ˝ A represents the functor M M .8). Therefore. The afﬁne monoid M is algebraic if and only if the k-bi-algebra O.M / is ﬁnitely presented. Here denotes the trivial monoid.
.A. h 2 G.m. (b) Let . . In other words. We conclude that to give an afﬁne group over k is essentially the same as giving a commutative bi-algebra over k with an inversion.
.A.inv/
M ? ? y
(35)
e
!
M
e
commutes. 2.g.M / is a contravariant equivalence from the category of afﬁne monoids over k to the category of commutative k-bi-algebras. . to give an afﬁne monoid over k is essentially the same as giving a commutative bi-algebra over k. m0 . and the identity element of G. Conversely.f ˝ f / ı D 0 ı f . m. m0 . e/ ! . : : :/ be commutative Hopf algebras over k. inv/ denotes the morphism whose composites with the projection maps are id and inv.

Therefore A hA is an equivalence from the category of monoid objects in Aopp to the category of monoid objects in A_ whose underlying functor to sets is representable (equivalently. A homomorphism f W .4. Now take A D Algk . Hopf algebras) correspond to algebraic monoids (resp. A hA W Aopp ! A_ is fully faithful. S / making . m0 . in particular. e.f ˝ f / ı D 0 ı f . and so A has ﬁnite products. groups). a group object in a category C with ﬁnite products is deﬁned to be an object M together with morphisms m. m.
8 For
. On comparing the diagrams (26) and opp (33). m. and inv deﬁne a group structure on Hom. (a) follows from (b) applied to the identity map. Now assume that A has ﬁnite direct sums. Tensor products in this category are direct sums (in the sense of category theory). A morphism of monoid objects is a morphism of the objects compatible with the additional data. Therefore f is a homomorphism of Hopf algebras. If . e/ and homomorphism W A ! A ˝ A. commutative Hopf algebras) to the category of afﬁne monoids (resp.A. S / a bi-algebra with involution.31). : : :/.A. . Thus.A. one can also deﬁne a group object to be a monoid object for which there exists a morphism inv such that the diagram (35) commutes. we see that a monoid object in Algk is just a commutative bi-algebra. m. e. The Yoneda lemma shows that inv is uniquely determined by m and e. . / is a bi-algebra (because an abstract monoid can have at most one neutral element). to the category of functors A ! Mon whose underlying functor to sets is representable). the functor A i 2I Fi . and therefore is a morphism of group valued functors (because any map of abstract groups sending products to products is a homomorphism). (b) Let G and G 0 be the afﬁne groups over k with Hopf algebras . and invW M ! M such that the diagrams (33) and (35) commute. According to the Yoneda lemma (2. e. a group object in Algk is just a commutative bi-algebra with an inversion.A/ is the product of the Fi . e. it would be more natural to deﬁne a commutative Hopf algebra to be a quadruple . then this morphism sends products to products. e/ ! . In summary: the functor A hA deﬁnes an equivalence from the category of commutative bi-algebras (resp. (b) There can exist more than one Hopf algebra structure on a commutative k-algebra (see 7. there exists at most one such that . . To give the structure of a monoid object on a functor M W A ! Set is to give a factorization of M through Mon.
any object T of C. and let A_ be the category of functors A ! Set. It follows from the deﬁnitions of direct sums and products. e/ is a commutative k-algebra and is a homomorphism A ! A ˝ A such that there exists a pair .A0 . Under the equivalence. groups).A.8 The same arguments as above show that A hA is an equivalence from the category of group opp opp objects in Algk to the category of afﬁne groups over k. Similarly. that the functor A hA sends direct sums to direct products. : : :/ and . / such that . Moreover. and so the above remarks show that A hA is an equivalence from the category of monoid opp objects in Algk to the category of afﬁne monoids over k. .A0 .
4g
Abstract restatement
Let C be a category with ﬁnite products and. m. 2 In view of the proposition. m. a ﬁnal object (the product over the empty set).A. e 0 / deﬁnes a morphism of functors G ! G 0 . Its essential image is (by deﬁnition) the subcategory of representable functors.1). Let A be a category.A.9 (a) For any commutative k-algebra .T.A. R EMARK 4. m. Afﬁne groups and Hopf algebras
33
P ROOF. m. the maps m. e. ﬁnitely presented bi-algebras (resp. For any ﬁnite family Q _ .Fi /i 2I of functors. m0 . M /. A monoid object in C is an object M together with morphisms mW M M ! M and eW ! M such that the diagrams (33) commute.

a/ D fR .Sf /R . As . b/ D fR . : : : .:::. Xnn . . i ¤ j : .a.a/ . : : : . X /: E XAMPLE 4. X2 / 2 kŒX1 .a. Moreover. X2 . f /R .G/ D kŒX.xij /
. y .1.D constant term of f /.f / is the (unique) element of O.11 For G D Gm .Sf /R .a/ D fR .G/ becomes identiﬁed with the coordinate ring of G G.G/.f2 /R . and so Ga Ga has coordinate ring kŒX1 .X1 C X2 /. and S
Let G be an afﬁne group over k.34
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
4h
Explicit description of
. E XAMPLE 4.xij / D 0.y/ D 1
S. kŒX ˝ kŒX ' kŒX1 .X / 2 kŒX acting as a 7! f .Ga / D kŒX I
in other words. X2 / D f . b/ D fR .G/ deﬁnes a function . f /.11) that an element f of the coordinate ring O.G/ such that .G/ is a family of functions fR W G.12 For G D GLn .X1 .0/ . and . is the homomorphism of k-algebras kŒX.a C b/ (see (36)). is the homomorphism kŒX ! k sending f .R/:
(38)
E XAMPLE 4.R/. b 2 G. Recall (2.R/ ! R by the rule: . O. f /R .12) that Ga has coordinate ring kŒX with f .f1 ˝ f2 /R . X 1 sending X to X 1 .a. X 1 ! kŒX.R/:
(36)
(37)
/.X1 . f D f . .xi i / D 1 . 1/.G/ ˝ O. b/ 7! F .ab/.R/ G.f1 /R . X 1 ! kŒX.R/ ! R of sets compatible with homomorphisms of k-algebras. X2 with F .a. X12 .n
kŒX11 . The ring kŒX ˝ kŒX is a polynomial ring in X1 D X ˝ 1 and X2 D 1 ˝ X.X. O.Sf /.X / D f .Y det. Moreover. X 1 / to f . a/.Xij / 1/
xij ˝ xj k
y D y ˝y
8 < . Y D kŒx11 .10 Recall (2. X 1 . we ﬁnd that . b/ on G. so that . X2 acting as .
for all R and all a 2 G.a
1
for all R and all a.G/ ˝ O. An element f1 ˝ f2 of O.f1 ˝ f2 /R W G.xij / D yaj i S.b/: In this way.R/ D R. is the homomorphism of k-algebras kŒX ! kŒX ˝ kŒX sending X to X ˝ 1 C 1 ˝ X.1/ (constant function).G/ such that .a. For f 2 O. X 1 ˝kŒX. f D f .G/ ˝ O. and Sf is the element of O. and S is the homomorphism kŒX. b/ D . X 1 sending X to X ˝X. O.a/ on Ga .G/ D and ( xi k D P
j D1.R/ G.y/ D det. f 2 O. xnn .

x/ is the matrix with ij th entry xij .4.F /k .cij // D ckl as claimed.
P
e D 1..e / D e
1
:
deﬁne a bi-algebra structure on A with inversion S. . Symbolically.F /k . then a k-algebra homomorphism A ! R must send one e to 1 and the remainder to 0. j D1.x/ where .R/ may be bigger than F ).A.13 Let F be a ﬁnite group. S is the given one. On comparing the ﬁrst and third diagrams and applying the Yoneda lemma.bij / P D j aij bj k P D .xj i /.
35
E XAMPLE 4.e / D X
.x/ D . . y/ 7! .
.xi k /: deﬁnition (36) as .R/ ' F .bij / .F /k be the associated algebraic group. Afﬁne groups and Hopf algebras where aj i is the cofactor of xj i in the matrix . We check the formula for .n xij ˝ xj k /R . The e ’s are a complete system of orthogonal idempotents for A: e2 D e .
e ˝e .
.
4i
Commutative afﬁne groups
A monoid or group G (resp. let e be the function that is 1 on and 0 on the remaining elements of F . The maps . x/ or x ˝ y 7! y ˝ x.xi k /R . and a co-algebra or bi-algebra C is co-commutative if the diagram at right commutes: G G ?
t
?? ?? m ? ?
G
/G G m
A˝A ?
t
?? ?? m ? ?
A
/ A˝A m
C ˝C o _?
t
? ? ? ? ? ?
C
C ? ˝k C
(39)
In each diagram.F /k . . .y. so that .F /k is called the constant algebraic group deﬁned by F (even though for k-algebras R with nontrivial idempotents. Let .aij /. an algebra A) is commutative if and only if the diagram at left (resp.x.bij / D . and one checks that the group structure provided by the maps . Therefore.aij /.xkl /R . More precisely.R/ D Homk-alg . For this reason.x/ ˝ .e / D
1 if D 1 .aij /. we can write the formula for as . we see that an afﬁne monoid or group is commutative if and only if its coordinate ring is co-commutative. Then A is a product of copies of k indexed by the elements of F . xi k /R .:::. with D
e e D 0 for
¤ . the middle diagram) commutes. 0 otherwise
S. t is the transposition map . R/: If R has no idempotents other than 0 or 1. and let A be the set of maps F ! k with its natural kalgebra structure. . .

i. e. the group acts continuously on ˝ ˝ V according to rule: 2 . An algebraic group G is said to be ﬁnite and ﬂat if O. To check that S _ is an algebra homomorphism. if
. m_ .15 The functor V ˝ ˝k V from vector spaces over k to vector spaces over ˝ equipped with a continuous action of is an equivalence of categories. then . See AG. m_ . . c 2 ˝. and let D Gal. we endow with the Krull topology. m. ! O. and so G _ is an algebraic group.G/ admits an inversion S .G/_ is the coordinate ring of a commutative ﬁnite ﬂat algebraic monoid G _ . which commutes precisely because G is commutative (the inverse of a product is the product of the inverses): O. we have check the diagram at left below commutes.G/ G x ? ?inv G
m
G
G x ? ?inv G:
inv
m
G
4k
Galois descent of afﬁne groups
In this subsection. _ . and so its dual O. and v 2 V:
2 i ci
. and v 2 V . If moreover .36
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
4j
Finite ﬂat algebraic groups. 0 P ROPOSITION 4. / is commutative (resp. or.4).15 or V. _ .G/ ˝ O. e _ / is co-commutative (resp. e.cv/ D . _ . called the Cartier dual of G. This corresponds (under a category equivalence) to the diagram at right. For any vector space V over k. .e.A. 1.c ˝ v/ D c ˝ v for all
P ROPOSITION 4. In other words.14 For any ˝-vector space V equipped with a continuous action of P is an isomorphism. 16. If O. e _ / is also a k-bialgebra (see 4c and Proposition 4. / is a bi-algebra over k and A is ﬁnitely generated and projective as a k-module. k is a ﬁeld. such that .G/ ? ? yS O. _ . When ˝ is an inﬁnite extension of k.S _ ˝ S _ / D S _ ı _ . Cartier duality
If .˝=k/. P ROOF. then .A_ .2 (the proof is quite elementary).G/ is ﬁnitely generated and projective as a k-module. equivalently. m.G/ ˝ O. c 2 ˝.G/ ? ? yS ˝S ! O. By an action of on an ˝-vector space V we mean a homomorphism ! Autk . The coordinate ring O. that ı S D . the map
˝ vi 7!
P
i ci vi W ˝ ˝k
V
!V
. commutative).V / such that each 2 acts -linearly. co-commutative). then S _ is an algebra homomorphism. we have to check that _ ı .G/ of a commutative ﬁnite ﬂat algebraic monoid is a commutative co-commutative bi-algebra. We say that the action is continuous if every element of V is ﬁxed by an open subgroup of [ 0 V D V (union over the open subgroups 0 of )..
..A.c/ .G/ Note that G __ ' G. i.S ˝ S/ ı .e.v/ for all 2 . Let ˝ be a Galois extension of the ﬁeld k.A_ .

g 2 O. no definition of “quantum group”. One interesting aspect of quantum groups is that. i. d modulo the relations ba D qab. The algebra Aq is usually referred to as the Hopf algebra of SLq . their Hopf algebras can be. ˝ ˝ V 0 / become identiﬁed with with certain sets of matrices. many statements in this section become more difﬁcult to prove. while it is still true that a bi-algebra admits at most one inversion. III 3).SL2 /. or even false. That the functor is essentially surjective follows from (4. there are different deﬁnitions for “Hopf algebra”.e. . For example. 2 Let G be an afﬁne group over k.15. e. c. ad D q
1
da D ad C . By a continuous action of on G we mean a continuous action of on O. bc D cb. For example. the only Hopf algebras seriously studied were either commutative or cocommutative. c ˝aCd ˝c c ˝b Cd ˝d
and with suitable maps and S . and the fully faithfulness of the functor follows from the fact that ˝ D k. Despite the name. ) e use “cog` bre” and “big` bre” for “co-algebra” and “bi-algebra”. When q D 1.
4m
Terminology
From now on.G/ preserving its Hopf algebra structure . Then Drinfeld and Jimbo independently discovered noncommutative Hopf algebras in the work of physicists.G/: P ROPOSITION 4.2/: For bi-algebras that are neither commutative nor cocommutative. V 0 / and Hom˝-lin . When we choose bases for V and V 0 .m.. a quantum group does not deﬁne a functor from the category of noncommutative k-algebras to groups. g// D . There is. Thus.f. Serre.a/ . Bourbaki and his school (Dieudonn´ . g/ for all 2 .14). Afﬁne groups and Hopf algebras
37
P ROOF. the notion of a bialgebra is not self dual. Aq becomes O. only examples. and so the Aq can be regarded as a one-dimensional family of quantum groups that specializes to SL2 when q ! 1. b. . “bialgebra” will mean “commutative bi-algebra” and “Hopf algebra” will mean “commutative bi-algebra that admits an inversion (antipode)” (necessarily unique). S /. and Drinfeld called them quantum groups. q
1
dc D qcd.q deﬁned by Ã Â Ã
/bc. Immediate consequence of Proposition 4.4. the composite of an inversion with itself need not be the identity map (Kassel 1995. . .˝ ˝ V. f.
2
P ROOF. For q 2 k . ca D qac.V.16 The functor G equipped with a continuous action of G˝ from afﬁne groups over k to afﬁne groups over ˝ is an equivalence of categories. . at present.
bc D 1:
This becomes a Hopf algebra with Â a b c d Ã Â
a b a b D ˝ .c/ .
4l
Quantum groups
Until the mid-1980s. e e
9 In
.d /
D D D D
a˝aCb ˝c a˝b Cb ˝d . f.b/ . while semisimple algebraic groups can’t be deformed (they are determined up to isomorphism by a discrete set of invariants). deﬁne Aq to be the free associative (noncommutative) k-algebra on the symbols a. we require that . then Homk-lin . c d c d ˆ ˆ :
8 ˆ ˆ <
. . db D qbd.9
the literature.

Q=Z/. _ p
'
E XERCISE 4-5 Let A be a Hopf algebra. y/ D f . /. .Gm / satisfy the axioms to be a Hopf algebra. .
. (Cf. (b) ı S D t ı S ˝ S ı where t .X / be the k-algebra of maps X ! k. (b) Show that maps R0 . p . /.a
1/ 1
D e.y/. / the structure of a Hopf algebra. / spanned by the functions arising in this way for varying n.b/W A ˝ A ! A ˝ A is a homomorphism of k-algebras.) (a) Show that R0 . Chapter 2. / into the subring R. let R.a ˝ b/ D b ˝ a: (c) ı S D : (d) The map a ˝ b 7! .gg 0 / f D f . and consider an element fi ˝ gi . From a homomorphism W ! GLn .Z=pZ/k ) (here . Cartier 2007.
E XERCISE 4-6 Show that there is no algebraic group G over k such that G. .x.) (b) Let be a group and deﬁne maps W R.Z=pZ/_ ' p (hence k . / of R. g 0 / D f . / is a k-subalgebra of R. f /.g/i. / into R0 . / ˝ R0 . and S deﬁne
Show that if maps R. /.x/g.j . /.Z=pZ/k . /.15). / ! R.
_ E XERCISE 4-4 If k has characteristic p ¤ 0.g/ D f . (The elements of R0 . in the exercises assume k to be a ﬁeld. then on R. let R.
E XERCISE 4-2 We continue the notations of the last exercise.1. / ! R.
e
1
D e.14).Y / ! R. Let R0 . (a) S ı S D idA . show that ˛p ' ˛p and .Sf /. / ˝ R. we obtain a family of functions g 7! . show that always maps R.1/ . / ˝ R.X Y / just deﬁned is injective. / are called the representative functions on . For a second set Y . 2. Let n be the exponent of .ab/
1
Db
1a 1.g. 3.Y / act on X Y by the rule (f ˝ g/. / ! k. in particular.X / as a k-vector space.g
1
/: . SW R. and assume that k contains n distinct nth roots of 1 (so. .38
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
4n
Exercises
To avoid possible problems. E XERCISE 4-7 Verify directly that O. / the structure of a Hopf algebra.k/.1. Hints: . and (2. Show that the Cartier dual of G is the constant algebraic group deﬁned by the dual group Hom. . and ˛p are the groups in (2. / be the k-subspace of R. 1 Ä i. Let be an arbitrary group. and S deﬁne on R0 . (c) If is ﬁnite. (c) Deduce that . E XERCISE 4-1 For a set X.R/ has two elements for every k-algebra R. Prove the following statements by interpreting them as statements about afﬁne groups. (2. (Hint: choose a P basis fi for R. W R.Ga / and O. / into R.16)). (a) Show that the map R. Abe 1980. n is not divisible by the characteristic of k).) E XERCISE 4-3 Let G be the constant algebraic group over k deﬁned by a ﬁnite commutative group . on G. /.X / ˝ R.a ˝ 1/ . j Ä n.X/ ˝ R.

V /
39
V ˝ C C C ˝ V and
(a) Show that the kernel of any homomorphism of coalgebras is a coideal. (c) Show that the image of a homomorphism ˛W A ! B of bialgebras is a sub-bialgebra ˛A of B. Show that any homomorphism of k-bialgebras A ! B whose kernel contains a factors uniquely through A ! A=a.4.f ˝ g/ D Ker. (b) Let V be a coideal in a k-coalgebra C .a/ a. Hopf ideal). Sweedler 1969. In other words. then Ker. (b) Let a be a bi-ideal in a k-bialgebra A. Afﬁne groups and Hopf algebras E XERCISE 4-8 Verify all the statements in 4.
.a/ D 0. and that ˛ deﬁnes an isomorphism of Hopf algebras A= Ker.a/ a: (a) Show that the kernel of any homomorphism of bialgebras (resp. If A and B admit inversions. an ideal a A is a bi-ideal if . and that ˛ deﬁnes an isomorphism of A= Ker.f / ˝ W C V ˝ Ker.V / D 0. Show that any homomorphism of k-coalgebras C ! D whose kernel contains V factors uniquely through C ! C =V . Show that an inversion on A induces an inversion on A=a provided that a is a Hopf ideal.˛/ onto ˛A. and it is a Hopf ideal if.
C . in addition. show that ˛A does also. Show that the quotient vector space A=a has a unique k-bialgebra structure for which A ! A=a is a homomorphism.1). E XERCISE 4-9 A subspace V of a k-coalgebra C is a coideal if C . A k-subspace a of a k-bialgebra A is a bi-ideal if it is both an ideal and a co-ideal. When A admits an inversion S . Hopf algebras) is a bi-ideal (resp. a bi-ideal a is a Hopf ideal if S. Show that the quotient vector space C =V has a unique k-coalgebra structure for which C ! C =V is a homomorphism.3.8). Hint: show that if f W V ! V 0 and gW W ! W 0 are homomorphisms of k-vector spaces.˛/ ! ˛A (hint: apply 4.g/: E XERCISE 4-10 (cf. 4.a/ a ˝ A C A ˝ a and
. S.13.

A nonempty topological space is irreducible if it is not the union of two proper closed subsets.3). 40
. A topological space V is noetherian if every ascending chain of open subsets U1 U2 eventually becomes constant.W / D fm j m 2 W g. we saw that algebraic groups over a ﬁeld k correspond to group objects in the opposite of the category of ﬁnitely presented algebras over k (see 4g). in particular.f / are a base for a topology on V whose closed sets are exactly the sets V . and the map a 7! V . in other words. A itself is reduced (CA 18. In this section we give a geometric interpretation of this opposite category as the category of afﬁne algebraic schemes over k. specm.5
Algebraic groups and afﬁne algebraic schemes
In the last section. Throughout this section. and the set V endowed with the Zariski topology is the max spectrum specm.a/ D fm 2 V j m ag: The sets D.10). then K ˝k A is reduced for all ﬁelds K containing k. When k is perfect.1). When A is a ﬁnitely generated k-algebra. The elements of this I are called the irreducible components of V . and let V be the set of maximal ideals in A.A/. every reduced ﬁnitely generated k-algebra is an afﬁne k-algebra (CA 18.4):
5b
The max spectrum of a ring
Let A be commutative ring. This is the Zariski topology. let V . or. If A is afﬁne. For an element f of A. to afﬁne algebraic group schemes. let D.
5a
Afﬁne k-algebras
An afﬁne k-algebra is a ﬁnitely generated k-algebra A such that k al ˝k A is reduced.a/ deﬁnes one-to-one correspondences radical ideals $ closed subsets prime ideals $ irreducible closed subsets maximal ideals $ one-point sets: T The ideal corresponding to a closed set W is I.A/ is irreducible if and only if N is prime. The nilradical N of A is the smallest radical ideal.A/ of A (see CA 12). This geometric interpretation provides additional insights.a/. and for an ideal a in A. k is a ﬁeld. and so it corresponds to the whole space specm.f / D fm 2 V j f … mg.A/ is noetherian. Therefore specm. We shall need to use some statements from commutative algebra (proved in CA). The tensor product of two afﬁne k-algebras is again an afﬁne k-algebra (CA 18. Every noetherian topological space V can be expressed as the union of a ﬁnite collection I of irreducible closed subsets: [ V D fW j W 2 I g: Among such collections I there is exactly one such that no subset in I contains a second (CA 12. Thus algebraic groups over k correspond to group objects in the category of afﬁne algebraic schemes over k.

fr /:
For a principal open subset D of V . and so OA .U / ! OA . then SD SD 0 . The set I in (40) can be replaced by any subset of B
.
VARIANT:
THE PRIME SPECTRUM
The (prime) spectrum spec.m/.U / into a functor on B satisfying the sheaf condition: for any covering D D i 2I Di of a D 2 B by Di 2 B.D. is a base for the topology. Algebraic groups and afﬁne algebraic schemes
41
Let 'W A ! B be a homomorphism of ﬁnitely generated k-algebras. The advantage of the functor A spec.5.'/W specm B ! specm A. These restriction maps make D S D. deﬁne OA .
which is continuous because the inverse image of D.A/ consisting of the closed points.A/ of a commutative ring A is the set V of prime ideals in A endowed with the topology for which the closed sets are those of the form V .A/ have the same topologies — only the underlying sets differ.i. deﬁne OA . B D fD.A/ onto the set of open subsets of specm.D \ D 0 / 0
D2I .D/ ! OA .48). Recall that the set of principal open subsets of V .'.f / j f 2 Ag. therefore specm.A/. but the inverse image of a maximal ideal need not be maximal when the rings are not ﬁnitely generated algebras over a ﬁeld. and so there is a canonical “restriction” homomorphism OA .A/ is a bijection from the set of open subsets of spec. Clearly. For any maximal ideal m in B. The advantage of specm.D/ ! OA .m/ is maximal in A.Di / OA . U OA .f //.p/ is the smallest closed subset containing the point p 2 V . Moreover. and it is not difﬁcult to check that it is a sheaf.A/ is that it is deﬁned on all commutative rings: the inverse image of a prime ideal is again prime.D/ ' Af (see CA 6). the map U 7! U \ specm.U / is a functor on the open subsets of V .A/ and spec. ' 1 .j /2I I
is exact.f1 fr / D D. and therefore ' deﬁnes a map specm. m 7! '
1
. and so specm.Di \ Dj /
i 2I .A/ is the subspace of spec.
5c
Afﬁne algebraic schemes
Let A be a commutative ring. When A is a ﬁnitely generated k-algebra.f /. the sequence Y Y OA .D /2I I
(40)
where I D fD 2 B j D U g. If D D D. B is closed under ﬁnite intersections because D. a an ideal in A:
Clearly V . If D D 0 . and let V D spec A.D 0 /. In this way.A/ is that it is more geometric.U / by the exactness of Y Y OA .f1 / \ : : : \ D. For an open subset U of V .D/ OA . then SD is the smallest saturated multiplicative subset of A containing f .f / is D.D/ D SD1 A where SD is the multiplicative subset S A p2D p of A. specm becomes a contravariant functor from the category of ﬁnitely generated k-algebras to topological spaces (CA p.a/ D fp 2 V j p ag.

V /. When k is algebraically closed. An afﬁne scheme .A/ D . Xn =a D kŒx1 .a1 .V. OA / is called the spectrum of A.42
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
covering U without changing OA . xn an /W V ! specm . This has the same underlying topological space as before (namely. Now consider Specm. Let V be a closed subset of k n .A/ (so the points of an afﬁne algebraic scheme are closed). O/ O. : : : . : : : . Y =. The functor A Spec. the x-axis in k k) but it should now be thought of as having multiplicity 2.spec A. the stalk of OA at a point p 2 V is A ' Ap O . Therefore.h/ of V where h is nonzero.h/ h. and we call O. an / 7! . and let kŒV D kŒX1 . xn : A pair of elements g.f // ' Af .Y 2 //. We often write V for .U / ' OA . P ROPOSITION 5.x1 a1 . from afﬁne algebraic schemes over k to ﬁnitely generated commutative k-algebras.V.V / is a contravariant equivalence from the category of afﬁne schemes to the category commutative rings (resp.V. then OA . and endow k n with the topology for which the closed sets are the zero-sets of families of polynomials.V. O/.V / the coordinate ring of V .D.V / D kŒV . E XAMPLE 5. O/ is an afﬁne algebraic scheme over k with O.A/ for some commutative ring A. note that A D OV . and . Y =. The ringed space Spec.U / D lim Op D lim !f …p f !f …p A !U 3p A (for the last isomorphism.U / be the set of regular functions on U . an afﬁne algebraic variety over k). OV / is a ringed space isomorphic to Spec. The afﬁne algebraic scheme Specm.V /. resp.4 — the map . : : : .kŒX.P /
!k
on the open subset D. A function f W U ! k on an open subset U of V is said to be regular if it is of this form in a neighbourhood of each point of U . this deﬁnition of an afﬁne algebraic variety over k agrees with that in AG.kŒX.3 Let k be an algebraically closed ﬁeld.1 The functor .V. When V D k n . P ROOF. then we say that V is an afﬁne algebraic scheme over k (resp.U /. and a morphism of afﬁne algebraic schemes over k is a morphism of spaces ringed by k-algebras. In the second case.
2
def
E XAMPLE 5.A/ with Specm. Chapter 3. A morphism of afﬁne schemes is morphism of ringed spaces. from afﬁne algebraic varieties over k to afﬁne k-algebras). or being a line thickened in another dimension.f // ' lim O . The reader should think of an afﬁne scheme as being a topological space V together with the structure provided by the ring O.kŒV / is a bijection because of the Nullstellensatz.2 Let k be an algebraically closed ﬁeld. : : : . if U D D. an afﬁne k-algebra).Y // can be identiﬁed with the scheme attached to the closed subset Y D 0 of k k in (5.f /. the scheme .2).D. Then U O. let a be the set of polynomials that are zero on V . and we replace Spec.
. h 2 kŒV with h ¤ 0 deﬁnes a function P 7!
g. see CA 6). Let k be a ﬁeld.A/ is a quasi-inverse. See AG 3. we often call it a regular map. Let O. If OV .P / W D.U / is a sheaf of k-algebras on V .V / is a ﬁnitely generated k-algebra (resp. In particular. O/ is afﬁne n-space An .

and so V W is the product of V and W in the category of afﬁne algebraic schemes. and it is geometrically reduced of VK is reduced for all ﬁelds K containing k. and so a geometrically reduced afﬁne algebraic scheme is an afﬁne algebraic variety. D Specm. V W / ' Hom. group scheme) over k. which is n.11) that the height ht. Therefore the category of afﬁne algebraic schemes has ﬁnite products. and so O.V /.T. the category of afﬁne algebraic varieties also has products. An afﬁne algebraic scheme V over k deﬁnes an afﬁne algebraic scheme VK over K with O. The dimension of an afﬁne algebraic scheme V is the Krull dimension of O. p.V /=N.VK / D K ˝k O. Thus V is geometrically reduced if and only if O.V W / D O.k/ is a ﬁnal object.40).T. V is reduced and irreducible ” O.V / is prime.11). the nilradical N of O. all afﬁne algebraic schemes T.W /. In this case. A monoid (resp. Similarly. Xn / over k. group) in the category of afﬁne algebraic varieties will
.V / has length dim V (CA 13. : : : .V / is an afﬁne k-algebra. W /. Then V is reduced ” N D 0I V is irreducible ” N is prime.8). group) in the category of afﬁne algebraic schemes over k will be called a monoid scheme (resp. and so monoid objects and group objects are deﬁned.23. A monoid (resp. The ﬁrst statement follows from the deﬁnitions.A/g.T. Then (16).V / is an integral domain. For example. and let V W be an afﬁne algebraic scheme such that O. the dimension of V is the transcendence degree over k of the ﬁeld of fractions of O. Therefore. the dimension of An is the transcendence degree of k.m/ j m 2 specm. p1 pd . shows that Hom.
5e
Algebraic groups as groups in the category of afﬁne algebraic schemes
Let V and W be afﬁne algebraic schemes over k. V / Hom. and the third statement follows from the ﬁrst two. Recall (CA 3.
A scheme is a locally ringed space that admits an open covering by afﬁne schemes.5. As the tensor product of two afﬁne k-algebras is again afﬁne. and every maximal chain of distinct prime ideals in O. the second statement has already been noted (p. When V is irreducible. Algebraic groups and afﬁne algebraic schemes
43
S CHEMES IN GENERAL .V / — this is ﬁnite (CA 13.V / is reduced.X1 .V /=N is an integral domain.
5d
Properties of afﬁne algebraic schemes
Let K be a ﬁeld containing k.V / ˝k O.p/ of a prime ideal p in a noetherian ring A is the greatest length d of a chain of distinct prime ideals p and that the Krull dimension of A is supfht. An afﬁne algebraic scheme V over k is reduced if O. Let N be the nilradical of O.V /. every maximal chain of distinct irreducible closed subsets of V has length dim V .

groups). not as topological spaces). with quasi-inverse G Specm. ﬁnitely generated Hopf algebras) over k. group variety). or just jGj. For an algebraic group G. O/ V sending an afﬁne algebraic scheme over k to its underlying set preserves ﬁnite products.jGj. jGj is a group when k is algebraically closed. S UMMARY 5. O/ V does not preserve ﬁnite products: the topology on V W is not the product topology. group) schemes to the category of ﬁnitely generated bialgebras (resp. there is a canonical map jG Gj ! jGj jGj. Let V 0 be the functor deﬁned by V . when k1 and k2 are ﬁnite ﬁeld extensions of k.V /.V. However. for ﬁnitely generated algebras A1 and A2 over an algebraically closed ﬁeld k. group) scheme G such that O. . (b) the opposite of the category of ﬁnitely generated bialgebras (resp.G/ is an afﬁne k-algebra. Monoid (resp.k1 / specm.k/.44
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
be called a monoid variety (resp.1) that V O. Hopf algebra) structures on O. Thus V O. For an afﬁne algebraic group G.10 and hence also monoid objects and group objects. and so the forgetful functor . denote the corresponding group scheme (or group variety). O.
I S THE SET jGj A GROUP ?
Not usually. The dimension of G is deﬁned to be the dimension of jGj.k2 / has only one. groups).R/ ' Rn for all k-algebras R.V. For example. groups) over k.1) and the Yoneda lemma that V V 0 is an equivalence from the category of algebraic schemes over k to the category of functors from k-algebras to sets representable by ﬁnitely generated k-algebras.A2 / (42)
(as sets. For an algebraic group G. An . (c) the category of monoids (resp. thus jGj D Specm.4 There are canonical equivalences between the following categories: (a) the category of algebraic monoids (resp. the map mW jGj jGj ! jGj is not usually continuous relative to the product topology.V / is an equivalence from the category of algebraic schemes over k to the category of ﬁnitely generated k-algebras. An afﬁne algebraic scheme V deﬁnes a functor R V . group) variety is a monoid (resp.
regarded as a functor to topological spaces.O.
def
(41)
from k-algebras to sets.G//. groups) in the category of algebraic schemes over k.O. To put it another way.V / is a contravariant equivalence from the category of monoid (resp. but the map jG Gj ! jGj deﬁned by m need not factor through it.A1 / specm. R/. specm.O. a monoid (resp. group) structures on V correspond to factorizations of V 0 through the category of monoids (resp.k/. Then the Nullstellensatz shows that jGj ' G.k1 ˝k k2 / may have many points whereas specm. Thus V V 0 is an equivalence from the category of monoid (resp. For example. group) structures on V correspond to bialgebra (resp. It follows from (5. the set specm. and so jGj inherits a group structure from G. we let . The problem is that the functor specm does not send sums to products. We saw in (5.
10 When
. Equivalently. group) schemes to the category of algebraic monoids (resp.R/ D Homk-alg .V /.A1 ˝k A2 / ' specm. ﬁnitely generated Hopf algebras). Monoid (resp.G//.G//.

k al / and G. and let D Gal. because the equalities La ı La show that La is an isomorphism. and so `a extends to an isomorphism O. Let e denote the identity element of G.k/.g/ D fR .G/ j fk . The homomorphism O.6 For each a 2 G.G/ ! k. the former need not. The isomorphism `a W O. For example.f /R .G/ deﬁned by La is the composite of the homomorphisms O.R/.G/
a˝O. Then O.k/ with the set of orbits consisting of a single point. An element a of G. While the latter inherits a group structure from G. thus ma D ff 2 O.G/ ˝k O.k/. Le D idG and La ı Lb D Lab .
.G/ ! O. Here e is the neutral element in G.G/. which we denote aR (or just a). the natural map La W G.5 For each a 2 G.k al / and G. P ROOF.k al =k/.G/ j fk .aR g/.a/ ¤ 0g: P ROPOSITION 5.G/me (because of the universal property of rings of fractions. we let ma denote the kernel of aW O. b 2 G. and so ma is a maximal ideal in O. The situation is worse with spec.G/ma ' O. In other words. Therefore.k/. and the ﬁrst follows from it. jGj can be identiﬁed with the set of -orbits in G.1).7 When k is algebraically closed.G/ ! O. Note that O.G/ma is the ring of fractions obtained from O. `a 1 me D ma . the local rings O.G/.11).G/m at maximal ideals m of O. 2 C OROLLARY 5.k/. The second statement is obvious.R/.k/ deﬁnes an element of G. Algebraic groups and afﬁne algebraic schemes
45
Assume k is perfect.
all a. all g 2 G.G/ are all isomorphic. O.G/ ! O. (42) fails for spec even when k is algebraically closed.G/ by inverting the elements of the multiplicative set ff 2 O.G/
1
g 7! aR g.5. Then jGj ' nG. P ROPOSITION 5.G/ corresponding (by the Yoneda lemma) to La is deﬁned by `a .R/ ! G.k/ ' G.G/ ' O.
(43)
For a 2 G.G/me : P ROOF.k/:
D Le D idG
2
! k ˝k O.R/ for each k-algebra R.k/.a/ D 0g (see the notations 2.k al / . CA 6.
5f
Homogeneity
Let G be an algebraic group over a ﬁeld k.G/=ma ' k.k/. Moreover.G/ma ! O. is an isomorphism of set-valued functors.

and so O.G/=N. p kŒX kŒX kŒX O.G/ ! k. O. Hence Ared ˝k Ared is also an afﬁne k-algebra. and so it is an afﬁne k-algebra. for the algebraic group 3 over Q. P ROPOSITION 5. 2
A
5.k/. the intersection of its maximal ideals is zero (CA 11.6) shows that all maximal ideals in O.G/ D k. Let A D O. let k be a ﬁeld of characteristic 2. CA 11. and it follows that there exists a unique structure of a Hopf algebra on Ared such that A ! Ared is a homomorphism of Hopf algebras. It contains k.G/=N such that O.
A
5. and the identity element in G is a homomorphism O. 2 The algebraic group Gred is called the reduced algebraic group attached to G. and so m D 0.X 3 1/ . Every maximal ideal of O. but ˛p is not the trivial group.X 1/ .G/ is a ﬁeld. P ROPOSITION 5. Any homomorphism H ! G with H a reduced algebraic group factors uniquely through Gred ! G. ˛p . For example.9 Let G be a reduced algebraic group over a ﬁeld k. it is reduced. 3 / D ' ' Q QŒ 3.R/ D fx 2 R j x 4 D ax 2 g is an additive commutative algebraic group. i..e. then G is the trivial algebraic group.G/=N is not a Hopf algebra quotient of O.K/ D f1g for every ﬁeld K containing k. i. For example. In particular.46
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROOF.G/=N is a homomorphism of Hopf algebras.5). Let Gred ! G be the corresponding homomorphism of algebraic groups.G/ and Ared D O. but O.G/.K/ D f1g for some algebraically closed ﬁeld K containing k.G/ is reduced.
.8 The corollary fails when k is not algebraically closed.G/ has only one maximal ideal m. and let N be the nilradical of O.G/ D k. When k is algebraically closed. Similarly. Any homomorphism from A to a reduced k-algebra factors uniquely through A ! Ared . 2 5. O..X 2 C X C 1/ p and so the local rings are Q and QŒ 3.G/ has no nilpotents. As O. If G. For example.
A
5g
Reduced algebraic groups
An algebraic group G is reduced if jGj is reduced.11 Let G be an algebraic group over a perfect ﬁeld k. and so O.G/ are of the form ma for some a 2 G. S and are deﬁned on Ared . and so the map A ! A ˝k A ! Ared ˝k Ared factors through Ared .10 The proposition is false for nonreduced groups.G/ (see Exercise 12-6 below). Then R G.G/ arises as the kernel of a homomorphism O. and let a be a nonsquare in k. from which the ﬁnal statement follows.G/ ! O. . a Hopf algebra structure on A need not pass to the quotient A=N. Then Ared is a ﬁnitely generated reduced algebra over a perfect ﬁeld.12 When k is not perfect.e. Therefore O. P ROOF. P ROOF.G/ ! K (Nullstellensatz.8). There is a unique Hopf algebra structure on O. the Nullstellensatz (CA 11.

5.H / is a quotient of O. if H ¤ G.G/m is regular for all m by homogeneity (5.18 An algebraic group G over an algebraically closed ﬁeld k is smooth if and only if O. 2 P ROPOSITION 5. Moreover. P ROOF.14 Let A be a local noetherian ring with maximal ideal m and residue ﬁeld k. and dim H < dim G if G is smooth and connected and H ¤ G. P ROOF. If G is smooth and connected.G//. for all r Ä n. For the converse.3). 2 P ROPOSITION 5. and so.O. let f be a nilpotent element of O.7).16). sf D 0 for some s 2 O.e.O.13 Let m be a maximal ideal of a noetherian ring A.16). and let n D mAm be the maximal ideal of the local ring Am .17 Let H be an algebraic subgroup of an algebraic group G. 5. Algebraic groups and afﬁne algebraic schemes
47
5h
Smooth algebraic schemes
We review some deﬁnitions and results in commutative algebra. then jGj is an algebraic variety by (5.15 A point m of an afﬁne algebraic scheme V is said to be regular if the local ring O. dim. an algebraic group G over an algebraically closed ﬁeld is smooth if and only if jGj is reduced.V /.m/ Ä dimk . then O. Hence G is smooth. then dim H < dim G by (CA 13. W O. P ROPOSITION 5.V /.m=m2 / (CA 16. If O. the map a C mn 7! a C nn W mr =mn ! nr =nn is an isomorphism (CA 6.G/..G/m is regular for m D me .V /m .V /m is regular.V / not contained in any maximal ideal. Because O. and V is said to be regular if all of its points are regular. If V is smooth. CA 3.15) that O. we have to show that Gk al is regular. If G is smooth.3).V / m. Then dim H Ä dim G. an algebraic variety).15).6). Gk al has a regular point. as f maps to zero in O. V itself is regular (CA 18. in particular. then VK is regular for all ﬁelds K containing k. P ROOF. ht. all of its points are regular. A regular afﬁne algebraic scheme is reduced. Then m=m2 is a k-vector space of dimension equal to the minimum number of generators of m (Nakayama’s lemma. 5. According to (5. then it follows from (5. If V is smooth.H // Ä dim. and so it equals O.19 An algebraic group G is smooth if and only if jGj is geometrically reduced (i. In particular.V / is an afﬁne k-algebra. Every afﬁne algebraic variety contains a regular point (CA 18.16 An afﬁne algebraic scheme V over k is said to be smooth if Vk al is regular.
5i
Smooth algebraic groups
An algebraic group G is said to be smooth if jGj is smooth. therefore the annihilator of f is an ideal O. To see this. and so V is an algebraic variety.G/ is an integral domain. by homogeneity (5.G/ ! k/.9). 2
.14). 5.5). 5. and when equality holds A is said to be regular.G/me is regular. Every regular noetherian local ring is an integral domain (CA 17.6). where me D Ker. then O.

G/me /.k/. Indeed.A. x ˝ y D xW ˝ y C xW 0 ˝ y.G/ is contained in m2 .48
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
C OROLLARY 5.G/me is regular. y/ 7! y p ax p W Ga Ga ! Ga .Gk al / D k al ŒX. so that char. Ker.23 Let .22 Let V and V 0 be vector spaces.21 A reduced algebraic group over a nonperfect ﬁeld need not be smooth.k/ D p ¤ 0 and there exists an element a of k that is not a pth power. O.G/ N of O. . Y =.G/.G/ are in natural one-to-one correN N spondence. Because jGj is a reduced.me =m2 / DKrull dim.
. The reduced subgroup . /.G/ and O. y 2 V 0 . Y .W ˝ V 0 / ˚ .Y p
1
contains the nilpotent element y a p x.G/ contains N. 2
A
5.G/=N where N is the nilradical of O. Let G be the reduced algebraic group attached to G (see 5.˛/ is (of course) deﬁned over k.
we see that x ˝ y 2 W ˝ V 0 if and only if xW 0 ˝ y D 0. L EMMA 5. / be a Hopf algebra over k. Every prime ideal element of G.me =m2 / D Krull dim. in particular. Choose a decomposition V D W ˚W 0 . e O. For example. and so O.Gk al /red of Gk al is the subgroup of 1 Ga Ga is deﬁned by Y D a p X . The hypothesis implies that me =m2 ! me =m2 N e e N N N N N is an isomorphism of k-vector spaces.G/me have the same N Krull dimension. and so N N dimk . x ˝ y 2 W ˝ V 0 ” x 2 W or y D 0: P ROOF. and x lies in W if and only if xW 0 D 0.G/ D kŒX.O.G/me and O. By homogeneity (5. O. G is smooth.20 An algebraic group G over an algebraically closed ﬁeld k is smooth if every nilpotent element of O. e N P ROOF. Therefore.G/. Therefore me and me have the same height.W 0 ˝ V 0 /.
5j
Algebraic groups in characteristic zero are smooth (Cartier’s theorem)
We ﬁrst prove two lemmas.x. For x 2 V . 2 L EMMA 5. and so O.
aX p is irreducible in kŒX. Y =. Then the subgroup G of Ga Ga deﬁned by Y p D aX p is reduced but not smooth. y
O.G/me /. Then x has a unique decomposition x D xW CxW 0 with xW 2 W and xW 0 2 W 0 . which holds if and only if xW 0 D 0 or y D 0. Then O. although Ker.G/me is regular. let k be such a ﬁeld. but aX p / D k al Œx.11).6).Y p which is an integral domain because Y p aX p /. and let I D Ker. N e N Therefore dimk .˛/red is not. and so the prime ideals of O.O. and let W be a subspace of V . As V ˝ V 0 ' . ˛ Note that G is the kernel of . which is not deﬁned over k (as a subgroup of Ga Ga ).G/m is regular for all maximal ideals m in O. and let e be the neutral N N N D O.

1 ˝ a C y/. . we ﬁnd that an D 0 in A but an 1 ¤ 0 in Am . Thus. /. c 2 A:
From the commutativity of the second diagram in (26).a/. We may replace k with its algebraic closure. Let a be a nilpotent element of A. . .
. let G be an algebraic group over an algebraically closed ﬁeld k of characteristic zero.. . 2.a/ C . 0D .a// 2 k ˚ I . P b ˝ c/ D P b0 ˝ c0
/.a ˝ 1 C 1 ˝ a C y/n . On replacing a with sa. a/n D .b ˝ c 00 b 0 ˝ c 0 / P 0 Á b ˝ c 0 mod I ˝ I .20).a/ D . Algebraic groups and afﬁne algebraic schemes (a) As a k-vector space.9) . (b) For a 2 A. we may suppose that there exists an n 2 such that an D 0 in Am but an 1 ¤ 0 in Am .a/ 2 k and a00 2 I: Let . / ı / .24 (C ARTIER 1962) Every algebraic group over a ﬁeld of characteristic zero is smooth.1 ˝ a/i y j
. and so a 2 m2 .a/ D P b ˝ c.
When expanded. and compose to the identity.
P ROOF. Now san D 0 in A for some s … m. (44) with y 2 m ˝k m. Thus.13). the right hand side becomes a sum of terms an ˝ 1. and therefore in A=m2 by (5. and so (see 5.a/ D a ˝ 1 C 1 ˝ a C y Because is a homomorphism of k-algebras.h C i C j D n. (b) For any a 2 I . i C j
2/:
.a/ a˝1 1˝a D P .mAm /2 .b ˝ c b 0 ˝ c b ˝ c 0 / P 00 D . / ı / .. . write a D a0 C a00 with a0 D .5. P P b0 ˝ c b ˝ c0 in k ˝k A in A ˝k k. and so P b 0 ˝ c 0 D 0 if a 2 I . according to (5.an
1
˝ 1/ .G/. we ﬁnd that 1˝a D a˝1 D Therefore . (a) The maps k ! A ! k are k-linear.a/ D a ˝ 1 C 1 ˝ a
49
mod I ˝ I . A D k ˚ I .
Now (cf.a/ D . /. n. it sufﬁces to show that a lies in m2 .a/ D . P ROOF.a ˝ 1/h . and let A D O. Therefore A D k ˚I and a 2 A decomposes as a D .23) . If a maps to zero in Am . p.
2
T HEOREM 5. Now a 2 m (because A=m D k has no nilpotents). Let m D me D Ker.28.an / D . then it maps to zero in Am =. b.a .

25 Let G be an algebraic group over a ﬁeld of characteristic zero.g. For each a in the nilradical N of O.G/ is a homomorphism of Hopf algebras. which is a contradiction. If G. such that ap D 0. As N is ﬁnitely generated. equation (44) shows that
1
˝ a 2 an
m ˝k A C A ˝k m2
(inside A ˝k A).G/. there exists an exponent. n is a nonzero element of k. we may suppose n n that an D 0 in O. G is reduced. an 1 … an 1 m.G/ but an 1 ¤ 0 in O. and (b) ex D x for all x 2 X (here e is the identity element of G). because if an 1 D an 1 m with m 2 m. and hence it is a unit in A. Therefore n is nonzero in k. there exists a single r that works for all a 2 N. it is reduced. a 2 m2 .G/. x 2 X.G/. g2 2 G.G/ ! O. As it satisﬁes (46).
(45)
As k has characteristic zero.G/:
r
The map f 7! f p W O. which is a contradiction.K/ D f1g for some algebraically closed ﬁeld K.g1 g2 /x D g1 . this would imply an 1 D 0 in Am .
In the quotient A ˝k A=m2 this becomes nan
1
˝ a 2 an N
1
m ˝k A=m2
(inside A ˝k A=m2 ). then .9. and so O.50 As an D 0 and the terms with i C j nan
1
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1) 2 lie in A ˝k m2 .
5k
Transporters
Recall that an action of a monoid G on a set X is a map .G/p D ff p j f 2 O. and so a p D 0. as 1 m is a unit in Am . and so (see 5.24 shows that a 2 m2 .
2
T HEOREM 5. On the other hand.24. x/ 7! gxW G such that (a) .26 An algebraic group G over an algebraically closed ﬁeld k of characteristic p ¤ 0 is smooth if O. as required. and the argument in the proof of Theorem 5. then .G/p is a Hopf subalgebra of O. According to the theorem.G/m . If pjn.G/ has the following property: a 2 O.22). ap D 0 H) a D 0: (46)
P ROOF. P ROOF. 2 C OROLLARY 5. and so we can apply Proposition 5. which we may take to r be a power p r of p.G/g
r r
O. As in the proof of Theorem 5. With this r. let O. 2 R EMARK 5.g2 x/ for all g1 . m D me .1 m/an 1 D 0. X !X
.a p /p D 0. then G is the trivial algebraic group.G/. Let a be a nilpotent element of O. Hence nan 1 … an 1 m.27 Let G be an afﬁne algebraic group over an algebraically closed ﬁeld k of characteristic p ¤ 0.

T
T
Y
Y ' hA
Y
Y . then TG .b/ D 0: Because Z ! Y is a closed immersion. for any kalgebra B and ˇ 2 Y . for any k-algebra B and ˇ 2 Y . 3d). and so the statement
2
L EMMA 5. Algebraic groups and afﬁne algebraic schemes
51
Now let G be an afﬁne monoid over k. Let Z Y be subfunctors of X . Z/ is represented by a quotient of O.A ˝ '/. Let b0 D A˝b c b.30 If Z ! Y is a closed immersion. then TG . and so b0 has the required property. P ROOF. for any map T ! Y of functors.. Then hA is obvious. Z/ ' G
˘B=k Y
˘B=k Z:
2
P ROOF. For example.G/. there exists an ideal b0 B such that. To prove that ˘A=k Z ! ˘A=k Y is a closed immersion we have to show that.'/. so also is T Y Z !T.R/ on the set X.ˇ/ 2 Y .c/ D 0. Z/ of Y into Z is the functor R fg 2 G. let B D O. L EMMA 5.X /. if Y D hB .R0 / for all R-algebras R0 .29 If Z ! Y is a closed immersion.A ˝ B/ if T and only if . then.R/ X. that gY Z as functors on the category of R-algebras.5. Let hA ! T be a map of functors. This follows from the next three lemmas.A ˝ R/ (cf.b0 / D 0.Y.R/ j gY Zg. and let X be a functor from the category of k-algebras to sets.28 Let G X ! X be an action of G on X . The transporter TG . we let ˘A=k X denote the functor R X. This follows directly from the deﬁnitions. there exists an ideal b B such that. ˇR 2 Z.c/ D 0 ” '.A ˝ '/.Y. For a k-algebra A and functor XW AlgA ! Set. i. To say that Z ! Y is a closed immersion means that.R/ for all k-algebras R.
where the condition gY Z means that gY .Y.R0 / Z. then. ˇR 2 Z. for any k-algebra A. for a homomorphism 'W B ! R of k-algebras.A ˝ B/ ” '.R/:
P ROPOSITION 5. so also is ˘A=k Z ! ˘A=k Y .31 In the situation of the proposition.e.R/ ! X.R/ ” '. P ROOF.
2
L EMMA 5.R/ is an action of the monoid G.B/. the ﬁbred k product hA Y Z is represented by a quotient of A. We say that a map Z ! Y in Alg_ is a closed immersion if. If X is representable. for every map hA ! Y . Then . and let Z Y be subfunctors of X such that Z ! Y is a closed immersion.b/ D 0 where ˇR D Y . then Z ! Y is a closed immersion if and only if Z itself is represented by a quotient of A. for a homomorphism 'W B ! R of k-algebras.
. An action of G on X is a natural transformation G X ! X such that G.A ˝ B/. there exists an ideal c 2 A ˝ B such that ˇR 2 Z.

A/ D Homk .H ! G/ D B A.N 0 ! N / is injective (CA 9.n/
is
Bn ˝A . (DG I 2. T HEOREM 5. Therefore Proposition 5. x maps to zero in Mn . and let n be a maximal ideal of B. Let ˛WH ! G be a homomorphism of afﬁne groups such that O. and let a D fb 2 B j bx D 0g.e.52
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
5l
Relaxing the conditions on k and G
In this subsection. when k contains a ﬁeld of characteristic zero. and so a contains an element not in n.N 0 ! N // is injective. and so sends bialgebras and Hopf algebras over k to monoid objects and group objects in the category of afﬁne k-schemes. Let k be a commutative ring.A/ ! F .G/ is free as a k-module. A functor F W Algk ! Set is said to be formally smooth if. Then p D ' a prime ideal in A. Therefore. Algebraic groups over k correspond to ﬁnitely presented afﬁne group schemes over k (which are called algebraic afﬁne group schemes over k). for any k-algebra A and nilpotent ideal n in A.8). Therefore.Spec A.
In the proof. and so it sufﬁces to show that k 0 ˝ B is faithfully ﬂat over k 0 ˝ A (CA 9. Chapter II). A k-scheme X is smooth over k if it is locally def ﬁnitely presented and the functor A X. we shall make use of the following criteria.9). There is the following criterion (SGA1.32).
5m
Appendix: The faithful ﬂatness of Hopf algebras
In this subsection. and Ap ˝A . The functor Spec from k-algebras to afﬁne k-schemes takes ﬁnite direct sums (i. For each maximal ideal n of B. Thus there is an equivalence of categories between afﬁne groups over k and afﬁne group schemes over k (group objects in the category of afﬁne k-schemes). Hence a D B.N 0 ! N / ' Bn ˝Ap . we prove a very important technical result (Theorem 5. Cartier’s theorem (5. tensor products) to ﬁnite products. the reader is assumed to be familiar with the basic theory of schemes (Hartshorne 1977.32 For any Hopf algebras A B over a ﬁeld k.30. II): a ﬁnitely presented morphism is smooth if it is ﬂat and its geometric ﬁbres are nonsingular algebraic varieties.29.
N 0 ! N be an injective homomorphism of A-modules.
11 Let 1 . Therefore we may suppose that k is algebraically closed. X / is formally smooth.Ap ˝A .11 (b) A ﬂat homomorphism 'W A ! B is faithfully ﬂat if every maximal ideal of A is of the form ' 1 . and so x D 0. For any ﬁeld k 0 k.A=n/ is surjective. and 5.n/ ! Bn is ﬂat for all maximal ideals n in B.24) shows that every ﬂat algebraic afﬁne group scheme over k is smooth. the map F . the homomorphism A ! k 0 ˝ A is faithfully ﬂat.n/ for some maximal ideal n of B (CA 9.N 0 ! N / has the property that Mn D 0.4). Let x 2 M .28 is true provided O. 5.. The lemmas 5.
.31 remains true with k a commutative ring provided A is free as a k-module. 7). and so the kernel M of B ˝A . B is faithfully ﬂat over A. (a) A homomorphism 'W A ! B is ﬂat if A' 1 .

the existence of commutative diagrams H ? ? y G
Lb
! H ? ? y
O. the map b ˝ b 7! Œb ˝ bW B ˝A B ! .G ! G 0 / H ' G
G0 G0 ˛
(49)
! G 0.G/ ! O. Thus U D jG ı j. More generally.˛/ H ! H
G
H
(47)
is an bijection — this just says that two elements of H with the same image in G differ by an element of the kernel. To show that O.
E LEMENTARY DIGRESSION
For any homomorphism ˛W H ! G of abstract groups. implies that it sufﬁces to prove this for a single b.9) to ﬁnd an open set of b’s for which O. According to the above criterion.11. its identity component G ı is irreducible.
In this case. CA 16.H /mb . more precisely. h/ 7! . the image of jH j ! jGj is dense in jGj. for each b 2 H.H /me x ? ? O. the sets U 1 and Ug 1 have nonempty intersection. this implies that.h/. This means that there exist u. and we can apply the generic ﬂatness (or freeness) theorem (CA 9.G/ma ! O.b/.H / is faithfully ﬂat.13).G/ ! O.nh. the map . This implies a similar statement for afﬁne groups: Ker.5. and the translates of U by points in the image cover jGj.G/ ! O. it sufﬁces to show that.H / is ﬂat.G/ma . H
H:
(50)
. there is an isomorphism Ker. For any g 2 jG ı j.B=IA B/ ˝k B is an isomorphism (IA is the augmentation ideal of A). h/ 7! . h/W Ker. it sufﬁces to show that the map (of sets) jH j ! jGj is surjective.H /mb is faithfully ﬂat over O. Because O. the map O.G/ becomes an integral domain.H /mb x ? ? O. Because of the correspondence between afﬁne groups and Hopf algebras. jGj is a smooth algebraic variety. the ring O.G/ma
La
! G
'
(see 5f). the image of jH j ! jGj contains a nonempty open subset U of jG ı j.˛/ H ! H
G
H
(48)
which becomes the map (47) for each k-algebra R. Similarly.n ˇ. v 2 U such that u 1 D vg 1 .k/. Homogeneity. in particular. Algebraic groups and afﬁne algebraic schemes
53
C ASE THAT A IS REDUCED AND A AND B
ARE FINITELY GENERATED .n. to show that 'W O. for any homomorphism A ! B of Hopf algebras. Because G is smooth. and so g D vu 2 U . for any homomorphisms of abstract groups H ? ? yˇ G the map . is ﬂat. When we replace G with its identity connected component. for any homomorphism ˛W H ! G of afﬁne groups. a D '. According to (CA 12.G/me
'
O. it meets every connected component of jGj. h/W Ker.˛/ H ! G is a bijection.H / is injective.n.

J / D M ˚ N (direct sum of B-submodules).aj /j 2J 7! j aj ej W A.J / maps to zero in B ˝A B.J / . we may suppose that k has characteristic p ¤ 0.G/. and B . According to (5. and so M IM .M / O. there n exists an n such that O.J /
'
!
B=IB ? ? y
onto
! B ˝A B
'
! . which implies that it maps to zero in .B=IB/ ˝k . We have shown that B . and consider the diagrams 1 ! N ? ? y ! M ! H ? ? y ! G ! G0 O.J / ! B is injective because A. B ˝k B=IB is free as a left B-module. Hence C D IC . Hence M D IM D I 2 M D D 0. consider diagrams A.J / mapping isomorphically onto B ˝A B.A=I /.J / D IM ˚ I N.G/p is a reduced Hopf subalgebra of O.G 0 /
.e.J / .IB/.B=IB/. say I n D 0. A diagram chase in A.. lying in IB.B=IB/.J / ! B ? ? y ! C ! 0
onto
! B=IB
shows that every element of C is the image of an element of B mapping to zero in B=IB.J / ! B ˝A B is contained in .J / is a direct sum of copies of A indexed by J . and so C D IC D I 2 C D D I n C D 0.24.J / . Therefore the kernel M of B . If b 2 B .N / x ? ? O.B=IB/
in which the bottom arrows are obtained from the top arrows by tensoring on the left with B and B=IB respectively.J / D I B .
C ASE THAT A IS FINITELY GENERATED
After Theorem 5.G 0 / D O. As B=IB is free as a k-module (k is a ﬁeld). Therefore there exists a B-submodule N of B .J / ! B (51) where A. We know that M I B .G/
faithfully
O.ej /j 2J of elements in B whose image in B=IB is a k-basis and consider the map P .J / ? ? y . For the injectivity. and this implies that A.J / ? ? y .54
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
C ASE THAT THE AUGMENTATION IDEAL OF A IS NILPOTENT
Assume that the augmentation ideal I D Ker. Let C be the cokernel of (51).J / B . and so B ˝A B is free (hence projective) as a left B-module. W A ! k/ of A is nilpotent. Let G 0 be the algebraic group n such that O.27).H / ﬂat x ?injective ? O. i. Choose a family .J / ! B is surjective.J / ! B ˝A B is injective.G/p .J / ? ? y B .J /
onto
!
B ? ? y
k .G 0 /
1
! G0
O. Recall (49) that B ˝A B ' B ˝k B=IB. Hence A. We shall show that (51) is an isomorphism (hence B is even free as an A-module). then it maps to zero in B=IB ˝k B=IB.

G 0 / O. C OROLLARY 5.H // ˝O. the proof given here follows Oort 1966. and give the elementary Hopf algebra proof in detail N OTES In most of the literature.H / is faithfully ﬂat (CA 9. As O.M / ! O. In extending the proof to all algebraic groups.G 0 / k is injective. Theorem 5.H / injective H) .N / x ? ? O. From the diagram O. Therefore. if a=c 2 B.O. The approach through functors can be found in Demazure and Gabriel 1970 and Waterhouse 1979. and an elementary but uninformative proof using Hopf algebras.N / ˝ O. the proof given here follows Waterhouse 1979.G/ ! O.G/ ! O.H // ˝O. In general. Chapter 14.G 0 / k
(20)
'
we see that O.H / ' O. 2
5n
Terminology
From now on “group scheme” and “algebraic group scheme” will mean “afﬁne group scheme” and “afﬁne algebraic group scheme”. and Springer 1998. cB \ A D cA for any c 2 A.H / is faithfully ﬂat.M / is nilpotent). L be their
P ROOF. and hence is faithfully ﬂat (because the augmentation ideal of O. The important Theorem 5.G/ ˝O. in particular.H / is faithfully ﬂat. a.G 0 / O.G 0 / O. we sketch the easy geometric proof for smooth algebraic groups.N / is injective.H // ˝O.5.G/ ! O.H /:
we see that .G 0 / O.G 0 / is reduced. and so O. this implies that O. Borel 1991. this implies that . and so a=c 2 A.G 0 / k x ? ? O.32 is proved entirely in the context of Hopf algebras in Takeuchi 1972. Because O. Humphreys 1975.H / ' O. Then B \ K D A.33 Let A B be Hopf algebras with B an integral domain.G/ ! O.G/ ˝O. “algebraic group” means “smooth algebraic group” in our sense.
As k is a direct summand of O.G 0 / ! O.M / ˝ O.H / ˝O.G/ ! O.O.M /
(20)
'
O.4).
A SIDE 5.23) below that A and B are directed unions of ﬁnitely generated Hopf subalgebras Ai and Bi such that Ai Bi .H / is faithfully ﬂat.H / injective. c 2 A. Algebraic groups and afﬁne algebraic schemes
55
where N and M are the kernels of the homomorphisms H ! G 0 and G ! G 0 respectively.24 was announced in a footnote to Cartier 1962.H / ˝O. for example.
G ENERAL CASE
We show in (7.
. From the diagrams N ?H ? y M H
(48)
'
H ?0H G ? y G
G0
(50)
'
H
O.34 Some statements have easy geometric proofs for smooth algebraic groups.H /. Because B is faithfully ﬂat over A.G 0 / ! O. one often has to make a choice between a nonelementary (sometimes difﬁcult) proof using algebraic geometry. A D B if K D L. Our deﬁnition of “algebraic group” is equivalent to “afﬁne group scheme algebraic over a ﬁeld”. then a 2 cB \ A D cA.O.H / x x ? ? ? ? O. and so the general case follows from the ﬁnitely generated case. and let K ﬁelds of fractions. the homomorphism O.

Now.H / implies that the two homomorphisms (52) are equal.H / ˝O. normal subgroup) of G.H / is surjective.R/ is a subset of G.G/ ! O.G/ O.32).R/ is injective for all k-algebras R.O. and so it is an isomorphism.H / is injective.H / In particular.R/ for all k-algebras R.G/ ! O. k is a ﬁeld. and ˘ condition (b) implies that O.G/ O. and so the subset of O.H // mapping to the same element in G.R/ ! G.G/ O. For the sufﬁciency. An injective homomorphism is also called an embedding.H / on which the two homomorphisms (52) agree is ˛ .H / ˝O. In other words. P ROOF. this means that H.
12 In
particular.6
Group theory: subgroups and quotient groups.3 A homomorphism ˛W H ! G of afﬁne groups is injective if and only if the map ˛ W O. and (b) the homomorphism ˛ W O.
2
6b
Subgroups. Throughout.H //.
In this section and in Section 8.H / (52) O. we ﬁnd that that ˛ is surjective. The conditions are obviously necessary.H /:
agree on O.R/W H. ˘ condition (a) with R D O.G/ ˝. injective homomorphisms
D EFINITION 6.O. the homomorphisms x 7! x ˝ 1 x 7! 1 ˝ x W O.R/ ! G.6).H / ˝O.
6a
A criterion to be an isomorphism
P ROPOSITION 6.O. (b) A subgroup (resp.G/ ! O. On combining these statements. P ROPOSITION 6.G/ O. and so deﬁne elements of H. normal subgroup) of an afﬁne group G is a subfunctor12 H of G such that H.H / ˝O.R/ for all R.G// by (CA 9.H / is a faithfully ﬂat O. ˛W H ! G is injective if and only if j˛jW jH j ! jGj is a closed immersion. note that the maps H
G
H
H !G
˛
give rise to homomorphisms of Hopf algebras O.R/ is a subgroup (resp.
56
.R/ is injective for all k-algebras R.G/-algebra (see 5.H / ! O.2 (a) A homomorphism H ! G of afﬁne groups is injective if the map H.G/.H / ˝O. we show how the basic deﬁnitions and theorems in the theory of abstract groups can be extended to afﬁne groups.1 A homomorphism of afﬁne groups ˛W H ! G is an isomorphism if and only if (a) the map ˛.

): The homomorphism ˛ factors into homomorphisms of Hopf algebras O. Gred is a subgroup of G (see 5. then Gred D .h/ D 1 for all h 2 H. Conversely. However.R/ is injective for all k-algebras R.O. so also is O.G/ is ﬁnitely generated as a k-algebra.H / and a 7! G. 2 P ROPOSITION 6.
. the map O. it need not be normal.G// D O. if a is a Hopf ideal in G. Because O.G/.G/ ! O. then any two homomorphisms O.11). Group theory: subgroups and quotient groups. If ˛ is injective.1 shows that the map H ! H 0 is an isomorphism.G/ ! k 0 ˝k O.2).R/ and all Rg is a Hopf ideal in G (it is the kernel of O.Z=2Z/k .H / ! R that become equal when composed with ˛ must already be equal. I.H 0 / ! O. then O. Then ˛ factors into H ! H 0 ! G.5 Let ˛W H ! G be a homomorphism of afﬁne groups. and so ˛ .H / is injective. P ROOF.G//. over a ﬁeld k of characteristic 3.a/ are inverse.Z=2Z/k on 3 .H / is surjective if and only if the map k 0 ˝k O.G/ ! O. P ROPOSITION 6.! O. then so also is ˛k 0 W Hk 0 ! Gk 0 for any ﬁeld k 0 containing k.4 A subgroup of an algebraic group is an algebraic group.R/ ! G.8 Every set of algebraic subgroups of an algebraic group G has a minimal element (therefore every descending chain of algebraic subgroups becomes stationary). and so. P ROOF.H /. For example. then the functor R fg 2 G.a/ of O.H / is surjective.Z=2Z/k for the nontrivial action of .6. then ˛ is injective. (: If ˛ is surjective.H /. 2 C OROLLARY 6. let G D 3 . P ROOF.H /. and the injectivity of ˛ implies that H. For a subgroup H of G. Let H 0 be the afﬁne group whose Hopf algebra is ˛ .G// .O.H /
57
(see Exercise 4-10). which is not normal in G (see SGA3 VIA 0. If H ! G is an embedding.G/ (it is represented by O. 2 6. if O. see Exercise 4-10).g/ D 0 for all f 2 ag
A
is an afﬁne subgroup G.O. The maps H 7! I.7 The subgroups of an afﬁne group G are in natural one-to-one correspondence with the Hopf ideals on O. and so H.H / D ff 2 O.G/ ! O.G/=a).R/ is injective.6 When k is a perfect ﬁeld.G/ ˛ . P ROOF.H / is surjective (this is simply a statement about vector spaces over ﬁelds). if ˛k 0 is injective for one ﬁeld k 0 containing k.R/ ! H 0 . 2 C OROLLARY 6. Proposition 6. Conversely.G/ j fR .R/ j fR . For any ﬁeld k 0 containing k.

Hj /g: Therefore.R/ lies in N. : : : .Z/k of Ga over Q is neither closed nor afﬁne.R/ D Ker. The ring O. Then an element hW O.a/ D fm j m
def
ag of jGj :2
P ROPOSITION 6. T
j 2J
\
j 2J
Hj .
P ROPOSITION 6. with coordinate ring O.9 For any algebraic subgroup H of an algebraic group G.12 In the realm of (not necessarily afﬁne) group schemes over a ﬁeld. then jH j is the subspace V .Hj /j 2J of subgroups of an afﬁne group G. there can exist nonafﬁne (necessarily nonclosed) subgroup schemes of an algebraic group.H /
h ˛
O.R/ j fR .g/ D 0 for all f 2 D Hom.R//.R/W H.G/ ! k be the identity element of G.R/ j fR .
def
Let W O.c.6). R/: [ I.G/ is noetherian (Hilbert basis theorem. CA 3.
6c
Kernels of homomorphisms
The kernel of a homomorphism ˛W H ! G of afﬁne groups is the functor R N. c/ with c n D 1).H /.
A
6.G/ ! O. the algebraic scheme jH j is closed in jGj.H / factors through : O.R/)
Hj of G.R/ D fg 2 G.Hj /.10 For any family .R/ D fg 2 G. We sometimes loosely refer to an injective homomorphism ˛W H ! G as a subgroup of G.g/ D 0 for all f 2 I. For example.58
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
2
P ROOF. the functor R is a subgroup ideals I. We have Hj . If a is the kernel of O. P ROOF.k/.G/=I.G/ ! O.R/ if and only if its composite with ˛ W O.11 The intersection of the algebraic subgroups SLn and Gm (scalar matrices) of GLn is n (matrices diag. the constant subgroup scheme .O. H.H / ! R of H.Hj /g
2
E XAMPLE 6.R/ ! G.˛.R/ (intersection inside G.G/
R
k:
.G/=I where I is the ideal generated by the
P ROOF.

G/ ! k (this is called the augmentation ideal). the augmentation ideal for Gm is .R/ D Ker. the kernel of det is the algebraic group SLn .H / ˝O. : : : : .Xij /W kŒX.15 Let N be the kernel of the determinant map detW GLn ! Gm .14 Consider the map g 7! g n W Gm ! Gm .O.det.aij //: As we just noted. X 1 because X n .e. Xij .X n
n.H / (see 3b). det.
1/:
In other words.H / ! R that are zero on IG O.Xij /.
.16 When k has characteristic zero. Group theory: subgroups and quotient groups. a homomorphism G ! H is injective if and only if G. Then the elements of N. This corresponds to the map on Hopf algebras X 7! det.det.aij / D det. Alternatively.X O.R// for all R. and so it is an
=.
59
Let IG be the kernel of W O.1/.
E XAMPLE 6.H.Xij / . R/: We have proved: P ROPOSITION 6. P ROPOSITION 6. as we would expect.Xij /
1
'
kŒ: : : .14).G/=IG / ' O.H / denote the ideal generated by its image in O.H /=IG O.Y 1/.g n / (cf. N. the kernel has coordinate ring kŒX. X because det.13 For any homomorphism H ! G of afﬁne groups.6.H /.aij / D X. p. and so
1 1
! kŒ: : : .det.Xij / 1/ 1/. Thus. Xij .X n
1/ ' kŒX =.O.Xij / 1/
In other words. This corresponds to the map on Hopf algebras Y 7! X n W kŒY.H /. Y 1 ! k sends f . (8). E XAMPLE 6. i. the kernel is the algebraic group
as we would expect.N / D kŒ: : : . det. : : : .H /. there is an afﬁne subgroup N of H (called the kernel of the homomorphism) such that N. Xij .R/ correspond to the homomorphisms O. Y 1 ! kŒX. note that the kernel of ˛ is the ﬁbred product of H ! G algebraic group with coordinate ring O. X
1
f1G g. and so the augmentation ideal for Gm is . its coordinate ring is O. and let IG O.G/ .R/ ! G.H /..R/ D Homk-alg .Y / to f .H /=IG O. The map W kŒY. : : : .k al / is injective.k al / ! H.g/ D g n D Y .H /=IG O.

k/. n 7! nW N ! N are distinct. so that its kernel N is nontrivial. Then the homomorphisms n 7! 1.H 0 / shows that ˇ D . p. the kernel N of the homomorphism has the property that N.G/=m D k. For example. then (6.R/ is injective for every reduced k-algebra R.R/ ! Ga . in each equivalence class of R monomorphisms with target G.G/ ? ? y R.
A SIDE 6. there is a commutative diagram O. S / ! Hom.3a) with R D O. A morphism uW S ! A is a monomorphism if f 7! u ı f W Hom. van Oosten.k/ is a homomorphism O. P ROOF. Let G be an afﬁne group. With the notations of Exercise 4-1. P ROPOSITION 6. and an equivalence class of monomorphisms is called a subobject of A.T. R EMARK 6. on the other hand. every monomorphism is regular (see. If D H.G/). let ˛W H ! G be a homomorphism of afﬁne groups. and let be a subgroup of G.k/ for an algebraic subgroup H of G.21). As we discussed 5e. For example. whose kernel we denote ma (a maximal ideal in O.uR / D Im. but the map x 7! x p W Ga . Is it true that every monomorphism in the category of afﬁne (or algebraic) groups is regular?
A
6d
Dense subgroups
Let G be an algebraic group over a ﬁeld k.9). A monomorphism that arises in this way is said to be regular. A homomorphism of afﬁne groups is a injective if and only if it is a monomorphism in the category of afﬁne groups. Basic Category Theory.
!
. a point a 2 G.G/ ! k. but they agree when composed with ˛.k al / ! H. There exists an algebraic subgroup H of G such that H. /.17 Proposition 6.k/ D if and only if is closed. and so ˛ is not a monomorphism.G/ ? ? y R. Suppose. then D jH j \ G.k al / is injective. for example. that ˛ is not injective.G/ such that O. In Grp. x 7! axa 1 W A ! A. every smooth algebraic subgroup of G arises in this way. Two monomorphisms uW S ! A and u0 W S 0 ! A are said to be equivalent if each factors through the other. A/ is injective for all objects T .16 is false for ﬁelds k of characteristic p ¤ 0. Exercise 42. the map a 7! ma W G. This is an equivalence relation on the monomorphisms with target A. It follows that.k al / D 0. When k is algebraically closed. By deﬁnition.19 In any category.u0 / for all k-algebras R. and so it is the trivial algebraic group (by 5. in which case there exists a unique smallest H with this property. the centralizer of an element a of a group A (which is not a normal subgroup in general) is the equalizer of the homomorphisms x 7! x.T. To see this.18 Let A be an object of some category A.G/ ˝ O. the homomorphism x 7! x p W Ga ! Ga has kernel ˛p . 2 6.k/. there is exactly one with H an afﬁne subgroup of G and with u the inclusion map.60
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROOF. Two monomorphisms uW H ! G and uW H 0 ! G are equivalent if and only if Im. W H 0 ! H agree when composed with ˛. and so is not injective.25).k/ with the subspace topology. /
G
! O. We endow G. If G. the equalizer of a pair of morphisms is a monomorphism. which is closed by (6.k/ ! jGj is injective with image the set of maximal ideals m of O. If ˛ is injective and the homomorphisms ˇ.20 Let G be an algebraic group over a ﬁeld k.

6. Group theory: subgroups and quotient groups.

61

which shows that maps into R. / ˝ R. /, and so .R. /; / is a Hopf algebra (ibid.). If def is closed, then it is the zero set of a D Ker.O.G/ ! R. //, and the subgroup H of G with O.H / D O.G/=a ' R. / has H.k/ D ; clearly, it is the smallest subgroup of G with this property. When k is algebraically closed and H is a smooth subgroup of G, then the group attached to D H.k/ is H itself. 2 R EMARK 6.21 When k is not algebraically closed, then not every smooth algebraic subgroup of G arises from an closed subgroup of G.k/. Consider, for example, the algebraic subgroup n Gm over Q. If n is odd, then n .Q/ D f1g, and the algebraic group attached to f1g is the trivial group. D EFINITION 6.22 Let G be an algebraic group over a ﬁeld k, and let k 0 be a ﬁeld containing k. We say that a subgroup of G.k 0 / is dense in G if the only algebraic subgroup H of G such that H.k 0 / is G itself. 6.23 It follows from the proof of (6.20) that G.k/ is dense in G if and only if f 2 O.G/, f .P / D 0 for all P 2 G.k/ H) f D 0: (53)

Equivalently, G.k/ is dense in G if noT nonzero element of O.G/ maps to zero under all homomorphisms of k-algebras O.G/ ! k, i.e., Kerk-alg .O.G/; k/ D 0. 6.24 If G.k/ is dense in G, then G is reduced (hence smooth if k is perfect). 6.25 If G is smooth, then G.k al / is dense in G by the Nullstellensatz (CA 11.5); in fact, G.k sep / is dense in G (AG 11.15). [In detail, assume that G is smooth, and let H be a subgroup of G. By deﬁnition, this means that O.H / D O.G/=a for some ideal a in O.G/. The elements of H.k sep / are the k-algebra homomorphisms O.H / ! k sep . To say that H.k sep / D G.k sep / means that every k-algebra homomorphism O.G/ ! k sep factors through O.G/=a, but this implies that the elements of a lie in every maximal ideal m of O.G/ such that O.G/=m is separable over k, and hence they are zero by (AG 11.15).] 6.26 If G.k/ is ﬁnite, for example, if the ﬁeld k is ﬁnite, and dim G > 0, then G.k/ is never dense in G. P ROPOSITION 6.27 If k is inﬁnite, then G.k/ is dense in G when G D Ga , GLn , or SLn . P ROOF. We use the criteria in (6.23). Recall (e.g., FT, proof of 5.18) that, because k is inﬁnite, no nonzero polynomial in kŒX1 ; : : : ; Xn can vanish on all of k n . This implies that no nonzero polynomial f can vanish on a set of the form D.h/ D fa 2 k n j h.a/ ¤ 0g; because otherwise hf would vanish on k n . The proposition for GLn follows from this, because GLn .k/ D fa 2 k n j det.a/ ¤ 0g:

2

def

h ¤ 0;

62

BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)

The proposition is obvious for Ga , and it can be proved for SLn by realizing O.SLn / as a subalgebra of O.GLn /. Speciﬁcally, the natural bijection A; r 7! A diag.r; 1; : : : ; 1/W SLn .R/ Gm .R/ ! GLn .R/ (of set-valued functors) deﬁnes an isomorphism of k-algebras O.GLn / ' O.SLn / ˝ O.Gm /; and the algebra on the right contains O.SLn /.

2

P ROPOSITION 6.28 Let G be an algebraic group over a perfect ﬁeld k, and let D Gal.k al =k/. Then acts on G.k al /, and H $ H.k al / is a one-to-one correspondence between the smooth subgroups of G and the Zariski-closed subgroups of G.k al / stable under . P ROOF. Combine (6.20) with (4.16). (More directly, both correspond to radical ideals a in the k al -bialgebra k al ˝ O.G/ stable under the action of ; see AG 16.7, 16.8.) 2

A SIDE 6.29 (a) Let G be a smooth connected algebraic group over an inﬁnite ﬁeld k. It is known that, if either k is perfect or G is reductive, then G.k/ is dense in G (Borel 1991, 18.3, p.220). However, there exists a one-dimensional (smooth connected) unipotent group U over an inﬁnite ﬁeld k such that U.k/ is ﬁnite and hence not dense (Rosenlicht 1957, p.46). (b) If G is unirational and k is inﬁnite, then G.k/ is dense in G. Borel 1991, 18.2, in fact proves that if k is perfect and G is reduced, or G is perfect, then G is unirational over k. In SGA3, IV 6.11 one ﬁnds: One knows (Rosenlicht) examples of forms of Ga over a nonperfect ﬁeld, which have only ﬁnitely many rational points, and therefore a fortiori are not unirational. Moreover Chevalley has given an example of a torus over a ﬁeld of characteristic zero which is not a rational variety. On the other hand, it follows from the Chevalley’s theory of semisimple groups that over an algebraically closed ﬁeld, every smooth connected afﬁne algebraic group is a rational variety. For a nonunirational nonconnected algebraic group, Rosenlicht gives the example of the group of matrices a b over R with a2 C b 2 D ˙1. For a nonunirational connected algebraic group, Rosenlicht gives the b a example of the subgroup of pa Ga deﬁned by Y p Y D tX p over the ﬁeld k D k0 .t / (t transcendental). p G On the other hand, if kŒ a; b has degree 4 over k, then the norm torus associated with this extension is a three-dimensional torus that is not a rational variety. [To be rewritten.] A SIDE 6.30 Before Borel introduced algebraic geometry into the theory of algebraic groups in a more systematic way, Chevalley deﬁned algebraic groups to be closed subsets of k n endowed with a group structure deﬁned by polynomial maps. Hence, effectively he studied smooth algebraic groups G with the property that G.k/ is dense in G.

6e

Normalizers; centralizers; centres

For a subgroup H of an abstract group G, we let NG .H / (resp. CG .H /) denote the normalizer (resp. centralizer) of H in G, and we let Z.G/ denote the centre of G. In this subsection, we extend these notions to a subgroup H of an afﬁne group G over k. For g 2 G.R/, let g H be the functor of R-algebras R0 g H.R0 / g

1

(subset of G.R0 /):

Deﬁne N to be the functor of k-algebras R fg 2 G.R/ j g H D H g:

**6. Group theory: subgroups and quotient groups. Thus, for any k-algebra R, N.R/ D fg 2 G.R/ j g H.R0 / g 1 D H.R0 / for all R-algebras R0 g \ D G.R/ \ NG.R0 / .H.R0 //: 0
**

R

63

P ROPOSITION 6.31 The functor N is an afﬁne subgroup of G. P ROOF. Clearly N.R/ is a subgroup of G.R/, and so it remains to show that N is representable. Clearly g H.R0 / g

1

D H.R0 / ” g H.R0 / g

1

H.R0 / and g

1

H.R0 / g

H.R0 /;

**and so, when we let G act on itself by conjugation, N D TG .H; H / \ TG .H; H /
**

1

(notations as in 5k). Proposition 5.28 shows that TG .H; H / is representable, and it follows from (6.10) that N is representable. 2 The subgroup N of G is called the normalizer NG .H / of H in G. It is obvious from its deﬁnition that the formation of NG .H / commutes with extension of the base ﬁeld, i.e., for any ﬁeld k 0 k, .NG .H //k 0 ' NGk0 .Hk 0 /. C OROLLARY 6.32 Let G be an algebraic group, and let k 0 k be such that G.k 0 / is dense in G. A subgroup H of G is normal in G if H.k 0 / is normal in G.k 0 /. P ROOF. The condition says that NG .H /.k 0 / D G.k 0 /, and so NG .H / D G. P ROPOSITION 6.33 If H is algebraic and H.k 0 / is dense in H for some ﬁeld k 0 NG .k/ D G.k/ \ NG.k 0 / .H.k 0 //: P ROOF. Let g 2 G.k/ \ NG.k 0 / .H.k 0 //. Because g 2 G.k/, g H is an algebraic subgroup of G, and so g H \ H is an algebraic subgroup of H . Because g 2 NG.k 0 / .H.k 0 //,

g 2

k, then

H .k 0 / D H.k 0 /;

(54)

and so .g H \ H / .k 0 / D H.k 0 /. As H.k 0 / is dense in H , this implies that g H \ H D H , and so gH D H. 2 P ROPOSITION 6.34 Let H be a subgroup of an afﬁne group G. For any ﬁeld k 0 containing k; NGk0 .Hk 0 / ' NG .H /k 0 CGk0 .Hk 0 / ' CG .H /k 0 : In other words, the formation of normalizers and centralizers commutes with extension of the base ﬁeld. P ROOF. It follows directly from the deﬁnitions that the statements are true for the functors; therefore they are true for the afﬁne groups. 2

37 If H is algebraic and H. It is an afﬁne subgroup of G. 2 The centre Z.H.k 0 /. When we let G act on itself by conjugation. it is isomorphic to Ga Gm . v/.H / to be nontrivial but for CG .64
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1) Let N be the normalizer of H . uv/.b.H /.H /. The centre of SLp is p . Then n 2 NG . and if G is algebraic and G. u/. 2 P ROPOSITION 6. The condition says that CG .R0 / ! H.k 0 / .H / and the automorphisms in and idH of H agree on an algebraic subgroup Z of H .R/ \ \
R0
CG.b.k/: This is an algebraic group because.
2
The subgroup C of G is called the centralizer CG . then Z.H /.R0 //:
P ROPOSITION 6.k 0 / .k 0 //.k 0 / D H.k 0 / to be trivial for all ﬁelds k 0 k.H / need not be smooth.h.G/.25). and we deﬁne C to be the functor of k-algebras R fn 2 N. Therefore Z D H .G/. \ CD TG .k/ \ Z.k al / is dense in G (6. v/ D . then
A
6. the centre of G is p .R/ D G. To see this. as a functor to sets. v/ if and only if up D 1.a C bup .k/ D G. u/ for all .10 imply that C is representable. u/ 2 R R .H / commutes with extension of the base ﬁeld. v/ D . C OROLLARY 6.G/ of an afﬁne group G is deﬁned to be CG . and we know that G.G.k/ D G.36 A subgroup H of a smooth algebraic group G is central in G if H.28 and 6. P ROOF. let G be the functor R R R with the multiplication . Let n 2 G.k 0 / D 1 for all ﬁelds k 0 containing k. which has the property that Z.G/.H /.k 0 //: k. C. .k al /. For a pair . here 0 ¤ p D char. Another example is provided by SLp over a ﬁeld of characteristic p. Therefore.k al /. Each n 2 N.a.R0 /
of H regarded as a functor from the category of R-algebras to sets.k al / D G. it is possible for CG .k 0 //: P ROOF.H / of H in G.
.
h2jH j
and so Propositions 5.R0 / . which is not smooth.H. and so in 2 CG .38 Even when G and H are smooth.a.k 0 / is dense in G.R/ j in D idH g: Thus. and so Z.k 0 / is dense in H for some ﬁeld k 0 CG .k/ \ CG. CG .k/ \ CG. It is obvious from its deﬁnition that the formation of CG .b. u/.R/ deﬁnes a natural transformation in h 7! nhn
1
W H.a.35 The functor C is an afﬁne subgroup of G. h/.k al / is central in G. P ROOF.b.a.H. For example.

If G ! Q is surjective.R/
9g
/
O
q:
_
In other words. Thus HN ' . E XAMPLE 6.
65
For N sufﬁciently large CG . let HN be the subgroup R of G. if H NG . equivalently.Q/ map to the same element of Q.23)).G/ is injective.k/ .R/ normalizes (resp.N /).k/ D 1 and CG.k/ is surjective whenever k is algebraically closed.G/ ? ?g y R0
faithfully ﬂat
O.
6f
Quotient groups. D EFINITION 6. Then HN . it would say that x 7! x n W Gm ! Gm is not surjective even though x 7! x n W Gm . For an integer N .a1 . i.41 A homomorphism G ! Q is surjective if and only if O. P ROOF.k/ ¤ Dn .6. Then HN .Q/ such that g and idO.Q//. For example.k/: An afﬁne subgroup H is said to normalize (resp.O.R0 / with R0 faithfully ﬂat over O. such that the diagram O. a homomorphism G ! Q is surjective if every q 2 Q.R0 / O / Q. Group theory: subgroups and quotient groups.k/ D CG .Q/ ! O.k/: (b) k is algebraic closed of characteristic p ¤ 0 and N is a power of p.HN .R/ lifts to G after a faithfully ﬂat extension. centralizes) N.40 A homomorphism G ! Q is said to be surjective (and Q is called a quotient of G) if for every k-algebra R and q 2 Q. A surjective homomorphism is also called a quotient map.k/ D CG . an / 2 GLn . centralize) an afﬁne subgroup N of G if H. there exists a g 2 G.k/ . surjective homomorphisms
What does it mean for a homomorphism of algebraic groups G ! Q to be surjective? One might guess that it means that G.Q/ ? ?id y O. : : : .R0 / mapping to the image of q in Q. T HEOREM 6. N D an D 1g.k/ ! n Gm .R0 /: G. Gm ! Gm is surjective according to the following deﬁnition.N / (resp.R/ ! Q. and CG.HN .Q/ 2 Q.R/ is surjective for all R. H CG .k// D GLn .HN / D Dn
(group of diagonal matrices) (see (13..HN /. but this condition is too stringent.k/ D f1g.Q/ O.R/.R/
/ Q. there exists a faithfully ﬂat R-algebra R0 and a g 2 G.k// D GLn .HN /.39 Let G D GLn over a ﬁeld k.Q/
. (a) k D Q and N odd.
N HN .R/ j a1 D n N/ .R/ D fdiag.R/ for all k-algebras R. In fact.R0 /. We consider two cases. ): Consider the “universal” element idO.R0 / O G.k/ ¤ Dn .e.

11) that an element f of O.2).32) O.R/ k al Q.Q/ is a family .Q/ ! R.k al / is surjective. then so also is ˛k 0 W Hk 0 ! Gk 0 for any ﬁeld k 0 containing k. If ˛ is surjective.44 Let G ! Q be a homomorphism of algebraic groups.k al /
g _ q
/ qR
/ gk al _ /q
For the converse. the converse is true if Q is smooth. If G ! Q is a quotient map.G/ is injective. As G.6).R/ lifts to an element g of G. The map O.Q/ n n nnn 0n
q
(55)
? _R
Then R0 is a faithfully ﬂat R-algebra because O. The map on coordinate rings is both surjective and injective. we may suppose that k is algebraically closed.k al / / Q.G/ ! k 0 ˝k O.G/ is faithfully ﬂat.R0 / maps to the image q 0 of q in Q.k/ ! Q. Conversely. a homomorphism ˛W G ! H of algebraic groups is surjective if.9 below).Q/ ! O.k al / ! Q.66
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
commutes. P ROOF.8). (: According to (5.k al /:
id
k al
/R
/(
G.G/ o
gD1˝q faithfully ﬂat
R0 D O. The commutativity of the square in (55) means that g 2 G.fR /R with fR a map Q.G/ ! O.Q/ R: O.Q/ R o
nnq vnnn
? _ O. Let q 2 Q. 2 P ROPOSITION 6.G/ is a faithfully ﬂat O.R/
/ G. is injective (CA 9. and so O. then G. Because k ! k 0 is faithfully ﬂat.G/ ˝O. Regard q as a homomorphism O.k/ ! Q.H /-algebra (apply CA 9.43 A homomorphism of afﬁne groups that is both injective and surjective is an isomorphism. Because Q is smooth. O. P ROOF. Let q 2 Q. being faithfully ﬂat.H / is injective (see CA 9.G/ ˝O. for some ﬁeld containing k.Q/ ! O. which implies that O.R/ ! R. and hence is an isomorphism. and form the tensor product R0 D O.7).Q/ is reduced.Q/ ! O. 2 P ROPOSITION 6. the image of G. the image of q in Q.k al /.k 0 / in H.R/.H / is injective if and only if k 0 ˝k O. and so f is determined by fk (CA 11.R0 /.
. 2 P ROPOSITION 6.
fk 2
k0
More generally. For some ﬁnitely generated k al -algebra R. the map O. f is determined by the composite G.k/ ! k. Recall (2.k al / maps to q 2 Q.k/ is surjective. then ˛ is surjective.G/ is injective. and the image of g in G.Q/ ! R0 .k al /
/ Q.R/.k 0 / is dense in H (8.1) shows that there exists a k al -algebra homomorphism R ! k al . Zariski’s lemma (CA 11. P ROOF.42 Let ˛W H ! G be a homomorphism of afﬁne groups. if ˛k 0 is surjective for one ﬁeld k 0 containing k.

A. Therefore. because both have the property that ˇ ı Â D Â.B. The map 1. then there is a unique homomorphism ˛W Q ! Q0 such that ˛ ı Â D Â 0 . Now ˛ 0 ı ˛ D idQ .R/.G/ ˝O.
. Therefore g and g 0 have the same image in Q0 .6. It is uniquely determined up to a unique isomorphism by the universal property in (6. Therefore G ! Q0 factors uniquely through G ! Q. and consider the homomorphism 1 ! ˛p where 1 denotes the trivial algebraic group. and so the map O.
A SIDE 6.48 A sequence 1!N !G!Q!1 is exact if G ! Q is a quotient map with kernel N .Q/ . Then any homomorphism G ! Q0 whose kernel contains N factors uniquely through Q: N G
0
Q
Q0 .k al / ! ˛p . but 1 ! ˛p is not a quotient map because the map on coordinate rings is kŒX =. Let k be a ﬁeld of characteristic p ¤ 0.G/ O. which is not injective. D EFINITION 6. P ROOF. and so ˛ and ˛ 0 are inverse isomorphisms. T / ! Hom. there are unique homomorphisms ˛W Q ! Q0 and ˛ 0 W Q0 ! Q such that ˛ ı Â D Â 0 and ˛ 0 ı Â 0 D Â.Q/ O. This shows that the composites of the homomorphisms G
QG
G ! Q0
are equal. T / is injective for all objects T .Q0 / ! O.45 The smoothness condition in the second part of the proposition is necessary.G/ O.k al / is f1g ! f1g.R/ with the same image in Q. From the theorem.46 Let G ! Q be a quotient map with kernel N . the composites of the homomorphisms O.Q0 /
are equal.49 A morphism uW A ! B in some category A is said to be an epimorphism if Hom. and is denoted G=N . The subring of O. We shall see later (7. ˛ ı ˛ 0 D idQ0 .G/. 2 When G ! Q is a surjective map with kernel N . ˛ is an isomorphism.Q/ (CA 9. 2 C OROLLARY 6. if g and g 0 are elements of G.G/ factors through uniquely through O.G/ on which the two maps coincide is O.
P ROOF. the afﬁne group Q is called the quotient of G by N .47 If ÂW G ! Q and Â 0 W G ! Q0 are quotient maps with the same kernel. T HEOREM 6.6). Note that.46). then g 1 g 0 lies in N and so maps to 1 in Q0 . Similarly. which is surjective. Group theory: subgroups and quotient groups.
67
A
6.R/.X p / ! k.R/.59) that quotients by normal subgroups always exist.! O. moreover.

then xv D 0. then xy n v can be computed as usual using that Œx.55) below. q/ 7! nqW N.qi /.5). one can show that any homomorphism of B \ SL2 has at most one extension to SL2 because any ﬁnite dimensional representation of sl2 can be reconstructed from the operators h and y. But then HI [fqg D HI \ Ker.g/
.R/ j aij D 0 for i > j g: Then Tn is the semidirect product of its (normal) subgroup Un and its subgroup Dn . This is a consequence of the next lemma. y D h.R/ for all k-algebras R (cf.R/ and Q. If it isn’t. and let H be a subgroup. According to (6.aij / 2 GLn . There exists a surjective homomorphism G ! Q containing H in its kernel and universal among homomorphisms with this property. let HI D i 2I Ker. denoted G D N Q. For example. For such a family.g/ 1 deﬁnes a map P1 ! G. then g 7! f f . According to a theorem of Chevalley (7.N / D 1 but q.qi / all any i . P ROOF. if hv D mv and y mC1 v D 0. which will be proved in (7.
qi
6h
Semidirect products
D EFINITION 6. if hv D mv and u D y n v.68
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
It is obvious from Theorem 6.! D GL2 is a nonsurjective epimorphism (any two homomorphisms from GL2 that agree on B are equal).q/ is properly contained in HI .46).8). the embedding Â Ã Â Ã BD 0 .R/ Q. Let W be the largest subspace of V ˝ V _ on which N acts trivially. 2 T HEOREM 6. Then N 0 acts nontrivially on W because it acts nontrivially on L ˝ L_ W .50 Let G be an algebraic group. Speciﬁcally. Exercise 5 to I 5). there exists a representation G ! GLV and a line L in V such that N is the stabilizer of L in V . let Tn be the algebraic group of upper triangular matrices. if N is normal in G and the map . Let f.R/ is a semidirect product of its subgroups N. then there exists a homomorphism qW G ! Q containing H in its kernel but not HI . We brieﬂy sketch the proof.G ! Qi /i 2I of surjective morphisms such that H Ker. then there exists a homomorphism qW G ! Q such that q. If N ¤ N 0 . which has image 1G because G is afﬁne and P1 is complete (see AG 7. If f jB D f 0 jB. in characteristic zero. there exists a family for Q which HI is minimal. I claim that the map from G to the image of .41 that a surjective homomorphism of afﬁne groups is an epimorphism. but it is false for afﬁne groups.N 0 / ¤ 1.qi /W G ! i 2I Qi is universal.53 An afﬁne group G is said to be a semidirect product of its subgroups N and Q. The converse is true for groups (MacLane 1971.R/ is a bijection of sets for all k-algebras R. Alternatively.
13 This 0 . f 0 be two homomorphisms GL2 ! G.
follows from the fact that GL2 =B ' P1 .7).51 Let N be a normal subgroup of an algebraic group G. In other words. L EMMA 6.n. ForT ﬁnite family .13
6g
Existence of quotients
P ROPOSITION 6. The universal surjective homomorphism G ! Q containing N in its kernel has kernel exactly N .R/ D f. so Tn .52 Let N N 0 be normal subgroups of an algebraic group G.R/ ! G. GT 3. For example. G is a semidirect product of the subgroups N and Q if G.

R0 / F . the product map is a bijection of functors N Q ! G.R/ as a semidirect product G.R/ ! N.Ri /i 2I . which is represented by O. it follows directly from the deﬁnition of product that Q Q Hom. )W By assumption.R/ D N.A..R/ ! X.R/ Q. (W Let 'W G ! Q be the given homomorphism. Then G is the semidirect product of N and Q if and only if there exists a homomorphism G ! Q whose restriction to Q is the identity map and whose kernel is N . If F is representable.
6i
Algebraic groups as sheaves
Some of the above discussion simpliﬁes when regard afﬁne groups as sheaves. Homk-alg . Then the functor R N.A. '. the canonical map F .R/ and Q. i 2I Ri / ' i 2I Hom.R/ of its subgroups N. R0 / is exact.R/ X. R/ ! Hom.R/ realizes G. R0 ˝R R0 /
2
R0 ˝R R0
.
69
P ROPOSITION 6.R/ on the set X. GT 3.A. Recall 5k that an action of G on X is a natural transformation ÂW G X ! X such that each map G.R/. n/ 7! ÂR .e. and it follows that Homk-alg .54 Let N and Q be afﬁne subgroups of an algebraic group G. For any k-algebra A. According to (CA 9.A. For each k-algebra R.N / ˝k O.R/
(cf.5). R/ ! Homk-alg .R0 ˝R R0 /
is exact (i.A. (F1).R/ ! F . and so Hom.A. and call it the semidirect product of N and Q deﬁned by Â.q. Now let N and Q be algebraic groups and suppose that there is given an action of Q on N .q.55 Let F be a functor from the category of k-algebras to sets. the ﬁrst arrow realizes F . as a functor to sets. n/ is a group homomorphism. i Ri / ! i F . the map n 7! ÂR . for each q. P ROOF. it is N Q.9) is an afﬁne group because.R/ N. then Q Q (F1) for every ﬁnite family of k-algebras .6. 2 Let G be an afﬁne group and X a functor from the category of k-algebras to sets.G/. P ROOF. Ri /.q. n/W Q. R0 / is injective for any k-algebra A.R/ such that.Ri / is bijective. the sequence F . the sequence R ! R0 is exact. (F2) for every faithfully ﬂat homomorphism R ! R0 of k-algebras. We denote it by N Â Q.R/.R/ as the equalizer of the pair of arrows). (F2). then it is injective.A. Group theory: subgroups and quotient groups. The composite of the inverse of this map with the projection N Q ! Q has the required properties.R/
ÂR
Q.R/ is an action of the group G. If R ! R0 is faithfully ﬂat. P ROPOSITION 6.

any natural tranformation from F to a sheaf will factor uniquely through F ! F= . P ROOF. If [ S. For a.R/.R0 / for some faithfully ﬂat R-algebra R0 . e
. P ROOF. and let S be the full subcategory of P consisting of the sheaves. Then is an equivalence relation on F . Now let F be a functor satisfying (S1). deﬁne F 0 .70
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1) A functor satisfying the conditions (F1) and (F2) is said to be a sheaf for the ﬂat topology14 .R/= satisﬁes (S1). there exists a sheaf aF and a natural transformation F ! aF that is universal among natural transformations from F to sheaves. P ROPOSITION 6.
P ROPOSITION 6.Ri ˝k Ri 0 /
Q We sometimes use (S1) to denote the condition that F . b 2 F . Easy exercise (cf.R0 ˝R R0 //:
where R0 runs over the faithfully ﬂat R-algebras.Ri /i 2I of k-algebras such that R ! i Ri is faithfully ﬂat.
for the fpqc (ﬁd` lement plat quasi-compacte) topology. II 1. For any k-algebra R.Ri /
Q
.R0 / 0
R a faithfully ﬂat R-algebra
(intersection inside S. P ROPOSITION 6.56 A functor F W Algk ! Set is a sheaf if and only if it satisﬁes the following condition: Q (S) for every k-algebra R and ﬁnite family . and that any natural transformation from F to a sheaf factors uniquely through F ! F 0 . P ROPOSITION 6. set a b if a and b have the same image in F . and let F be a subfunctor of S. P ROOF.R/ D lim Ker. then S is the sheaf associated with F . the sequence F . the functor aW P ! S preserves direct limits and ﬁnite inverse limits.i 0 /2I I
F .R/ \ F .
2
Q
i 2I
F .58 Let S be a sheaf.5).R/ D S.59 The inclusion functor i W S ! P preserves inverse limits.F .R/ ! i 2I F . and the functor R F .R/ ! is exact.R0 / ! F . 2 The sheaf aF is called the associated sheaf of F .R0 /).
2
Let P be the category of functors Algk ! Set.i. Moreover. Milne 1980.
14 Strictly.R/. Obviously any natural transformation F ! F 0 with F 0 a sheaf extends uniquely to S .57 For any functor F W Algk ! Set. One checks easily that F 0 is a sheaf.Ri / is injective and (S2) for the condition that its image is subset on which the pair of maps agree.

For some index set I .R/ ! P .R0 / \ G. : : :/ D 0 and N. Group theory: subgroups and quotient groups. the map P . then it already satisﬁes the conditions in RI (because R ! R0 is injective).R0 ˝R R0 //
where the limit is over all faithfully ﬂat R-algebras.60 Let G ! Q be a surjective homomorphism of afﬁne groups with kernel N .R0 / ? ? y Q. then P 0 ' Q: For (a). 2
A SIDE 6. (b) Let P 0 . the map P .a. By deﬁnition Hom.
71
P ROOF. it follows directly from the deﬁnition of “surjective’ (see 6.R0 / P . To show that a preserves ﬁnite inverse limits. Then Q represents the sheaf associated with the functor R P ROOF.6.P . i.R/ D N. / ' Hom.
.R/ ! Q.R0 / ! Q. But if an element of RI satisﬁes the conditions when regarded as an element of R0I . Let P be the functor R show: G. . When we pass to the limit over R0 .R/:
G.R0 / ! 0
R
P . For any k-algebra R0 . and so the two vertical arrows induce an injective homomorphism K.R/ D lim Ker. This implies (immediately) that i preserves inverse limits and a preserves direct limits.R0 / is the subset of R0I deﬁned by the same polynomial conditions.R0 /). which follows from the construction of a.: : : . it sufﬁces to show that it preserves ﬁnite products and equalizers. and we shall
(a) For any injective homomorphism R ! R0 of k-algebras. 2 P ROPOSITION 6. N. /.61 Discuss why it is useful to regard the afﬁne groups as sheaves rather than presheaves. consider the diagram K.R/=N.R0 ˝R R0 /
where K. Then P commutes with products.R0 / is injective.R0 / ! P . we have to prove that N.R0 ˝R R0 / ? ? y Q.R/ is the equalizer of the bottom pair of maps.40) that this map becomes surjective. For (b).R/ is the subset of RI deﬁned by some polynomial conditions fj . and so a and i are adjoint functors.R0 / ! Q.R0 / is the equalizer of the top pair of maps — we know that Q.R/ (intersection inside G.R0 / is injective.R/.R/. Xi . //.R/=N.

It is easy to see that F D lim F in the category of functors from k-algebras to groups. 6. For example.109).R/ (cf. are more difﬁcult. Moreover. but when G1 and G2 are algebraic groups. the quotient afﬁne group G=N is the sheaf associated with the functor R G. or just limit).R/ is the inverse limit of the functor i Fi . P ROOF. 2 In particular. and it is the inverse limit of F in the category of algebraic groups. the functor R lim Fi . P ROOF. and let AB be the sheaf associated with the subfunctor R A.68) that.R/ B. even ﬁnite direct limits.I /
(MacLane 1971.R/ need !I not be a sheaf. and the equalizer of the two homomorphisms is equal to their ﬁbre product. it can be constructed by forming the naive direct limit in the category of functors. the functor R G1 .60). then the functor R lim Fi .R/=N. p. V 2 Theorem 2. when N is a normal subgroup of an afﬁne group G.R/ from I to the category of (abstract) groups. and so that also is an afﬁne group (see 3b). for a functor F W I ! C from a small category I to a category C. there is the notion of an inverse limit of F (also called a projective limit.62 Let F be a functor from a small category I to the category of afﬁne groups over k. This generalizes the notions of a limit over a directed set and of a product. and then forming the associated sheaf. Roughly speaking.R/ (58)
is an algebraic group. T HEOREM 6. Both products in (57) are algebraic groups. Both products are afﬁne groups. But F is equal to the equalizer of two homomorphisms Y Y Fi Ftarget. the sum of two groups is their free product.
6k
Exercises
E XERCISE 6-1 Let A and B be subgroups of the afﬁne group G. p. inverse limits of algebraic groups exist as afﬁne groups.R/ G2 .21) we shall see that every afﬁne group arises in this way.
. thus F .72
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
6j
Limits of afﬁne groups
Recall (MacLane 1971. and it will follow that F is the inverse limit in the category of afﬁne groups once we show that it is an afﬁne group. when the direct limit of a system of afﬁne groups exists. III 4.I / u2ar.u/ (57)
i 2ob. Denote the functor (56) by F .
2
Direct limits. For example.63 Let F be a functor from a ﬁnite category I to the category of algebraic groups over k. T HEOREM 6. Later (7.R/ will generally be far from being an algebraic group. then the functor R lim F .R/ of G. and it is the inverse limit of F in the category of afﬁne groups.R/ (56)
is an afﬁne group.

i. \ .e. Show: (a) C \ A is a normal subgroup of C \ B. E XERCISE 6-3 (Dedekind’s modular laws).. for any k-algebra R.B/ deﬁned by the map a.R0 / . B. E XERCISE 6-2 Let A. an element G.A/ ˝ O.R0 / for some faithfully ﬂat R-algebra R0 .
73
(a) Show that AB is representable by O.AB/ .R0 / B.R/ if and only if its image in G. then B D A. Let A. Show: (a) B \ AC D A. 0
R
(c) Show that AB is a subgroup of G if B normalizes A. (b) Show that. Group theory: subgroups and quotient groups.B \ C /. B. b 7! abW A B ! G (of set-valued functors).6.R/ lies in .B \ C /I (b) if G D AC .R/ D G.R0 / lies in A.R/ \ A.
.R0 / B.G/=a where a is the kernel of homomorphism O. (b) CA is a normal subgroup of CB.G/ ! O. C be subgroups of an afﬁne group G such that A is a normal subgroup of B. C be subgroups of an afﬁne group G such that A is a subgroup of B.AB/ .

R// ? ? y ! EndR0 -lin .R/ is a functor from the category of k-algebras to monoids and R AutV . or. Deﬁne V to be the k-vector space of maps X ! k. x 2 X:
This deﬁnes a representation of G on V . R a k-algebra.R/ ? ? y G. the only difference between a left module and a right module is one of notation: it is simply a question of r whether we write rm or mr (or better m). k is a ﬁeld.R/ D V ˝ R.7
Representations of afﬁne groups
One of the main results in this section is that all algebraic groups can be realized as subgroups of GLn for some n. Note that for a commutative ring R. : : : . A representation of G on a k-vector space V is a homomorphism of groups G ! Autk-lin . ng. f 2 V . and the map G ! Autk-lin . and so this gives a homomorphism G ! GLn . Before looking at the case of algebraic groups.V /.R//.R/ ! EndR-lin . for every homomorphism R ! R0 of k-algebras.
Then R EndV .V . For a k-algebra R.V . A linear representation of G on a k-vector space V is a natural transformation rW G ! EndV of functors Algk ! Mon. because it says that all possible multiplications in algebraic groups are just matrix multiplication in disguise. Let G be an afﬁne monoid or group over k. we often work with algebraic monoids rather than groups since this forces us to distinguish between “left” and “right” correctly. The vector space V has a basis consisting of the maps that send one element of X to 1 and the remainder to 0. In other words.R0 // 74
rR0
. it will sometimes be convenient to regard R-modules as right modules. For example. When we take X D G.R/ is a functor from the category of k-algebras to groups. AutV . At ﬁrst sight. and tensor products are over k unless otherwise noted.
7a
Deﬁnition of a representation
Let V be a vector space over k.x/ D f . it is a family of homomorphisms of monoids rR W G.k/ in 1. Let X G ! X be a right action of G on a ﬁnite set X. In this section. (59)
such that.V / is injective.27). in other words. the representation we get is called the regular representation. (R-module) (monoid under composition) (group under composition). Throughout. /W Sn ! GLn . for the symmetric group Sn acting on f1.V .V .gf /.R//. EndV . which is faithful if G acts effectively on X. and write V ˝k R instead of R ˝k V . an action G V ! V of G on V in which each 2 G acts k-linearly.xg/ for g 2 G. and let G act on V according to the rule: .k/ where n is the order of X. the diagram G.R/ D AutR-lin . this gives the map 7! I. this is a surprising result.V . we let V .R//.R0 /
rR
! EndR-lin . Let G be a ﬁnite group.R/ D EndR-lin . With the terminology of (2. we should review how to realize a ﬁnite group as a matrix group. AutV D EndV . 2. In this section.

and so rR . group) G on V is a homomorphism G ! . We write V also for the functor R V .Mn . A linear representation is said to be ﬁnite-dimensional when V is ﬁnitedimensional as a k-vector space. p185. G ! GLn ).V 0 .g/. When G is an afﬁne group.W . We shall see later (7.R/. j
0.V . See Abe 1980. A linear representation of an afﬁne monoid (resp. / (resp.2 Let G D GLn .V. When V D k n .R/ deﬁne a linear representation of G on Mn . Note that (61) implies that i ı 1 D . When k n has characteristic zero.g/ y yR V .14) that all ﬁnite-dimensional representations of n Ga are of this form. 1 t /n =nŠ D exp. : : : .R//
i 0
(so rR .v/ ˝ ct
i ).k/ acting on Mn .i C 1/ i C1 . t 0 2 R. Representations of afﬁne groups
75
commutes.R/ ? ? ? ?r 0 .R/ and all k-algebras R. The orbits of G.t 0 / for all t. r takes values in AutV and is a natural transformation of group-valued functors. and let Mn denote the vector space of all n n matrices with entries in k. E XAMPLE 7.R/ ! Mn .R/ Mn .R/
commutes for all g 2 G. A homomorphism of linear representations . there are more possibilities.Mn .
If
0
i
ı
j
D idV D i Cj i
i Cj
all i. r 0 / is a k-linear map ˛W V ! V 0 such that V . For t 2 R.P.
. EndV is the monoid R . / and AutV D GLn . which are represented by the Jordan matrices when k is algebraically closed.t / D X .7.R// W . let X i rR .1 Let G D Ga . Let V be a ﬁnite-dimensional k-vector space.R/ deﬁned by V .
(61)
then rR . i .k/ are the similarity classes.g/ rR . 1 t /:
When k has nonzero characteristic.t / D i t 2 End.t C t 0 / D rR . and it is faithful if all the homomorphisms rR are injective. A/ 7! PAP 1 W G. G V ! V.t / C rR .R/
! V 0 . r/ ! . and so rR is a representation. The actions . Then a linear representation of G on V can also be deﬁned as an action of G on V. : : : be a sequence of endomorphisms V such that all but a ﬁnite number are zero. and so 1 D nŠ n .R/ ! V 0 .R/ acts R-linearly on V . A subspace W of V is a subrepresentation if rR .R/ for all k-algebras R and all g 2 G. this implies that 1 is nilpotent and that n D 1 =nŠ.R/ ˛.t/.R/.R/
˛.v ˝ c/ D
P
i . (60)
such that each g 2 G.R/. and let 0 . E XAMPLE 7.

R/: (62)
This is called the regular representation. / ! . D EFINITION 7.gf /R .R/ R ˝ O.C. which extends by linearity to a map G.
7b
Terminology
From now on.R/ O. A comodule is said to be ﬁnite-dimensional if it is ﬁnite-dimensional as a k-vector space. On reversing the directions of the arrows.4 Let . A right C -comodule15 is a k-linear map W V ! V ˝ C (called the coaction of C on V ) such that the diagrams V ? ? y V ˝C commute.R/. i.V 0 .A. we obtain the notion of comodule over a coalgebra.x/ D fR . “representation” will mean “linear representation”. for all g 2 G. In more detail: the formula (62) deﬁnes a map G.G/ ! R ˝ O. e/ be a k-algebra. not necessarily commutative.G/ ! R ˝ O.
15 It
. except that it is right comodules that correspond to left representations of monoids.V.e.76
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
E XAMPLE 7. Recall that a left A-module is a k-vector space V together with a k-linear map W A ˝ V ! V such that the diagrams V x ? ? A˝V
A˝
A˝V x ? ?m˝V A˝A˝V
V
A˝V x ? ?e˝V
'
(63)
V
k ˝V
commute.
7c
Comodules
Let .3 There is a unique linear representation r of G on O.V ˝ / ı . ˝C/ı D V: ! V ˝C ? ? yV ˝ V ! V ˝C ? ? yV ˝
'
(64)
˝C
! V ˝C ˝C
V
! V ˝k
A homomorphism ˛W .V ˝ / ı D . . such that .G/. Because we consider right comodules we are more-or-less forced to write V ˝ R where elsewhere we write R ˝ V . 0 / of C -comodules is a k-linear map ˛W V ! V 0 such that the diagram V ? ? y V ˝C
˛
!
V0 ? ? 0 y
˛˝C
! V 0 ˝C
commutes. / be a k-coalgebra.G/..
would be more natural to consider left comodules. f 2 O.G/.G/ (regarded as a k-vector space) such that . m. x 2 G.xg/.

C.V1 .v ˝ V / D v for all v 2 V because X 0 . (c) If .C. the equations (66) show that the k-subspace spanned by the cij is a subcoalgebra of C .6 Let . then V CV . there is a smallest subcoalgebra of C through which the co-action of C on V factors.ei /i 2I for V . Clearly. Choose a basis . note that the restriction of the co-identity of C to V is an element V of V _ and that 0 . The map V ! V ˝C
V ˝˛
! V ˝C0
provides V with the structure of a right C 0 -comodule.ei /i 2I for V .28. (c) Let . for any k-vector space V . j 2 I: (66) .V //. /. Representations of afﬁne groups
77
E XAMPLE 7. The next remark shows that for a comodule V over a coalgebra C . Homk-lin .V. and so it is independent of the choice of the basis. Then . (b) Recall that for a ﬁnite-dimensional k-vector space V .ej / X D ej .V2 . (a) When we choose a k-basis . the image of A in Endk . / is a right comodule if and only if P . V ˝ WV ˝C ! V ˝C ˝C is a right C -comodule (called the cofree comodule on V ). through which the action of A on V factors. then X ei ˝ ci H) . it is the smallest subspace of C such that . / is a sub-comodule of .
. and write X . V ˝ C / ' Homk-lin . 2 / be comodules over coalgebras C1 and C2 respectively.cij / D ıij For a module V over an algebra A.v/ D
i 2I
0
. / is a right C -module (compare (26).5 (a) The pair .V. Therefore CV is the image of 0 W V ˝ V _ ! C . p. 0 . To see this. for each j .ei /ci :
_ In particular. C /: If $
0
under this isomorphism. (65)
i 2I
(ﬁnite sum. / is a C -comodule.V.V ˝ V _ . there is a smallest quotient of A. almost all cij ’s are zero).cij /
i 2I
(by (26). namely. R EMARK 7.V. which we denote CV . More generally. and let W V ! V ˝ C be a k-linear map. so.idC ˝ / . p. / be a right C -comodule.cij / D k2I ci k ˝ ckj all i.ej ˝ ei / D cij (notation as in (65)).ej / D .v ˝ f / D
X
i 2I
f .ej ˝ V / D .ej /cij i2I
D . (d) Let V be a k-vector space.28)
D ej
(by (66)). ˝ idC / .V. cij 2 C.7. The map V1 ˝ V2
1˝ 2
! V1 ˝ C1 ˝ V2 ˝ C2 ' V1 ˝ V2 ˝ C1 ˝ C2
provides V1 ˝ V2 with the structure of a C1 ˝ C2 -comodule. (b) Let . 1 / and .ej / D ei ˝ cij .V / V ˝ CV . with (64)). and let ˛W C ! C 0 be a homomorphism of coalgebras.

78
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
R EMARK 7.8 Every comodule .vi ˝ ej ˝ ek / D . Let V be a k-vector space.vk / ˝ ek
i. P ROOF.j.k
rij k 2 k. jW / is itself a C -
P ROPOSITION 7. V /. 2. Recall from (64) that . Write X .
i
(ﬁnite sum. Deﬁne to be the composite of C_ ˝V
C _˝
! C_ ˝V ˝C ' V ˝C_ ˝C
V ˝ev
! V ˝k ' V
where evW C _ ˝ C ! k is the evaluation map. _ /. As a sum of sub-comodules is a sub-comodule. we obtain (67). it sufﬁces to show that each element v of V is contained in a ﬁnite-dimensional sub-comodule. When C and V are ﬁnite-dimensional.C _ .
C OROLLARY 7.
We shall show that . 7! is a bijection Homk-lin .C. V ˝ C / ' Homk-lin . / is a right C -comodule if and only if . One can check that .4). and let X . W ˝ C . When k is noetherian. In the general case.1). For any ﬁnite-dimensional sub-comodule V of C . V CV C
2
(see 7.
.ei / D rij k .V.e. P ROOF.7 Recall (4c) that the linear dual of a coalgebra . not every C _ -module structure arises from a C -comodule structure. ˝ C / ı : On applying each side of this equation to v. every comodule over a k-coalgebra C is a ﬁltered union of ﬁnitely generated subcoalgebras (Serre 1993.vk / D X
i. where V runs over the ﬁnitedimensional C -comodules.ei /i 2I be a basis for C as a k-vector space. vi 2 V.9 A coalgebra C is a union of its sub-coalgebras CV .C _ ˝ V.6).v/ D vi ˝ ei . but it is known which do (Sweedler 1969. / is an associative algebra with identity .k k
(inside V ˝ C ˝ C /:
2
On comparing the coefﬁcients of 1 ˝ 1 ˝ ek in these two expressions. / is a ﬁltered union of its ﬁnite-dimensional sub-comodules. _ . we ﬁnd that X X rij k . Then .V. .ej ˝ ek /.W / comodule. Let .V. 1.j
rij k vi ˝ ej
(67)
from which it follows that the k-subspace of V spanned by v and the vi is a subcomodule containing v.V..10 The main deﬁnitions in this subsection require only that k be a commutative ring.V ˝ / ı D . and let W V ! V ˝ C be a k-linear map. and so this follows from the proposition. A k-subspace W of V is a subcomodule if .
j. / is a left C _ -module.W. only ﬁnitely many vi are nonzero).
A SIDE 7. i. and so there is a one-to-one correspondence between the right C -comodule structures on V and the left C _ -module structures on V .

G/-comodule structures on V .3.7.rij /. 8. there is a natural one-to-one correspondence between the linear representations of G on V and the O. m. Speciﬁcally. if . a bialgebra structure on C endows Comod.V1 . V / ! .G/ is the same as giving a family P .2.rij / $
.C / be the category of ﬁnite-dimensional comodules over a k-coalgebra C .4) hold in comodules over C . C ˝ V _ /.j /2I I of regular functions on G: rR .1. On reviewing the deﬁnition of and (4h). the inversion can be used to turn the left comodule structure on V _ into a right comodule structure.V.C / with tensor products.77.R/:
The natural transformation r is a morphism of afﬁne monoids if and only if P rij D k2I ri k rkj all i. V =W / is a homomorphism. a Hopf algebra structure on C endows Comod. For any k-vector space V . to give a k-linear map W V ! V ˝ O. Therefore every sub-comodule arises as the kernel of a homomorphism of comodules.C / with tensor products and duals. when .V. V ˝ C / ' Homk-lin . The choice of a basis . 8. For example. 1 / and . the standard isomorphism theorems (cf.A. V /. holds. W / is a sub-comodule of .R/ / D ıij (Kronecker delta)
(68)
On the other hand.i. S UMMARY 7. When C is a bialgebra with an inversion. . /.5c).V _ .g/ ej D P
i 2I
rij
R
. e. V1 ˝ V2 has a natural structure of C ˝ C -comodule. / is deﬁned to be a comodule over the coalgebra . 2 / are C -comodules.j Än of elements of O.g/ ei
for all k-algebras R and g 2 G. A bialgebra structure .1G. The category of comodules over C is abelian and the forgetful functor to k-vector spaces is exact.W. . Therefore r $ . e/ on C deﬁnes a tensor product structure on the category of comodules over C . 8. right comodule structures on V correspond to left comodule structures on V _ .G/ (write . With the obvious deﬁnitions. /. then the quotient vector space V =W has a (unique) comodule structure V =W for which . Then Comod. and is a co-action if and only if (66).ei /i 2I for V identiﬁes EndV with Mn and the natural transformations of set-valued functors rW G ! EndV with the families of .5b).V =W.V. p.
7e
Representations and comodules
A comodule over a bialgebra . Moreover. the homomorphism of coalgebras eW k ! C ' k ˝ C deﬁnes a C -comodule structure on k — this is the trivial comodule. 8. P ROPOSITION 7.m.A. as we observed (7.12 Let G be an afﬁne monoid over k. C / ' Homk-lin .11 Let Comod. j 2 I: rij .C / is an abelian category. Representations of afﬁne groups
79
7d
The category of comodules over C
Let C be a coalgebra over k.rij /1Äi.V ˝ V _ . Assume that V is ﬁnite dimensional. Under the canonical isomorphisms Homk-lin . The space V _ with this structure is called the dual or contragredient of .V2 . We ﬁrst describe the correspondence in the case that V is ﬁnite dimensional. one sees that the conditions (66) and (68) correspond.V. the homomorphism of coalgebras mW C ˝ C ! C turns this into a C -comodule structure (7.ej / D ei ˝ rij ).

gh/ for g. Let g 2 G. gh is the map A ! A ˝ A ! R.R/. In the proof of Proposition 7.A.h/
.R/ ! V .g/rR . we construct a canonical correspondence between the representations and the comodule structures.u/ NN def DrA .R/ whose restriction to V is .v//. h 2 G.. and let rW G ! EndV be a natural transformation of set-valued functors. a comodule structure on V determines a representation r such that.V ˝ g/ ı .R/ is V ! V ˝ A ! V ˝ R: These operations are inverse.k/ is A ! k.
v7!v˝1
V NNNv7!v˝1 / V ˝ A NNN NNN NNN & V ˝A
V˝
+/
V ˝k
V ˝k
V˝
/ V ˝ k.A/ D EndA-lin .h/ D rR .A/ is a comodule structure W V ! V ˝ A on V .g/ because rA . Let V be a vector space over k.g/
V ˝g
The k-linear map determines rR . A/ maps to an element of EndV . the diagram can be used to extend any k-linear map W V ! V ˝ A to a natural transformation r of set-valued functors. Moreover.g/ to be the linear map V .1G.A. and it remains to show that r is a representation of G if and only if is a comodule structure on V .V ˝ g/. Next consider the condition that rR .R/ ! V .g/ is the unique R-linear map making the right hand square commute.u/ is the unique A-linear extension of to V ˝ A and rR .R/. Recall that the identity element 1G. for g 2 G. we get a one-to-one correspondence r $ between natural transformations of setvalued functors r and k-linear maps . and in Proposition 7.V . the restriction of rR . P ROOF ( OF P ROPOSITION 7.12 below. We prove the following more precise result: Let rW G ! EndV be a representation. for R a k-algebra and g 2 G. and consider the diagram:
/ V ˝A V NNN NNN NNN rA .G/comodule structures on V .12) Let A D O.g. To say that rk .A/ ' def Homk-alg .c .g/W V . c 2 R: (69)
In this way.R/ D Homk-alg .A// whose restriction to V V . the correspondence becomes that just described.k/ / D idV ˝k means that the following diagram commutes. rR . for all g 2 G.17 we show that.
. Thus. v 2 V .v ˝ c/ D . By deﬁnition (see (34)). R/ and deﬁne rR .G/.e.k/ of G. Conversely. R/.R/ to V V .R/ D Homk-alg . once a basis is chosen.
i. namely.80
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
gives a one-to-one correspondence between the linear representations of G on V and the O. the “universal” element u D idA in G. that the right hand diagram in (64) commutes.u/jV N' V ˝A
v7!v˝1 V ˝g V ˝g
/ V ˝R / V ˝ R:
rR . Thus the map determines the natural transformation r.g/.A.R/.

When A D O.v/.V.R/ ! V .3).Ga / be a ﬁnite-dimensional O. this implies that only a ﬁnite number of the i are nonzero.g/
(73)
(equality in V .
.
! V ˝A˝A
V ˝. The k-vector space O.g/.v/ D i . X 2 .v/ ˝ X .16 Every representation of G is a union of its ﬁnite-dimensional subrepresentations. Representations of afﬁne groups and so rR . Choose a basis .
E XAMPLE 7.ej ˝ 1/ D P
i 2I ei
˝ g. P ROOF.h/
! V ˝ R:
(70)
! V ˝R
˝R
! V ˝A˝R
V ˝.ei /i 2I for V . h if and only if the ﬁrst diagram in (64) commutes.R/.G/ (7.ej / D ei ˝ aij . this comodule structure on A corresponds to the regular representation of G on O..h/
! V ˝ R:
(71)
2
The maps (70) and (71) agree for all g.R/. /. for each g 2 G. /.13 Recall (7.g. aij 2 O. and write X . X.5) that.1) form a complete set. : : : and so we can write X i . A subspace W of V is a subrepresentation if and only if it is a subcomodule.e.g/ is a map V .Ga / ' kŒX has basis 1. It follows that the representations constructed in (7.V. v 2 V:
i 0
As is k-linear.G/.aij / D
P
i 2I ei
˝ aijR . P ROOF.15 Let rW G ! EndV be the representation corresponding to a comodule . P ROPOSITION 7. rR .
2
P ROPOSITION 7. i . recall that aijR is a map G.id/
! V ˝ R. In view of (7. for each v. E XAMPLE 7. and as the sum is ﬁnite. this is simply a restatement of Proposition 7. as V ! V ˝A
˝A V ˝h V˝
81
! V ˝A˝A
V ˝. Routine checking.14 Let W V ! V ˝ O. the map W A ! A ˝ A is a comodule structure on A.Ga /-comodule.g. so also is each map v 7! i .17 Let rW G ! EndV be the representation corresponding to a comodule .gh/ acts on V as V ! V ˝A On the other hand.12) and (7.8.R/). for any k-bialgebra A. rR .R/ ! R and that rR .7.g.15).G/: (72)
i
Then. As V is ﬁnite-dimensional.
2
P ROPOSITION 7.h/ acts as V ! V ˝A i.v/ is zero except for a ﬁnite number of i .g/rR .

R0 /gI (c) fv 2 V . and so the proposition says that the homomorphism O.G/ deﬁned by r sends Xij to aij . and so if W is the solution space over k. then v 2 V H if and only if r. say v D ej (as in the statement of the proposition).g/ ˝ rR .g/v D v for all g 2 H. V .G/ and g 2 G.82 P ROOF. The tensor product of two representations . Nothing in this section requires that k be a ﬁeld (provided one assumes V to be free).
.idV ˝g/. We may suppose that v ¤ 0.R/. /. .g/ D rR . This proves the equality of the sets in (a) and (c).19 Let G. The condition . P ROOF.R/ j . and let Repk . C OROLLARY 7.18.v/ Á v ˝ 1 mod V ˝ a. Conditions (a) and (b) each hold for ej if and only aij ıij 2 a for all i .G/=a.V ˝ V.Mn / ! O. r. and we let V H denote the subspace of ﬁxed vectors in V .V. Let H be a subgroup of G.g/: In the last step. : : : .18 Let rW G ! EndV be the representation corresponding to a comodule .r ˝ r 0 /R . L EMMA 7. . it is an abelian category and the forgetful functor to k-vector spaces is exact and faithful. rR .aij / P D i ei ˝ aijR .k/ (because there is a largest subgroup of G ﬁxing v). Xnn .G/ be the category of representations of G on ﬁnitedimensional vector spaces. 2
7f
The category of representations of G
Let G be an afﬁne monoid.V 0 .R/ (see 2.ej ˝ 1/ D .ej // P D .
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
rR . r 0 / is deﬁned to be .Mn / D kŒX11 . X12 .R/ are equal: (a) V H ˝ R. The following conditions on a vector v 2 V are equivalent: (a) for all k-algebras R and all g 2 H. i ei ˝ aij / P D i ei ˝ g. and so is part of a basis for V .V. The following submodules of V .g/. P ROOF.G/-comodules (see 7.v/ Á v ˝ 1 mod V ˝ a is linear in v. If H.vR0 / D vR0 for all R-algebras R0 and g 2 H. According to (69). r/ and . and let R be a k-algebra.vR / D vR I (b) . 2 We say that v is ﬁxed by H if it satisﬁes the equivalent conditions of the corollary.R/ j rR0 . and H be as in the corollary.H / D O.f / D fR .g/. then W ˝k R is the solution space over R.g/.idV ˝g/.g/.11). we used that g. and let O.v/ Á v ˝ 1 mod V ˝ a ˝ Rg. (b) fv 2 V .g/ for f 2 O.12).
2
The coordinate ring of the monoid Mn is O.k/ is dense in H . Therefore the equality of the sets in (b) and (c) follows by taking k D R in Corollary 7. r ˝ r 0 / 0 where . As this is essentially the same as the category of ﬁnite-dimensional O.

P ROOF. r/ and .25 A Hopf algebra whose augmentation ideal is ﬁnitely generated is itself ﬁnitely generated.R/. If A is an integral domain and its ﬁeld of fractions is ﬁnitely generated (as a ﬁeld) over k.g/. then .23 Let A be a Hopf subalgebra of the Hopf algebra B. 2 C OROLLARY 7. every afﬁne monoid (resp.aij / D k ai k ˝ akj (see (66). By (7. groups) in which the transition maps are quotient maps. r/ has contragredient (or dual) is deﬁned to be . Then A and B are directed unions of ﬁnitely generated Hopf subalgebras Ai and Bi such that Ai Bi . T HEOREM 7. v 2 V . T HEOREM 7.v/ D f . group) is an inverse limit of algebraic monoids (resp.R/.V 0 . group) is an inverse limit of its algebraic quotients. The comodule structures on V ˝ V 0 and V _ deﬁned by r ˝ r 0 and r _ are those described in 7d. r 0 / be representations of G. Representations of afﬁne groups
83
When G is a group.
.rR .20 Let . r _ / where.ei / be a basis for V . and let and 0 be the corresponding comodule structures on V and V 0 .ej / D i ei ˝ aij . and write . P ROOF. C OROLLARY 7.A0 / A0 . p.L/ L ˝ L. . and so they are equal (5.v/ D f . a representation . It follows that A is the directed union A D A0 of its ﬁnitely generated sub-bialgebras. f 2 V _ .22 Every bialgebra (resp.A0 / A0 ˝ AS and so it is a ﬁnitely generated sub-bialgebra of A.21 Every afﬁne monoid (resp.21 in the following equivalent form. 2 C OROLLARY 7. P ROOF.
P ROPOSITION 7. the k-subalgebra A0 generated by L and SL satisﬁes S. and the subspace L of A spanned by the ei and aij 0. Hopf subalgebras) over k. Any ﬁnite subset S of A is contained in a ﬁnitely generated Hopf subalgebra A0 of A.g
1 v/). P Then . satisﬁes .
2
7g
Afﬁne groups are inverse limits of algebraic groups
It is convenient at this point to prove the following theorem. Easy exercise for the reader. then A is ﬁnitely generated. g 2 G.Sa/ D Sci ˝ S bi (Exercise 4-5b).V. We prove Theorem 7. then A0 and A have the same ﬁeld of fractions.8). P P Suppose that A has an inversion S . In particular.V / V ˝ A. If .77).f / .33). Let . Therefore. Hopf algebra) over k is a directed union of its ﬁnitely generated sub-bialgebras (resp.V _ . The k-subalgebra A0 generated by L satisﬁes . It follows that A is the directed union of its ﬁnitely generated Hopf subalgebras. every ﬁnite subset of A is contained in a ﬁnite-dimensional P k-subspace V such that .R/ (more succinctly.a/ D bi ˝ ci .g 1 /v/. and so it is a ﬁnitely generated Hopf subalgebra of A. Let A be a k-bialgebra.7. When S is chosen to generate the ﬁeld of fractions of A.V. _ rR .24 Let A be a Hopf algebra over a ﬁeld k.gf /.

the image of O. r/ be a faithful representation of an algebraic group G. choose i 2I a faithful ﬁnite-dimensional representation . 2 C OROLLARY 7. and the family is faithful.
16 As
a B-module.26) is nontrivial.84
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROOF. Let . every sufﬁciently large ﬁnite-dimensional subrepresentation of the regular representation will be faithful. We have to show that a Hopf subalgebra A of a ﬁnitely generated Hopf algebra B is ﬁnitely generated. Xn that are not ﬁnitely generated. Write G as an inverse limit G D lim Gi of algebraic groups.ej / D i ei ˝ aij . 2 P ROPOSITION 7. the subring kŒX.17). : : : . : : : . The inclusion A0 ! A corresponds to a quotient map G ! G 0 whose kernel has Hopf algebra A ˝A0 A0 =IA0 ' A=IA0 A D A=IA ' k. and because B is ﬂat over A.26 Every quotient of an algebraic group is itself an algebraic group.27 There are many subalgebras of kŒX1 . cm ˝ bm g. : : : of kŒX. X Y 2 .28). Then V is a union of ﬁnite-dimensional faithful representations.ei /aij .29 Every afﬁne group admits a faithful family of ﬁnite-dimensional representations. According to (7. there exists an injective homomorphism G ! GLn . For some n. Let A be a Hopf algebra. If IA is ﬁnitely generated. and write . Hence (7. and let V be a ﬁnite-dimensional subcomodule of A containing a set of P generators for A as a k-algebra.GLV / ! A is surjective. ˝ idA / .28 Let G be an algebraic group. and so A0 ' A. and the map . P ROOF.Vi .Vi . because W A ! k is a co-identity (see (26). Proposition 6. this implies that IA is ﬁnitely generated. ri / of Gi . which means that G ! GLV is injective. p. T HEOREM 7. Let A D O. : : : . the ideal IA B is ﬁnitely generated.GLV / ! A contains the aij . it therefore equals A. Y is not even noetherian.ei /1Äi Än be a basis for V . and.a1 . Therefore IA ˝ B is a ﬁnitely generated as a B-module. We have shown that O.1 shows that G ' G 0 . the map IA ˝A B ! A ˝A B ' B is an isomorphism of IA ˝A B onto IA B. then there exists a ﬁnitely generated Hopf subalgebra A0 of A containing a set of generators for IA . Each .30 Let .G/.V. ri / can be regarded as a representation of G.
. P ROOF. for each i 2 I . and as B is faithfully ﬂat over A.
7h
Algebraic groups admit ﬁnite-dimensional faithful representations
In fact. X ej D .ej / D .16 2
A SIDE 7. X Y.
i
and so the image contains V . Because B is noetherian. IA ˝A B has a ﬁnite set of generators fc1 ˝ b1 . 2 P ROPOSITION 7. am / 7! ˙ai ci W Am ! IA
is surjective because it becomes surjective when tensored with B. for example. P ROOF. But.

and let rA be the regular representation of G on A.ei /i 2I be a basis for V ..x C x 0 . z/ 7! @0 1 y A 0 0 1 is an embedding of G into GL3 .31 Let G be the functor sending a k-algebra R to R R R with the (noncommutative) group structure . Because A is ﬁnitely generated as a k-algebra.V.R/ and u 2 V _ . Because . only ﬁnitely many aij ’s are need to generate it.R/ ! R are natural in R. Z has more than one Hopf algebra structure.V. and so G . Representations of afﬁne groups
85
P P ROOF. y C y 0 .x 0 . then it is projective. P ROPOSITION 7.
. rV .v/R .x. y 0 .G/.7. and hence automatically ﬂat. z 0 / D .). When k is a Dedekind domain and G is ﬂat. rV / be a representation of G. z C z 0 C xy 0 /: This is an algebraic group because it is representable by kŒX. 2 The theorem says that every algebraic group can be realized as an algebraic subgroup of GLn for some n. the maps x 7! hu. The map 0 1 1 x z . rV . Note that the functor R R R R also has an obvious commutative group structure (componentwise addition).V. let hu. the k-algebra A is generated by the aij (7.
A SIDE 7. an argument using restriction of scalars proves the statement for any reductive group (mo22078). Z. For a ﬁxed v 2 V and u 2 V _ . and so deﬁne an element of O. y. rV / ! .32 Does Theorem 7. r/ is faithful.G/. Y. and write . Using (7. For v 2 V . This does not mean that we should consider only subgroups of GLn because realizing an algebraic group in this way involves many choices. z/ . and hence a direct summand of a free ﬁnitely generated k-module L. aij 2 A.A. and so there exists a ﬁnitely subset J of I such that the aij ’s appearing in .ej / for some j 2 J generate A.
7i
The regular representation contains all
Let . the module M is torsion-free. i.28 still hold when k is allowed to be a commutative ring? Apparently.V.L/ .v/ 2 R.x/vi (in R) for all x 2 G.x.x/viW G.v/
2 O. Thus.G/ such that
D hu. which shows that the k-algebra kŒX. there exists a
u . vi D u. Y. one sees by the above arguments that an afﬁne group scheme G of ﬁnite type over a noetherian ring k has a faithful representation on a ﬁnitely generated submodule M of the regular representation.e. r/ containing fej j j 2 J g is faithful. As every split reductive group scheme over a ring k arises by base change from a similar group over Z (Chevalley). this is not known even when k is the ring of dual numbers over a ﬁeld and G is smooth (mo22078.ej / D i 2I ei ˝ aij . Let . E XAMPLE 7. Since every ´ reductive group splits over a ﬁnite etale extension of the base ring (SGA3).x/ u .R/.10). rA /. If M is ﬂat over k. y.33 The map
u
is a homomorphism of representations .! GLrank. such group schemes admit embeddings into GLn .17 et seq. every ﬂat group scheme of ﬁnite type over a Dedekind domain admits an embedding into GLn for some n. Brian Conrad).R/:
Let A D O. Every ﬁnite-dimensional subrepresentation of .

p.x/. Let A D O.xg/vi D
u .R/.v/. rV . is injective. The k-vector space V0 ˝ A becomes a comodule (isomorphic to a direct sum of copies of A) with the coaction17 V ˝ W V ˝ A ! V ˝ A ˝k A: The commutative diagram (see (64).x/
D .
D .2
is the cofree comodule on V0 — see (7. hun . P ROOF. Note that
u .g/ ı .76) V ? ? y V ˝A
17 This
!
V0 ˝ A ? ? yV0 ˝
!
An ? ? y
n
˝1
! V0 ˝ A ˝ A
! .G/ through the second factor.34 If u1 .
u /R
for all g 2 G.x/
D hu. It is obviously injective. vi/.G/. : : : .v/R . Here G acts on V0 ˝ O. Then W V ! V0 ˝ O.v//W V
! An
(74)
D hu.v// .g/ as required.G/ is an injective homomorphism of representations. .R/ and all R:
For any v 2 V . vi.5a).rV .xg/ u . We give a second proof of this.hu1 . rV .G/ is a direct sum of copies of O. P ROPOSITION 7. P ROOF.g/vi D hu. and so V0 ˝ O. V embeds into a ﬁnite sum of copies of the regular representation.
2
P ROPOSITION 7.v// .
. and let V0 denote the underlying vector space with the trivial representation.R/ and x 2 G.G/.rA .R/ ! Rn
of (74) with the map “evaluate at 1” is v 7! .1/ u1 . vi. Let be the coaction of O.
2
Thus. : : : . un . : : : .35 Let V be a ﬁnite-dimensional representation of G.G/ on V .v/. un span V _ . then the k-linear map v 7! .86 P ROOF.R/ ! An . which is injective by our choice of the ui ’s.
u /R ı rV .x/rV .g/v/R . We have to show that .x/ D
u .v//R .LHS.A ˝ A/n
says exactly that the map W V ! V0 ˝A is homomorphism of comodules. and so the composite V .g/
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
D rA .RHS.

V.ej / D i 2I i 2I 0
Then . 2 L EMMA 7.ei /i 2I 0 for V and V 0 . aen /W A ! V n is injective because V is faithful. ei ˝ aij : V . note that AV ˝AV 0 is the sub-coalgebra attached to the A˝Acomodule V ˝ V 0 .i 0 /2I ˝I 0 is a basis for V ˝k V 0 . Representations of afﬁne groups
87
7j
Every faithful representation contains all
Recall that. Any CV -comodule .ae1 .ej / D V 0 .
i 0 .)2 L EMMA 7. and P 0 V ˝V 0 . it sufﬁces to prove the following statement: let A be a ﬁnite k-algebra and let V be a ﬁnite-dimensional faithful left A-module. Choose k-bases .7. it is ﬁnitely generated.37 Let A be a bialgebra. We may replace C with CV .6). Every ﬁnite-dimensional CV -comodule (considered as a C -comodule) is isomorphic to a quotient of a sub-comodule of V n for some n. and it is the smallest such sub-bialgebra.36 Let V be a ﬁnite-dimensional C -comodule. : : : .i. : : : . But W is isomorphic to a quotient of the free module Am for some m.39 Let A be a bialgebra. / over a coalgebra (or bialgebra) C . By duality (7. and let V and V 0 be ﬁnite-dimensional A-comodules. For any ﬁnite-dimensional C -comodule V . CV _ D S CV . P ROOF.W. j 2 I i AV 0 D haij j i. then a 7! . and so it sufﬁces to prove that A is isomorphic to a submodule of V n for some n. en span V as a k-vector space. If e1 . As AV D haij j i.i 0 .aij ai 0 j 0 / (see 7d).
0 P ROOF.ei ˝ ei 0 /. for a comodule . and that AV ˝V 0 is the image of this by the multiplication map mW A ˝ A ! A. P ROOF. (Alternatively.V / D AV ˝n A:
n 0 2
Then A. then every ﬁnitedimensional A-module W is isomorphic to a quotient of a submodule of V n for some n. and write X X 0 0 0 ei ˝ aij . j 0 2 I 0 i. Let X A. P ROPOSITION 7. j 2 I.6). Then AV ˝V 0 D AV AV 0 . and let V be a ﬁnite-dimensional A-comodule.
. W / can be regarded as a C -comodule with the coaction W ! W ˝ CV
W
W ˝. j 2 I 0 i
0 AV ˝V 0 D haij ai 0 j 0 j i.38 Let C be a Hopf algebra (with inversion S).ei ˝ ei 0 / ˝ . and we showed that it is a sub-coalgebra of C (see 7.V / is a sub-bialgebra of A containing AV and 1.ej ˝ ej 0 / D i.ei /i 2I and .!
! W ˝ C:
L EMMA 7.
the statement is clear. This follows easily from the deﬁnition of the contragredient V _ (see 7d). we deﬁned CV to be the smallest subspace of C such that .V / V ˝ CV .

G/V and 1 and stable under S . it is ﬁnitely generated.36.V ˚ V _ /. G D lim GV S because A D A.G 0 / is a sub-bialgebra of O.W ˝ R/ D W ˝ Rg is an afﬁne subgroup of G.. P ROPOSITION 7. we let MV denote the quotient afﬁne monoid of M with coordinate ring A.42 A representation rW G ! GLV is faithful if and only if V is a faithful O. then every ﬁnitedimensional representation of G is a quotient of a sub-representation of a direct sum of representaN N N tions n .38) and (7. P ROOF.
7k
Stabilizers of subspaces
P ROPOSITION 7.e.V ˚ V _ /.V / 1 ). It follows from the lemma that AV ˝n D AV
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
AV
(n factors). Moreover. we let GV denote the quotient afﬁne group of G with coordinate ring A. Similarly. and so r is a faithful representation. i.V / is the subalgebra of A generated by AV and 1. This follows from Lemma 7.
. Then O.G/ containing O.V /. M D lim MV . D EFINITION 7. P ROOF.V /.G/comodule. The functor def R GW .V / is trivial.V ˚ V _ / is a sub-bialgebra of A containing AV and 1 and stable under S. It may happen that GV is closed in EndV (and not only in GLV ).G/. Let G 0 be the image of G in GLV . and it is the smallest such sub-bialgebra. If V is faithful.R/ j g.44 Let G ! GLV be a representation of G.V / D C.G 0 / D O.V / because A D V AV (see 7.41 An A-comodule V is said to be faithful if A.
2 2
When M is an afﬁne monoid with coordinate ring O.
2
and so it is clear that A.9). then Proposition 7.40 The algebra A. This follows easily from (7. when G is an afﬁne group. S S Note that A D V A.V ˚ V _ / D A.40 shows that it O.R/ D fg 2 G.40). P ROOF.88 P ROOF. therefore if V is a faithful comodule. Now assume A has admits an inversion S. 2 P ROPOSITION 7.M / D A.
P ROPOSITION 7.43 Let G ! GLV be a representation of G. and let W be a subspace of V . that C. The converse is similar. Both MV and GV act faithfully on V . This is the case if GV SLV — then det.V ˚ V _ / (and hence also of a direct sum of representations n V m det.

G/:
Let g 2 G.45 Let G ! GLV be a representation of G.
. Let . If v D 0. Let H be an algebraic subgroup of G. 2 The afﬁne group GW is called the stabilizer of W in G.46 (C HEVALLEY ) Every algebraic subgroup of an algebraic group G is the stabilizer of a subspace in some ﬁnite-dimensional representation of G.O.
T HEOREM 7.ej / D P
i 2J tI ei
89
˝ aij .W ˝ R/ W ˝ R if and only if g. R/. Then H is the stabilizer of the subspace def W D a \ V in V .G/.R/ D fg 2 G.v ˝ 1/ D v ˝ 1g
def
˝ aij .G/ as a k-algebra. Then H is the stabilizer of a in the regular representation of G on O. Write .G/.ai i0 / D 1 if i D i0 0 otherwise.aij /:
Thus.
aij 2 O.aij / D 0 for j 2 J .ei /i 2J be a basis for W .H /.R/ ﬁxes v ˝ 1 if and only if g. The functor R is an afﬁne subgroup of G. and let v 2 V . P ROPOSITION 7. Then (see 7. there exists a ﬁnite-dimensional k-subspace V of O.R/ j g.g/ (7)).aij /R . and let a be the kernel of O.G/ by the ideal generated by faij j j 2 J.aij / D .ei /i 2I for V with ei0 D v for some i0 2 I . 2 L EMMA 7. then Gv D G and there is nothing to prove.ej / D P
i 2J tI ei
Gv . the subspace can even be taken to be one-dimensional and the representation faithful. P ROOF. Then the def V stabilizer of W in G is equal to the stabilizer of D D d W in G. P ROOF.G/ ! O.G/ that is G-stable and contains a generating set for a as an ideal and a generating set for O.7. and extend it to a basis .2
Therefore Gv is represented by the quotient of O. According to (7.G/ by the ideal generated by fai i0 The afﬁne group Gv is called the isotropy or stablity group of v in G. g. Write . ıi i0 j i 2 I g. choose a basis . Hence GW is represented by the quotient of O.47 Let V be a G-module. i 2 I g (recall that g. Representations of afﬁne groups P ROOF.R/ D Homk-alg .G/:
An element g 2 G. i 2 I .16).17) gej D P
i 2J tI ei
˝ g. That W can be taken to be one-dimensional follows from the next lemma.ei /J tI for V .
aij 2 O. and let W be a G-submodule of V of dimension d . Otherwise.

: : : .C /. Let D be a sub-coalgebra of C . if . We let D _ denote the full subcategory of Comod. H is the stabilizer of a line D in a ﬁnite-dimensional representation V of G. we can take V 0 D V _ . P ROOF. 2
7l
Sub-coalgebras and subcategories
Let C be a coalgebra over k. : : : . Let S be a replete subcategory of Comod. d g/. and so. em is a basis for W \ gW . In this case. if . Let D _ be the dual of D with H acting contragrediently. we have to show that:
. choose a basis e1 .e1 ^ ^ em ^ ^ ed / D c e1 ^ ^ em ^ ed C1 ^ ^ ed
2
V for some c 2 k and so. then V becomes C -comodule with the coaction V ! V ˝ D . in case (a).V.! V ˝ C: We get in this way an exact fully faithful functor Comod.50 The map D 7! D _ is a bijection from the set of sub-coalgebras of C onto the set of replete subcategories. and e1 . According to Chevalley’s theorem.49 A full subcategory of an abelian category is replete if it is closed under the formation of ﬁnite direct sums. If gW D W . In this case. : : : ed is a basis for W . : : : . e2d m is a basis for gW . (b) every character of H deﬁned over k extends to a similar character of G. : : : . It is obvious that D _ is replete. and let Comod. then H will be the subgroup of G ﬁxing any nonzero vector in D ˝ D _ V ˝ V 0 . If we can ﬁnd a representation V 0 of G containing D _ as an H -stable subspace. which in case (b) extends to a character of G. and the inclusion functor is exact.
V 2S
To prove the theorem. then m D d and gW D W . we can take V 0 D D _ . / is a comodule over D. : : : .48 A subgroup H of an algebraic group G is the subgroup of G ﬁxing a vector in some faithful ﬁnite-dimensional representation of G in each of the following two cases: (a) all the representations of H are semisimple. P ROOF. In particular. T HEOREM 7. The action of H on D deﬁnes a character of H . A replete subcategory is an abelian category. then certainly . Certainly D _ occurs as a quotient of V _ .D/ ! Comod. D EFINITION 7. As we noted earlier. subobjects. em . Let g 2 GL. d g/.C / denote the category of ﬁnite-dimensional right C -comodules. en for V such that e1 .C /.S/ D CV (sub-coalgebra of C ).
C OROLLARY 7. and let X C.D/ D. For the converse. and quotient objects.C / whose objects are isomorphic to a comodule in the image of this functor. it also occurs as a direct summand of V _ (regarded as an H -module). e1 .90
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
V P ROOF.V /. ed C1 . em . Then V . d g/.D/ D D. every object isomorphic to an object in a replete subcategory also lies in the subcategory.

recall that D D DV and that DV DV 0 D DV ˝V 0 (see 7. D contains DV0 D k. w 0 2 W 0 .G/. Because D _ contains V0 D k.v/ D c 1 v. We let Q_ denote this subcategory of Rep. T HEOREM 7. (b) Use the formula AV _ D S.
P ROPOSITION 7. and write V D hvi ˚ V 0 . where hv ˝ wi D hvi ˝ hwi. The essential image of the functor consists of the representations of G containing Ker. W D hwi ˚ W 0 :
.53 Let ˛ and ˇ be endomorphisms of vector spaces V and W respectively. (a) If D is a subalgebra of A. and let v and w be nonzero vectors in V and W . Rep. Let v and w be nonzero elements of vector spaces V and W .51 Let A be a bialgebra over k. Representations of afﬁne groups ˘ C. A representation rW Q ! GLV deﬁnes a representation G ! Q ! GLV of G. (b) Assume A has an inversion S.
91
2
The ﬁrst statement follows from the fact that D is a union of its subcoalgebras DV (see 7. v 0 2 V 0 . and the dictionary between Hopf algebras and their comodules and afﬁne groups and their representations.G ! Q/ in their kernel. In particular. and ˘ C. Then V ˝ W D hv ˝ wi ˚ hvi ˝ W 0 ˚ V 0 ˝ hwi ˚ V 0 ˝ W 0 . (a) A sub-coalgebra D of A is a sub-bialgebra of A with 1 if and only if D _ is stable under tensor products and contains the trivial comodule k.v/ D cv and ˇ. ˇw D bw C w 0 . and the second follows from (7. and so D is closed under products.G/ denotes the category of ﬁnite-dimensional G-modules. For the converse.G/.AV /.37). P ROOF.G/ closed under the formation of tensor products (including the empty tensor product) and under passage to the contragredient.9). Obvious from (7.D _ / D D for all sub-coalgebras D of C .7.Q/ ! Rep.C /. we see that v ˝ w ¤ 0 (because hvi ˝ hwi ¤ 0). Now write ˛v D av C v 0 .52 The map Q 7! Q_ is a bijection from the set of isomorphism classes of quotients of G to the set of replete subcategories of Rep.S/_ D S for all replete subcategories S of Comod.51). S certainly D _ is stable under tensor products and contains then k. then there exists a c 2 k such that ˛. P ROOF. Let r G ! Q be a quotient of G. We get in this way an exact fully faithful functor Rep. A sub-bialgebra D is stable under S if and only if D _ is stable under the contragredient functor. P ROOF.50). (7. 2
7m
Quotient groups and subcategories
For an afﬁne group G over k. 2
7n
Normal subgroups and subcategories
L EMMA 7. If v ˝ w is ﬁxed by ˛ ˝ ˇ.36).

and let V N be the subspace on which N acts trivially. as required. and let g 2 G. as required.R/.46).53 shows that the stabilizer of L0 D hv ˝p i is still N . rR0 . Suppose ﬁrst that G is smooth.V.v ˝ w/.v ˝ w/ C av ˝ w 0 C v 0 ˝ bw C v 0 ˝ w 0 : If . Let U be the smallest subspace of V stable under G.. we see that U decomposes into a direct sum U D 2 U in which U is a subspace of U on which N acts through the character . r. (75)
2
L EMMA 7. r/ be a representation of G.V.rR .˛ ˝ ˇ/.11) shows that if V is a sum of simple representations. P ROOF.v ˝ w/ D ab.n0 /w D r. then it is a direct sum of such representations.n/. the stabilizer of V N is G).g/w/R0 . It sufﬁces to prove this after an extension of the base ﬁeld k.
2
By a character of an afﬁne group G. but it is not ﬁxed by N 0 (since the line L is not ﬁxed by N 0 and we can apply Lemma 7. and so we may assume that k is algebraically closed.v ˝ w/ D ab. Replace V by V ˝p and v by v ˝p . In particular.g/v 2 V N . A standard argument (GT 7. (75) implies that w 0 D 0 D v 0 . The action of N on L0 factors through its
.k/.V ˝ U _ /N (which Lemma 7. For any g 2 G.gn0 /w D r. Let .R0 /.g/w/ D r.54 Let N be a normal afﬁne subgroup of an afﬁne group G. for any R-algebra R0 and n 2 N. then ab D 1 and av ˝ w 0 D 0 D v 0 ˝ bw: As av ¤ 0 ¤ bw. Let N act on the line L D hvi through the character 0 . and so (7. N acts on the line gL through the conjugage character g 0 .G 0 / is reduced (see 5. A one-dimensional representation of G deﬁnes a character of G. Let w 2 V N .g/w/R0 D . Lemma n def 7. P ROOF.g/w: When N.rR . r/ of G and a line L in V such that N is the stabilizer of L. we mean a homomorphism G ! Gm . then the characteristic of k is p ¤ 0. which implies that G stabilizes V when G. In the general case. The same argument shows that.55 For any distinct normal afﬁne subgroups N N 0 of an afﬁne group G.e.k/ is Zariski dense in G.54 shows to be a representation of G).R/ for some k-algebra R.n/.92 Then
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
. For any n 2 N. this implies that r.G/p D O. we let w 2 V N .r.ng/w D r. there exists a representation of G on which N acts trivially but N 0 acts nontrivially. U contains a line L D hui on which N acts through 0 . I claim that N 0 acts nontrivially on W D . there exists a representation . U _ decomposes into a direct sum L _ 1 _ _ U_ D 2 U 1 .g/r. A one-dimensional representation is obviously simple. According to Chevalley’s theorem (7. and there exists an n n n n such that O.k/.˛ ˝ ˇ/. Correspondingly.19) shows that rR .R/ and g 2 G. P ROPOSITION 7. As U is spanned by the L lines gL.k/ is Zariski dense in N . Then V N is stable under G (i. If G is algebraic but not smooth.g/w 2 V N .53).27). Now v ˝ u is ﬁxed by N .k/.

the kernel of G ! Q is the normalizer of H .57 Let N be a normal afﬁne subgroup of an afﬁne group G. This implies that N D N 0 . In particular.G/N D Q_ . and the theorem implies that N is the kernel of G ! Q. P ROOF. 2 T HEOREM 7.G/N of Rep.62 This subsection needs to be rewritten. and the kernel N of G ! Q is a normal subgroup of G. well-deﬁned up to isomorphism. Waterhouse 1979. 2 T HEOREM 7.60 Perhaps prove: let H be a subgroup of G. there exists a quotient map with kernel N . Then Rep. P ROOF. 0 0 (: Let N 0 be the kernel of G ! Q.46.63 For a representation G ! GLV . The previous argument shows that U _ contains n a line L0_ D hu0 i such that v ˝p ˝ u0 is ﬁxed by N but not by N 0 . A SIDE 7.G/N ).61 Add a discussion of the correspondence between normal subgroups of an afﬁne group G and normal Hopf algebras (Abe 1980.V. and to which we can apply the previous argument.G ! Q/ if and only if Rep. the sketch of proof of (6. 2 C OROLLARY 7. then Lemma 7. 16. ): According to Theorem 6. The subcategory Rep.4. 2 L EMMA 7. The maps S 7! N and N 7! Rep.55 shows that there exists a representation . for any ﬁeld k 0 G 0 . Then N D Ker.G/N are inverse.G/N D Q_ . there will be an algebraic quotient in which the images of N and N 0 are distinct. add a deﬁnition of V G .G/N1 but not of Rep.55) needs to be rethought.G/N2 then N1 D N2 .G/N2 .3.52).V G /k 0 ' Vk 0 k . p123). 2
A SIDE 7. and let Q be a quotient of G.59 For any normal afﬁne subgroup N of an afﬁne group G. the proof of (7. This implies the existence of normalizers (and perhaps centralizers). P ROOF.G/N D Rep. p179).7. Representations of afﬁne groups
93
quotient in G 0 (cf. A SIDE 7. P ROOF. Show that.G/ closed under tensor products and passage to the contragredient. A SIDE 7. and let Q be the quotient of G such that Q_ D Rep. If Rep. Let S be a replete subcategory of Rep. and also of the correspondence between normal Hopf ideals and Hopf subalgebras (ibid.G/H . When G is not algebraic. in the case that k is algebraically closed and the Hopf algebras are assumed to be reduced).G/ closed under tensor products and passage to the contragredient. The S D Q_ for some quotient Q of G.58 The map N 7! Rep. which contradicts the hypothesis. r/ of G and a v 2 V ﬁxed by N1 but not by N1 N2 . Then V N1 is an object of Rep.7.G/N is a bijection from the set of normal afﬁne subgroups of G to the set of replete subcategories of Rep. Therefore Rep. Cf. p207. a representation rW G ! GLV factors through Q (and so lies in Q_ ) if and only if r maps N to 1 (and so lies in Rep. and so Rep.
. k.56 Let N1 and N2 be normal afﬁne subgroups of G.G/N1 D Rep. 4.G/N .G/ is replete and closed under tensor products and passage to the contragredient. If N1 ¤ N2 .G/N D Q_ for some quotient Q of G.

V ˚ V _ / D A. Serre 1993.V. Show that.5. p.94
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
7o
Exercises
EXERCISE 7-1 Let . for a suitably deﬁne “determinant” element ıV of A. / be a comodule over a bialgebra A. A.61).V /Œ ı1 (cf. V
. 3.

18 8. in which case the map G=H ! . N is a normal subgroup of HN . The image of a homomorphism ˛W G ! G 0 is a subgroup ˛G of G 0 . 8. and ˛ deﬁnes an isomorphism from G= Ker.4 (Correspondence theorem). every homomorphism is the composite of a quotient map and an embedding.G=N /=. an embedding (injective homomorphism). we show that the isomorphism theorems in abstract group theory extend to afﬁne groups.59. appropriately interpreted.H \ N / 7! hN W H=H \ N ! HN=N is an isomorphism.2 (Homomorphism theorem). and third isomorphism theorems. A subgroup H of G containing N is normal if and only if H=N is normal in G=N . we have the notions of subgroup. which are often referred to collectively as the isomorphisms theorems. in particular.
18 Statements
2
95
.3 (Isomorphism theorem). P ROOF. H \ N is a normal subgroup of H . Thus the normal subgroups of an afﬁne group are exactly the kernels of homomorphisms. to algebraic groups).4) are sometimes called the ﬁrst. we shall see that.4) and the second is (8.2).H=N / deﬁned by the quotient map G ! G=N is an isomorphism. 8. and every normal subgroup N of G arises as the kernel of a quotient map G ! G=N .3). there exists a quotient map with kernel N . all these notions and statements extend to afﬁne groups (in particular.
(8.˛/ onto ˛G. This is a restatement of Theorem 7. second. The map H 7! H=N deﬁnes a one-to-one correspondence between the set of subgroups of G containing N and the set of subgroups of G=N . but the numbering varies. (8.
8b
The existence of quotients
T HEOREM 8. Let N be a normal subgroup of G. 8.8
Group theory: the isomorphism theorems
In this section. a normal subgroup. and of a quotient map (surjective homomorphism). In Noether 1927. and the map h. then HN is a subgroup of G. Let H and N be subgroups of G such that H normalizes N . there are the following basic results. the ﬁrst isomorphism theorem is (8.
8a
Review of abstract group theory
For a group G (in the usual sense). and (8.1 (Existence of quotients). The kernel of a quotient map G ! Q is a normal subgroup of G. In this section.3).5 For every normal subgroup N of an afﬁne group G. Moreover.

R0 / is in the image of PSLn . therefore. just that it is an ideal.T n a/ in which it does become an nth power.2a) and (6. the image of g in PGLn .96
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
E XAMPLE 8.g0 / D 1.2a). I claim that (76) is an isomorphism of algebraic groups.R/.P // D 0 for all k-algebras R and all P 2 G.40).k/
(77)
is injective.R0 /.
8c
The homomorphism theorem
A homomorphism ˛W G ! G 0 of afﬁne groups deﬁnes a homomorphism ˛ W O.R0 / mapping def to q.R/. The image of g in GLn .k/=
n . If a is not an nth power in R0 . For some faithfully ﬂat R-algebra R0 . the image of q in PGLn . but not in general surjective: not every invertible n n matrix can be written as the product of a matrix with determinant 1 and a scalar matrix (such a matrix has determinant in k n ). and the image of g in GLn .R/ ! PGLn .g/ D 0 for f 2 ag.R0 /=Gm . then g D g0 t with det. In the general case.k/
(76)
! GLn .
.˛R .R0 / x ? ? PSLn .6 Let PGLn be the quotient of GLn by its centre. Let q ¤ 1 2 PSLn .R0 / is ¤ 1.19 Thus a D ff 2 O.g/ is an nth power. there exists a g 2 SLn .R/g: The afﬁne subgroup H of G 0 corresponding to a (see 6. this follows from the fact that (77) is an isomorphism when k D k al (apply 6. and so deﬁnes a homomor-
PSLn ! PGLn : Is this an isomorphism? Note that SLn .R/ j fR .40).G 0 /.R0 / (because q ¤ 1/.R0 / is in the image of SLn .7) is called the image of ˛ (and often denoted ˛G).R0 /. Let q 2 PGLn .44). and let PSLn be the quotient of SLn by its centre: PGLn D GLn =Gm . which implies that the image of q in PGL. PSLn D SLn = n : The homomorphism SLn ! GLn ! PGLn contains phism
n
in its kernel.k/=Gm .R0 / mapping to q in PSLn . Thus H.R0 /.G/ of Hopf algebras. we replace R0 by the faithfully ﬂat (even free) algebra R0 ŒT =.R0 / x ?injective ? ! PGLn .G 0 / j fR . say a D t n .R/:
We have checked condition (6. We have checked condition (6. Hence.R0 / is not in Gm .R0 /= n .
19 In
fact. In characteristic zero.R/ D fg 2 G. Nevertheless.16 and 6. we have to check the conditions (6. If a D det.G 0 / ! O. whose kernel a is a Hopf ideal in O. we don’t need to use that a is a Hopf ideal.R/ is ¤ 1: PSLn . For some faithfully ﬂat R-algebra R0 . there exists a g 2 GLn .

k 0 / in G 0 . and ˛ factors in a natural way into the composite of a surjection. there is a unique isomorphism G=N ! ˛G such that the composite G ! G=N ! ˛G is G ! ˛G (apply 6.R0 / 0
R ˛ 2
O.G.9 A homomorphism ˛W G ! G 0 of algebraic groups is surjective if.R/ \ Im ˛.k/ is dense in G 0 .R/
Moreover.R/ \ Im ˛. the image ˛G of ˛ is a subgroup of G 0 .k al /
.
2
Let ˛W G ! G 0 be a homomorphism of algebraic groups.k/ \ Im. Then G. 0
R
and so H
˛G.R/.R/ for all k-algebras R.26). As ˛. the image of G.k/ D G 0 .
Therefore ˛G represents the sheaf associated with R Im.60).˛.˛G/.R//: H. for any ﬁxed k-algebra R. P ROOF.R/ for all k-algebras R. C OROLLARY 8.R/ D G.k al / is surjective (see 6. an isomorphism.k/ \ .k al / ! G 0 . and so .˛G/.
2
C OROLLARY 8.˛G/.R0 / D . If H is an afﬁne subgroup of G 0 such that H.
P ROOF.k 0 // . The ﬁnal statement follows from (7. then.˛G/.8 For any k-algebra R.G 0 /=a
O. P ROOF.8.44). and so the ﬁrst statement follows from (6.47). As G ! G=N and G ! ˛G are both quotient maps with kernel N .R/ G. the kernel N of ˛ is a normal subgroup of G. The map G ! ˛G is a quotient map. for some ﬁeld k 0 containing k.k al / ! . then so also are G=N and ˛G.k 0 /. The factorization O.k al //:
˛. Group theory: the isomorphism theorems
97
T HEOREM 8.k al / D G 0 . the condition implies that ˛G D G.k 0 / G.G. [ . ˛G is the intersection of the subgroups H of G 0 with the property that Im ˛.7 For any homomorphism ˛W G ! G 0 of afﬁne groups.˛G/. and an injection: G ? ? surjectivey G=N
˛
! G0 x ?injective ? ! ˛G:
isomorphism
If G is an algebraic group.G/ of ˛ deﬁnes a factorization G ! ˛G ! G 0 of ˛ into a surjection followed by an injection. [ H.˛G/.G 0 /
(R0 runs over the R-algebras).R/ Im ˛.

11 We have .HN /.R/: Moreover. 8.k al /.k/ D G.R/ and N.R/:
2
(78)
The statement now follows from (6.HN /. We deﬁne NH D HN to be the image of this homomorphism. The natural map H=H \ N ! HN=N is an isomorphism.R/ on N. and N D Gm (scalar matrices in G). The following statements are obvious from 8c.R/ whose determinant is not an nth power is not the product of a scalar matrix with a matrix of determinant 1. for all k-algebras R.HN /. Then GLn D SLn Gm . and (78) becomes the isomorphism SLn =
n
n
! GLn =Gm :
.98
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
8d
The isomorphism theorem
Let H and N be algebraic subgroups of G such that H normalizes N .k al / N. and so .k/ \ . HN is the interesection of the subgroups G 0 of G such that.R/N.12 It is not true that .R/ for all k-algebras R. but a matrix A 2 GLn . or by passing to the associated sheaves.R/ deﬁnes an action Â of H on N by group homomorphisms. We have an isomorphism of group-valued functors H. and multiplication deﬁnes a homomorphism N Â H ! G.k al / D H. consider the algebraic subgroups SLn and Gm (nonzero scalar matrices) of GLn .R/N.R/ contains both H.10 For any k-algebra R.k al / N.R0 /N. P ROOF. The natural action of H.R/ ! H.R/=. E XAMPLE 8.R/ D H.H \ N /. H D SLn .k al //:
A
8. 8.R/ consists of the elements of G.H.R/=N.HN /.R/=N.HN /.R/ that lie in H. Then N \ H D (obviously).13 Let H and N be algebraic subgroups of the algebraic group G such that H normalizes N .60).R/ G. Therefore HN represents the sheaf associated with the functor R H.R0 / for some ﬁnitely generated faithfully ﬂat R-algebra R0 . . T HEOREM 8.R/ .R/ N.14 Let G D GLn . For example.R/. HN D GLn (by the arguments in 8. G 0 .6).

(80)
.H1 \ N2 /.H1 \ H2 / \ N1 .R/ ! . The ﬁrst of Dedekind’s modular laws (Exercise 6-3a) with A D H1 \ N2 .H1 \ N2 / (see Exercise 6-2).R/=H. For any subgroup H of G.H2 \ N1 / is normal in N2 . Group theory: the isomorphism theorems
99
8e
The correspondence theorem
T HEOREM 8.N1 \ H2 / are normal algebraic subgroups of the algebraic groups N1 .8.H1 \ N2 / .H1 \ H2 / . For the second statement.17 (B UTTERFLY L EMMA ) Let H1 N1 and H2 N2 be algebraic subgroups of an algebraic group G with N1 and N2 normal in H1 and H2 .G=N /=.R// deﬁned by the quotient map G.H2 \ H1 / respectively. B D H1 \ H2 . The map H 7! H=N deﬁnes a one-to-one correspondence between the set of algebraic subgroups of G containing N and the set of algebraic subgroups of G=N .H1 \ H2 \ N1 / D . The subgroup H1 \ H2 of G normalizes N1 .R//=. The ﬁrst statement follows from the fact that the analogous statement holds for Hopf algebras (cf.H1 \ H2 / and N2 .N1 \ H2 / P ROOF.H1 \ N2 /. This isomorphism is natural in R.H1 \ H2 / \ N1 . we have that H1 \ H2 D . and so the isomorphism Theorem 8. Similarly.R/=N.
(79)
8f
The Schreier reﬁnement theorem
L EMMA 8. Then N1 .H=N / deﬁned by the quotient map G ! G=N is an isomorphism.R/ ! G.H1 \ N2 / D .H2 \ H1 /.H1 \ H2 / ' N1 .N1 \ H2 /.H1 \ N2 / and N2 .G.H1 \ N2 / N1 .H.H1 \ H2 / D N1 . and C D N1 becomes . in which case the map G=H ! .R/=N.H1 \ N2 /.H1 \ N2 / As H1 \ N2 H1 \ H2 . and when we pass to the associated sheaves.H1 \ H2 / N1 .H1 \ N2 /. P ROOF. note that the map G. Deduce that if H 0 is normal in H .15 (Third isomorphism theorem). and so N1 . An algebraic subgroup H of G containing N is normal if and only if H=N is normal in G=N . then H 0 N is normal in HN .H1 \ N2 / N2 . which corresponds to HN . we obtain the isomorphism (79). Exercise 4-10). 2
A SIDE 8.H1 \ H2 / N2 . The algebraic group H1 \ N2 is normal in H1 \ H2 and so N1 . N2 .R/=N.H1 \ H2 / is normal in N1 .R/ is an isomorphism. and there is a canonical isomorphism of algebraic groups N1 .H1 \ N2 / H1 \ H2 ' : .13 shows that .16 Let qW G ! G=N be the quotient map. Let N be a normal algebraic subgroup of G. qH is a subgroup of G=N .H1 \ H2 / .

.i / =H .Hj i / of .H1 \ N2 / A symmetric argument shows that H1 \ H2 N2 .H2 \ N1 / and so N1 . and so the reﬁnement . .N1 \ H2 / N1 .H1 \ N2 / .i /C1 .H2 \ N1 /
2
A subnormal series in an algebraic group G is a ﬁnite sequence of subgroups.
2
A subnormal series is a composition series if no quotient group Gi has a proper nontrivial normal subgroup.i C1 . : : : .N1 \ H2 / N2 . sg such that Gi =Gi C1
are said to be equivalent if s D t and there is a permutation H .H1 \ H2 / ' .Hi /.H1 \ N2 / N2 . T HEOREM 8. only one of the quotients Gi C1 .Gi / is equivalent to the reﬁnement .100 Therefore (80) is an isomorphism
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
H1 \ H2 N1 . Let Gij D Gi C1 .Hj C1 \ Gi / is nontrivial 2
. such that each subgroup is normal in the preceding subgroup. 2. According to the butterﬂy lemma Gij =Gi.i /C1
of f1.H1 \ H2 / ' : N1 .18 Any two subnormal sequences in an algebraic group have equivalent reﬁnements. 2.i / =H .j C1 ' Hj i =Hj. Two subnormal sequences G D G0 G D H0 G1 H1 Gs D f1g Ht D f1g of f1.H1 \ H2 / N2 .Gij / of . for each i . : : : .H1 \ N2 / . P ROOF.Hj \ Gi / and let Hj i D Hj C1 .Gi \ Hj /. beginning with G and ending with 1.
T HEOREM 8. sg such that Gi =Gi C1 is isomorphic to H
P ROOF.19 For any two composition series G D G0 G D H0 s D t and there is a permutation for each i.H1 \ H2 / ' : .Hj \ Gi /=Gi C1 . G1 H1 Gs D f1g Ht D f1g. Use that.

HN / is equal to the kernel of the composite O. Moreover. The importance of allowing nilpotents was pointed out by Cartier (1962) more than forty years ago. Every morphism in the category has both a kernel and cokernel (6.21 Fortunately. “algebraic group” means “smooth algebraic group”. except for Demazure and Gabriel 1970 and Waterhouse 1979. 4. Humphreys 1975. many of the results in this section become false. Show that the kernel of O. in much of the expository literature (e. then Nred need not arise from a smooth algebraic group over k. and in Takeuchi 1972.355.
A SIDE 8. For example.21 is proved purely in the context of Hopf algebras in Sweedler 1969. it is a fairly direct consequence of allowing the coordinate rings to have nilpotent elements. which sometimes makes it difﬁcult to understand their statements.20.g. the kernel of ' need not be deﬁned over k. this point-of-view has not been adopted in the expository literature.G/ ! O.
. the product G1 G2 of two commutative algebraic groups is both a product and a sum of G1 and G2 . because these books use a terminology based on Weil’s Foundations. See SGA3. With this terminology. with our (or any reasonable) deﬁnitions.5).22 Theorem 8.21 The ﬁnitely generated co-commutative Hopf algebras over a ﬁeld form an abelian category. but.16.24. 8. and the canonical morphism from the coimage of the morphism to its image is an isomorphism (homomorphism theorem. Contrast our treatment of the isomorphism theorems and the Schreier reﬁnement theorem with that in Rosenlicht 1956.G/ ˝k O. 2 C OROLLARY 8. p.4. Chapter XVI. Borel 1991.H / ˝k O.G/ ! O. Corollary 8. one ﬁnds the following statement: “for a homomorphism 'W G ! G 0 of k-groups. Group theory: the isomorphism theorems
101
8g
The category of commutative algebraic groups
T HEOREM 8. for ﬁnite-dimensional co-commutative Hopf algebras. P ROOF.218. 8.” By this. The Hom sets are commutative groups.13. the kernel of a homomorphism of algebraic groups over k is certainly an algebraic group over k. he means 0 the following: form the kernel N of 'k al W Gk al ! Gk al (in our sense). for ﬁnitely generated cocommutative Hopf algebras. Of course. Thus the category of commutative algebraic groups over a ﬁeld is additive. 21 The situation is even worse.4.20 is generally credited to Grothendieck but.N /:
N OTES As noted earlier. in the category of smooth groups.G/ ! O.7). because of Theorem 5. as we have seen.. Springer 1998). 7.8. this is only a problem in nonzero characteristic. p. in Humphreys 1975. the homomorphism H=H \ N ! HN=N is a purely inseparable isogeny of degree q where q is the multiplicity of H \ N in the intersection product H N .20 The commutative algebraic groups over a ﬁeld form an abelian category. Therefore the category is abelian. DG III 3. VIA . 5.
8h
Exercises
E XERCISE 8-1 Let H and N be subgroups of the algebraic group G such that H normalizes N . and the composition of morphisms is bilinear.
20 For example.

x. Clearly. I mean a continuous homomorphism from the group to the circle group fz 2 C j z z D 1g. Let ˛ be a k-linear map A ! A for which the three diagrams k
e
! A ? ?˛ y ! A
A ˝k A ? ? y˛˝˛ A ˝k A
m
! A ? ?˛ y ! A. .R/:
(82)
The commutativity of the third diagram says that. Recall that g 2 G. More generally. and so. the dual of a compact abelian group is a discrete abelian group. In this section. the canonical homomorphism G ! G __ is an isomorphism (Pontryagin duality). again.x y/ for f 2 A (see 4h). f /R . because any character will be trivial on the derived group. y/ D fR .. A ﬁnite abelian group G can be recovered from its group G _ of N characters because the canonical homomorphism G ! G __ is an isomorphism.x/ D fR .1 Let G be an afﬁne monoid over a ring k.x.R/.RHS/R .
A ? ?˛ y A
! A˝A ? ? y1˝˛ ! A˝A
k
e
m
commute. “abelian” is required in the above statements. Tannaka showed that it is possible to recover a compact nonabelian group from its category of unitary reprsesentations. The commutativity of the ﬁrst two diagrams says that ˛ is a homomorphism of A as a k-algebra. and so (by the Yoneda lemma) there exists morphism W G ! G of set-valued functors such that .x/ D fR .g/. .
R x/
all f 2 A. a locally compact abelian topological group G can be recovered from its character group because. I discuss an analogue of this for algebraic groups. x 2 G.x
. ı ˛/. the study of compact abelian topological groups is equivalent to that of discrete abelian groups. y/ D .R/ acts on f 2 A according to the rule: .LHS/R . for f 2 A. .x. and let A D O. P ROOF.˛f /R . Moreover.gf /R . Then there exists a unique g 2 G. f /R .˛f /R . for x.G/. However. y/ D .x
y//
R y/: 22
D fR .f / D . Jordan decompositions
By a character of a topological group.x g/ all x 2 G.k/ such that ˛ D rA .R/: (81)
L EMMA 9.1 ˝ ˛/ ı /. y 2 G.x y/ D fR .x. and let rA W G ! EndA be the regular representation.9
Recovering a group from its representations. 102
R y/ R .f /: Recall that .
9a
Recovering a group from its representations
Let G be an algebraic monoid with coordinate ring A. Therefore.

x. for x 2 G.˛f /R . Recovering a group from its representations.x/.g/. shows that there exists a g 2 G.x/ D .16) that every representation of G is a ﬁltered union of ﬁnite-dimensional representations.R/:
R .xg/ f 0 .g/f / . Hence ˛ D rA .R/.
22 In
detail. f /R . The map mW A ˝k A ! A is equivariant for the representations r ˝ r and r. It follows that.
P ROOF. applied to the afﬁne monoid GR over R. Now Lemma 9. The conditions (a. y/ D D D D P P P
i i
fi ˝ ˛gi
R .g/ for all V .1) commutes with ˛ replaced by A .idR ˝˛/ ı
V
V.x.xg/ f 0 .x/ D .
V ˝W
D
V
˝
W.
.
2
T HEOREM 9. and so the third diagram in (9. there exists a unique R-linear endomorphism V of R ˝k V such that V jW D W for each ﬁnite-dimensional subrepresentation W V . Recall (7. then .e/.G/. W . Suppose that we are given. y 2 G.x/ .g// . for
.R/ such that
D rV .
23 We
check that.c) will continue to hold for the enlarged family.x.x/ D f . .r. an Rlinear map V W R ˝k V ! R ˝k V .x g/ D .f ˝ f 0 /.r. (a) for all representations V.g/ ı m/ .xg/ D f .g/f 0 /.g/. If the family .x/ D . y/
fiR . The uniqueness of g follows from the faithfulness of the regular representation. Jordan decompositions Hence
R .. Let A D R ˝k O.g/ ˝ r.RHS/R .
(b) 1 D id1 (here 1 D k with the trivial action) 1 1 1 (c) for all G-equivariant maps ˛W V ! W .
i
P
D . for each representation rV W G ! GLV of G (possibly inﬁnite dimensional).y/.
then there exists a unique g 2 G. for each ﬁnite-dimensional representation rV W G ! EndV of G. V / satisﬁes the conditions.f ˝ f 0 /. Similarly.23 and so the ﬁrst two diagrams in (9.ff 0 //.m ı r.G/.2 Let G be an afﬁne monoid (or group) over a ﬁeld k.x/ .
Therefore. we ﬁnd that f 2 A and x 2 G.
i fi ˝ gi R y/: R y/ R y/ R .˛gi /R .y/ fiR .ff 0 /.x.82/
D x g with g D
R .
all x.gf /R .xg/ .r. and let A W A ! A be the R-linear map corresponding to the regular representation r of G on O. let f D
P
fi ˝ gi .r.
W
ı .xg/:
.b.g/.9.1) commute with ˛ and ˛ ˝ ˛ replaced by A and A˝A D A ˝ A respectively. and let R be a k-algebra.x/ D fR .R/ such A D r.idR ˝˛/ D .81)
On setting y D e in the last equation. the map W A ! A ˝ A is equivariant for the representation 1 ˝ r on A ˝k A.x
103
y/ D x
R .1.x/ giR .R/.

V / (because this is true of the family . rV / be a ﬁnite-dimensional representation of G.
u .g/:
As this holds for all u 2 V _ .!/ denote the functor R 7! End˝ .k/. and let End˝ . For each representation . vi
is equivariant (see 7.4 Let .!/. which clearly contains H .5 Let !R be the forgetful functor RepR .!R /.c) of Theorem 9.b. we have a function u on G.g/
D hu.104
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1) Let . instead of all representations of G.R/ ! End˝ .G/ in terms of the representations of G. and the uniqueness follows the fact that G admits a faithful family of ﬁnite-dimensional representations (see 7. For this reason.R/ of course deﬁnes such a family. and let u 2 V _ . 9. 2 We close this subsection with a series of remarks.G/ of representations of G on ﬁnite-dimensional k-vector spaces we can recover G. 9.V.R/ ﬁxing all tensors in all representations of G ﬁxed by H .G/ ! ModR .R/ for any k-algebra R. It is possible to give an explicit description description of O.
u .7 Let H be a subgroup of an algebraic group G.!R / is an isomorphism.G/.g/i 2 k:
Then u 2 O. this can be written G ' Aut˝ . and the family satisﬁes the condition _ V _ D .!R / be the set of natural transformations W !R ! !R commuting with tensor products — the last condition means that satisﬁes conditions (a) and (b) of the theorem. each V is an isomorphism. rV / of G (over k/ and u 2 V _ .g//). and so
uı V
D
Aı u
D r.G/ can be identiﬁed with a ring of k-valued functions on G.29).G/ arises in this way (cf. 9. 9. The functor R H 0 . and hence the group G itself. Now let End˝ .R/ is representable by a subgroup H 0 of G. let H 0 . The theorem says that the canonical map G.k/.6 Suppose that k is algebraically closed and that G is reduced.V. When G is an afﬁne group (rather than just a monoid). V / be a family satisfying the conditions (a.!/. Springer 1998.G/.7).3 Each g 2 G. Thus. this implies that V D rV . and every element of O. For each k-algebra R. 9.V ˚ V _ / (by 7.v/
D hu.g/ ı
u
D
u ı rV .rV .43).2 is often called the reconstruction theorem.
.R/ be the subgroup of G. When G is a group. then G ' End˝ .33). it sufﬁces to choose a faithful representation V and take all quotients of subrepresentations of a direct sum of representations of the form ˝n . 9. and Exercise 4-2).4. rV . p. It follows from the theorem that H0 D H. so that O. from the category Repk .8 In (9.39. This proves the existence of g. Then
uW V
! O.g/. Theorem 9.

then V is a direct sum of its generalized eigenspaces: M V D V ai . if and only if the subring kŒ˛ ' kŒT =. Recall that an endomorphism ˛ of a vector space V is nilpotent if ˛ m D 0 for some m > 0.2 holds with k a noetherian ring and Repk . and that it is unipotent if idV ˛ is nilpotent. Recall that an endomorphism ˛ of a vector space V is diagonalizable if V has a basis of eigenvectors for ˛. Cf 6. From linear algebra.49.V ˚ V _ /˝n is ﬁxed by B. Recovering a group from its representations. ai 2 k:
For each eigenvalue a of ˛ in k.V / generated by ˛ is separable. if ˛ is nilpotent. we can omit it if some nonzero multiple of every homomorphism H ! Gm extends to a homomorphism G ! Gm . Therefore. the generalized eigenspace is deﬁned to be: Va D fv 2 V j . Jordan decompositions
105
9.1 and its proof are valid with k a commutative ring.8). the subgroup B D 0 of GL2 acting on V D k k and suppose that a vector v 2 . we can’t omit “quotients of” from (9. 9. then its minimum polynomial divides T m for some m.9 In general.10 Lemma 9. Clearly. we know that ˛ is semisimple if and only if its minimum polynomial m˛ . in other words. For example. that 0 0 N is sufﬁciently large for the partial ordering M Ä N ” M divides N:
24 Consider
. and so any such map is trivial.
i
˚ « for example.e.11 If ˛ has all of its eigenvalues in k. From linear algebra.T // of Endk . 24 By this I mean that there exists an N such that the statement holds for all positive integers divisible by N .
9b
Application to Jordan decompositions
J ORDAN DECOMPOSITION OF A LINEAR MAP
In this subsubsection.9. or with k a Dedekind domain.G/ the category of representations of G on ﬁnitely generated projective k-modules (or even ﬁnitely generated free k-modules). and so B 0 D B.T ar /nr . But GL2 =B ' P1 . and that it is semisimple if it becomes diagonalizable after an extension of the base ﬁeld k. we review some linear algebra. one sees that Theorem 9. i.m˛ .. Let ˛ be an endomorphism of a ﬁnite-dimensional vector space V over k.10). we know that the converse is also true. and so the eigenvalues of ˛ are all zero. N sufﬁciently divisible25 g:
P ROPOSITION 9. Then g 7! gv is a regular map GL2 =B ! .˛ a/N v D 0. even in k al . the linear map x 7! AxW k n ! k n deﬁned by an n n matrix A is diagonalizable if and only if there exists an invertible matrix P with entries in k such that PAP 1 is diagonal.T / D . G a ﬂat group scheme. Therefore (using 7. and so ˛ is unipotent if and only if its eigenvalues in k al all equal 1. and it is semisimple if and only if there exists such a matrix P with entries in some ﬁeld containing k.V ˚ V _ /˝n of algebraic varieties (not afﬁne). 24 However. We say that ˛ has all of its eigenvalues in k if the characteristic polynomial P˛ .T a1 /n1 . and Repk . v is ﬁxed by GL2 .T / has distinct roots.T / of ˛ splits in kŒX : P˛ .G/ the category of representations of G on ﬁnitely generated k-modules.

106

BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)

P ROOF. Let P .T / be a polynomial in kŒT such that P .˛/ D 0, and suppose that P .T / D Q.T /R.T / with Q and R relatively prime. Then there exist polynomials a.T / and b.T / such that a.T /Q.T / C b.T /R.T / D 1: For any v 2 V , a.˛/Q.˛/v C b.˛/R.˛/v D v, (83)

which implies immediately that Ker.Q.˛// \ Ker.R.˛// D 0. Moreover, because Q.˛/R.˛/ D 0, (83) expresses v as the sum of an element of Ker.R.˛// and an element of Ker.Q.˛//. Thus, V is the direct sum of Ker.Q.˛// and Ker.P .˛//. On applying this remark repeatedly, we ﬁnd that M V D Ker.T a1 /n1 ˚ Ker..T a2 /n2 .T ar /nr / D D Ker.T ai /ni ;

i

as claimed.

2

T HEOREM 9.12 Let V be a ﬁnite-dimensional vector space over a perfect ﬁeld. For any automorphism ˛ of V , there exist unique automorphisms ˛s and ˛u of V such that (a) ˛ D ˛s ı ˛u D ˛u ı ˛s , and (b) ˛s is semisimple and ˛u is unipotent. Moreover, each of ˛s and ˛u is a polynomial in ˛. P ROOF. Assume ﬁrst that ˛ hasL of its eigenvalues in k, so that V is a direct sum of the generalall ized eigenspaces of ˛, say, V D 1Äi Äm Vai where the ai are the distinct roots of P˛ . Deﬁne ˛s to be the automorphism of V that acts as ai on Vai for each i. Then ˛s is a semisimple automorphism of V , and ˛u D ˛ ı ˛s 1 commutes with ˛s (because it does on each Va ) and is unipotent (because its eigenvalues are 1). Thus ˛s and ˛u satisfy (a) and (b). Because the polynomials .T ai /ni are relatively prime, the Chinese remainder theorem shows that there exists a Q.T / 2 kŒT such that Q.T / Á a1 mod .T Q.T / Á a2 mod .T : Then Q.˛/ acts as ai on Vai for each i , and so ˛s D Q.˛/, which is a polynomial in ˛. Similarly, ˛s 1 2 kŒ˛, and so ˛u D ˛ ı ˛s 1 2 kŒ˛. It remains to prove the uniqueness of ˛s and ˛u . Let ˛ D ˇs ı ˇu be a second decomposition satisfying (a) and (b). Then ˇs and ˇu commute with ˛, and therefore also with ˛s and ˛u (because they are polynomials in ˛). It follows that ˇs 1 ˛s is semisimple and that ˛u ˇu 1 is unipotent. Since they are equal, both must equal 1. This completes the proof in this case. In the general case, because k is perfect, there exists a ﬁnite Galois extension k 0 of k such that ˛ has all of its eigenvalues in k 0 . Choose a basis for V , and use it to attach matrices to endomorphisms of V and k 0 ˝k V . Let A be the matrix of ˛. The ﬁrst part of the proof allows us to write A D As Au D Au As with As a semisimple matrix and Au a unipotent matrix with entries in k 0 ; moreover, this decomposition is unique.

def def

a1 /n1 a2 /n2

9. Recovering a group from its representations; Jordan decompositions

107

Let 2 Gal.k 0 =k/, and for a matrix B D .bij /, deﬁne B D . bij /. Because A has entries in k, A D A. Now A D . As /. Au / is again a decomposition of A into commuting semisimple and unipotent matrices. By the uniqueness of the decomposition, As D As and Au D Au . Since this is true for all 2 Gal.K=k/, the matrices As and Au have entries in k. Now ˛ D ˛s ı ˛u , where ˛s and ˛u are the endomorphisms with matrices As and Au , is a decomposition of ˛ satisfying (a) and (b). Finally, the ﬁrst part of the proof shows that there exist ai 2 k 0 such that As D a0 C a1 A C C an

n 1 1A

.n D dim V /:

The ai are unique, and so, on applying , we ﬁnd that they lie in k. Therefore, ˛s D a0 C a1 ˛ C Similarly, ˛u 2 kŒ˛. The automorphisms ˛s and ˛u are called the semisimple and unipotent parts of ˛, and ˛ D ˛s ı ˛u D ˛u ı ˛s is the (multiplicative) Jordan decomposition of ˛. P ROPOSITION 9.13 Let ˛ and ˇ be automorphisms of vector spaces V and W over a perfect ﬁeld, and let 'W V ! W be a linear map. If ' ı ˛ D ˇ ı ', then ' ı ˛s D ˇs ı ' and ' ı ˛u D ˇu ı '. P ROOF. It sufﬁces to prove this after an extension of scalars, and so we may suppose that both ˛ and ˇ have all of their eigenvalues in k. Recall that ˛s acts on each generalized eigenspace Va , a 2 k, as multiplication by a. As ' obviously maps Va into Wa , it follows that ' ı ˛s D ˇs ı '. Similarly, ' ı ˛s 1 D ˇs 1 ı ', and so ' ı ˛u D ˇu ı '. 2 C OROLLARY 9.14 Every subspace W of V stable under ˛ is stable under ˛s and ˛u , and ˛jW D ˛s jW ı ˛u jW is the Jordan decomposition of ˛jW: P ROOF. It follows from the proposition that W is stable under ˛s and ˛u , and it is obvious that the decomposition ˛jW D ˛s jW ı ˛u jW has the properties to be the Jordan decomposition. 2 P ROPOSITION 9.15 For any automorphisms ˛ and ˇ of vector spaces V and W over a perfect ﬁeld, .˛ ˝ ˇ/s D ˛s ˝ ˇs .˛ ˝ ˇ/u D ˛u ˝ ˇu : P ROOF. It sufﬁces to prove this after an extension of scalars, and so we may suppose that both ˛ and ˇ have all of their eigenvalues in k. For any a; b 2 k, Va ˝k Wb .V ˝k W /ab , and so ˛s ˝ ˇs and .˛ ˝ ˇ/s both act on Va ˝k Wb as multiplication by ab. This shows that .˛ ˝ ˇ/s D ˛s ˝ ˇs . Similarly, .˛s 1 ˝ ˇs 1 / D .˛ ˝ ˇ/s 1 , and so .˛ ˝ ˇ/u D ˛u ˝ ˇu . 2 C an

1˛ n 1

2 kŒ˛:

2

108

BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)

9.16 Let k be a nonperfect ﬁeld of characteristic 2, so that there exists an a 2 k that is not a square p 0 in k, and let M D a 1 . In kŒ a, M has the Jordan decomposition 0 ÃÂ Âp p Ã a 0 0 1= a p p : MD 0 a a 0 Since these matrices do not have coefﬁcients in k, the uniqueness shows that M does not have a Jordan decomposition in M2 .k/.

A

**I NFINITE - DIMENSIONAL VECTOR SPACES
**

Let V be a vector space, possibly inﬁnite dimensional, over a perfect ﬁeld k. An endomorphism ˛ of V is locally ﬁnite if V is a union of ﬁnite-dimensional subspaces stable under ˛. A locally ﬁnite endomorphism is semisimple (resp. locally nilpotent, locally unipotent) if its restriction to each stable ﬁnite-dimensional subspace is semisimple (resp. nilpotent, unipotent). Let ˛ be a locally ﬁnite automorphism of V . By assumption, every v 2 V is contained in a ﬁnitedimensional subspace W stable under ˛, and we deﬁne ˛s .v/ D .˛jW /s .v/. According to (9.12), this is independent of the choice of W , and so in this way we get a semisimple automorphism of V . Similarly, we can deﬁne ˛u . Thus: T HEOREM 9.17 For any locally ﬁnite automorphism ˛ of V , there exist unique automorphisms ˛s and ˛u such that (a) ˛ D ˛s ı ˛u D ˛u ı ˛s ; and (b) ˛s is semisimple and ˛u is locally unipotent. For any ﬁnite-dimensional subspace W of V stable under ˛, ˛jW D .˛s jW / ı .˛u jW / D .˛u jW / ı .˛s jW / is the Jordan decomposition of ˛jW .

**J ORDAN DECOMPOSITIONS IN ALGEBRAIC GROUPS
**

Finally, we are able to prove the following important theorem. T HEOREM 9.18 Let G be an algebraic group over a perfect ﬁeld k. For any g 2 G.k/ there exist unique elements gs ; gu 2 G.k) such that, for all representations .V; rV / of G, rV .gs / D rV .g/s and rV .gu / D rV .g/u . Furthermore, g D gs gu D gu gs : (84)

P ROOF. In view of (9.13) and (9.15), the ﬁrst assertion follows immediately from (9.2) applied to the families .rV .g/s /V and .rV .g/u /V . Now choose a faithful representation rV . Because rV .g/ D rV .gs /rV .gu / D rV .gu /rV .gs /; (84) follows.

2

The elements gs and gu are called the semisimple and unipotent parts of g, and g D gs gu is the Jordan decomposition of g.

9. Recovering a group from its representations; Jordan decompositions

109

9.19 To check that a decomposition g D gs gu is the Jordan decomposition, it sufﬁces to check that r.g/ D r.gs /r.gu / is the Jordan decomposition of r.g/ for a single faithful representation of G. 9.20 Homomorphisms of groups preserve Jordan decompositions. To see this, let ˛W G ! G 0 be a homomorphism and g D gs gu a Jordan decomposition in G.k/. For any representation 'WG 0 ! GLV , ' ı ˛ is a representation of G, and so .' ı ˛/.g/ D ..' ı ˛/.gs // ..' ı ˛/.gu // is the Jordan decomposition in GL.V /. If we choose ' to be faithful, this implies that ˛.g/ D ˛.gs / ˛.gu / is the Jordan decomposition of ˛.g/.

N OTES Our proof of the existence of Jordan decompositions (Theorem 9.18) is the standard one, except that we have made Lemma 9.1 explicit. As Borel has noted (Borel 1991, p88; Borel 2001, VIII 4.2, p169), the result essentially goes back to Kolchin 1948, 4.7.

Y of C.e.Y / such that the only subobject Y 0 of Y such y 2 !. and that the duals of locally compact abelian groups are exactly the locally compact abelian groups. and let !W C ! Veck be an exact faithful k-linear functor. This it does by showing that the canonical map G ! G __ is an isomorphism. at least. there exists a smallest subobject Y of X such that !.X 0 /. 2.5. For a slightly different proof.
110
. the maps Hom. or Saavedra Rivano 1972. and 10. write x Ä x 0 if there exists a morphism X ! X 0 (necessarily unique) giving a commutative triangle. we characterize the categories that arise as the category of representations of an algebraic or afﬁne group. For X in C. 10. The lattice of subobjects of Y is obtained from the collection of monomorphisms by identifying two monomorphisms x and x 0 if x Ä x 0 and x 0 Ä x. then the kernel of
x x0
WX
X0 ! Y
projects isomorphically onto each of X and X 0 (because it does after ! has been applied).
10a
Characterization of categories of comodules
An additive category C is said to be k-linear if the Hom sets are k-vector spaces and composition is k-bilinear. which we call the subobject of X generated by S . F b/ deﬁned by F are required to be k-linear. T HEOREM 10. i.F a.10
Characterizations of categories of representations
Pontryagin duality has two parts.e.
26 If
x
x0
!. !Y /. In Theorems 10. and hence has ﬁnite dimension.5. because ! is faithful.Y 0 / is Y itself. C is a small category (or. Hom. we follow Serre 1993. Let X be an object of C. i.X. it characterizes the abelian groups that arise as dual groups.˛/ is a monomorphism (resp. then so also is ˛. see Deligne and Milne 1982. Throughout.8. Hence X has ﬁnite length. An object Y is monogenic if it is generated by a single element. we let hX i denote the full subcategory of C whose objects are the quotients of subobjects of direct sums of copies of X. an epimorphism). For any subset S of !. We shall frequently make use of the fact that.a. Y / is a subspace of Hom. k is a ﬁeld. b/ ! Hom. For monomorphisms X ! Y and X 0 ! Y with the same target. Recall that Veck denotes the category of ﬁnite-dimensional vector spaces over k. For objects X . admits a set of representatives for its isomorphism classes of objects). For example. it shows that the duals of discrete abelian groups are exactly the compact abelian groups.1. Similarly ! maps the lattice of quotient objects of Y injectively to the lattice of quotient spaces of !Y .Y / S . Then there exists a coalgebra C such that C is equivalent to the category of C comodules of ﬁnite dimension.1 Let C be a k-linear abelian category.
For the proof. Functors of k-linear categories are required to be k-linear.!X. Secondly. In this section. 2.G/. The functor ! maps the lattice of subobjects of Y injectively26 to the lattice of subspaces of !Y .X / D !. First it shows that a locally compact abelian group G can be recovered from its dual. there exists a y 2 !... In 9 we showed how to recover an algebraic group G from its “dual” Rep. if !.X /.

f / sends p to x.p. it is exact.Q/ contains . We have y 2 !.z1 .. because P is generated by p.P.Y / generate Y . Am ! An ! M ! 0
27 Let
˛
A be an m
1
of N onto X m
Ã a1 : : : a m is invertible. : : : . P ROOF.P / generate P . and so it sufﬁces to prove the lemma for Z.1
The case that C D hXi for some object X.Y / can take only ﬁnitely many values when Y is monogenic. 1 m matrix such that
Â
1
deﬁnes an isomorphism
. X / is an equivalence from C to the category of right Amodules of ﬁnite dimension over k. By hypothesis.e.Y / Ä n2 : P ROOF. L EMMA 10.3 (a) The pair . Therefore. let Q be the smallest subobject of P X such that !. Let Z be the subobject of Y1 generated by y1 .X /. The composite of its inverse with the second projection Q ! X is a morphism P ! X sending p to x. It remains to prove that it is essentially surjective. zm / in !.X /. we may suppose that Y X m for some m.P // is maximal.!. L EMMA 10. Then AW X m ! X m A (because it does after ! has been applied).P / has its largest possible value. The morphism Q ! P deﬁned by the projection map is surjective. and so Y N because y1 generates Y . (a) Let X be an object of C. Let n D dimk !. The uniqueness follows from the fact p generates P .Q/ dimk !. we have to prove that there exists a unique morphism f W P ! X such that !. and so Q ! P is an isomorphism. The image of Z in Y D Y1 =Y2 is Y . Therefore Z embeds into X m 1 . but because dimk . Let y D . the functor is fully faithful. i.P. (b) The object P is projective because ! is exact. Continue in this fashion until Z X m with m Ä n.Y1 / whose image in !. dimk !. and choose a ﬁnite presentation for M .2 For any monogenic object Y of C. and let x 2 !. there exists a monogenic P for which dimk !.X m / D !. Because P is a generator. equality must hold. But the ai deﬁne a surjective morphism X m ! X.P / — it is a k-algebra of ﬁnite dimension as a k-vector space (not necessarily commutative). On the other hand. If m Ä n. Suppose therefore that m > n.P /. (b) The object P is a projective generator for C. such that P ai zi D 0. if N is the kernel of this morphism. Let p 2 !.10.4 The functor X Hom. Y D Y1 =Y2 where Y1 is isomorphic to a subobject of X m for some m. there exist ai 2 k. and because P is projective. its composite with the forgetful functor is !.X/m . Let y 2 !.. dimk !. then one sees easily27 that N X m 1 . Let M be a ﬁnite-dimensional right A-module.X /m . p/ represents the functor !. Since m > n. not all zero.N /. and let y1 be an element of !. L EMMA 10. To prove the existence. the lemma is obvious. x/.Y / is y. P ROOF. y1 2 !.2 As dimk !.Y / !. Characterizations of categories of representations
111
P ROOF OF T HEOREM 10. and it is a generator because ! is faithful. 2 Let A D End.

and the category of right A-modules is equivalent to the category of left C -comodules (see 7. Consider the functors
Mod.R/:
R ı !/
For M 2 ob.P /.CX /: We deﬁne a partial ordering on the set of isomorphism classes of objects in C by the rule: ŒX Ä ŒY if hX i hY i.M /. X / ' Hom. hP . and that if ŒX Ä ŒY .˛ ı ˇ. For an object X of C. Then ! deﬁnes an equivalence of categories C ! ! Comod.k/
R
V 7!R˝k V
! Mod.P. Clearly ˇ ı˛ D id. and let CX D A_ . We have !.5 Let C be a k-linear abelian category and let !W C ! Veck be a k-linear exact faithful functor.X /:
2
As A is a ﬁnite k-algebra.
which we shall show to be an isomorphism by deﬁning an inverse ˇ. E XAMPLE 10. Let ˇ. and so we have a map ˛W R ˝k A ! End. this completes the proof in this case. hP . When we pass to the limit over the isomorphism classes. Then ! deﬁnes an equivalence of categories X hXi ! Comod.B/ carrying ! into the forgetful functor. X / M . / D A . its linear dual C D A_ is a k-coalgebra. An element of R ˝k A deﬁnes such a family.1 ˝ m/ for 1 ˝ m 2 R ˝ M.A//. //M . ŒY Ä ŒX ˚ Y .
Note that ŒX .A/
! forget
! Vec.X // ' hP .1 ˝ 1/. This matrix deﬁnes a morphism P m ! P n whose cokernel X has the property that Hom.A. let AX be algebra of endomorphisms of !jhX i.6 Let A be a ﬁnite k-algebra (not necessarily commutative). so that we get a directed set.4).X // D Hom. T HEOREM 10.Mod.P.
R ı !/.
is a family of R-linear maps
M0 ! R ˝k M0 . we obtain the following more precise form of the theorem.X / ' Hom.hP . and so we only have to show ˛ ıˇ D id. An element
M W R ˝k
of End. Let B D lim End. The A-module A˝k M0 is a direct sum of copies of A.
. Therefore
M .112
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
where ˛ is an m n matrix with coefﬁcients in A. and the additivity of implies that A˝M0 D A ˝ idM0 . and let R be a commutative k-algebra.1/ ˝ m
D . and hence R ˝k A ˝k M0 ! R ˝k M ? ? ? ˝id ? y A M0 y M R ˝k A ˝k M0 commutes.
functorial in M . The general case. let M0 D !.!jhXi/_ . The map a ˝ m 7! amW A ˝k M0 ! M is A-linear.7). then restriction deﬁnes a homomorphism AY ! AX .1 ˝ m/
! R ˝k M
D
A . Together with (10.

5) we take C D Mod. u ˝ v 7! v ˝ u: U. With the notations of the previous subsection.
2 EndR-linear . The tensor structure will enable us to deﬁne an algebra structure on A.!X.U ˝k V / ˝k W.V ˝k W / ! . and let ! be an exact faithful k-linear functor C ! Veck satisfying the following conditions: (a) (b) (c) (d) ˛ is an isomorphism if !.BX /. Let B be a coalgebra over k and let ! be the forgetful functor Comod.!.V T HEOREM 10.10.
113
In particular. so that C D hAi. A ! End.6 shows that B D lim End.V /R /. and 1 1 1 1/ W ı !. and ! deﬁnes an equivalence of tensor categories over k. Characterizations of categories of representations i.X / ˝k !.8 Let C be a k-linear abelian category. Y. let BX D A_ .1 D k and the canonical isomorphisms 1 1/ !.B/ ! Vec.. Then ! deﬁnes an equivalence of X categories from hX i to Comod. W 2 ob. V 1
in C such that X ˝ X
1
D1 1. We say that a morphism !X ! !Y lives in A if it lies in Hom. there exists an object 1 in C such that !.
A SIDE 10.5.˛/R D !. W .
'
10b
Characterization of categories of representations of afﬁne groups
Let !W A ! B be a faithful functor.C / .X / ' !.1 for every identity object of 1 of C.1 ˝ !.˛/R ı V for all arrows ˛ in C. let ˝ be a k-bilinear functor C C ! C.u ˝ v/ ˝ w W U ˝k V ! V ˝ U.V.X ˝ Y / D !. the isomorphisms !X.
C ! Rep.!Y. Assume C has a tensor product.e. for all X. u ˝ .!jhXi/_ .W W U
˝k . !Y /. The discussion in Example 10. We deduce easily that every functor !
.1 1/ 1/ live in C.X / ˝ !. V.C/.A/. plus a uniqueness statement.!/. Y / Hom. For k-vector spaces U.X / ' !. ˛ ı ˇ D id. such that ˘ ˘ ˘ D V ˝ W for all V. Then there exists a coalgebra C in C together with a coaction of C on each object of A such that every exact faithful k-linear functor ! to Veck deﬁnes an equivalence A ! Comod!. !. (e) if !. then the equivalence of categories obtained is the canonical equivalence (7. then there exists an object X For each k-algebra R. Y . and it follows that.!Z and !X. let G. there are canonical isomorphisms
U.k/.
Then G is an afﬁne group over k.Y /I for all X.X. if in (10.˛/ is. Z.R/ be the set of families .v ˝ w/ 7! .7 Need to rewrite this section to obtain the following result: Let C be a k-linear abelian category for which there exists an exact faithful k-linear functor.G/:
We have already deﬁned the coalgebra of G in Theorem 10.C/ . D id!.7).
V ˝W V /V 2ob.X / has dimension 1.!Y live in C.

B. R/.forget/.R/: We have shown that. there is an identity object in .Y / (as a k-vector space).7).B/ Comod.! G / is represented by G.B/ sending . R/ D Homk-lin .X ˝ Y / D !. Y / D X ˝k Y as k-vector spaces.9). let B be a coalgebra over k.C.10 Let .6). satisﬁes the conditions of (10. u / if and only if the multiplication deﬁned by u on B is associative and commutative. (Cf.B/ Comod.5) except now assume that C is a tensor category and ! is a tensor functor. condition (e) is omitted.
A SIDE 10. !/ and B be as in (10. The natural associativity and commutativity constraints on Veck induce similar constraints on .2: Aut˝ . with .8) in this case one recovers Theorem 9. mo3131. Using (10.X.R/ D Homk-alg . and hence.! ˝ !jh.)
.!/ D Aut˝ .B/ such that .X.114
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
unique homomorphism B ! B 0 . thus End˝ .B/.B/ ! Comod.
X˝ Y
Comod. Thus ! the ﬁrst statement of the proposition follows from (10.11 Add discussion of how much of this section extends to base rings k. the structure of an associative commutative k-algebra with identity element on B.!/.8.! ˝ !/.
An element 2 Homk-lin .Repk . Again.!/ (apply (e) to the highest exterior power of an object). Thus B lacks only a coinverse map S to be a k-bialgebra (in our sense) and G D Spec B is an afﬁne monoid scheme.!/. then one can conclude that End˝ .R/ commuting with the tensor structure if and only if is a k-algebra homomorphism. u / with underlying vector space k if and only if B has an identity element. R/ D G.X / ˝ !. Y /i/_ D B ˝ B.5).BX .
def
R ı !/
D lim Homk-lin . R EMARK 10. if in the statement of Theorem 10.B/ ! Repk .B/.!/.B 0 / carrying the forgetful functor into the forgetful functor arises from a
! X ˝B ˝Y ˝B ' X ˝Y ˝B ˝B
X ˝Y ˝u
! X ˝Y ˝B
P ROPOSITION 10. for any k-algebra R. Y / to X ˝k Y with the coaction X ˝Y (see 7.Comod.!/ is representable by an afﬁne monoid G D Spec B and that ! deﬁnes an equivalence of tensor categories
C ! Comod.B/ ! Comod.C.
It is easy to see that (e) implies that End˝ .BX . A coalgebra homomorphism uW B ˝k B ! B deﬁnes a functor u W Comod. The pair . The tensor structure on C induces a similar structure on Comod. R/ corresponds to an element of End.G/.B/.B/ ! Comod. !/ be .B/.X.R/ D End. 2 Let . On following through the proof of (10. The remaining statements are easy.Comod.Comod.9 The map u 7! u deﬁnes a one-to-one correspondence between the set of homomorphisms B ˝k B ! B and the set of functors W Comod. because of (10.B. P ROOF. which completes the proof.G/.6) we ﬁnd that. ! ˝ !/. End. and lim End.B/ Comod.

12) says that the map A ! m2S A=m is surjective with kernel m2S m. For a ﬁnite k-algebra A. the minimum polynomial f . Obvious.2 A ﬁnite k-algebra satisfying the equivalent conditions of the proposition is said to be separable. (b) A ˝k k al is a product of copies of k al .12). in which case ŒO.6. k is a ﬁeld and all k-algebras are commutative.
2
D EFINITION 11. and.X / D . We say that an algebraic group is a p-group if it is ﬁnite and its order is a power of p. we let ŒAW k denote the degree of A (dimension of A as a k-vector space). and so A has only ﬁnitely many maximal ideals.
11a
Etale groups
S EPARABLE k. P ROPOSITION 11. An algebraic group G is ﬁnite if O. and quotients of separable k-algebras are separable. jS j Ä ŒAW k.X p / with f 2 kŒX (see FT 3. P ROPOSITION 11. et seq. then m2S m is the nilradical N of A (CA 11.G/ is a ﬁnite k-algebra.X p //. then k has characteristic p ¤ 0 and there exists an element ˛ of k 0 whose minimum polynomial is of the form f . For any ﬁnite set S of maximal ideals in A. From the primitive element theorem (FT 5. Now A ˝k k al ' . Y k al ŒX =.f / ' k al ŒX =. tensor products.f .kŒX =.X p / is a pth power in k al ŒX . If k 0 is not separable over k.3 Finite products. 115
.ALGEBRAS
Let A be a ﬁnite k-algebra. We may suppose that A itself is a separable ﬁeld extension of k.).X ˛i / ' k al
i
k al .G/W k is called the order of G. in k al ŒX. and so A=N is a ﬁnite product of ﬁelds.f /. Because kŒ˛ is separable over k.f .
(b))(c).1 The following conditions on a ﬁnite k-algebra A are equivalent: (a) A is a product of separable ﬁeld extensions of k.X p // ˝k k al ' k al ŒX =.X / of ˛ is separable. Hence A ˝k k al is not reduced. Let k 0 one of the factors of A.X ˛i /. Therefore the above discussion shows that it is a ﬁnite product of ﬁelds. we know that A D kŒ˛ for some ˛. and so A is reduced. (c))(a). If S is the set of all maximal ideals in T A.11
Finite algebraic groups
Throughout this section. The map a 7! a ˝ 1W A ! A ˝k k al is injective. ˛i ¤ ˛j for i ¤ j. (c) A ˝k k al is reduced. which is not reduced because f . which means that Y f .kŒX =. the Chinese remainder theQ T orem (CA 2. In particular.1).f // ˝k k al ' k al ŒX =. (a))(b). P ROOF. according to the Chinese remainder theorem (CA 2. Now kŒ˛ ˝k k al ' .8).

For any subﬁeld K of k sep . Let X be a ﬁnite set with an action of
X ! X: We say that the action is29 continuous if it factors through k sep ﬁnite and Galois over k.E. for example. let ! Gal. This is obvious from the condition (b). and let k al be the algebraic closure of k in k al .K=k/ .5 If A is separable over k. Let be the group of k-automorphisms of k sep .K=k/ for some subﬁeld K of
F . and so . an W A1 ˝k An is the image of the map ˝k An ! B. ﬁnite and Galois over k. ˛/ where E is a subﬁeld of k sep containing k and ˛ is homomorphism E ! k sep whose restriction to K is 0 .4 The composite of any ﬁnite set of separable subalgebras of a k-algebra is separable.A/ through its action on k sep : .
is surjective.
0 28 Let
. P ROOF. an easy Zorn’s lemma argument28 shows that 7! jKW ! Gal. then A ˝k k 0 is separable over k 0 for any ﬁeld extension k 0 of k.A/ D Homk-alg . k al / D Homk-alg . Let Ai be separable subalgebras of B.a/ D .A/.k al ' k 0al k al / ˝k al k 0al k 0al :
2
! k 0al x ? ? ! k al
C LASSIFICATION OF SEPARABLE k. 2 .
C OROLLARY 11.
2
P ROPOSITION 11. Then A1 a1 ˝ and so is a quotient of A1 ˝k ˝ an 7! a1 ˝k An . and so the action of on F .A ˝k k 0 / ˝k 0 k 0al ' A ˝k k al ˝k al k 0al ' . a 2 A:
The images of all homomorphisms A ! k sep will lie in some ﬁnite Galois extension of k. of characteristic zero. 29 Equivalently.a//. Apply Zorn’s lemma to the set of all pairs .K=k/.A. If k is perfect. then k sep D k al .A.f .ALGEBRAS
Let k sep be the composite of the subﬁelds k 0 of k al separable over k. f /. k sep /: Then acts on F .
2 Gal. Then k0 x ? ? k is commutative. f 2 F .A/ is continuous. For a separable k-algebra A.116
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
2
P ROOF. P ROOF. the action is continuous relative to the discrete topology on X and the Krull topology on (FT 7). Let k 0al be an algebraic closure of k 0 .

X // is the decomposition of A into a product of ﬁelds. namely.8).6 The map A 7! F . Assume this.X / has distinct roots in k al .G/ is an equivalence from ´ the category of etale algebraic groups over k to the category of group objects in the category Aopp (see 4g).f . ´ Let X be a group of order 3.8 The functor G G.X / is monic.10). and let f1 .6.7 (a) When k has characteristic zero.X / D Z=2Z. On combining this statement with Theorem 11. we obtain the following theorem. ´ T HEOREM 11.X / D 1 when X is a group of order 1 or 2. and let 0 be the subgroup of ﬁxing the elements of K.fi . Then each fi is stable under .k sep / of elements ﬁxed by 0 : consisting of the (see FT 7. ´ (b) Algebraic geometers will recognize that an algebraic group G is etale if and only if the ´ morphism of schemes jGj ! Spec k is etale.A/ decomposes into r orbits under the action of .K=k/ ' Z=2Z
and all nontrivial such homomorphisms arise in this way (see FT 7).G/ is a separable k-algebra. the order of G is the order of the (abstract) group G.G/ is ﬁnite-dimensional as a k-vector space. and is left as an exercise to the reader (the indolent may see Waterhouse 1979. and that Q A D 1Äi Är kŒX =. there is exactly one etale algebraic group of order 1 and one of order 2 over k (up to isomorphism). : : : . 7! jKW ! Gal. A k-algebra homomorphism A ! k sep is determined by the image of x. Then A is separable if and only if f . ´ R EMARK 11. It follows that f D f1 fr is the decomposition of f into its irreducible factors over k.K/ is the subgroup G. and (for simplicity) that f . Thus.A/ can be identiﬁed with the set of roots of f . We say G is etale if in addition O. ´ Since Aut. Let K be a subﬁeld of k sep containing k. and
0
E XAMPLES
´ For an etale algebraic group G. Therefore the etale algebraic groups of order 3 over k correspond to homomorphisms ! Z=2Z factoring through Gal.
´ C LASSIFICATION OF E TALE ALGEBRAIC GROUPS
Recall that an algebraic group G is said to be ﬁnite if O.K=k/ for some ﬁnite Galois extension K of k. Suppose F . Then K is the subﬁeld of k sep of elements ﬁxed by it follows that G. and so has coefﬁcients in k (FT 7.3). The functor G O. there ´ is exactly one etale algebraic group G K of order 3 over k for each separable quadratic extension
.24 says that every ﬁnite algebraic ´ group is etale. Let A be the category of separable k-algebras. up to isomorphism.11. 2 Suppose that A is generated by a single element. 6. fr be the monic polynomials whose roots are the orbits.A/ is a contravariant equivalence from the category separable k-algebras to the category of ﬁnite sets with a continuous action of . This is a restatement of the fundamental theorem of Galois theory (FT 3).k al /. A separable quadratic extension K of k deﬁnes such a homomorphism. which can be any root of f in k sep .k sep / is an equivalence from the category of etale algebraic groups over k to the category of ﬁnite groups endowed with a continuous action of . Finite algebraic groups
117
T HEOREM 11. Therefore.X // D kŒx. so A D kŒX =. F . Theorem 5. P ROOF. Such a group is cyclic and Aut.

k/ has order 3.QŒ 3 1/ D 3. For G0 .G/ ' kŒx1 . xr of IG =IG .G/. : : :. V e ´ 6 (a k-algebra A is etale if A ˝k K K n for some integer n 0). then W D Zp . G p p p p QŒ 1. For any 2 basis x1 .G/ from the category of ﬁnite commutative algebraic p-groups to the category of triples . : : : . Demazure 1972. for example. G K .10 Let G be a ﬁnite group scheme over a ﬁeld k of characteristic p ¤ 0.k/ has order 1 butp K . xr =. G0 . Omitted for the moment.G/ ' kŒx1 . 3 must be the p group corresponding to QŒ 3 1. F.11 Let G be a ﬁnite group scheme over a ﬁeld k of characteristic p ¤ 0. : : : . plus the constant group G0 . : : : .M.c m/ D c F m V . xr of IG =IG .Q/ D 1 but 3 .12 There exists a contravariant equivalence G M. c m/ D c V m F V D p idM D VF: The order of G is p length. align more with Bourbaki A.mod p/: T HEOREM 11. QŒ p. Since 3 . xr /. xr /: p e1
er
2 p p
P ROOF. : : : . 14. p.G/W k D p r ). There are inﬁnitely many distinct quadratic extensions of Q. Omitted for the moment (Waterhouse 1979.4). : : :.
0 Ä mi < p
form a basis for O. The Frobenius automorphism of W is the unique automorphism such that a Á ap .
A SIDE 11. er 1 such that
p O. For any perfect ﬁeld k 0 containing k. Thus W is a complete discrete valuation ring with maximal ideal generated by p D p1W and residued ﬁeld k. For example.6.M.G// .
2
C LASSIFICATION
Let k be a perfect ﬁeld of characteristic p.9 Should replace “separable k-algebra” with “´ tale k-algebra”. Add proof of Theorem 11.G/ as a k-vector space (and so ŒO. T HEOREM 11.K/ has order 3. and 2 suppose that x p D 0 for all x 2 O. This generalizes. For G K . and let W be the ring of Witt vectors with entries in k. P ROOF.k/ M. For any basis x1 .G/: P ROOF. the monomials
m x1 1 m xr r . QŒ 3. there is functorial isomorphism M. : : : .118
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
K of k. QŒ 2. V / in which M is a W -module of ﬁnite length and F and V are endomorphisms of M satisfying the following conditions (c 2 W .
2
. III 7. xr =.x1 .Gk 0 / ' W . : : : . : : : .
11b
Finite group schemes in characteristic p ¤ 0
P ROPOSITION 11.69. The proposition says that O.x1 . if k D Fp . m 2 M ): F .k 0 / ˝W . there exist integers e1 .

R/ G.
p/
119
F D 1.R/ ! R : 1
T HEOREM 11. we describe G _ as a functor. m.
11c
Cartier duality revisited
Let G be a ﬁnite commutative algebraic group with bialgebra .13 For an element f of A ˝ R satisfying (a).O. x/ 7! ei .11. (86) is the pairing kŒX. L EMMA 11. GmR / . In this subsection.A_ . and (b) eR .G/.! Homk-alg .Z=pZ/ D W =pW. On the other hand.ei /i 2I of A.
V D 0I V D pI V D 0:
D W =pW. R/ ' HomR-lin .f / D 1.
homomorphisms GR ! GmR correspond to elements f 2 A ˝ R such that (a) R . it deﬁnes the bijection (85). for any k-algebra R.G/_ .O.˛p / D W =pW. /. and (c) f 2 . m_ . k-algebra homomorphisms correspond to elements f 2 A ˝ R such that (a) R .y.A ˝ R/ .y/ei .x/W G _ . A ˝ R/.f / D f ˝ f .GR . . Finite algebraic groups For example: M. Let R be a k-algebra. GmR /: Let G_ G ! Gm (86) (85)
2
be the natural transformation corresponding to the k-algebra homomorphism ! A_ ˝ A P _ sending X to the element ı of A_ ˝ A such that ı D ei ˝ ei for any basis . X
1
. A ˝ R/ ' A ˝ R. In other words. e _ /. the conditions (b) and (c) are equivalent. We conclude that there is canonical bijection G _ .
M. Omitted for the moment (easy calculation).R/ ' Hom. F D 0. _ . Under the inclusion Homk-lin .kŒX. under the inclusion Hom. Let A D O.14 The pairing (86) is bi-multiplicative.GR . F D 0.A_ ˝ R.R.! HomR-lin . _ . e. R/ .f / D f ˝ f . Recall (4j) that the Cartier dual G _ of G is the algebraic group with bialgebra . X
_ . M.G/. P ROOF.
.

120 In other words.R/ ! R where exp. Finite ﬂat group schemes. y2 . G _ represents the functor R
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
Hom. 121–154. G _ ' ˛p . 4 are there over R (up to isomorphism)?
.G _ / D O.
E XAMPLE 11. Let 1. GmR /:
2
P ROOF.G/ D kŒX =. x p 1 .
11d
Finite ﬂat group schemes over commutative rings
Include a discussion.] E XERCISE 11-2 How many ﬁnite algebraic groups of orders 1.ab/ D 1 C ab . Let A be a k-algebra and M an A-module. in particular. 1995).X p / D kŒx. y. 2. MA. Springer. Modular forms and Fermat’s last theorem (Boston. New York.p 1/Š
C
A SIDE 11. see: Tate. John.
11e
Exercises
E XERCISE 11-1 Show that A is separable if and only if there are no nonzero k-derivations DW A ! k.ab/W ˛p . A k -derivation is a k-linear map DW A ! M such that D.GR . so that O. 1997. Omitted for the moment (easy calculation).ab/p 1 .g/ C g D. In fact. [Regard A as a left A-module by left multiplication.15 Let G D ˛p . yp 1 be the basis of O.R/ ˛p .f / (Leibniz rule). . : : : . b 7! exp. Then y i D i Šyi . 3.16 The omitted proofs can be found in the notes on Cartier duality on Ching-Li Chai’s webpage. y p D 0.ab/2 C C 1Š 2Š . For the present. x.G/_ dual to 1. : : : .fg/ D f D. and the pairing is a.

Write 0 . sb m g is a complete set of orthogonal idempotents. s 2 A.sb m / D .ab/m D 0. . and so a C b D A. we can write 1 D ram C sb m for some r.ram / and V . Throughout. The maximal connected subspaces of X are called the connected components of X. en g be a complete set of orthogonal idempotents in A. 0 and 1 are both idempotents — they are called the trivial idempotents. Any set of orthogonal idempotents fe1 . no prime ideal contains both a and b.sb m /.a/ and V . 2 P ROPOSITION 12. A is a commutative ring. .a/ D V . it is connected. en g is a complete set of orthogonal idempotents in A. Conversely.G/ is a normal connected subgroup G ı of G. we see that V .ram / \ V . : : : .A/ is disconnected if and only if A contains a nontrivial idempotent. : : : . if it contains a nonempty proper closed-open subset. : : :/.. Therefore.f / ¤ spec A. L EMMA 12. For a prime ideal p. ram C sb m D 1.b/ D V .a/ V . 1/ form a complete set of orthogonal idempotents. For a topological group G. Otherwise.e1 / t : : : t D. say. A set fe1 . k is a ﬁeld.5). : : : . we extend these statements to afﬁne and algebraic groups. For example. 0.e/ D spec A.b/. Now . : : : . 0.sb m / D 0.f /.
12a
Some algebraic geometry
Throughout this subsection. e2 D .sb m /. the map A ! A=p must send exactly one of e or f to a nonzero element. similarly. then the elements e1 D .0. 0.1 The space spec. 0 .e/ and D. Then spec A D D. .ram /. any prime ideal containing b m contains b. Let e be a nontrivial nilpotent. suppose that specm. Similarly.2 Let fe1 . : : :/.ram /. GL2 . V . : : : . en g of orthogonal idempotents is complete if e1 C C en D 1.0. Any prime ideal containing am contains a. then Aei becomes a ring with the addition and multiplication induced by that of A. en D . In this section. equivalently. called the identity (connected) component of G.1. For example. and A ' Ae1 Aen . which shows that the ideal they generate is A. if fe1 . This shows that spec A is a disjoint union of the sets D. and the kernel of G ! 0 . 1. which is therefore nilpotent. the identity component consisting of the matrices with determinant > 0 and another connected component consisting of the matrices with determinant < 0. .ram / and V . An element e of A is idempotent if e 2 D e. and so fram . Moreover. then e is nilpotent (CA 2. As ab 2 a \ b. : : : .G/ is again a group. D. and X is a disjoint union of them. say . Conversely.. thus no prime ideal contains both am and b m . en g can be completed by adding the idempotent e D 1 . each of which is open.R/ has two connected components.X / for the set of connected components of X (for compact spaces it is ﬁnite). Any sum of orthogonal idempotents is again idempotent. and so neither ram nor sb m is zero.e1 C C en /. and hence 0.ram /2 D . If D. P ROOF.b/ V .en / 121
. If A D A1 An (direct product of rings). As V . en are orthogonal if ei ej D 0 for i ¤ j .1 sb m / D ram . namely. Idempotents e1 . and let f D 1 e. Because the union is disjoint. Thus a C b D 1 for some a 2 a and b 2 b.A/ is disconnected. the disjoint union of two nonempty sets V .12
The connected components of an algebraic group
Recall that a topological space X is disconnected if it is a disjoint union of two nonempty open subsets. all prime ideals contain ab.

and so we can assume that k is algebraically closed. : : : . 2 Let A be a ﬁnitely generated k-algebra. then it is connected for the complex topology. After the above discussion. whose closed sets are the zero sets of collections of polynomials.4 On Cn there are two topologies: the Zariski topology. Similarly. When A is Jacobson. In particular. every such decomposition arises in this way. The converse is false. corresponds to a decomposition A D A1 An of A into a product of rings. Basic Algebraic Geometry.1. Clearly Zariski-closed sets are closed for the complex topology.0.
12b
Separable subalgebras
L EMMA 12.5 Let A be a ﬁnitely generated k-algebra. D. Let K be a ﬁeld containing k. and so D. The degrees of the separable subalgebras of A are bounded (in fact. 1994.ei / is a ﬁnite union of its irreducible components.5). each Ui open.
A SIDE 12. Thus the next result is a surprise: If V Cn is closed and irreducible for the Zariski topology. the resulting space is not connected for the complex topology but it is connected for the topology induced by the Zariski topology (a nonempty Zariski-open subset of C can omit only ﬁnitely many points). the elements e1 D . Then a separable subalgebra is of the form k k. for a Jacobson ring A. Because specm A is noetherian. er D . In a Jacobson ring. 0.A/ of A containing all other separable subalgebras. P ROOF. Moreover.e1 /. A separable subalgebra of A will give a separable subalgebra of the same degree of A ˝k k al . and that every ﬁnitely generated algebra over a ﬁeld is Jacobson (see CA 12. and for n > 2 it can be proved by induction. It follows that a subset of Cn that is connected in the complex topology is connected in the Zariski topology. and the complex topology.4). it is a ﬁnite union of its irreducible components (CA 12. P ROOF. 2 R EMARK 12. When n D 2 this follows from the lemma. : : : . The composite of two separable subalgebras of A is separable (11. VII 2. We shall need to know that it is the largest separable subalgebra. there are natural one-to-one correspondences between the decompositions of specm.A/ into a ﬁnite disjoint unions of open subspaces. : : : . and so the complex topology is the ﬁner than the Zariski topology. : : :/.122
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
is a decomposition of spec A into a disjoint union of open subsets.3 Recall that a ring A is said to be Jacobson if every prime ideal is an intersection of maximal ideals.
. For such a subalgebra. “prime ideal” can be replaced by “maximal ideal” and “spec” with “specm” in the above discussion. and the complete sets of orthogonal idempotents in A. the nilradical is an intersection of maximal ideals. each open subsets D. For the proof. the decompositions of A into a ﬁnite direct products of rings. there is a separable subalgebra 0 . and so. Therefore r is at most the number of irreducible components of specm A. 0. 1/ form a complete set of orthogonal idempotents.er / are disjoint open subsets covering specm A. Then 0 . by the number of irreducible components of specm A ˝ k al ). For example. each of which is an irreducible component of A. see Shafarevich. if we remove the real axis from C.10). it remains to prove that every decomposition specm A D U1 t : : : t Un .A/ ˝k K is a separable subalgebra of A ˝k K (see 11.3 et seq. because of the lemma.).

and a connected algebraic variety remains connected under extension of the algebraically closed base ﬁeld. 0 . and we can deﬁne on 0 . Cf.A ˝ k / .9 Let G ! 0 . The map W A ! A ˝k A is a k12.A/.A ˝k A/ D 0 .A ˝ k al / has a basis in A.7 Let A and A0 be ﬁnitely generated k-algebras. and let K be a ﬁeld containing k. After (12. We shall need to know that it is the largest separable subalgebra. Similarly. 0 . all the elements ci P pr pr pr pr lie in k.12.A/ ! A. But ej D ej . then k has characteristic p ¤ 0 and k al is purely inseparable over it. we have that 0 .G/. then A ˝k k sep is an integral domain (in particular. T HEOREM 12. If 0 . P ROPOSITION 12. in which case the proposition follows from the fact that a product of connected algebraic varieties over an algebraically closed ﬁeld is connected.A/ into 0 . Hence specm A is an algebraic variety over an algebraically closed ﬁeld.A/ to be the restriction of on A. and so sends 0 .A ˝k
K/:
P ROOF. we prove the statement when k and K are both algebraically closed. S W A ! A sends 0 .A ˝k
A0 / D
0 .6) we may suppose that k is algebraically closed. 2
A SIDE 12. ı D Ker. 11.A/ D k. 2 Let A and A0 be ﬁnitely generated k-algebras. Then
0 . The connected components of an algebraic group
123
P ROPOSITION 12. Therefore. : : : .6 Let A be a ﬁnitely generated k-algebra. We may suppose that 0 . em be a basis of idempotents P pr for 0 . Let e1 .
12c
The group of connected components of an algebraic group
Let G be an algebraic group with coordinate ring A D O. The subalgebra 0 .A/ ˝k
0 .A ˝ k sep / ˝k k sep (cf.G ! ı (b) Let G 0 G/.A/ ˝k 0 .A/ ˝k 0 .A/ becomes a Hopf subalgebra of A. For some r. we ﬁrst prove the statement with K D k sep . Since sep sep D sep 0 .A ˝ k al /.A/. then 0 . the result (AG 11.5): let A be a ﬁnitely generated k-algebra.6).A/ ˝k k We next prove the statement when k is separably closed and K D k al . With these maps 0 . Write ej D ai ˝ ci with ai 2 A and ci 2 k al .
0G
be the quotient map corresponding to the inclusion of bialgebras
´ (a) Every quotient map from G to an etale algebraic group factors uniquely through G ! 0 .A ˝k k sep / D k sep ). and so 0 .
. Finally.7 algebra homomorphism.G/.k sep =k/. it sufﬁces to prove the proposition with K algebraically closed. and hence equals 0 .A
0
/:
P ROOF.A/ into 0 .A/ ˝ L will not be maximal in A ˝ L for any ﬁeld L containing K.3). Then
0 .A/ 0 . However.8 Need to rethink this. Since specm A D specm A=N we may suppose that A is reduced. It k al ¤ k.A ˝ k /. and then ej D ai ˝ ci 2 A.A0 / is a separable subalgebra of A ˝k A0 (see 11. ´ ii) G=G ı is etale. Then 0 .A/ ˝k
KD
0 . assume that A is an integral domain and that 0 .A ˝ k sep / of A ˝ k sep is stable under the action of Gal.A/ ˝ K is not maximal in A ˝ K.A/ D k. Then G is the unique normal algebraic subgroup of G such that i) 0 G ı D 1.

and so the image of the homomorphism is def contained in 0 . and (by deﬁnition) G=G ı ' 0 G.1 e/O.Gk 0 /
'
0 .13). P ROPOSITION 12. G 0 /ı ' G ı G 0ı . and so we get a commutative diagram 1 ! Gı ? ? y ! H ! G !
0G
! 1
? ? y ! G ! G=H ! 1
1
with exact rows.H / ! O. 0 G/ into a direct product O.
0 G/ .G/ .O.G/.G/ ' O.Gk 0 /ı ' .1
e/O.10 The subgroup G ı is called identity component of G. then (by (a)) the homomorphism G ! G=H factors through 0 . k D 0 .G/k 0
and .R/ ! .G/
0 . If H is etale.
2
P ROOF. As O.O.G
0 .G/
with eO.R/ gives. and
Let e D .
The exactness of (87) shows that G ı is the kernel of H ! and so if 0 H D 1. The coordinate ring O.G ı /k . Clearly. This proves (a). The similar diagram with each an exact sequence
replaced with . and so this follows from (12.G/ ˝ O.
0 . If G=H is etale.1 e/.G ı / (see 6.R/:
(87)
Since this functorial in R. ´ Therefore 0 G ı D 1. is .1. ´ Suppose H is a second normal algebraic subgroup of G.R/ ! H. this follows from (12.G/ ˝k k 0 . then O.6). (a) A quotient map G ! H corresponds to an injective homomorphism O.Gk 0 / ' O. its kernel is H : therefore G ı ' H .G/ D O.G/ D eO. 0 G/.H / is separable.eO.12 For any algebraic groups G and G 0 . 2
D EFINITION 12.G
0 / and . the functors G
0G
k.7).
and G
G ı commute with extension of the base ﬁeld.G
G 0 / ' O.
2
. Then the augmentation ideal of O.
O. (b) The k-algebra homomorphism W O.11 For any ﬁeld extension k 0 In other words. which is etale. 0 G/ ! k decomposes O. for each k-algebra R.124
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROOF. P ROPOSITION 12.G 0 /.G/=. P ROOF. 0/.
0 G/
Dk
0 G/
B.G
G 0/ '
0 . it gives a sequence of algebraic groups 1 ! Gı ! H !
0 G: 0 G.
1 ! G ı .G// D O.G/ ´ of k-bialgebras.G ı //.G// ' 0 .
This map factors through
0H .

a connected algebraic group is geometrically connected. there must be only one. Write specm O. specm A is irreducible if and only if the nilradical of A is prime (see 5d). and so is a ﬁeld. T HEOREM 12.. Therefore. The connected components of an algebraic group
125
12d
Connected algebraic groups
0 . An integral afﬁne algebraic variety V over a ﬁeld k is geometrically connected if and only if k is algebraically closed in O. For any ﬁeld k 0 containing k.G// is connected. Y =. for any algebraic group G. which is certainly the case if there exists a k-algebra homomorphism O. so also is Gk al . Let N0 be the nilradical of O.(c).X 2 C Y 2 / is an integral domain.X 2 C Y 2 / ' CŒX.15 In particular.
Theorem 12.Gk al /=N0 is an integral domain.V /. we obtain (d). Y =. there is a unique exact sequence 1 ! Gı ! G ! such that G ı is connected and
0 . The existence of the k-algebra homomorphism W O. (b))(a).G/
D EFINITION 12. In other words.Gk al /=N0 is injective. Proposition 12.
P ROOF.e. Proposition 12. all points have this property.G/ — we have shown that O.k/). the topological space specm. the topological space specm.Gk al / as a union of its irreducible components.O. (a))(d). By homogeneity (5.G/ ! k (the neutral element of G.G//
D k). but becomes a disjoint union of the two lines X C ˙iY D 0 over C — the ring RŒX. This is false for algebraic varieties: for example. (b) Let G and G 0 be algebraic groups over k.12 shows that G G 0 is connected if and only if both G and G 0 are connected.Gk al / D k al ˝k O.3 implies that 0 O. If G is connected.12. Y =.O.14 (a) Let G be an algebraic group over k. then so also is Gk al .16 The following four conditions on an algebraic group G are equivalent: (a) (b) (c) (d) G is connected. Y =. 2
.9 says that.5).G/=N is an integral domain.
The reason for the difference is the existence of the homomorphism W O. and so Gk al is irreducible. and so the irreducible components are disjoint.X iY /. R EMARK 12.G// D k.G// is irreducible.O.X C iY / CŒX. (c))(b). Remark 12.
R EMARK 12.Gk al / is connected.G/ has no nontrivial idempotents.G/ ! k implies that 0 .O.V / ! k (AG 11. but CŒX.G/ 0 . X2 C Y 2 D 0 is connected over R (even irreducible).
0 .G/
!1
´ is etale. In general. Trivial. the ring O.13 An algebraic group G is connected if
D 1 (i. if an algebraic group G over a ﬁeld is connected. No irreducible component is contained in the union of the remainder. As the canonical map O.5).G/=N ! O. As specm O.11 shows that G is connected if and only if Gk 0 is connected. there exists a point that lies on exactly one irreducible component. (d).

A SIDE 12. If N and Q are connected. As we noted in the proof of (6.Q/. E XAMPLE 12. E XAMPLE 12. and therefore has image f1g.q// D f˙1g (constant algebraic group).4)).R/ (of set-valued functors) deﬁnes an isomorphism of k-algebras O. It sufﬁces to prove this after replacing k with k al .q/ is connected. the natural isomorphism A.r. For example. so also is G. so this map factors through G ! Q (see 6. 1/W SLn . E XAMPLE 12. in which case we write SO.C/ is connected for the complex topology. Dn are connected because in each case O. we could for example deduce that GLn is a connected algebraic group from knowing that GLn . Assume N and Q are connected.126 P ROPOSITION 12. the algebraic group SO.17 Let
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
1!N !G!Q!1 be an exact sequence of algebraic groups. 0 . and so we may suppose 2 2 that q is the standard quadratic form X1 C C Xn . 2.C/ is connected from knowing that GLn is connected (of course.C/ is connected for the complex topology.
. which is isomorphic to the polynomial ring in the symbols Xij .q/.q/ ! f˙1g with kernel SO. Therefore O.Xij j i > j /. E XAMPLE 12. Thus. this requires the serious theorem stated in (12.q/ and 0 .Gm /.23 According to (12.G/. it is easier to deduce that GLn . The determinant deﬁnes a quotient map O. and the algebra on the right contains O. and so is an integral domain. Conversely. and G ı D Dn .q/ı D SO.19 For the group G of monomial matrices (2. P ROOF. so also is Q.k/ ¤ 2. r 7! A diag.2.9).22 The symplectic group Sp2n is connected (for some hints on how to prove this.4) and (12. since G maps onto 0 . 1 Ä i Ä j Ä n. In particular.21 Assume char.SLn / is a subring of O.23). kŒTn D kŒGLn =.O. Then N is contained in the kernel of G ! 0 .GLn /. if G is connected. GLn .G/ is an integral domain.18 The groups Ga .SLn /. Un (strictly upper triangular). 2
12e
Examples
E XAMPLE 12.GLn / ' O. conversely.q/ D SOn . However. : : : .SLn / ˝ O.V.13)). with the product X11 Xnn inverted.G// is a product of copies of k indexed by the elements of Sn .O. an algebraic group G over C is connected if and only if G.16).R/ Gm .20 The group SLn is connected. see Springer 1998. 1. The latter is shown to be connected in the exercise below. For any nondegenerate quadratic space . it must be trivial if G is connected.R/ ! GLn . Tn (upper triangular). 0 G D Sn (regarded as a constant algebraic group (4. Thus. q/.27). O.46).

Deduce that G is not isomorphic to the semi-direct product of G ı and 0 . Show that Gred is not a subgroup of G unless the extension splits
.In C A/ 1 . and that A is an integral domain.12. Show that G. i.
E XERCISE 12-3 (Springer 1998.R/ being connected.
12f
Afﬁne groups
For an afﬁne group G.R/ be the set of skew-symmetric matrices.In A/ deﬁnes a bijection from a nonempty open subset of SOn .Gi /i 2I is the family of algebraic quotients i 2I of G (see 7. Show that the k-bialgebra structures on A are in natural one-to-one correspondence with the group structures on S.R/ D f1g is certainly connected for the real topology.
0 Gi :
D lim
12g
Exercises
E XERCISE 12-1 What is the map O. and 3 is not connected as an algebraic group.SLn / ! O. the matrices A such that At D A. show that the following conditions are equivalent: (a) O. E XERCISE 12-5 For a ﬁnite algebraic group G. For any k-algebra R.G/ has two. (c) Deduce that SOn is connected.e. GL2 is connected as an algebraic group.) E XERCISE 12-7 Let k be a ﬁeld of characteristic p.Gred / is separable. we write G D lim Gi where . and we deﬁne G ı D lim
0G i 2I i 2I
Giı .k al /. but 3 . and conversely.24 For an algebraic group G over R.k al / onto an open subset of V .20? E XERCISE 12-2 Prove directly that
0 . (a) Show that the functor R 7! V . but GL2 . let V .. E XERCISE 12-4 Let A be a product of copies of k indexed by the elements of a ﬁnite set S .
E XERCISE 12-6 Let k be a nonperfect ﬁeld of characteristic 2.On //
Dk
k. 2.21). (b) Show that A 7! . The connected components of an algebraic group
127
A
12.R/ is not connected for the real topology.k/ has only one element but 0 .G/=N is not a Hopf algebra. so that there exists an a 2 k that is def not a square.GLn / deﬁned in example 12.2).R/ D fx 2 R j x 4 D ax 2 g becomes a ﬁnite commutative algebraic group under addition.O.2. G may be connected without G. Show that the extensions 0!
p
! G ! Z=pZ ! 0
with G a ﬁnite commutative algebraic group are classiﬁed by the elements of k =k p (the split extension G D p Z=pZ corresponds to the trivial element in k =k p ). (Hence 12-5 shows that O.R/ is represented by a ﬁnitely generated k-algebra A. For example. Show that the functor R G. (b) Gred is an algebraic subgroup of GI (c) G is isomorphic to the semi-direct product of G ı and
0 G.G/. Assume k has characteristic ¤ 2.

every afﬁne algebraic group has a composition series with the quotients listed at right: afﬁne G ´ j ﬁnite etale connected G ı j semisimple
solvable
j
unipotent
torus
j unipotent f1g We have constructed the top segment of this picture.
.128
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
12h
Where we are
As discussed in the ﬁrst lecture. we put everything together. Next we look at tori and unipotent groups. the semisimple ones. and ﬁnally. Then we study the most interesting groups.

denoted X.mult ı . Therefore the group-like elements form a submonoid of .2 For characters . P ROOF. then
and so . X .G/. .a/a. tori
In this section we study the algebraic groups that become diagonalizable over an algebraically closed extension ﬁeld. If a is a group-like element in A. According to (4.3 There is a canonical one-to-one correspondence between the characters of G and the group-like elements of O. 0 .x/ D x n 1
n .Gm / ! O.b/ D .g/ D . c are group-like elements in A and c is a unit.X / D X ˝ X . then 1 D . Thus the group-like elements of A are exactly the units such that . if a is a unit in A such that a D .c/
1
/D
D .b ˝ b/ D ab ˝ ab . characters of G correspond to homomorphisms of k-algebras O.R/ D R O. Conversely.A. To give a homomorphism of k-algebras O. / is group-like if .X / D 1 S. and a subgroup of A if A is a Hopf algebra. .G/ (the image of X).S ˝ idA / ı /. deﬁne C by .X n 1/ D kŒx .g/:
Then C is again a character.a ˝ a/.8).a/ D S.G/ amounts to giving a unit a in O.e ı / . hence a is group-like.c ˝ c/
1
Dc
1
˝c
1
when a.a/ . . Let A be a bialgebra.1 An element a of a k-coalgebra .2 A character of an algebraic group G is a homomorphism G ! Gm .a/. 129
.g/
.R/
13a
Group-like elements
.a/ D a ˝ a.x/ D 1 S.
(27)
. Let A be a Hopf algebra. and the homomorphism respects if and only if a is group-like.a/ D a ˝ a. idA / ı /.A.13
Groups of multiplicative type..G/. n / D kŒX =.a/ D 1.Gm / D kŒX.x/ D x ˝ x .a/ D .X / D X 1
1
D f 2 R j n D 1g O. P ROPOSITION 13.Gm / ! O. Because is a k-algebra homomorphism.a/a. We state for reference: Gm .c
1
. and the set of characters is an abelian group. and so a is a unit in A with a
1 (31)
D S. b. C
0 0 0
W G.a/ D . The correspondence in the proposition is an isomorphism of groups.a/ D 1.ab/ D .R/ ! R
0
/.
13b
The characters of an algebraic group
D EFINITION 13. /.a/ D a ˝ a and
D EFINITION 13.G/ respecting .

6 For any ﬁnitely generated abelian M . .M / with a ﬁnite product of copies of Gm and various n ’s.
and kŒM becomes a k-algebra (called the group algebra of M ) when we set P
i ai mi
P
Á P bj nj D i.M / R Hom.m ˝ m/ D .j /2I J . for m an element of the basis M .mult ı .
mi 2 M.m/ D m ˝ m. and so it is ﬁnitely generated.m/ D 1. ˝ id/.m ˝ m/ D 1 D . . ˝ id/.id ˝ /.m/ D m
1
.ei /i 2I and . m2 / $ m1 ˝ m2 W kŒM $ kŒM1 ˝k kŒM2 is an isomorphism of k-vector spaces. kŒM ' kŒGm . P ROPOSITION 13.m// D m ˝ .4 Let M be a cyclic group.j ai bj mi nj : j
It becomes a Hopf algebra when we set . kŒM ' kŒ n . Then the elements of kŒM are sums a0 C a1 e C C an 1 e n 1 with the obvious addition and multiplication (using e n D 1). . .130
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
13c
The algebraic group D.mult ı . .e/ D e 1 .5 If W and V are vector spaces with bases . the elements of kŒM are ﬁnite sums P
i
ai mi . the functor D.fj /j 2J . S.m// D 1 ˝ m. R / (homomorphisms of abelian groups)
is an algebraic group.e/ D e ˝ e. then W ˝k V is a vector space with basis .m ˝ m/ ˝ m D .id ˝S //.id ˝ /.m 2 M /
because.e/ D e n 1 . and let kŒM be the k-vector space with basis M .e/ D e ˝ e. .e/ D 1. . P (a) Case e has inﬁnite order. Thus. and S.e/ D 1.
ai 2 k.m//. generated by e.m// D m ˝ 1. and one checks easily that it respects the Hopf k-algebra structures. . and .M.m1 . . Therefore.i. E XAMPLE 13. E XAMPLE 13. (b) Case e is of order n.S ˝ id//. and . with coordinate ring kŒM .ei ˝ fj /.m ˝ m/: Note that kŒM is generated as a k-algebra by any set of generators for M . This shows that if M D M1 M2 . Therefore.M /
Let M be a ﬁnitely generated abelian group (written multiplicatively). . Then the elements of kŒM are the ﬁnite sums i 2Z ai e i with the obvious addition and multiplication. then .
. S. The choice of a basis for M determines an isomorphism of D. .

tori
131
P ROOF.4.ei / D i ci ei ˝ ei : The ei ˝ ej are also linearly independent.13. To give a k-linear map kŒM ! R is the same as giving a map M ! R.e/ D i ci . 2 R EMARK 13.M /
L EMMA 13. we see that the ci form a complete set of orthogonal idempotents in the ﬁeld k. and n is etale unless pjn).
. ci 2 k: We may even suppose that the ei are linearly independent. P ROOF.8 The group-like elements in any k-coalgebra A are linearly independent.9 The group-like elements of kŒM are exactly the elements of M . The map kŒM ! R is a k-algebra homomorphism if and only if M ! R has image in R and is a homomorphism M ! R .M / is represented by kŒM . and so a D mi for some i . Then P a D ci mi for some ci 2 k. and it is therefore an algebraic group. 2 ci 0 if i D j otherwise.e/ D ci . 2 L EMMA 13. Let a 2 kŒM be group-like.
deﬁnes a decomposition of k-bialgebras kŒM kŒGm ˝k ˝k kŒGm ˝k kŒ
n1 ˝k
˝k kŒ
nr
(13. P ROOF.M / is smooth except when k has characteristic p ¤ 0 and M has ´ p-torsion (because Gm is smooth. and so this implies that ci cj D We also know that . This shows that D.
13d
Characterizing the groups D. Groups of multiplicative type. and so one of them equals 1 and the remainder are zero.e/ D 1 P P . mi 2 M: The argument in the above proof shows that the ci form a complete set of orthogonal idempotents in k.13.j ci cj ei ˝ ej P P .ei / D ci : On combining these statements. If not.5). it will be possible to express one group-like element e as a linear combination of other group-like elements ei ¤ e: P e D i ci ei .e/ D e ˝ e D i.7 The group D. this proves the second statement. Now P . A decomposition of abelian groups M Z˚ ˚ Z ˚ Z=n1 Z ˚ ˚ Z=nr Z. Since every ﬁnitely generated abelian group M has such a decomposition. which contradicts our assumption that e is not equal to any of the ei .

M / is a contravariant equivalence from the category of ﬁnitely generated abelian groups to the category of diagonalizable algebraic groups (with quasiinverse G 7! X.11 An algebraic group G is diagonalizable if O.M 00 / ! D.! O.m/
! R D Gm .M // ' M: The character of D. As D sends direct sums to products.R/ D Hom. M 0 / ! Hom.e. e.10 shows that the functor is essentially surjective.G/. then 1 ! D. R /
def
f 7!f . The remaining cases are similarly easy.M /.M. T HEOREM 13. and so the inclusion M . and (88) is an isomorphism.M 0 /.M // (88) fdiagonalizable groupsg:
is an isomorphism for all M.G/). 2
13e
Diagonalizable groups
D EFINITION 13. that Hom. The group-like elements of kŒM span it by deﬁnition. P ROOF. . (c) Subgroups and quotient groups of diagonalizable algebraic groups are diagonalizable.M // D Hom..D.M. M 0 . for example.M 0 /.12 (a) The functor M D. suppose the group-like elements M span O.D.Z. Conversely. Then they form a basis for O. then Hom. / can be checked on the basis elements m 2 M . D.G/ span it as a k-vector space.G/ of vector spaces.D. P ROOF.M / for some M if and only if the group-like elements in O. and so it is an equivalence. Z/ ' Z. (a) Certainly.M / corresponding to m 2 M is D.10 An algebraic group G is isomorphic to D. Gm / D fX i j i 2 Zg ' Z. D.M. i. That this isomorphism is compatible with the bialgebra structures .M 0 / ! 1 is an exact sequence of algebraic groups.
. where it is obvious.M / ! D.G/ is spanned by group-like elements.132 Thus
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
X. (b) If 1 ! M 0 ! M ! M 00 ! 1 is an exact sequence of ﬁnitely generated abelian groups. Hom. If.R/:
def
P ROPOSITION 13.Gm .G/ (as a k-vector space). we have a contravariant functor DW fﬁnitely generated abelian groupsg We ﬁrst show that D is fully faithful.G/ extends to an isomorphism kŒM ! O. Proposition 13. it sufﬁces to do this when M. M 0 / D Hom.m. M and M 0 are both inﬁnite cyclic groups. M 0 are cyclic.

): As Dn is diagonalizable. and write P . In more down-to-earth terms.M / ! D. and so are isomorphic (6.13.G/.H / is surjective. the theorem says that an algebraic subgroup G of GLn is diagonalizable (in the sense of deﬁnition 13.idV ˝ / ı to v. (W Let W V ! V ˝k O. The family . tori
133
(b) The map kŒM 0 ! kŒM is injective. Let v 2 V . and so if the grouplike elements of O. Let M 0 be the kernel of M ! M 00 . Therefore M ! M 00 deﬁnes an isomorphism kŒM =IkŒM 0 ! kŒM 00 .! GLV (see 7. 2
13f
Diagonalizable groups are diagonalizable
Recall that Dn is the group of invertible diagonal n n matrices.G/ ! O. 0 P ROOF. and let H be its kernel. 80 19 0 > ˆ < = B C :: P gP 1 2 @ A : : ˆ > : .M 0 / and D. and the second shows that the set of ui ’s arising in this way span V (and hence include a basis). and so kŒM =IkŒM 0 is the quotient ring obtained by putting m D 1 for all m 2 M 0 . for all k-algebras R and all g 2 G. Groups of multiplicative type.47).ui / D ui ˝ ei 2 hui i ˝k A.u/ 2 hui ˝k O.11) if and only if there exists an invertible matrix P in Mn .k/ such that. the same is true of O.G/ span it. where IkŒM 0 is the augmentation ideal of kŒM 0 (see 6.M 0 / is a quotient map (by deﬁnition). But IkŒM 0 is the ideal generated the elements m 1 for m 2 M 0 . then O. ˝ idA / ı D idV :
ui ˝ e i ˝ e i D vD
X
i
. Its kernel is represented by kŒM =IkŒM 0 .76) . and let G be an algebraic subgroup of GLV .12c).M / ! D. we ﬁnd that P
i
D . There exists a basis of V for which G Dn if and only if G is diagonalizable.R/. so also is any subgroup of it (13. thus Dn ' Gm Gm ' D. that there exists a basis for V consisting of vectors u such that .M / ! Q be a quotient map. and so D. Then D. equivalently.ei /i 2I of group-like elements in O.H /.13 Let V be a ﬁnite-dimensional vector space.M 00 / for some quotient M 00 of M .v/ D i ui ˝ ei : On applying the identities (p.G/ is a basis. Then H D D.Zn /: „ ƒ‚ …
n copies
T HEOREM 13. Let D.idV ˝ / ı .M / ! Q are quotient maps with the same kernel.
2
. (c) If H is an algebraic subgroup of G.12).G/ be the comodule corresponding to the representation G .ui / ˝ ei
P
ui :
The ﬁrst equality shows that . We have to show that V is a direct sum of one-dimensional representations or.13).

Then X.14 An algebraic group is a split torus if it is isomorphic to a product of copies of Gm .t/ 1 and subtract it from (89). let V be the largest subspace of V on which T acts through the character .t /v1 C
C
m 1 .R/. (89)
On applying t 2 T . we get a new relation
1 . with quasi-inverse T 7! X. If the sum is not direct. Then M V : V D
2X. and consider V 1 C C V m . In other words.T / V . For example. there exists a t 2 G. A character W T ! Gm deﬁnes a representation of T on any ﬁnite-dimensional space V : let t 2 T .k/ such that m . m be distinct characters of T such that V i is nonzero.t /vm
D 0:
(90)
As m ¤ m 1 and T is smooth. For example. this means that the image of composite of is contained in the centre Gm of GLV and is the
T ! Gm . The functor M 7! D.t /.V / be a representation of a split torus on a ﬁnite dimensional vector space V . m2 / 2 Z ˚ Z is m m . vi ¤ 0.t /vm 1 C
m .134
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
13g
Split tori and their representations
D EFINITION 13.m1 . the split tori are the diagonalizable groups D. Let 1 . deﬁnes a representation of T on k n by 0 1 . then GLV D Gm . Continuing in this fashion. 2
. and the character corresponding to . In proving that the sum is direct.! GLV : If V is 1-dimensional. let T D Gm Gm . we may replace k by its algebraic closure. This will give us a new relation of the same form but with fewer terms.15 Let rW T ! GL . A quotient group of a torus is again a torus (because it corresponds to a subgroup of a free abelian group of ﬁnite rank).R/ act on R ˝k V as multiplication by .T /
P P ROOF.13 shows that V D 2X.k/ to (89). but a subgroup of a torus need not be a torus. vi 2 V i .t / ¤ m 1 .T /.R/.M / is a contravariant equivalence from the category of free abelian groups of ﬁnite rank to the category of split tori. then there exists a relation v1 C C vm D 0. T HEOREM 13.t1 . For example.t / 2 R .t / Let W T ! GLV be a representation of T . and so T always acts on V through some character.M / with M torsion-free. and it is a torus if it becomes a split torus over k al . Theorem 13.t /v all t 2 T . For each character .t /v D .T / ' Z ˚ Z.R/ ! Gm . : : : . we arrive at a contradiction.t / 0 B C :: t 7! @ A: : 0 . v 2 R ˝k V: More precisely. t2 / 7! t1 1 t2 2 W T . n is a subgroup of Gm (the map n ! Gm corresponds to Z ! Z=nZ). We say that T acts on V through if . Multiply (90) by m .

and each of its generators is deﬁned over a ﬁnite extension of k. Then V decomposes into a direct sum of subspaces V. P ROOF.G/ is a contravariant equivalence from the category of algebraic groups of multiplicative type over k to the category of ﬁnitely generated abelian groups with a continuous action of .16. Let M be a ﬁnitely generated abelian group. For any K G.
. tori
135
For example.k/ acts on V.R ˝k K/ .17 The functor G X .16 An algebraic group G is of multiplicative type if Gk sep is diagonalizable.m2 / . and so there are two possible actions of on X . and G ' Gm . To give a continuous action of on ˝ŒM is the same as giving a continuous action of on M . then it becomes diagonalizable over a separable extension (see Waterhouse 1979.m1 .k / (families of elements of Z indexed by the k-homomorphisms K ! k al ). Chapter 7). and let rW T ! GL.G/ ! k sep commuting with the actions of K . which is compact. such that . and the notation means the G. and let D Gal. and let X .m1 .R/ D R .K=k/ for some ﬁnite Galois extension K of k contained in k al . Then Aut. Then G.G/: (a) Trivial action. Then G.18 Take k D R.12 and Proposition 4. m2 / 2 m m Z Z.
K
2
where K is the subgroup of of elements ﬁxing K.G/. Let T be the functor R 7! .G/ D Hom. E XAMPLE 13.R/ D fz 2 C j z z D 1g.k sep =k/. Then T is an algebraic group. in fact. For an algebraic group G. N E XAMPLE 13.20 Need to add a proof that if G becomes diagonalizable over some ﬁeld containing k.V / be a representation of T on a ﬁnitedimensional vector space V . .Z.K/ D Hom. C / elements of C ﬁxed under the following action of Ã. so that is cyclic of order 2. (b) The generator Ã of acts on Z as m 7! m.R/ D Hom.K.
A SIDE 13.X .G/.Gk sep . all but a ﬁnite number of the subspaces V.13.m1 . A continuous action of on M is a homomorphism ! Aut. T HEOREM 13. t2 / 2 T . let T D Gm Gm . Gm /: Then acts continuously on X . we deﬁne X . and so this follows from Theorem 13. Groups of multiplicative type.19 Let K be a ﬁnite extension of k.
13h
Groups of multiplicative type
D EFINITION 13.t1 .m2 / are zero).M / factoring through Gal. k sep / k sep . the group of multiplicative type corresponding to the -module al ZHomk .G/ D Z. because X . Let G be a group of multiplicative type over k.K/ equals the group of homomorphisms X . Then can replace k al with k sep everywhere.m2 / as t1 1 t2 2 (of course. Ãz D z N
1
consists of the
:
Thus G.m1 .G/ is ﬁnitely generated.Z/ D Z D f˙1g.

2 id
! M=mM
P ROPOSITION 13. because G is connected. Because of the category equivalence T X.T /.˛/W X .T / D X . P P ROOF. which is simpler and doesn’t require G to be smooth (Waterhouse 1979. Then ˛ D id.k/ acts trivially on the X. A similar argument proves the theorem when H is ﬁnite (hence a ﬁnite product of groups of the form n ).g. 2 L EMMA 13. When H D n . that mjaij for i ¤ j and mjai i 1. It sufﬁces to show that X .
n/
' Hom.˛/ induces the identity map on the quotient X .ej / D i aij ei .
n/
Hom.7). i.T /=NX .136
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
13i
Rigidity
Later we shall need the following result. this implies that . 7. Let ˛W T ! T be a homomorphism whose restriction to TN is the identity map.
. and write ˛.(Add that TN is dense in T for every algebraic group T of multiplicative type. P ROOF. Choose a basis ei for M ..TN / for all N . and so G acts trivially on it.
n .24 Need to add the Hopf algebra proof. T HEOREM 13. The hypothesis is that. P ROOF. Clearly. We may suppose that k is algebraically closed. Gm /
' Z=nZ
(see 11c).23 Let T be a torus. which is trivial. and because g acts trivially on the kernel of mW T ! T it acts trivially on X.) 2
A SIDE 13. an action of G on M deﬁnes a map G ! Aut. if G is smooth). . and let TN be the kernel of N W T ! T .aij / D In .T / is the identity map. it sufﬁces to show that g 2 G. We can now apply the following elementary lemma.
n.22 Let M be a free abelian group of ﬁnite rank. for every integer m. but the hypothesis says that X .39. for 6.T / ! X .T /.k/ is dense in G (e. Then ˛ is the identity map.T /.aij / Á In mod m.T /=mX.e.21 Every action of a connected algebraic group G on a group H of multiplicative type is trivial. and let ˛W M ! M be a homomorphism such that M ! M ? ? ? ? y y M=mM commutes for all m. aij 2 Z. S and so this follows from the lemma. We next prove the theorem when G. The kernel of x 7! x m W T ! T is a product of copies of m .. We may suppose that H is a torus T .

Hint: To recognize the elements of Aut.13. E XERCISE 13-2 Let k 0 =k be a cyclic Galois extension of degree n with Galois group by . (b) Show that 80 9 1ˇ ˆ a1 a2 : : : an ˇ > ˆ ˇ > ˆB > < an a1 : : : an Cˇ = B Cˇ End .8). a2 a3 a1 ˇ generated
. m / ' . see the proof of (13. tori
137
13j
Exercises
E XERCISE 13-1 Show that Aut.Gm /k 0 =k . (a) Show that X . m /.X .G/ ' ZŒ (group algebra Z C Z C C Z n 1 of ). and let G D .R/ as complete systems of orthogonal idempotents.Z=mZ/ (constant group deﬁned by the group of invertible elements in the ring Z=mZ). Groups of multiplicative type.G// D B : ai 2 Z : : : Cˇ : : Aˇ ˆ@ : > : : ˇ ˆ : > ˆ > : .

x1 .14
Solvable algebraic groups
1
Let G be an abstract group. let S be a set of commuting n n matrices. The (ﬁrst) derived group G 0 (or DG) of G is the subgroup generated by commutators. Every automorphism of G maps commutators to commutators. it is normal). and let S be a set of commuting endomorphisms of V . : : : . In this section we extend this theory to algebraic groups.e. yn / 7! Œx1 . a basis e1 . Recall that the commutator of x. i. : : : . There exists a basis of V for which S is contained in the group of upper triangular matrices. : : : . : : : . and so DG is the union of the images of these maps. Because every element of S commutes with ˛.. and so G 0 is a characteristic subgroup of G (in particular. : : : . Note that the map G 2n 2 ! G factors through G 2n ! G.ˇx/: The induction hypothesis applied to S acting on Va and V =Va shows that there exist bases e1 . xn . then Sn (symmetric group on n letters) is not solvable because its derived series Sn An terminates with An . ˛. y D xyx y
1
D . : : : .ˇx/ D ˇ. Throughout.
14a
Commutative groups are triangulizable
We ﬁrst prove a result in linear algebra. yn W G 2n ! G
has image the set of elements of G that can be written as a product of at most n commutators. xn
1 . yn 1 :
A group G is said to be solvable if the derived series G DG D2 G
terminates with 1. Œx. yn 1 . en is a basis for V satisfying (91): N 138
. : : : . : : : . it is the smallest normal subgroup such that G=G 0 is commutative. 1.1 Let V be a ﬁnite-dimensional vector space over an algebraically closed ﬁeld k. y1 .yx/
1
:
Thus. y1 Œxn . there exists an ˛ 2 S and an eigenvalue a for ˛ such that the eigenspace Va ¤ V . y D 1 if and only if xy D yx. : : : . if n 5. emCi i for all i Ä n N N
Let emCi D emCi C Va with emCi 2 V . en for V =Va such that N N ˛. y1 .x1 . ei i for all i Ä m m:
2
hemC1 . and G is commutative if and only if every commutator equals 1. Otherwise. Then e1 . ei i/ he1 .he1 . y 2 G is Œx.he1 . : : : . We prove this by induction on the dimension of V . ei i/ ˛. y1
Œxn
1 .xy/. en such that ˛. : : : . emCi i/ N N he1 . : : : . The map (not a group homomorphism) .hemC1 . In fact. . If every ˛ 2 S is a scalar multiple of the identity map. yn 1 /
7! . then there is nothing to prove. For example.x1 . ei i for all i: (91)
In more down-to-earth terms. em for Va and emC1 . y1 . Va is stable under the action of the elements of S : for ˇ 2 S and x 2 Va . then there exists an invertible matrix P such that PAP 1 is upper triangular for all A 2 S . P ROOF. k is a ﬁeld.˛x/ D ˇ.ax/ D a. L EMMA 14. xn
1 . : : : . 1/
7! Œx1 .

Thus. P ROOF.R/) Gu
. if G is commutative. Solvable algebraic groups
139
P ROPOSITION 14.V /
Tn . there exists a basis for V such that the map GL. There exists a basis of V for which G is contained in Tn . First note that the subgroups Dn and Un of Tn have trivial intersection. 2
14b
Decomposition of a commutative algebraic group
D EFINITION 14.g/ is semisimple (resp.k/ D G. However.R/ D fIn g for all R.V / ! GLn .gh/u ¤ gu hu in general. unipotent) elements in G. However. and let G be a commutative smooth algebraic subgroup of GLV . (inside Tn .k/ as a product of subgroups. because Jordan decompositions do not respect products. . An element g of G. P ROOF. this will not in general be a decomposition of groups. g D gu ). Therefore G Tn . According to the lemma.k/ D G. for example.k/s (resp.k/ into Tn . g is semisimple (resp.k/ D G. unipotent) for all representations of G.k/ D G. then G G
multiplication
!G
is a homomorphism of groups.k/. unipotent) if g D gs (resp.4 Every commutative smooth algebraic group G over an algebraically closed ﬁeld is a direct product of two algebraic subgroups G ' Gs such that Gs .3 Let G be an algebraic group over a perfect ﬁeld k.18 shows that G.k/u (cartesian product of sets) (92)
where G. and so G \ Tn D G (see 6d).k/s G. in this case (92) realizes G.k/ deﬁned by it sends G.G \ Tn /. Thus.k/ is semisimple (resp. because Dn .14.R/ \ Un . Theorem 9.k/
Now G \ Tn is an algebraic subgroup of G such that .k/
GLn . We can do better. and so it does respect the Jordan decompositions (9.k/.k/s and Gu .k/: G. unipotent) if and only if .k/ GL.k/u ) is the set of semisimple (resp.k/u . T HEOREM 14.20). G.2 Let V be a ﬁnite-dimensional vector space over an algebraically closed ﬁeld k.

e.N1 \ N2 /. G. 6). there exist maps of functors G 2 ! G 4 ! ! G 2n ! G: Let In be the kernel of the homomorphism O.R/ and . Because G is commutative.R/ ' G.R/=N2 .k/Gu . (93) is surjective (6.k/ D G.41).R/ (GT 1. embed G in GLn for some n..G 2n / of k-algebras (not Hopf algebras) deﬁned by G 2n ! G: Then I1 I2 In T and we let I D In .50). which is analogous to the description of the derived group as the subgroup generated by commutators. Gs Gu ! G (93)
is a homomorphism with kernel Gs \ Gu (cf. we obtain an isomorphism G=N1 \ N2 ' G=N1 G=N2 :
Therefore.k/ and G is smooth. Because Dn \ Un D f1g as algebraic groups. P ROOF.R/ and N2 .R/=.8 The quotient G=DG is commutative (hence DG is the smallest normal subgroup with this property). In general. D EFINITION 14. for any ﬁnite T family . Then N1 .7 The derived group DG (or G 0 or G der ) of G is the intersection of the normal algebraic subgroups N of G such that G=N is commutative. However. G.R/.8). On varying R and passing to the associated sheaves. Let N1 and N2 be normal subgroups of G. 2 R EMARK 14.R/ D N1 . and it sufﬁces to consider ﬁnite families by (6.140
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
Choose an embedding G . As for abstract groups. A matrix A is unipotent if and only if 1 is its only eigenvalue. i.
A SIDE 14. and so (93) is injective. and so this is a polynomial condition.5)
14c
The derived group of algebraic group
Let G be an algebraic group over a ﬁeld k. Similarly.5 Let G be a smooth algebraic group over an algebraically closed ﬁeld k.R/ are normal subgroups of G.R/ G.Ni /i 2I of G= Ni is commutative if each quotient G=Ni is commutative.R/ \ N2 .
. if and only if its characteristic polynomial is . Gs \ Gu D f1g. G=N1 \N2 is commutative if G=N1 and G=N2 are commutative. But the coefﬁcients of the characteristic polynomial of A are polynomials in the entries of A.T 1/n . therefore it is an isomorphism.! Tn for some n.k/s will not be closed for the Zariski topology.G/ ! O.k/u is closed. because Gs . P ROPOSITION 14. and let Gs D G \ Dn and Gu D G \ Un . To see this.6 In fact every commutative afﬁne algebraic group over a perfect ﬁeld decomposes into a product of a group of multiplicative type and a unipotent group (Waterhouse 1979.R/=N1 .N1 \ N2 /. 2 We shall need another description of DG. 9. Therefore G.

so also is G 2n (5.k al / is the zero-set of In .9 The coordinate ring of DG is O.G/=I factors through O.k al /. nilpotents). As G is smooth and connected.1. and . P ROOF. and a descending sequence of prime ideals in a noetherian ring terminates.G/=I2n x ? ? O.k al /.G/=I .12 Let G be a smooth connected algebraic group.G/=I .G/ O.G 2n / is injective. so that the zero-set of In is DG. smooth). Solvable algebraic groups P ROPOSITION 14. Clearly.G/=In for some n. It follows that W O.G/=In . Choose n as in the ﬁrst part.
P ROOF.11 If G is connected (resp.G/ G 2n ? ? y G !
mult
141
G 4n ? ? y G
!
(because O. which corresponds to the smallest algebraic subgroup G 0 of G such that G 0 . Its closure in G.G/=In ˝ O.R/ contains all the commutators for all R.G/ ˝ O. C OROLLARY 14. so is gU 1 .14. Vn contains a dense open subset U of its closure (CA 12. From the diagram of set-valued functors G 2n ? ? y G we get a diagram of k-algebras O.DG/ D O. which must therefore meet U . then so would O.G/=In is the image of O. Let Vn be the image of G 2n . and so O. DGK D .k al /:
m
It remains to show that U U D Let g 2 DG.G 4n / ). P ROOF.G/ ! O. 2 C OROLLARY 14. nilpotents). 2 C OROLLARY 14.G/=I had a nontrivial idempotent (resp. and deﬁnes a Hopf algebra structure on O.G/ in O.10 For any ﬁeld K k. If O.k al /.13). nilpotent).k al // D Vm DG.G/ has no nontrivial idempotents (resp.G. 12.k al //. This proves the ﬁrst part of the statement (CA 16.k al /.DG/K :
2
P ROOF. smooth) if and only if O.G/=In for some n.k al /. forcing g to lie in U U . this is also the smallest normal subgroup such that G=G 0 is commutative. each ideal In is prime.G. Then O.DG/.G 2n / has no nontrivial idempotents (resp. smooth).5).k al / in G. then DG is connected (resp. The deﬁnition of I commutes with extension of the base ﬁeld. Recall that an algebraic group G is connected (resp.12).
.G/=I ˝ O. 2
1
DG. Because U is open and dense DG.! O. If G is connected (resp. then so also is G 2n .k al / D D. smooth). but (by deﬁnition) the homomorphism of k-algebras O.G/=I . Then [ U U 1 Vn Vn V2n D.G/=In x x ? ? ? ? O. Therefore.G/ ! O. Being the image of a regular map.

Therefore .rCi C1 .G/k be the algebraic group such that .T11 1. there is an algebraic subgroup Gr of G0 of matrices . and Ga .aij / 7! .17 The group Tn of upper triangular matrices is solvable.
A
14d
Solvable algebraic groups
Write D2 G for the second derived group D.14 An algebraic group G is solvable if the derived series G terminates with 1. the theory of solvable algebraic groups includes the theory of solvable ﬁnite groups.k/ when k is a nonperfect separably closed ﬁeld of characteristic p dividing n. P ROOF. In the ﬁrst case. so Gn Dn G. . and Gi =Gi C1 commutative.15 An algebraic group G is solvable if and only if it admits a sequence of algebraic subgroups G D G0 G1 Gn D f1g (94) DG D2 G
with Gi C1 normal in Gi for each i .G/k is solvable if and only if G is solvable. which is already quite complicated. E XAMPLE 14. L EMMA 14.rC2 . Tnn 1/.aij / such that aij D 0 for 0 < j i Ä r. the quotients are Gm in the second the quotients are Gm Gm Gm . the functor R G0 .13 Let G D PGLn . Ga Ga .G. D2 .DG/. and
demonstrate that T2 and T3 are solvable. and let . E XAMPLE 14. and so on.D2 G/k .G/k D . : : : .G/k . and so on.G/k D . .R/ D f.DG/k .16 Let G be a ﬁnite group. . Conversely.aij / j ai i D 1 for all i g
def
is an algebraic subgroup of Tn because it is represented by O. given a sequence as in (94).D2 G/. ˚ 1 « ˚ 10 « f. D3 G for the third derived group D. In particular.R/ D G for any k-algebra R with no nontrivial idempotents. so G2 D2 G. Then DG D G but D. . Then D. 2 A sequence of algebraic subgroups (94) such that Gi C1 is normal in Gi for each i and Gi =Gi C1 is commutative is called solvable series. G1 DG. The functor . : : : .k// ¤ G.142
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
14. If G is solvable. More generally. then the derived series is such a sequence. 0 /g 0 1 01 and n
0 0 0
Áo
n
1 0 1 0 0 1
oÁ
n
10 01 0 00 1
Áo
n
100 010 001
Áo Gm and Ga . D EFINITION 14.Tn /=. For example.a1. Similarly. : : :/
. ai.

we say that G acts on a subspace W of V through a character if W is stable under G and G acts on W through . Solvable algebraic groups
143
is a homomorphism from Gr onto Ga Ga with kernel GrC1 . we also call V an eigenspace for G with character .e. axn /. P ROPOSITION 14. Note that this means. then for every g 2 G.k/.k/: if w 2 W . Now let rW G ! GLV be a representation of G on V .R/ consists of the automorphisms preserving the ﬂag. xn / 7! . More generally. G ! Gm
r
GLV
then r is a character of G. If V is a sum of spaces V .. E XAMPLE 14.G ! GLVi =Vi i /: Then G0 is a normal algebraic subgroup of T with quotient isomorphic to Gn . p. Let G0 be the algebraic subgroup of G of ˛ acting as id on the quotients Vi =Vi i .134)..
14e
Independence of characters
Let Gm be the subgroup of GLn of scalar matrices.x1 . Gm is the centre of GLV .18 The group of n n monomial matrices is solvable if and only if n Ä 4 (because Sn is solvable if and only if n Ä 4.ax1 . Similarly.19 For each character of G.g/w is a scalar multiple of w. that the elements of W are common eigenvectors for the g 2 G. For this reason. Alternatively. in particular.
. Let rW G ! GLV be a representation of G on V . Let T be the algebraic subgroup of GLV such that T. the Gm ’s correspond. we can work abstractly. If r factors through the centre Gm of GLV . In fact.R/ D R acts on Rn as . more precisely.14. : : : . : : : . let V be the largest subspace of V on which G acts through .33). it is the subgroup deﬁned by the equations Xij D 0 for i ¤ j I X11 D X22 D D Xnn : Then a 2 Gm . In other words. i. i. Under the isomorphism GLV ! GLn deﬁned by any basis of V . GrC1 is a normal algebraic subgroup of Gr with quotient isomorphic to a product of copies of Ga . such that ˛. If G acts on subspaces W and W 0 through a character . Thus the sequence of algebraic subgroups Tn G0 G1 Gn D f1g exhibits Tn as a solvable group.R/ acts on R ˝k V by the homothety v 7! av. Therefore. GT 4. then it is a direct sum.Vi ˝ R/ Vi ˝ R for all k-algebras R. GLV contains a subgroup Gm such that a 2 Gm . A full ﬂag F in a vector space V of dimension n is a sequence of subspaces V D Vn Vi Vi
1
V1
f0g
with Vi of dimension i . then it acts on W C W 0 through .e. Y G0 D Ker. When we take F to be the obvious ﬂag in k n . and we say that G acts on V through the character r (cf. Now deﬁne Gr to be m the algebraic subgroup of G0 of elements ˛ acting as id on the quotients Vi =Vi r 1 : Again. there is a largest subspace V of V on which G acts through . vectors lying in eigenspaces corresponding to different ’s are linearly independent. r. G D Tn .

ˇ ˚ ab « ˇ a. P Suppose V D i 2I V i with the i (i 2 I ) distinct characters of G. then G is commutative. because then we can apply induction on the dimension of V (cf. According to (14.vi / D
P
i 2J
vi ˝ a. Then 0D P
i 2J
.G \ Tn / . it will be useful to see that the hypotheses are really needed. the proof of 14.144
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROOF. Let W be the sum of the nonzero eigenspaces V for N . For example.
A SIDE 14. As we saw in 13.k/ have a common eigenvector.2). k algebraically closed If G. The only common eigenvectors of D2 . a2 C b 2 D 1 b a is an commutative algebraic group over R that is not triangulizable over R. for some character of N . Unless k is algebraically closed. the sum is direct.20 This section duplicates part of 13g.19). namely. and so we can assume that the elements of N have a common eigenvector. P ROOF. then there exists a basis for V such that G Tn . i /
2
which contradicts the linear independence of the a.
. Also. If $ a. i / (see 13. and we proved the result in (14.
14f
The Lie-Kolchin theorem
T HEOREM 14. and solvable. connected The group G of monomial 2 2 matrices is solvable but not triagonalizable. If the derived series has length zero.
We saw in 13 that if G is a split torus.k/. M W D V and so the set fV g of nonzero eigenspaces for N is ﬁnite. e1 D . and k is algebraically closed.k/ are e1 D 1 and e2 D 0 . but the monomial matrix 0 1 0 1 interchanges e and e . vi ¤ 0.21 Let G be an algebraic subgroup of GLV . It should be moved somewhere earlier. solvable As Tn is solvable (14. this condition is necessary.k/.k/ Tn . J a ﬁnite subset of I .. and so there is no common eigenvector for the elements of 1 2 10 G. We prove this by induction on the length of the derived series G. Its derived series is shorter than that of G.k/ (because then .1 0 : : : 0/t .k/ D G. i. then the coaction of O.v/ D v ˝ a. then the elements of G. for example. It sufﬁces to show that there exists a basis for V such that G. it sufﬁces to show that the elements of G. b 2 R. If the sum is not direct.k/.G/ on V is given by . SLn has no nontrivial characters. In general. an endomorphism need not have an eigenvector. this will be far from true.k/ Tn .17) and a subgroup of a solvable group is obviously solvable.1). /.k/ have a common eigenvector. If G is connected.8). and.e. the space V (for N ) is nonzero. smooth.k/ G. V is always a sum of the eigenspaces V .G/. characters of G correspond to group-like elements of O. then there exists a relation P vi 2 V i . Before proving this. which implies that G T). Let N D DG. and so G \ Tn D G. (95) i 2J vi D 0. /.

Thus. But p r .w//. this shows that N. the argument shows that the character takes values in m Gm where m D dim V .w/'w0 for some c. it is closed for the Zariski topology on G. Choose a and let H be the stabilizer of V in G. Then there exists a basis of V for which G is contained in Un (in particular.
it is the complement of ﬁnite set of cosets.12).n/ 2 k: But each element n of N.w/ D c.f . and so 'w D c.w/ is a G-homomorphism.k/ stabilizes V .k/=N.k/ is connected for the Zariski topology (12. Let G act on End.w/ 2 k. and so G. and so. Moreover.k/ in fact acts trivially31 on V . On applying this equality to f0 .k/. then m is etale.k/ is a product of commutators Œx. it is either zero or an isomorphism. or characteristic p and p 6 jm.g
1
145
ng/x D g
.1).W / hw0 i.k/ permutes the ﬁnite set fV g. as claimed.k/. Since X ¤ 0. 5. This shows that G. If pjm.24 Let V be a ﬁnite-dimensional vector space.23 Let G ! End.k/. more detail.W / be a simple representation of an abstract group G on a ﬁnitedimensional vector space W over an algebraically closed ﬁeld k. GT 7. and so the image of W G ! Gm is ﬁnite. Thus.k/ D H .
. we ﬁnd that f0 .W / contains an element f0 W W ! W such that f0 .23 or II. is trivial.k/ D 1. For any w 2 W . g 2 G. An element n 2 N.W / has dimension one. As both X and W are simple.28). which is commutative.W / by the rule: .f0 / D w0 . Then 'w0 ı 'w is a G-endomorphism of X. there exists a w0 2 W such that 'w0 is an isomorphism. and let G be a subgroup of GL.k/ on V is trivial. Then the k-algebra of G-endomorphisms of X is a division algebra (Schur’s lemma.14. L EMMA 14. and let g 2 G. For n 2 N.V / consisting of unipotent endomorphisms.k/.T
P ROPOSITION 14.3).k/.k/ because it is the set where the characters and 0 coincide. If k has ´ characteristic zero. Hence G. ngx D g.k/ acts on V as the homothety x 7! .16). and so the action of N. Solvable algebraic groups Let x be a nonzero element of V for some .k/ (see 14. . In this case. Because N is connected.22 Is smoothness needed?
14g
Unipotent groups
We begin by proving two results from linear algebra.w/ D g. This shows that .gf /. We may suppose that X is simple. P ROOF.W /.g
1
ng/ gx
For the middle equality we used that N is normal in G. and so f0 . 2 Recall that an endomorphism of a vector space is unipotent if its characteristic polynomial is 1/dim V .n/x. w 2 W: Then every nonzero G-subspace X of End. But every closed subgroup of ﬁnite index of a topological group is open30 . and so n acts on V as an automorphism of determinant 1. Let f0 2 X be such that 'w0 . each of which is also closed. we know there is a common eigenvalue (14. because N is connected. f 2 End.k/. the argument only shows that takes values in p r for p r the power of p dividing m. But G.w/w0 .n/dim V D 1. and hence equals k (GT 7. gx lies in the eigenspace for the character 0 D .g
1
ng/x D .k/ acts on V through the quotient G. G is solvable). This shows that W D V . the map 'w W X ! W sending f to f . and so G.
31 In 30 Because
. and so H is closed and open in G. H is a subgroup of ﬁnite index in G. 1 Let w 2 W .n 7! . y of elements of G.g 1 ng// of N . 2
A SIDE 14.

ei /1Äi Än for V .k/ is dense in G. (c))(d). and (c). Then .G/-comodule). which contradicts the fact that f0 2 U . (d))(c).g 0 1/jW 2 U for all g 0 2 G. TrW . and . and so TrW . and so U is nonzero if G acts nontrivially on W .G/ is coconnected. Obvious. the conditions are equivalent to: (d) all elements of G. there exists a basis of V for which the image of G is contained in Un .k/ is dense in G.e.V / of G. and so has a solution in k n if and only if it has a solution in . P ROOF. the proof of 14. Because G.3.k/ Udim V . The lemma then shows that U contains an element f0 such that f0 . Any quotient of a coconnected Hopf algebra is coconnected (the image of a ﬁltration satisfying (96) will still satisfy (96)).g 0 1// D TrW . Such an f0 has TrW f0 ¤ 0.W / has dimension one. Cr D A. According to Proposition 14. We ﬁrst prove the equivalence of (c) and (d) when G.W / j TrW . Choose a faithful representation G ! End. (c) G is isomorphic to an algebraic subgroup of Un for some n.k/ are unipotent.25 The following conditions on an algebraic group G are equivalent: (a) every nonzero representation of G has a nonzero ﬁxed vector (i. Therefore. 8. Clearly W is simple.gg 0 / TrW .g/ D 0:
Let U D ff 2 End. there exists a basis of V for which G.146
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROOF. (c))(a).Xij / D Xij ˝ 1 C 1 ˝ Xij C Xi r ˝ Xrj :
i <r<j
. (b). Prove (b) by induction on the dimension of V .1). 2 T HEOREM 14. (a))(b).g.V / of G. A Hopf algebra A is said to be coconnected if it satisﬁes the following condition: there exists a ﬁltration C0 C1 C2 of A by subspaces Ci such that [ X C0 D k. It sufﬁces to show that there exists a nonzero subspace of W on which G acts trivially.Un / D kŒXij j i < j . because then we can apply induction on the dimension of V to the vector spaces W and V =W to obtain a basis of V with the required property (cf.gf / D 0 for all g 2 Gg. (b) for every representation G ! End. We conclude that G acts trivially on W .k al /n . (b))(c). When G. We now prove the equivalence of (a).P Choose a basis .).. and then show that (a) holds for any algebraic group G such that O. For each g 2 G.G/ is coconnected. (Following Waterhouse 1979.24. and so it sufﬁces to show that O.Cr / Ci ˝ Cr i : (96)
r 0 0Äi Är
We ﬁrst show that (c) implies that O. and that X .Un / is coconnected.k/ is dense in G. we may suppose that k is algebraically closed. Recall that O. a nonzero v 2 V such that . Let W be a nonzero subspace of V of minimal dimension among those stable under G. The condition that a vector v D ai ei is ﬁxed by all g 2 G is linear in the ai . Apply (b) to a faithful representation of G.v/ D v ˝ 1 when V is regarded as a O.k/.g/ D dim W . this implies that G Udim V .

v/ 2 V ˝ Cr g. so that a monomial Xijij will have weight Sij .v/ D v 0 ˝ 1 for some vector v 0 . For the Xij it is obvious. and let W V ! V ˝ A be a comodule.VrC1 / V ˝ CrC1 .14. It sufﬁces to check the third condition in (96) on the monomials. then the same is true of every quotient of G. 2 D EFINITION 14. and so v is ﬁxed. Then V is a union of the subspaces Vr D fv 2 V j .
. If the condition holds for monomials P . on applying .v/ D v ˝ 1 is a linear condition on v (or. so also does G. We now use that .VrC1 / V˝ X
i
Ci ˝ Cr
i:
Hence VrCi maps to zero in V ˝ A=Cr ˝ A=Cr . if Gk 0 satisﬁes (a). We have (see the proof of 14. 2 C OROLLARY 14. If G can be realized as a subgroup of Un . P ROOF. An algebraic group G over k is unipotent if and only if Gk 0 is unipotent. Clearly.Q/ lies in Á X Á X X Cj ˝ Cr j Ci Cj ˝ Cr i Cs j Ci ˝ Cr i X Ci Cj ˝ CrCs i j . If every representation of G contains a nonzero ﬁxed vector. rather.P / .G/ coconnected H) Gk 0 unipotent.25): G unipotent H) O. then .id ˝ / ı D .27 Every subgroup and quotient group of a unipotent group is unipotent.V ˝ A=Cr / ˝ A=Cr is injective.28 Let k 0 be a ﬁeld containing k. and so . and Ci Cj Ci Cj . Now assume that A is a coconnected Hopf algebra. and on applying ˝ id we ﬁnd that V ! . Solvable algebraic groups
147
Q n P Assign a weight of j i to Xij . because . Hence VrC1 D 0. If V0 contains a nonzero vector v.G/ coconnected H) k 0 ˝ O.k al / consists of unipotent elements.25).id ˝ / . and let n Cm be the subspace spanned by the monomials of weight Ä m. we ﬁnd that v D v 0 .26 An algebraic group G is unipotent if it satisﬁes the equivalent conditions of (14. it is a linear condition on the coefﬁcients of v with respect to some basis for V ).PQ/ D .j i /. 2 C OROLLARY 14. .29 A smooth algebraic group G is unipotent if G. then . Q of weights r. C0 D k. r 0 Cr D A. We complete the proof by showing that Vr D 0 H) VrC1 D 0: By deﬁnition. We proceed by induction on weight of a monomial. then so also can any subgroup of G. The map V ! V ˝ A=Cr deﬁned by is injective because Vr D 0. ˝ id/ ı . P ROOF. C OROLLARY 14. Conversely. s.

4) Q ' Qu Qs . which implies that G itself is unipotent. Certainly Gu is a normal algebraic subgroup of G satisfying (a) and (b). over a ﬁeld k of characteristic zero. We may suppose that the base ﬁeld is algebraically closed. an algebraic group G is deﬁned to be unipotent if admits a composition series over k al whose quotients are isomorphic to subgroups of Ga . there exists a nontrivial homomorphism H ! Ga . 3.(v)). Both these conditions are equivalent to the condition that G be isomorphic to an afﬁne algebraic subgroup of Un for some n (DG IV.35 Let G be a connected solvable smooth group over a perfect ﬁeld k. SGA3 XVII. P ROOF. so that (14. P ROOF. more precisely.k al / can be unipotent without G being unipotent.31 A unipotent group need not be smooth.30 An algebraic group that is both unipotent and of multiplicative type is trivial.1. 2
A A
14.32 For an algebraic group G.485. Let Q D G=DG. We next prove that Gu is connected. but those over a ﬁeld of characteristic zero are classiﬁed by their Lie algebras. the functor G Lie.G/ is an equivalence from the category of unipotent algebraic groups over k to the category of nilpotent Lie algebras over k (see II 3. in characteristic p. p. 2. For example. the algebraic group p has p . 0 1 14.3.148
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
2
P ROOF. The formation of Gu commutes with change of the base ﬁeld.5. A nontrivial algebraic group of multiplicative type contains a subgroup ` for some prime `.
P ROPOSITION 14. We ﬁrst prove this when k D k al . the subgroup of U2 consisting of matrices 1 a with ap D 0 is not smooth (it is isomorphic to ˛p ). but it is easy to see that there is no nontrivial homomorphism ` ! Ga . For example.487. 2. an algebraic group G is deﬁned to be unipotent if.27). 2.34 The unipotent algebraic groups over a ﬁeld of characteristic p ¤ 0 are complicated. Embed G into Tn for some n.
. all elements of G.5(i). A SIDE 14.k al / D 1. 2.
14h
Structure of solvable groups
T HEOREM 14. p. It is commutative. in characteristic p. The condition implies that Gk al is unipotent. There exists a unique connected normal algebraic subgroup Gu of G such that (a) Gu is unipotent. 1.
A SIDE 14. for every nontrivial subgroup H . but it is not unipotent.33 In DG IV. In SGA3 XVIII. and construct 1 ! Un x ? ? ! Gu ! Tn x ? ? ! G ! Dn x ? ? ! T ! 1
1
! 1
where Gu D Un \ G and T is the image of G in Dn . (b) G=Gu is of multiplicative type.

DG Gu . For the uniqueness.N / ! 0 . and so (12.Gk al /u implies that it is stable under . so would Q). Omitted for the present (cf. the closure of a connected set is connected.k al / (and apply a previous result). 19.
2
14j
Exercises
E XERCISE 14-1 Give a geometric proof that G connected implies DG connected.
!
0 .Q/ ! 1 (in an obvious sense).N / ! 0 .
.4).Gu /. DG connected H) Gu connected. If T and T 0 are maximal tori in G. 19. P ROOF. say). As G=Gu is commutative.G/ ! 0 . this shows that
D Qu . or that Gu .G/ need not be injective.Gu /
! 1
? ? y ! Q ? ? y ! 1
1
! DG
! Q= Gu ? ? y 1
0 . note that Gu is the largest connected normal unipotent subgroup of G. which clearly has the required properties.2
14i
Tori in solvable groups
P ROPOSITION 14.k al / consists of the unipotent elements of G.28). so also is 0 . Give an example to show that 0 . Omitted for the present (cf. Humphreys 1975. and a nested union of connected sets is connected sets is connected. the uniqueness of .36 Let G be a connected smooth solvable group over an algebraically closed ﬁeld. and the diagram 1 ! DG ! Gu ? ? y ! G ? ? y T ? ? y 1 shows that T ' Q=
0 . then apply the criterion (12.k/. When k is only perfect. Humphreys 1975.] E XERCISE 14-2 Show that if 1 ! N ! G ! Q ! 1 is exact. then T 0 D gT g 1 for some g 2 G.17)
Qu . [Show that the image of connected set under a continuous map is connected (for the Zariski topology.14. and hence arises from a unique algebraic subgroup Gu of G (6.16). P ROOF. Solvable algebraic groups
149
´ This shows that Qu is connected (if it had an etale quotient.Gu /
Since .2).Gu /
Qu .37 The centralizer of any torus in a connected smooth solvable group G is connected.
2
P ROPOSITION 14.

ei . q1 / and . Throughout this section.
A smooth connected algebraic group is semisimple if it has no smooth connected normal commutative subgroup other than the identity. /.
A SIDE 15. q/ by . x/
for some symmetric bilinear form q.V1 . / is connected and almost-simple.V. y/ for all x. x.ei . . Note that q is zero (i. By . Often I’ll write (rather than q ) for the associated symmetric bilinear form and denote . .x2 /. . Let . q/ is the determinant of the matrix . A basis e1 . / which we now construct. the most important case. q/ consisting of a ﬁnitedimensional vector space and a quadratic form q. For example. The special orthogonal group SO. A quadratic space is a pair . y 2 V )..V.1 This section needs to be rewritten. k is a ﬁeld not of characteristic 2 and “k-algebra” means “associative (not necessarily commutative) k-algebra containing k in its centre”.
15a
Quadratic spaces
Let k be a ﬁeld not of characteristic 2. which is. x 1 2 V1 . q / or . ej // by a nonzero square.x. and the next chapter is largely devoted to their study and classiﬁcation.y/ C 2
qW V q . we conﬁne ourselves to constructing an important class of semisimple groups. q2 / be quadratic spaces. : : : . and study their basic properties. en for V is said to be orthogonal if . q1 / ˚ .V.V1 .x1 / C q.V2 .V / D 0) if and only if is zero (i.V. the n n matrices with entries in k become such a k-algebra Mn . V / D 0).x/ ¤ 0. In this section. 150
.x/ D 0 and anisotropic if q. P ROPOSITION 15.15
The classical almost-simple groups
In this section.x C y/ D q. and so the discriminant is an element of k=k 2 . A quadratic form on V is a mapping qW V ! k such that q. ej / D 0 for all i ¤ j . .
V ! k.ei .k/ once we identify an element c of k with the scalar matrix cIn . in fact.V. y/ D . Note that (97)
and so q is uniquely determined by q.2 Every quadratic space has an orthogonal basis (and so is an orthogonal sum of quadratic spaces of dimension 1).x.e. The choice of a different basis multiplies det. q. The discriminant of . x/ D q1 . q/ be quadratic space.e. ej // where e1 . Let be a nondegenerate bilinear form on a k-vector space V .x/ C q. : : : .x/ for all x 2 V (equivalently.x/ D
q . At the moment it only (partially) treats the spin group. q2 / we mean the quadratic space . An isometry is an injective k-linear map W V1 ! V2 such that q2 . x2 2 V2 :
Let .V.x1 C x2 / D q.. q/ with V D V1 ˚ V2 q. y/. en is a basis of V .V. and let V be a ﬁnite-dimensional k-vector space. A nonzero vector x in V is isotropic if q. This is the most important class of algebraic groups.V2 . we construct the split classical almost-simple algebraic groups. and it has a 2-fold covering Spin.V.x.

x/ 2 2 . en / e q. Finally. hai? D fx 2 V j .
ci D q. . we can even scale the ei so that each ci is 0 or 1. In the plane this is the line orthogonal to x y. q/ is said to be regular32 (or nondegenerate. then every basis is orthogonal. Then there exists a composite of reﬂections V ! V extending .. for example.V. and so det. Moreover. q 0 /. q.a/ D q. x/ a. : : : .x1 . and. : : : . : : : .. e2 is again a basis for V .V / D 0. Apply
An orthogonal basis deﬁnes an isometry . and let be an isometry from a subspace W of V into V . T HEOREM 15. xn / D c1 x1 C 2 C cn xn .e.x/ D x 2 .e/ . x/ D 0g is a hyperplane in V (i. ˘ totally isotropic if every nonzero vector is isotropic. if x y is anisotropic.a/2
and so Ra is an isometry. x/ C q. Suppose ﬁrst that W D hxi with x anisotropic.e.. i. : : : .e. let e 2 V be such that q. and ˘ anistropic if it is not isotropic. Geometry in the plane suggests that we should reﬂect in the line x C y.Ra . For an anisotropic a 2 V . and let x D y. .a.e.x/ D 0 for some x ¤ 0. . Then for any nonzero a 2 V . then Rx y .a.a. if k is algebraically closed.k n .V. . en a basis for W . the
Then Ra sends a to a and ﬁxes the elements of W D hai? . P ROOF. a subspace of dimension dim V reﬂection in the hyperplane orthogonal to a is deﬁned to be Ra . and extend it to a basis e.3 Let . en for V . Let .x// D q. . 1.
15b
Theorems of Witt and Cartan-Dieudonn´ e
A quadratic space . ) if for all x ¤ 0 in V . relative to a basis a. q/ ! . i.x/ D y
32 With
the notations of the last paragraph. it is singular.e. q. If q.ei / 2 k:
If every element of k is a square. if q. its matrix is diag.x. W / D 0. : : : .
. e2 . Otherwise. Also.Ra / D 1. if q. there exists a y such that .. Then e.V. en q. 1. x/ 4 .a/ q.e/ ¤ 0.. e2 / e. Otherwise. : : : . 1/.a/
def
def
1).x/.15.V. q/ is regular if c1 : : : cn ¤ 0. q/ is ˘ isotropic if it contains an isotropic vector.x/ D 0 for all x.x/ D 0 implies x D 0. . i. if q.e. x/2 q. y/ ¤ 0.e/
2
1 vectors span a subspace W for which . where
2 q 0 .V. The classical almost-simple groups
151
P ROOF.e.a. q/ be a regular quadratic space. and the last n induction to W .a. en with e2 .V. e2 . q/ be a regular quadratic space.

and we let W 0 D hxi? \ W . As m.y/.x y. moreover.x/
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
y.x
y/ D y.W / 1.W 0 / D m. By the argument in the ﬁrst paragraph.x/ D q.x y. x/ . then 4q. we see that x 2 W 0? . there exist reﬂections (one or two) of the form Rz .
x/ D y:
x. and so W D hxi ˚ W 0 (orthogonal decomposition).x y/ D q. y/ .x .x. Let W 0 be a subspace of V mapped isomorphically onto W _ .W / D dim W C 2 dim.
1
˙ 0 w 0 / D .
. y/ ¤ 0. x y/ D 2 . for any w 0 2 W 0 . x/I
Dx
2 .
y is isotropic.x C y/ C q.cx C w 0 / D ˙ 0 .5. there exists a y 2 W 0 such that . k/ be the dual vector space.x/ can apply induction to obtain a composite ˙ 0 of reﬂections such that ˙ 0 jW 0 D jW 0 . if x C x 0 2 W C W 0 with x 0 ¤ 0. x/ D
.W \ W ? /: C ASE W NOT TOTALLY ISOTROPIC : In this case.x. In this case. then there exists a y 2 W such that 0 ¤ . note that . w 0 / D . Then.152 as required.˙ 0 ı ˙ 00 /. the argument in the proof of (15. x y/
y/ D x
.x/ D q.w.x
y. we q. and so . for w 2 W . z 2 W 0? . if x ¤ 0. Indeed.x. Because ˙ 00 acts as the identity on W 0 .˙ 0
def 1 .2) shows that there exists an anisotropic vector x 2 W . and consider the surjective map ˛W V
x7! . y/ D .x because q.x y. whose composite ˙ 00 maps x to y.cy C w 0 / D c x C w 0 : C ASE W TOTALLY ISOTROPIC : Let V _ D Homk-lin .x hence Rx If x
y .x C y/
and so x C y is anisotropic. 5.x C x 0 . w . To see this.x/ x 2 W 0 . w 0 / D 0.x 0 . .V. y/ on W ).x. ˙ 0 ı ˙ 00 is the map sought: .
and so y D ˙ 0 1 x 2 W 0? . /
!V_
f 7!f jW
!W_
(so x 2 V is sent to the map y 7! .x/ D Rx We now proceed33 by induction on m.x . Chapter 1.x. RxCy ı Rx . From the deﬁnition of W 0 . x C y/ D 0 y. Then W \ W 0 D f0g and we claim that W C W 0 is a regular subspace of V .
33 Following
Scharlau 1985. y/
y. y/.

x 0 / C f 0 .x 0 . q/ is a composite of reﬂections.x 0 //W W C W 0 ! W ˚ W _ is an isometry. W /_ : Now the map W CW 0 ! W ˚W _ is an isometry extending . q1 / and . q1 / ˚ .V1 . by Theorem 15. As m. : : :/W W C W 00 ! W ˚ . 2 R EMARK 15. and suppose that dim W1 Ä dim W2 .4 Every isometry of .V1 . Now 2 1 W .6 All maximal totally isotropic subspace of . q/ have the same dimension. q1 / ˚ . q/ D . Then . the map x C x 0 7! . f 0 / 7! f . q2 / 0 0 0 0 with .V1 . W /_ ! W C W 00
(99)
V
2
2
? V to an isometry of V . q/ has orthogonal decompositions
0 0 0 0 .x.V2 . P ROOF. q2 / are isometric.W / we can apply induction to complete the proof.3 it extends to an isometry W V ! V .5 (W ITT CANCELLATION ) Suppose . Witt’s theorem says simply that there exists an isometry extending to V (not necessarily a composite of reﬂections).V2 . C OROLLARY 15.V.
. Therefore.W ˚ W 0 / D 2 dim W < 3 dim W D m. .x/.V. f /. and the Cartan-Dieudonn´ e theorem says that every isometry is a composite of at most dim V reﬂections. It will map V2 D V1 isometrically 2
C OROLLARY 15. Relative to this bilinear form. W1 D 2 dim W1 D dim 1 W2 D dim W2 . q2 / and .V2 . P ROOF.7 In the situation of Theorem 15. Because W1 is maximal. C OROLLARY 15.V1 . When V is anisotropic. and so totally isotropic subspace of V containing W1 . (98) ˚
_ 1
153
(98)
(99)
! W ˚ .x. ˛. but the general case is considerably more difﬁcult — see Artin 1957. Let W1 and W2 be maximal totally isotropic subspaces of V . 0 P ROOF. The same argument applied to W gives a subspace W 00 and an isometry x C x 00 7! .3 shows this. Then there exists an injective linear map W W1 ! W2 V .V.x.V. q2 / D . The classical almost-simple groups Endow W ˚ W _ with the symmetric bilinear form .V2 . the proof of Theorem 15. This is the special case of the theorem in which W D V .3. which is automatically an 1 W is a isometry. q1 / regular and isometric. Extend an isometry V1 ! V1 0 0? onto V2 D V1 .15.

20). T HEOREM 15..T / D ˙1.8 The (Witt) index of a regular quadratic space . e2m i? . D EFINITION 15.5).q/. q/ is regular and isotropic of dimension 2I (b) for some basis of V . /). Va is uniquely determined up to isometry. : : : . : : : .V.q/. q/ with Witt index m has an orthogonal decomposition V D H1 ˚ ˚ Hm ˚ Va (100)
with the Hi hyperbolic planes and Va anisotropic. O.ei . : : : . 2
15c
The orthogonal group
Let . C1 C0 C1 . q/ is the maximum dimension of a totally isotropic subspace of V .Ci / Di for i D 0.emCi / D 0 for i Ä m.T /2 D 1. the matrices T such that T t M T D M: Thus.e.V. det.10 (W ITT DECOMPOSITION ) A regular quadratic space . emC1 .
. Relative to a basis for V .V. em . The uniqueness of Va follows from the Witt cancellation theorem (15.q/ consists of the automorphs of the matrix M D . and so det. q/ is a hyperbolic plane if it satisﬁes one of the following equivalent conditions: (a) . P ROOF.
15d
Super algebras
Recall (2d) that a superalgebra (or Z=2Z-graded algebra) over k is k-algebra C together with a decomposition C D C0 ˚ C1 of C as a k-vector space such that k C0 .q/ is an algebraic subgroup of GLV (see 2. : : : .154
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
D EFINITION 15. q/ be a regular quadratic space. called the special orthogonal group SO. and let e1 . C0 C1 C1 . The subgroup of isometries with det D C1 is an algebraic subgroup of SLV . em be a basis for W . O. called the orthogonal group of q (it is also called the orthogonal group of . q/. Let T 2 O. emCj / D ıij (Kronecker delta) and q. emCi i and Va D he1 .ei . Then V decomposes as (100) with34 Hi D hei .9 A quadratic space . A homomorphism of super k-algebras is a homomorphism 'W C ! D of algebras such that '.V.q/ to be the group of isometries of . . the matrix of the form is 0 1 . As det M ¤ 0. ej //. i. C0 C0 C0 .V. moreover. and denoted O. e2m such that . 10 (c) V has dimension 2 and the discriminant of q is 1 (modulo squares). One easily extends the basis to a linearly independent set e1 . Let W be a maximal isotropic subspace of V .V. Deﬁne O. C1 C1 C0 :
Note that C0 is a k-subalgebra of C . 1:
34 We
often write hS i for the k-space spanned by a subset S of a vector space V .

e ˝ 1/: Therefore. O iD W D ! C ˝D. C. cn 2 k. cn /: O 1 /˝C.c1 /˝C. ci 2 Ci . 1/j k . dj 2 Dj etc. c 7! c ˝ 1 d 7! 1 ˝ d
have the following universal property: for any homomorphisms of k-superalgebras f W C ! T. : : : .dj /f . 1 ˝ e.ci ck ˝ dj dl0 /
1 0 ˝ dl /
ci 2 Ci .c1 . The super tensor product of C O and D. gW D ! T
O there is a unique superalgebra homomorphism hW C ˝D ! T such that f D h ı iC .c1 .c1 /˝ O ˝C. cn / to be the k-algebra with generators e1 .1 ˝ e/ D e ˝ e D .ck 0
D .c2 / ' C.15. C. O E XAMPLE 15. C ˝D.e ˝ 1/2 D e 2 ˝ 1 D c1 1 ˝ 1 . The classical almost-simple groups
155
E XAMPLE 15. When we set C0 and C1 equal to the subspaces i i C0 D he11 : : : enn j i1 C
C in eveni C in oddi.11 Let c1 .C1 ˝ D0 /
0 D .c1 . and .ci / ' C. is deﬁned to be the k-vector space C ˝k D endowed with the superalgebra structure O C ˝D O C ˝D
0 .12 As a k-vector space. C. 1gg. cn /.
i i As a k-vector space. : : : . e ˝ e. whose images anticommute in the sense that f . cn / has basis fe11 : : : enn j ij 2 f0. 1/ij g. then it becomes a superalgebra. O C. : : : .
. en and relations 2 ei D ci . and so has dimension 2n . O C.1 ˝ e/2 D 1 ˝ e 2 D c2 1 ˝ 1 .c2 / has basis 1 ˝ 1.ci /. : : : . : : : . : : : . : : : .
C1 D
i i he11 : : : enn
j i1 C
of C. : : : .e ˝ 1/.c1 .ci ˝ dj /.C1 ˝ D1 / D . c2 / e ˝ 1 $ e1 1 ˝ e $ e2 : Similarly.
The maps O iC W C ! C ˝D.c1 .dj / D . ej ei D ei ej (i ¤ j ). Let C D C0 ˚ C1 and D D D0 ˚ D1 be two super k-algebras.c1 .C0 ˝ D0 / ˚ .. dj 2 Dj .C0 ˝ D1 / ˚ .c1 /˝C. ci and so.cn / ' C.c1 . ci /. e ˝ 1. by induction. Deﬁne C. g D h ı iD .ci /g.1 ˝ e/.

V.V / is the k-algebra of noncommuting polynomials in e1 .e.q/ :
.X /L i H i .
def L E XAMPLE 15. we are regarding k as a subﬁeld of R.x/ . V ˝n D V ˝
where
˝ V .15 The Clifford algebra C.x/2 D q.x 2 V /.x/2 D q. L There is a k-linear map V ! T . : : : .
Let .V / ! R.V / ! C. R/ and H. namely. then A ˝k B D A˝k B. it is necessary to write (101) as . all x 2 V: Note that if x is anisotropic in V . : : : .14 Let X be a manifold. R/.
15f
The Clifford algebra
be the corresponding bilinear form on V .X /1 D i H 2i C1 .x/ is invertible in C. V D V ˝1 . Then H. Then is k-linear.X. The tensor algebra of V is T .V.V / of V by the two-sided ideal I.. q/ is the quotient of the tensor algebra T .x/. q.X. and let
D EFINITION 15.V.x/ (101)
35 For a k-algebra R.
Let V be a k-vector space.V /. and even a superalgebra with H.156
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
E XAMPLE 15.V. If Y is a second manifold.x/ . B are both k-algebras. q/ be a quadratic space.X /˝H. the K¨ nneth formula says that u
H. If V has a basis e1 .x/ D 1.
O Y / D H.X. i. because (101) shows that .q/ generated by the elements x ˝ x q.V / and the quotient map T .V.X /0 D i H 2i . then T .v1 ˝ ˝ vm / . and any other k-linear map from V to a k-algebra R extends uniquely to a k-algebra homomorphism T . Let W V ! C.13 Every k-algebra A can be regarded as a k-superalgebra by setting A0 D A and O A1 D 0. q/. then . R/ becomes an L D R-algebra under cup-product. V ˝1 D V.Y /
15e
Brief review of the tensor algebra
L
n 0V ˝n . q/. and35 .x/ 1C. . If A.! n 0 V ˝n . q/ be the composite of the canonical map V ! T . When one regards a k-algebra R as a ring with a k ! R.V. em .n copies of V /
with the algebra structure deﬁned by juxtaposition.X (super tensor product).V / D V ˝0 D k. em .vmC1 ˝ ˝ vmCn / D v1 ˝ ˝ vmCn :
It is a k-algebra.

17 If q D 0. E XAMPLE 15.x/.V / by the ideal generated by all squares x 2 .x/2 q. .
2
˝ xn / D r.x1 ˝ As r 0 .fi /W C. 0 D . namely. In C.x/ C .y/ .fj / . q/.x/ .x/ C .V.y/2 D .V. then . . q/ is the exterior algebra on V . C. . y/. r 0 .y/ . fn is an orthogonal basis for V . : : : .fi /2 D q.x/ and so .15.fi / D .V. /: (103)
.x/ D 2 .y/ C .C.16 If V is one-dimensional with basis e and q.x/ .y/ D .x/. q. / is uniquely determined up to a unique isomorphism by the universal property in the proposition.x ˝ x r 0 factors uniquely through C. Then there exists a surjective homomorphism ei 7! .V / D kŒe.V.x C y/2 D q. : : : . and I.y/ . q/.x C y/ D q.V.x.x// D r. q/ is the quotient of T .
T HE MAP C.V.e/ D c.x/ C q. q/ ! D such that r ı D r: N N V
r
C.c1 .18 Let r be a k-linear map from V to a k-algebra D such that r.y/2 : On comparing this with . if f1 . T .e.fi /. r extends uniquely to a homomorphism of k-algebras r 0 W T . q/. i. C. .e 2 c/.y//2 D .fi / .x/ C .q/ D .x/ C .c/.V.V.xn /. x 2 V . then T .x/ . y/: In particular.y/ C .V / is a polynomial algebra in one symbol e.y//2 D .V.
P ROPOSITION 15. q/
r N
D: P ROOF. then C. cn / ! C.V / ! D. we ﬁnd that .x1 /
r.x C y/2 D .i ¤ j /: (102)
(101)
Let ci D q. cn / ! C. q/ C.y/ .x/2 C .x/ D 0.fj / .x/ .c1 . According to the universal property of the tensor algebra. : : : . The classical almost-simple groups
157
E XAMPLE 15.y/ C . Therefore.y/ C . Then there exists a unique homomorphism of k-algebras rW C.x/2 C . ..x/2 D q.V.x.
As usual.y/ C 2 .y/ .fi /.x/ . q/
Because is linear.

q/ \ T . q2 /: Then the k-linear map V x has the property that r.x2 /:
D .x1 / ˝ 1/. q2 / 1 .158
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
T HE GRADATION ( SUPERSTRUCTURE ) ON THE C LIFFORD ALGEBRA
Decompose T . In more down-to-earth terms.x1 / ˝ 2 .x1 / C q.
T HE BEHAVIOUR OF THE C LIFFORD ALGEBRA WITH RESPECT TO DIRECT SUMS
Suppose .I. q2 /. 1 .q.1 ˝
2 . q1 /˝C.V2 .V /i :
Clearly this decomposition makes C. and so C.x2 //
D V1 ˚ V2 D .q/ D . q1 /˝C.V /0 / ˚ .V.q/ \ T .x/. because . q/:
O C. q/ ! C.x2 //. q/ into a super algebra.V /0 .V.V /1 M T .
(104)
. 1 . q/ D C0 ˚ C1 with Ci D T .x2 /
D .x1 / ˝ 1 C 1 ˝ D q.V /1 / .V1 .V.x1 / ˝ 1//:
Therefore.V / D T .q/ is generated by elements of T .V /0 D V ˝m
m even
T .V1 .1 ˝ 1/
D
1 . x2 /
! 7!
r
O C.V /i =I.x1 / ˝ 1 C 1 ˝ 2 .q/ \ T . q/ D .I.V2 . it factors uniquely through C.x2 //.V /1 D As I.x1 .1 ˝
2 . 1 .V1 . and C1 is spanned by products of an odd number of vectors.V.x2 // 2 2 .V2 .x/2 D .V /0 ˚ T .
M
m odd
V ˝m :
I. C0 is spanned by products of an even number of vectors from V . q1 / ˚ .V.

q1 /˝C. R EMARK 15.
.V.V.V. in C.V.V2 / D 2n :
From an orthogonal basis for . cn /. Certainly.c1 . q/ ! C. q/ is an isomorphism. .V. q1 /˝C.15. q/.V. q// D 2n .C. and that n D dim V > 0.V / < n.V1 / ˝ 1 and O 1 ˝ 2 .. we shall regard V as a subset of C.V2 . q/ where q 0 is quadratic form deﬁned by
0. 1/
n. : : : . we shall omit ).V.e1
en lies in the centre of C.V.V1 . q/.V 0 .19) that C. cn / ! C. q2 / is an isomorphism. Let D .c1 . q 0 / ' L ˝k C. (b) For every orthogonal decomposition . q/ D . e1 en /2 D .V. we can decompose . From now on.
en / ei D .C.V.V.V.C. 1/n i ci .n 1/ 2
c1
cn D .V. 1/i
ci . we deduce that the homomorphisms (103) and (104) are isomorphisms. : : : . Moreover. Therefore.e1 Therefore.n 1/ 2
det.V1 .V2 . dim. q2 / in such a way that dim.19 Let . : : : .e1 ei . If n D 1.c1 .20 The map W V ! C.e.V1 / 2dim. The classical almost-simple groups
159
E XPLICIT DESCRIPTION OF THE C LIFFORD ALGEBRA
T HEOREM 15. q/ D . : : : . .
T HE CENTRE OF THE C LIFFORD ALGEBRA
Assume that .e1 .V2 . the homomorphism (104) O C. P ROOF. Then
0
extends uniquely to an L-bilinear form
WV 0
V 0 ! L. q2 /. q/ as a k-vector space is 2n . 2 C OROLLARY 15. and let q. en be an orthogonal basis for . q// 2dim. q/: Note that. q/ (i. q/.ei / D ci . Let e1 .
and C.16). q2 /.ei . q1 / ˚ . (a) For every orthogonal basis for .
en / D .e1
ei ei
1 ei C1 1 ei C1
en / en /. q1 / ˚ .V2 . Assume inductively that they are true for dim. q// Ä 2n : It follows that dim. 1/ We saw in (15. q/ is injective. ej //. the homomorphism (103) C.V1 .
1
n.V. we get a surjective homomorphism (103).V.21 Let L be a ﬁeld containing k. The homomorphism (104) is surjective because its image contains 1 . q/ if and only if n is odd.V. all three statements are clear from (15. and so dim.
V 0 D L ˝k V.V2 / which generate C.V. By comparing dimensions.V. (c) The dimension of C. q/ is regular. q/ a quadratic space of dimension n.Vi / < n.V1 . cn / ' C.

28) below that the condition (c) is implied by (a) and (b)..x1 for x1 . Thus. Chapter 9.V. q/opp D ba in C. and. (b) No nonzero element of C1 centralizes C0 .V. C.V. and then ﬁnd the monomials that centralize the ei (or the ei ej ). the centre of C. C0 \ Centre. Note that. P ROOF.V.C.V. there exists an isomorphism W C. it is of degree 2 over k.V.V. for x 2 V . isomorphic to a tensor product of quaternion algebras. q/ is a p central simple algebra over the ﬁeld kŒ . (b) We shall see in (15.q/ consists of the elements t of C0 . q/opp inducing the identity map on V .V.x/. q// D k. Let C. q/: If n is even. The map W V ! C. (c) the map x 7! txt
1W V
! V has determinant 1:
R EMARK 15. : : : .V. q/.22 (a) If n is even. q/ is k.
.V. there is the following description of the complete structure of C.ab/ D b a and a D 1. q/opp is k-linear and has the property that .160
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
P ROPOSITION 15.24 (a) The condition (a) implies that t is invertible in C0 . generated by e1 en : In particular.
T HE INVOLUTION
An involution of a k-algebra D is a k-linear map W D ! D such that .V. q/ is a central simple algebra over k.V. xr 2 V . the centre of C. q/. q/ as a k-vector space but ab in C. First show that a linear combination of reduced monomials is in the centre (or centralizes C0 ) if and only if each monomial does. For example.e. q/opp D C. q/ such that (a) t t D 1.V. q/ ! C. We regard xr / D xr x1
as an involution of A.
15g
The Spin group
Initially we deﬁne the spin group as an abstract group.V.x/2 D q. If n is odd.V. q/.k/.23 The group Spin. 2. D EFINITION 15. and so (b) makes sense. i.V. (b) tV t 1 D V . then C.10. x x D q. q/opp be the opposite k-algebra to C. and which therefore has the property that .V. 2 In Scharlau 1985. if is not a square in k. C. q/ is generated over k by the element e1 en whose square is . if n is odd.x/. M 7! M t (transpose) is an involution of Mn .

an element t 2 Spin. b anisotropic elements of V . because q.q/.q/ is generated by elements Ra Rb with a. such a t must lie in the centre of C. As det. the image is the identity component of SO. For example.a/q.t /V t For t 2 . if k D R. q/ that acts as 1 on C0 . such elements generate SO. Then the mapping x 7! txt
1
1W V
!
/ D . P ROOF.txt
1 1
D V . q/.a/ D 1 D q. As we noted above.q/ deﬁnes an element x 7! txt T HEOREM 15.q/ ! SO.a/
Moreover.25 The homomorphism Spin. The kernel consists of those t 2 Spin.q/.b/ D 1.
15h
Write
The Clifford group
for the automorphism of C.a. and it is surjective if k is algebraically closed. we can even scale a and b so that q.15. SO.26 The Clifford group is . q/ and as 1 on C1 .b/: Therefore.q/
Let t be an invertible element of C.ab/ ab D baab D q.
1W V
.Ra1 Ram / D .V. Now axa
1
D . each element of SO.t / denote the homomorphism x 7! . If k is algebraically closed.txt
1 2
/ D tx 2 t
1
D t q.t /xt
1
D V g: !V. Since it is also in C0 . it must lie in k.q/. x/. For an anisotropic a 2 V .q/.q/ ! SO.x/. As V generates C. if q. According to Theorem 15.V.V.
D EFINITION 15.a/q. 1/m . x// a Â Ã 2 . q/ such that tV t V is an isometry. then Spin.q/ is connected but SO.x/:
1
as .a. and so SO.q/
1 of
just deﬁned has kernel of order 2.q/ can be expressed D Ra1 Ram for some ai . q/ j t invertible and .
Therefore.a/ D Ra .q/ such that txt 1 D x for all x 2 V . the homomorphism is not surjective.V.a.V. The classical almost-simple groups
161
T HE MAP Spin. let ˛.q/ will have two connected components when is indeﬁnite.3.q/ ! SO.V. q/.ax C xa D 2 . In this case. then Ra Rb is in the image of Spin.q/ D ft 2 C.b/. 2 In general. see (102)) as a2 D q. .x/t
D q. x/ D x a q.V. q/. Now the condition t t D 1 implies that t D ˙1. let Ra be the reﬂection in the hyperplane orthogonal to a. xa C 2 .q/ when k is algebraically closed. we see that m is even.

.q/ is superﬂuous.t
1
x . According to the proposition. the condition (c) in the deﬁnition of Spin . and the sequence 1!k ! .27 For all t 2
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1) .22a). i.t /
.V.q/ ! 1
˛
is exact (no condition on k).22b).t /xt 1 D x for all x 2 V .q/ can be expressed in the form
with c 2 k and the ai anisotropic elements of V .
P ROOF.˛/ D k : It remains to show that ˛ is surjective. q/. because and act as 1 and 1 on V . Now.3. Conversely. . If t 2 C0 .t / x t and so . and that t1 centralizes C0 .˛.q/. k is in the kernel of ˛.28 For an invertible element t of C0 .t / . Such an element acts as Ra1 has determinant . As V generates C.105/
D t
1
xxt t t
1
D q. and hence in k (15. q/ such that tV t of x 7! txt 1 W V ! V is one.q/. ˛.x/:
As k is in the centre of .˛.x// D . and so det. we ﬁnd that .
(105)
x .q/ ! O. t1 x D xt1
for all x 2 V . Let t 2 . We have shown that Ker.t/. and
2 1
D V .t /V D V t .˛. .V. ˛.t / D Rt ..25)). then m is even.t /V D V t. P ROOF.q/. let t D t0 C t1 be an invertible element of C.t /.t /.t /xt
1 .t / D 1. 1/m . Hence.x// D t
1 1
D
. q/ these equations imply that t0 lies in the centre of C. the determinant
.t / is an isometry of V . q/ such that . every element t 2 t D ca1 am Ram on V . such that t0 x D xt0 . Therefore the surjectivity follows from Theorem 15. 2 C OROLLARY 15.e. and hence is zero (15.t
1
/xt .y/ D tyt 1 and so (see the proof of (15.t /x D xt t: We use this to prove that ˛.x// .t /.V.V.q/.t / x t
1
/ D
. ˛.162 P ROPOSITION 15.V. On applying and to . and so t 2 .t / is an isometry: q. For t 2 V . q/.t // D .

III 6).q/ of GL. q/ be quadratic space over k.V. Clearly 1 ı 2 D e ı e and id D id. Show that. generalizing the terminology adoped by physicists in a special case for the rotation group of the three dimensional space” (C. q/ ! C. q/ ? ? yQ ! C.q/ ! SO.V. q/.x//2 D .29 An element of O.15.V / (see 6.q/ giving the homomorphism in (15.qR / that coincides with the previous functor when R is a ﬁeld. Consider ı W V ! C. The Construction and Study of Certain Important Algebras. q/ rendering V ? ? y V ! C.q/ on Spin. and much later.V.V.30 There exists a naturally deﬁned algebraic group Spin. “When E.q/ for the algebraic group Spin. Chevalley.K/ ' Spin.25) for each ﬁeld K containing k.q/ on Spin. Alternatively.21).
15j
Restatement in terms of algebraic groups
Let . q/ described in (15. the action of O.
N OTES A representation of a semisimple algebraic group G gives rise to a representation of its Lie algebra g. P ROOF. q/ by an invertible element of C C .q/ deﬁnes an automorphism of C. he discovered a new representation of the orthogonal Lie algebra [not arising from the orthogonal group].x// 1 D .q/. 2 In future. there is a unique homomorphism Q W C. C. we shall write Spin. T HEOREM 15.qK / for all ﬁelds K containing k. q/.V. and let qK be the unique extension of q to a quadratic form on K ˝k V .V. q/. /. deﬁne a functor R Spin.25) deﬁnes an action of O. Then .20) has these properties.V. q/ as follows.V.q/.V. it is known that Q is an inner automorphism of C. 1955.q/ on C. the algebraic group attached to the subgroup Spin. he called the elements on which this new representation operates spinors. by the universal property.q/.x/ 1 for every x 2 V . Finally. q/
1 e commutative.V.V. and all representations of g arise from G only if G has the largest possible centre. and so g D Q 1 . Hence. As we noted in (15.q/ over k such that Spin. But he did not give a speciﬁc name to it. qK / D K ˝k C. .q/
15.
. Moreover.V. The classical almost-simple groups
163
15i
Action of O. and so Q is an automor1 2 phism. Cartan classiﬁed the simple representations of all simple Lie algebras. This explains the origin and name of the Spin group. If 2 SO. . there is a homomorphism of algebraic groups Spin. when k is inﬁnite. .

16
The exceptional almost-simple groups
A SIDE 16.xy/ D N. A Hurwitz algebra over k is a ﬁnite k-algebra A (not necessarily commutative) together with a nondegenerate quadratic form N W A ! k such that N. A Hurwitz algebra of dimension 8 is also known as an octonion or Cayley algebra. we can recover them from the Lie algebras (see Chapter II). In characteristic zero.1 Need to explain how to construct them explicitly (perhaps without giving all proofs). y 2 A: The possible dimensions of A are 1.
164
.y/ for all x. 2.R ˝k A/
is an algebraic group over k of type G2 . the functor R Autk . 4. and 8. For such an algebra A.
16a
The group G2
Let k be a ﬁeld of characteristic zero.x/N. (To be continued).

(c) Let N and H be algebraic subgroups of G with N normal. 12.15).. (b) An algebraic group G is solvable if it contains a normal subgroup N such that N and G=N are solvable (i.1 (a) Subgroups and quotient groups of solvable algebraic groups are solvable.2 For any algebraic group G over a perfect ﬁeld k. has kernel . Therefore H \ Gi C1 is a normal subgroup of H \ Gi . k is a ﬁeld. (a) Let G D G0 G1 Gn D f1g be a solvable series for G. but may not commute with extension of scalars. Therefore H H \ G1 H \ Gn D f1g
is a solvable series for H . and so G G1 Gm .
17a
The radical of an algebraic group
P ROPOSITION 17. In other words.RG/u . (b) Let G=N D Q0 N D N0 Q1 N1 Qm D f1g and Nn D f1g
be solvable series for G=N and N . then HN is solvable (resp.3 Discuss what happens when k is not perfect. . and let Gi be the inverse image of Qi in G. and the quotient H \ Gi =H \ Gi C1 injects into Gi =Gi C1 . If H and N are solvable (resp. The ﬁrst statement follows from (c) of the proposition.D N / N1 Nn
is a solvable series for G.17
General algebraic groups
Throughout this section. connected). (c) The quotient HN=N is solvable (resp. there exists a unique largest connected normal smooth solvable subgroup (called the radical RG of G). and a unique largest connected normal smooth unipotent subgroup (called the unipotent radical Ru G of G). take Ru G D . P ROOF. which is commutative. Let H be a subgroup of G.13). connected). and let Gi be the image of Gi in Q. N Let Q be a quotient group of G. 2
A SIDE 17.Ru G/k al ¤ Ru Gk al . 2 C OROLLARY 17. for the second.H \ Gi / \ Gi C1 D H \ Gi C1 . P ROOF. in general. and so this follows from (b) (resp. Then Gi =Gi C1 ' Qi =Qi C1 (see 8. connected) because it is isomorphic to H=H \ N (see 8. H \ Gi ! Gi =Gi C1 . extensions of solvable groups are solvable).17). The restriction of the quotient map Gi ! Gi =Gi C1 to H \ Gi .
165
.e. Then it follows from the isomorphism theorems that N N N Q D G0 G1 Gn D f1g is a solvable series for Q. The subgroup Ru G exists.

166
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
17b
Semisimple and reductive groups
D EFINITION 17. P ROOF. and so the groups are normal in G by (6. commutative. Obvious. we use that. 0
A SIDE 17. to GLm GLn . over a ﬁeld of characteristic 2. P ROOF.5 Let G be a smooth connected algebraic group over a perfect ﬁeld k.6 Let G be a smooth connected algebraic group over a ﬁeld k: (a) If G is semisimple. then a semisimple group may have such a subgroup.RG/ D2 . i. (b) If G is reductive. (b) Similar.7 If one of the conditions. P ROPOSITION 17.e. SL2 has the connected normal commutative subgroup 2 . Finally. The unipotent radical of G is the 0 subgroup of matrices I B .V / is said to be semisimple (or completely reducible) if every stable subspace W has a stable complement W 0 (so V is a direct sum V D W ˚ W 0 of representations).
B E XAMPLE 17. The quotient of G by Ru G is isomorphic to the reductive group of 0 I 0 invertible matrices of the form A C . For this. This remains true over k al . SL2 SL2 is semisimple. (a) G is semisimple if and only if RG D 1. For example. and GLn is reductive. Moreover. or. This is obvious from their deﬁnitions: RG is the largest connected normal solvable algebraic subgroup and DG is the smallest normal algebraic subgroup such that G=DG is commutative. (a) Only the converse statement is nonobvious. i. the converse is true if k is perfect.RG/ Dr . The radical of this is Gm Gm .4 An algebraic group over an algebraically closed ﬁeld is reductive (resp. semisimple) if it is smooth and connected and has no smooth connected normal unipotent (resp. Therefore the chain G RG D.. is dropped. P ROPOSITION 17. for any algebraic group G. SLn .e.RG/ 1
2
is preserved by every automorphism of G.8 Let G be the group of invertible matrices A C . 2 R EMARK 17.
. RG and DG are characteristic subgroups. but it has the connected normal subgroup f1g SL2 .9 A representation G ! GL. semisimple) if it becomes reductive (resp. solvable) subgroup other than the identity. SL2 has the commutative normal subgroup f˙I g and the commutative connected subgroup U2 .32). smooth. An algebraic group over a ﬁeld is reductive (resp. the converse is true if k is perfect. connected. then it has no smooth connected normal commutative subgroup other than the identity. (b) G is reductive if and only if Ru G D 1. For example. SOn . semisimple) over the algebraic closure of the ﬁeld. then the only smooth connected normal commutative subgroups it contains are tori. every automorphism of G maps RG onto RG and DG onto DG.. normal. Spn are semisimple.

Z. At the moment this is only proved at the end of II 4.G/ı G der D G. this shows that R. However. and Z.G der / Z. Since G der .15).G der /. General algebraic groups
167
equivalently.21) implies that the action of G on RG by inner automorphisms is trivial.G/).k/ consists of commutators (14.G/ı G der is a normal algebraic subgroup of G such that G=. then the derived group G der of G is semisimple.G/ is semisimple. In general.G/.RG/u D 0. this shows that Z. Choose an embedding G .G/ \ G der is the (ﬁnite) centre of G der . For example. By deﬁnition. then 0 :: : 0 0 1
C 0A Ar
with each Ai nonzero and scalar.G/ı is commutative. T HEOREM 17.G/ is surjective with ﬁnite kernel.k/ consists of the matrices 0 A1 B @0 0 13. It sufﬁces to prove this with k D k.Vi / contains only ﬁnitely many scalar matrices. We next show that Z. Note that Z. the representations of a connected algebraic group are all semisimple if and only if it is a torus. if V is a direct sum of simple (i. a connected algebraic group is reductive if and only if all of its representations are semisimple (see later).G/ \ G der Z.G/ı of G is a torus. Since the reverse inclusion always holds. : : : .12). 2
is.G/ı G der : Therefore G der ! G=R.
17c
Reductive groups
For simplicity. but.G/ı . in this subsection I assume that k has characteristic zero. As G=R.G/ı D torus.
37 Because 36 That
. G D DG if G is semisimple. because G D Z. of the form diag.10 If G is reductive. and so RG Z.a. in characteristic zero. a semisimple group has no commutative quotients. it consists of such matrices with determinant 1. irreducible) representations (those with no proper nonzero subrepresentations). 0 1 y y is not semisimple because the only stable one-dimensional subspace is the x-axis (the map is a shear). representations of unipotent groups are not semisimple. moreover. When we choose bases for the Vi . a/ with a ¤ 0.e.G/ D Z.G/ı (see Z. Certainly Z.G/ı \ G der is ﬁnite. and so (14. . As SL. the action of U2 on k 2 . the connected centre Z.! GLV . nor should you expect the representations of a group containing a normal unipotent group to be semisimple.G/ı G der and Z. and write V as a direct sum V D V1 ˚ ˚ Vr of eigenspaces for the action of Z. Â ÃÂ Ã Â Ã 1 a x x C ay D .G/ı G der / is commutative (being a quotient of G=G der ) and semisimple (being a quotient of G=R. In characteristic p. Z. Hence37 G D Z. N P ROOF.G/ı \ G der is ﬁnite.17. so also is G der .36 and so its centralizer in GLV consists of the matrices of this shape with the Ai arbitrary..35) shows that RG is a torus T . Z.G/ı . Rigidity (13. In other words.

A SIDE 17. and a homomorphism 'W Z. In particular. G can be recovered from . Z. in other words. an algebraic group contains a unique largest smooth subgroup.G 0 / ! Z. Also discuss things over a ring k: A SIDE 17. '/ (and G 0 can even be chosen to be simply connected).11 From a reductive group G. and maybe even allow a base ring. G is pseudo-reductive if G has no smooth connected normal unipotent subgroup. (a) See 12.
. Moreover. (b) If k is perfect. (d) A smooth connected solvable algebraic group G over a perfect ﬁeld contains a unique connected normal unipotent subgroup Gu such that G=Gu is of multiplicative type. Z. but Gk al does). T HEOREM 17.168
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
R EMARK 17. and Prasad on pseudo-reductive groups (those groups G such Ru G D 0 but G is not reductive.z/
!Z
G 0 ! G ! 1:
(106)
Clearly.G 0 ..G 0 /
z7!. A SIDE 17.
17d
Structure of a general algebraic group
Recall that “algebraic group” means “afﬁne algebraic group”. (c) See 14.35.9.
2
A SIDE 17. (b) The radical of G is such a subgroup (see the above discussion). i. we obtain a semisimple group G 0 (its derived group). '/ as the quotient Z. at least in the smooth case. discuss the work of Conrad.G 0 . (c) A smooth connected algebraic group G over a perfect ﬁeld contains a unique smooth connected normal solvable subgroup RG such that G=RG is a semisimple group. a group Z of multiplicative type (its centre). Ru Gk al ¤ 0.12 (a) An algebraic group G over a ﬁeld k contains a unique normal connected ´ algebraic subgroup G ı such that G=G ı is an etale algebraic group.e.z/
1 .14 Explain the Barsotti-Chevalley-Rosenlicht theorem. P ROOF.'.15 Somewhere explain anti-afﬁne groups.16 Explain what is true when you drop “smooth” and “perfect”. Gabber. every reductive group arises from such a triple .13 Explain the connected components for a nonafﬁne algebraic group.

G 0 //. then Repk .X 0 /.25) and (14.5 Let f W G ! G 0 be a homomorphism of afﬁne groups over k. If G is ﬁnite.4 The identity component G ı of an algebraic group G over a ﬁeld of characteristic zero is reductive if and only if Repk . a group G is connected if and only if there is no non-trivial epimorphism G ! G 0 with G 0 ﬁnite.1 An algebraic group G is ﬁnite if and only if there exists a representation .G/. X 0 2 ob.21).G/ of representations G ! GL. (b) f is a closed immersion if and only if every object of Repk . then G D Spec B where B is the linear dual of the ﬁnite k-algebra AX .1). and the second condition shows
V n is a direct sum of n copies of V .G/ is isomorphic to a subquotient of an object of the form of ! f . A smooth algebraic group over k is unipotent (resp. P ROOF. stable one-dimensional subspace). ´ P ROOF. this is equivalent to Repk . is not stable under ˝.18
18a
Tannakian categories
Properties of G versus those of Repk .52).Repk . to a quotient of a subrepresentation).W / factoring through G 0 . with the notations of (10a). 2 P ROPOSITION 18.3 Let k be an algebraically closed ﬁeld.34) says that it has the required property. See (14. and (7. An algebraic group G is connected if and only if. According to (7. (a) f is faithfully ﬂat if and only if ! f is fully faithful and every subobject of ! f .G 0 / with a full subcategory of Repk . the full subcategory of Rep.G/ of subquotients of V n .G/. and let ! f be the corresponding functor Repk .
2
P ROPOSITION 18. P ROOF. (a) If G ! G 0 is faithfully ﬂat.G/ is semisimple. Therefore.2 Assume k has characteristic zero.
2
P ROPOSITION 18. for X 0 2 ob. P ROOF. It is therefore obvious that ! f has the stated properties. V / such that every representation of G is a subquotient38 of V n for some n 0. is isomorphic to the image of a subobject of ! f .G/ determines the other. P ROOF.X 0 /. then the regular representation X of G is ﬁnite-dimensional. solvable) if and only if every nonzero representation of the group has a nonzero ﬁxed vector (resp.G 0 / ! Repk .G/ having no non-trivial subcategory of the type described in (18. all ﬁnite groups are etale. for every representation V on which G acts nontrivially.r. Conversely. Conversely if.X /. P ROPOSITION 18. 2 P ROPOSITION 18. See later.G/ D hXi. and therefore an epimorphism. and subquotient means any representation isomorphic to a subrepresentation of a quotient (equivalently. we should be able to see properties of one reﬂected in the other. if ! f is fully faithful.G 0 / can be identiﬁed with the subcategory of Repk .
38 Here
f
169
.Repk .G/
Since each of G and Repk . Repk .G//. In characteristic zero. it deﬁnes an equivalence of Repk . n 0.

X 0 /i/_ ! ! lim End.!jh! f . and C D Repk .G/ is isomorphic to a direct factor of an object of the form ! f .X /i.EndG . and so B 0 ! B 00 is surjective. End. and assume G ı is reductive. (c) every H -isotypic component of any representation of G is stable under G. Let G D Spec B and G 0 D Spec B 0 . P ROOF. if C D Repk .!jh! f . 2.G 0 //.G/
correspond to homomorphisms of k-coalgebras B 0 ! B 00 ! B where G D Spec B and G 0 D Spec B 0 . it is also a faithful representation of G. Conversely. Compositio Math. Moreover.V //:
P ROOF.r ı .G/ be the functor .29). then (10b) shows that B 0 D lim End. moreover.32). (b) f is an embedding if and if every object of Repk .! 0 jhX 0 i/ is injective. !
and B ! B 0 being injective implies that G ! G 0 is faithfully ﬂat (5.EndH . these conditions imply that Z. 2. B 00 D B and B 0 ! B is surjective. An argument as in the above above proof shows that B 00 ! B is injective. Then: (a) f is a quotient map if and only if ! f is fully faithful. V / 7! . for X 0 2 ob.6 Let G and G 0 be algebraic groups over a ﬁeld k of characteristic zero. 82 (1992).V // Z.G 0 / ! C ! Repk . If f is a closed immersion. G is reductive. Yves. 1–24. [Take a faithful representation of G 0 .170
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
that.G 0 //.X /i/ ! End.!jhX i/_ D B. Let f W G ! G 0 be a homomorphism.G/.
2
P ROPOSITION 18. and let ! f W Rep. Mumford-Tate groups of mixed Hodge structures e and the theorem of the ﬁxed part. 1. The functors
Repk .! 0 jhX 0 i/_ D lim End.] 2 P ROPOSITION 18. Lemma 1).G/ whose objects are isomorphic to subquotients of objects of the form of ! f . for X 0 2 ob.Repk . then B 0 ! B is surjective and it follows that B 00 ! B. (b) for every tensor space T D V ˝m ˝ V _˝n and k-character of H . Omitted for the present (Andr´ . If.X 0 /. no. V /.G 0 / ! Rep.7 Let G GLV be a connected algebraic group over a ﬁeld k of characteristic zero..21. (b) Let C be the strictly full subcategory of Repk . 2
. etc. hX 0 i is equivalent to h! f . the subspace T on which H acts through is stable under G.V /.G/. The following conditions on a subgroup H of G are equivalent: (a) H is normal in G.r.Repk . Omitted for the present (Deligne and Milne 1982.

1. 18.
U.U ˝ V / ˝ W v ˝ w ! w ˝ vW V ˝ W ! W ˝ V: (107)
Any free R-module U of rank one together with an isomorphism U ! U ˝ U (equivalently.
and a commutativity constraint is a natural isomorphism
V.F. see Deligne and Milne 1982.W W F .U ˝ V / ˝ W
U ˝V.C/.V / ˝ F .V.U..C/:
Compatibility means that certain diagrams.W / ! F . 1. Tannakian categories
171
18b
Tensor categories
18. 18. An associativity constraint is a natural isomorphism
U.11). A tensor functor C ! C0 is a pair .e. for example. an object U together with an isomorphism uW U ! U ˝ U such that V 7! V ˝ U is an equivalence of categories.8.9 A tensor category over k is a k-linear category together with a k-bilinear functor ˝W C C ! C and compatible associativity and commutativity contraints ensuring that the tensor product of any unordered ﬁnite set of objects is well-deﬁned up to a well-deﬁned isomorphism. ˝/ and . It is trivial to check the compatibility conditions for this to be a tensor category. i.W
! .12 The category of ﬁnite-dimensional representations of a Lie algebra or of an algebraic (or afﬁne) group G with the usual tensor product and the constraints (107) is a tensor category.W U ˝ . U ˝ .8 A k-linear category is an additive category in which the Hom sets are ﬁnite-dimensional k-vector spaces and composition is k-bilinear.W. 18.
V.C.C0 . c/ consisting of a functor F W C ! C0 and a natural isomorphism cV.
.U ˝ V / ˝ W.W W U
˝ . Functors between such categories are required to be k-linear. We use 1 to denote a neutral object of 1 C. the choice of a basis for U ) is a neutral object. See Deligne and Milne 1982.u ˝ v/ ˝ wW U ˝ . W 2 ob.V ˝ W / ! .V ˝ W / ! .V ˝ W / ? ?id ˝ y U V.13 Let .U ˝ V / ? ? y W. ˝/ be tensor categories over k.18. W 2 ob.v ˝ w/ 7! .V. The required commutativities follow immediately from (18.. Then F commutes with ﬁnite tensor products up to a well-deﬁned isomorphism. For a complete deﬁnition.V ˝ W / compatible the associativity and commutativity constraints and sending neutral objects to a neutral objects.W W V
˝ W ! W ˝ V.W ˝ U / ˝ V. V.V ! . E XAMPLE 18. and that there exists a neutral object (tensor product of the empty set).W ˝ V /
U.W
˝idV
commute.
U.V
! . E XAMPLE 18.W
! W ˝ .U ˝ W / ˝ V
U.11 The category of ﬁnitely generated modules over a ring R becomes a tensor category with the usual tensor product and the constraints u ˝ . induce k-linear maps on the Hom sets.10 An object of a tensor category is trivial if it is isomorphic to a direct sum of neutral objects.e. i.

W
! !R .e. W 2 ob.1 1/
!R .R/ is an afﬁne group over k. T HEOREM 18.V ˝ W / ? ? y V ˝W
!R .R/ consists of the natural transformations W !R ! !R such that the following diagrams commute !R . ev/ of V is uniquely determined up to a unique isomorphism.9).f ˝ v/ D f .C. k/ with ev.
2
18.
ı ˛R D ˛R ı
V
C. Then ıV is uniquely determined.C/.14 Let C be a tensor category over k.V.8. 10.W
! !R .V / ˝ !R . Such a functor ! is called a ﬁbre functor over k. then G.1 ˝ 1 1 1/
cV.R/ be the set of families D.u/
for all objects V .V /R /.
G.18 Let !R be the functor V !. and let V be an object of C..W /
cV.1 for every neutral object of 1 of 1 1/ W V /V 2ob.8. and for all arrows ˛W V ! W in C. C is a neutral tannakian category over k and ! is a ﬁbre functor over k. such that ˘ ˘ ˘ Then R k. i.
18c
Neutral tannakian categories
18.
V ˝W D V ˝ W for all V.1 u/.W / ? ? y V˝ W !R .!. For example.16 A neutral tannakian category over k is an abelian k-linear category C endowed with a rigid tensor structure for which there exists an exact tensor functor !W C ! Veck . a module M over a ring admits a dual if and only if M is ﬁnitely generated and projective (CA 10. W of C and all identity objects .G/:
P ROOF.15 A tensor category is rigid if every object admits a dual. A pair .1 ˝ 1 1 1/ ? ? y 1 1 1˝1 ! !R . V
2 EndR-linear .V _ . V _ ˝ V ! 1 1/ is called a dual of V if there exists a morphism ıV W 1 ! V ˝ V _ such that the composites 1 V V_
ıV ˝V V ˝ev ev
!
V ˝V _ ˝V
!
V
V _ ˝ıV
! V _ ˝V ˝V _
ev ˝V _
! V_
are the identity morphisms on V and V _ respectively. and ! deﬁnes an equivalence of tensor categories over
C ! Rep.
. Similarly.v/ — here ıV P is the k-linear map sending 1 to ei ˝ fi for any basis .172
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
18.fi /. the category Veck of ﬁnite-dimensional vector spaces over k and the category of ﬁnite-dimensional representations of a Lie algebra (or an algebraic group) are rigid.V ˝ W /
!R . For each k-algebra R. More generally. 1 1 D id!.V _ .ei / for V and its dual basis . 1. For example. !/ be a neutral tannakian category over k.V / ˝ !R . a ﬁnitedef dimensional k-vector space V has as dual V _ D Homk .17 Let . and the dual . let G.V / ˝ R. 18. the contragredient of a representation of a Lie algebra or of an algebraic group is a dual of the representation. This is a restatement of Theorem 10.C/ .u/
! !R .1 1/ ? ? y 1 1 !R .

R/ by (9.25 Let C D Rep. we get a natural one-to-one order-reversing correspondence ftannakian subcategories of Cg $ fnormal algebraic subgroups of Gg (a tannakian subcategory is a full subcategory closed under the formation of duals. and subquotients). (b) (Tannaka correspondence.
CH. 18.G/ consisting of the representations of G on which N acts trivially.G/.) For a collection S of objects of C D Rep. Conversely.C.g/v D
V .!/ (functor of tensor automorphisms of !) for the afﬁne group G attached to the neutral tannakian category .v/
if g D .V / deﬁned by rV .
1W1
.20 A tannakian category C is said to be algebraic if there exists an object V such that every other object is a subquotient of P . Tannakian categories
173
18.G/ is algebraic.S / and H 7! CH form a Galois correspondence fsubsets of ob.
V/
2 G.G/ is the identity functor.2). this can be used as a deﬁnition of G=N ). direct sums. then (7.43) show that Rep. tensor products.S / D Ker. namely.R/ and v 2 V .19 I explain the ﬁnal statement of (18. An exact tensor functor F W C ! C0 such that ! 0 ı F D ! deﬁnes a homomorphism G 0 ! G.rV W G ! Aut. with generator V .C0 .R/:
18.G/. E XAMPLE 18.C0 /
7! .17). then G is algebraic because G GLV .C/ W G
0
.G/ is algebraic. The group attached to C and the forgetful functor is G=N (alternatively.
F V /V 2ob. 18.
V /V 2ob. denote it by . In this way.C. !/. Y .R/ ! G.V / endowed with this action of G is an equivalence of categories C ! Rep. and C ! Rep. Alternatively. let CH denote the full subcategory of C whose objects are those on which H acts trivially.28) and (7.22 If C is the category of ﬁnite-dimensional representations of an algebraic group H over k and ! is the forgetful functor..S / denote the largest subgroup of G acting trivially on all V in S .C/g falgebraic subgroups of Gg.21 It is usual to write Aut˝ .24 Let . !/ and . For each V in C.R/:
The functor sending V to !.R/ ' H. both maps are order reversing and S and H. and let C be the subcategory of Rep.V. It follows that the maps establish a one-to-one correspondence between their respective images.18. thus \ H. 18. (a) For an algebraic subgroup H of G. let H. Then CH is a neutral tannakian category whose Tannaka dual is G=N where N is the smallest normal algebraic subgroup of G containing H (intersection of the normal algebraic subgroups containing H ).23 Let N be a normal subgroup of an algebraic group G. If G is an algebraic group. !/ — we call it the Tannaka dual or Tannaka group of C.S /
i. E XAMPLE 18.e. if Rep. .G/ for some algebraic group G. then G.C.V //:
V 2S
Then the maps S 7! H.CH / H for all S and H . ! 0 / be neutral tannakian categories with Tannaka duals G and G 0 . V _ / for some P 2 NŒX. there is a representation rV W G ! GL!.

!/: a homomorphism corresponds to the M -gradation such that X m is the subobject of X on which D. 1/ (b) Let M.C/ to a commutative group deﬁnes a gradation on C if and only if f . and if X has dual X _ . Serre 1973.ŒX _ / D f . and let I. Chevalley (book).M / acts through the character m (Saavedra Rivano 1972.C/ is a universal tensor map.. write x x1 ˝ ˝ xm if X is a direct factor of X1 ˝ ˝ Xm . 18. then f . 18.C/ such that a.16 ).C/ represented by simple objects X.C/ ' I.C/ by a a0 ” there exist x1 . 4 (AJM 91.C/ ' N . so that Z D D. !/ be a neutral tannakian category such that C is semisimple and End. and let G be its Tannaka dual. xm of I. it is a tensor map.26).
A SIDE 18.!/.C/ ! M such that x x1 ˝ x2 H) f . For elements x. 1127–1140). : : : .C. The following statements are obvious. (a) Let M be a commutative group.C/ be the set of isomorphism classes of simple objects in C.e.V / D k for def every simple object in C. To give an M -gradation on C is the same as to give a central homomorphism D. An M -gradation on a tensor category C is an M -gradation on each object L of C compatible with all arrows in C and X with tensor products in the sense that . Breen). x1 . Therefore. !/ be as in (18. p71. Krein (cf.C/ ! N is universal. X1 . to give an M -gradation on C is the same as giving a homomorphism D.C/ modulo the relations: x D x1 C x2 if x x1 ˝ x2 . In more detail: let X be an object of C.C.27 Let C be a semsimple k-linear tensor category such that End. Hochschild and Mostow 1969.
.29 Let . if X is simple. 5. To give an M -gradation on C is the same as to give a map f W I.N / with N the group of characters of Z. M. : : : . On the other hand.30 Discuss the prehistory: Tannaka (cf. and deﬁne an equivalence relation on I. : : : . i. and every other tensor map I. Because C is semisimple. For such a map.28 Let .C/= . xm 2 I. Let Z be the centre of G D Aut˝ . Let .X / D k for every simple object X in C.X ˝ Y /m D rCsDm X r ˝ X s .ŒX /. The obvious map I. a0 x1 ˝ ˝ xm :
A function f from I.x2 /:
A map from I.C/ ! M.x1 / C f . Xm . and the tensor map ŒX 7! nW I.C/ to a commutative group satisfying this condition will be called a tensor map.a/ D f . According to (18. !/ be a neutral tannakian category.X m /m2M such that X D m2M X m .26 Let M be a ﬁnitely generated abelian group. (18. 18.174
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
18d
Gradations on tensor categories
18. An M -gradation on an object X of an abelian L category is a family of subobjects .C/ be the free abelian group with generators the elements of I. a homomorphism N ! M .1 D 0. f .28). G ı is reductive (II.M / ! G. remark).27) shows that to give an M -gradation on C is the same as giving a homomorphism M.a0 / whenever a a0 . or.C/ ! M. Therefore M. equivalently. 5).C/ ! M factors uniquely through it.C/ is surjective.x/ D f .M / ! Z.C. Note that I. then Z acts on X through a character n of Z. Assume (for simplicity) that Z is split. and so Z is of multiplicative type. Deligne and Milne 1982.C/ ! M .

.18. Tannakian categories
175
E XERCISES
E XERCISE 18-1 Use the criterion (18.1) to show that the centralizer of a torus in a connected algebraic group is connected.

.

I. AND M ILNE . Paris. Inc. Naz. Schemata over local rings. AND G AVARINI . and Shimura varieties. 1962). Cambridge. Quantum groups. (9) 39:307–317.07851. Th´ orie des e e Groupes Alg´ briques (Bruxelles. (2) 78:256–266. New York. 1977. J. Geometric algebra. 1972. Tannakian categories. Ann. volume 155 of Graduate Texts in Mathematics. Mat. Berlin. Editeur. D ELIGNE . Rend. M. C. (2) 73:624–648. SpringerVerlag. 2008. J. Springer-Verlag. M. New York. J. Hopf algebras. a Cl. 1980. pp. C HEVALLEY. AND G ABRIEL . R. M. Soc. e e e F IORESI . Groupes alg´ briques. e C ARTIER . Linear algebraic groups. 2002. J. P. and geometry. 17:1–18.. 1982. P. D EMAZURE . A primer of Hopf algebras. Lecture Notes in Mathematics 900. Englewood Cliffs. New York. New York-London. P.Bibliography
A BE . Masson & Cie. II. 1962. 177
. 1995. B OREL . e e e e ´ g´ n´ ralit´ s. Ann. C ARTIER . H UMPHREYS .. A RTIN . Cambridge University Press. Springer-Verlag. New York. A. 1991. Ramanujan Math. K ASSEL . H ARTSHORNE . Springer-Verlag. Tome I: G´ om´ trie alg´ brique. 1975. G REENBERG . Sci. 1970. J. Interscience Publishers. volume 74 of Cambridge Tracts in Mathematics. Schemata over local rings. Nat. B. motives. pp. of Math. Math. 1991. 101–228. F. 537–615. C ONRAD . 1961. II. G REENBERG . Linear algebraic groups. Translated from the Japanese by Hisae Kinoshita and Hiroko Tanaka. M. E. E. C. Chevalley supergroups. D EMAZURE . M. BARSOTTI . In Hodge cycles. (8) 18:43–50. Algebra. Berlin. arXiv:0808. NJ. In Colloq. 1963. e e e e Pures Appl. Fis. of Math. Lincei. J. Une d´ monstration d’un th´ or` me sur les groupes alg´ briques. 87–111. Springer-Verlag. In Frontiers in number theory. Groupes alg´ briques et groupes formels. 1960. Atti Accad. 1955. Lectures on p-divisible groups. Preprint available at IHES. R. P. Springer. E. pp. A RTIN . groupes commutatifs. Prentice Hall Inc. Springer-Verlag. Algebraic geometry. A modern proof of Chevalley’s theorem on algebraic groups. Berlin. S. Louvain. Un teorema di struttura per le variet` gruppali. physics. 2007. 1957. Librairie Universitaire.

2004. 1948. Vol. ROSENLICHT. (2) 49:774–789.. 1971. S ERRE . Graduate Texts in Mathematics. Manuscripta Math. volume 11 of Courant Lecture Notes in Mathematics. VARADARAJAN . (2) 39:33–85. Math. M ILNE . F. Springer-Verlag. volume 270 of Grundlehren der Mathematischen Wissenschaften. e S PRINGER . M. On certain concepts in the theory of algebraic matric groups.. Springer-Verlag. 1969. e S CHARLAU . ROSENLICHT. 1957. V. volume 33 of Princeton Mathematical Series. Ann. 1966. New York University Courant Institute of Mathematical Sciences. Berlin. New York. MA. R. J. T. volume 9 of Progress in Mathematics. New York. New York. Berlin. M. J. 1993. Ann. 1972. WATERHOUSE . 7:251–270. 78:401–443. Enseign. A. of Math. Mathematische Annalen 96:26–61. Princeton University Press. Springer-Verlag. Cat´ gories Tannakiennes. 1956. Etale cohomology. Mathematics Lecture Note Series. Abstrakter Aufbau der Idealtheorie in algebraischen Zahl. Algebraic group schemes in characteristic zero are reduced. W.-P. Invent. (4) 43:25–50. W. Pura Appl. Math. S. Boston.
. A. Inc. 2:79–80. Hopf algebras. Categories for the working mathematician. 1927. M. volume 66 of Graduate Texts in Mathematics. S. E. Linear algebraic groups. 1998.J. 1985. E. Some rationality questions on algebraic groups. S WEEDLER .178
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
KOLCHIN . TAKEUCHI . G` bres. 1972. M. Supersymmetry for mathematicians: an introduction. E. Quadratic and Hermitian forms. S AAVEDRA R IVANO . C. 1980.und Funktionenk¨ rpern. Mat. Princeton. Springer-Verlag. N. New York. S. N. o O ORT. N OETHER . Introduction to afﬁne group schemes. 1979. Amer. W. A correspondence between Hopf ideals and sub-Hopf algebras. 5. Benjamin. Some basic theorems on algebraic groups. J. Birkh¨ user a Boston Inc. Math. M AC L ANE .

110. 7. 5 contragredient of a representation. 5. 15 diagonalizable. 50 acts through a character. 28 co-commutative. 150 bi-algebra. 17 unipotent. 129 closed immersion.Index of deﬁnitions
action. 125 constant ﬁnite. 124 components irreducible. 171 centralize. 117 graded. 156 algebraic group connected. 88 ﬁnite-dimensional. 43 almost simple. 35 coaction. 173 decomposition Jordan. 154 reduced. 5 special orthogonal. 31 automorphs. 6 antipode. 77 contragredient. 79 dual. 39 comodule cofree. 43
179
. 40 Clifford. 154 symplectic. 6. 7. 51 co-algebra. 40 connected algebraic group. 171 neutral tannakian. 147 algebraic space regular. 107. 7. 115 tensor. 134 algebra afﬁne. 172 rigid tensor. 76 coconnected. 46 reduced attached to. 142 special linear. 39 category linear. 166 smooth. 156 etale. 76 component identity. 64 character of an algebraic group. 79 right. 83 correspondence Tannaka. 51 continuous. 5 multiplicative type. 65 centre of an afﬁne group. 132 general linear. 30 commutative. 30 bi-ideal. 108 dimension Krull. 47 algebraic variety afﬁne. 16 general linear GLn . 17. 154 opposite. 76 over a bialgebra. 135 of monomial matrices. 18 orthogonal. 47 smooth. 146 coideal. 79 faithful. 154 basis orthogonal. 160 separable. 47 solvable. 135 of a monoid. 166 semisimple. 46 reductive. 16 special linear SLn . 172 tensor.

36 of a representation. 173 trivial algebraic. 172 ﬁbred product. 11 tensor. 65 homomorphisms of representations. 160 irreducible. 146 ideal augmentation. 6 isogeny. 174 graded. 23 representable. 105 element group-like. 35 derived. 56 endomorphism diagonalizable. 161 constant algebraic. 40 normalize. 121 index Witt. 139 embedding. 44 height. 18 of superalgebras. 129 semisimple. 108 unipotent. 29 of comodules. 154 surjective. 113 quotient. 15 group algebra. 83 Tannaka. 43 dual Cartier. 138 Tannaka. 76 of monoids. 105 locally ﬁnite. 14 afﬁne. 30 of co-algebras. 150 function representative. 139 unipotent. 173 eigenspace generalized. 150 kernel of a homomorphism. 65 tensor. 59 Hopf. 12 of bialgebras. 39 idempotent. 121 orthogonal. 154 inversion. 140 multiplicative. 44 of an algebraic space. 174 transposition. 105 has all its eigenvalues. 40 isogenous. 108 plane
. 67 ﬂag full. 105 unipotent. 121 idempotents complete set of orthogonal. 56
BASIC T HEORY OF A LGEBRAIC G ROUPS (ALA1)
of afﬁne groups. 130 group space. 121 trivial. 38 functor ﬁbre. 105 semisimple. 18 afﬁne algebraic. 43 monoid variety. 154 group additive. 171 part semisimple. 110 trivial. 6 isometry. 58. 75 Hopf algebra coconnected. 143 form quadratic. 105 exact sequence. 20 Clifford. 31 involution. 65 object dual. 18 monoid space. 59 map lives in. 29 monoid. 172 monogenic. 18 afﬁne. 107. 44 noetherian. 8 solvable. 138. 14 of monomial matrices. 171 gradation. 108 nilpotent. 43 homomorphism injective. 43 group variety. 107.180
of an algebraic group.

47 product almost direct. 7. 150 isotropic. 174 tensor product super. 20 supergroup afﬁne algebraic. 46 geometrically. 67 of an algebraic group. 75 superalgebra. 142 subnormal. 76 semisimple. 78 subgroup afﬁne. 151 nondegenerate. 151 regular. 154 point regular. 90 subcoalgebra. 151 quotient by N . 29 subcomodule. 151 singular. 151 isotropic. 110 subrepresentation. 155 topology Zariski. 11 representation faithful. 20 tannakian category algebraic. 56 algebraic. 70 simple. 6 space quadratic. 75 ﬁnite-dimensional. 166 simple. 173 TBA. 19. 89 subcategory replete. 154 commutative. 40 prime. 6 semidirect. 100 derived.Index of deﬁnitions
hyperbolic. 134 split. 41 stabilizer. 43 reﬂection. 100 sheaf for the fpqc topology. 134 transporter. 43. 165 reduced. 69 quadratic space anisotropic. 42 regular. 51 vector anisotropic. 150 spectrum max. 74. 150
181
. 47 represent. 56 subobject generated by. 75 linear. 40 torus. 65 radical. 47 series composition. 151 totally isotropic. 167 ring coordinate. 74 regular. 151 regular. 68 semidirect deﬁned by a map. 165 unipotent.