§ 1. Primary decomposition of linear representations
1. Decomposition of a family of endomorphisms
Let V be a vector space, S a set, and r a map from S to End(V). Denote by P the set of maps from S to k. If λ∈P, denote by Vλ(S) (resp. Vλ(S)) the set of v∈V such that, for all s∈S,r(s)v=λ(s)v (resp. (r(s)−λ(s))nv=0 for n sufficiently large). The sets Vλ(S) and Vλ(S) are vector subspaces of V, and Vλ(S)⊂Vλ(S). We say that Vλ(S) is the eigenspace of V relative to λ (and to r), that Vλ(S) is the primary subspace of V relative to λ (and to r), and that V0(S) is the nilspace of V (relative to r). We say that λ is a weight of S in V if Vλ(S)=0.
In particular, if S reduces to a single element s,P can be identified with k; we use the notations Vλ(s)(s) and Vλ(s)(s), or Vλ(s)(r(s)) and Vλ(s)(r(s)), instead of Vλ({s}),Vλ({s}); we speak of eigenspaces, primary subspaces and the nilspace of r(s); an element v of Vλ(s)(s) is called an eigenvector of r(s), and, if v=0,λ(s) is called the corresponding eigenvalue (cf. Algebra, Chap. VII, §5).
For all λ∈P, the following relations are immediate:
Vλ(S)=s∈S⋂Vλ(s)(s)(1)
Vλ(S)=s∈S⋂Vλ(s)(s)(2)
Let k′ be an extension of k. The canonical map from End(V) to End(V⊗kk′) gives, by composition with r, a map r′:S→ End(V⊗kk′). Similarly, every map λ from S to k defines canonically a map, also denoted by λ, from S to k′. With these notations, we have the following proposition:
(V⊗kk′)λ(S)=Vλ(S)⊗kk′ and (V⊗kk′)λ(S)=Vλ(S)⊗kk′.
Let (ai) be a basis of the k-vector space k′. If v∈V⊗kk′,v can be expressed uniquely in the form ∑vi⊗ai where (vi) is a finitely-supported family of elements of V. For all s∈S,
The assertions in (ii) follow immediately. In case (iii), (q(s)−λ(s)μ(s))B(v,v′)=B((r(s)−λ(s))v,r′(s)v′)+B(λ(s)v,(r′(s)−μ(s))v′) for s∈S,v∈V,v′∈V′, hence by induction on n
The sum ∑λ∈PVλ(S) is direct. The sum ∑λ∈PVλ(S) is direct.
The second assertion is a consequence of the first; hence it suffices to prove that. We distinguish several cases.
a)S is empty. The assertion is trivial.
b)S is reduced to a single element s. Let λ0,λ1,...,λn be distinct elements of k. For i=0,1,...,n, let vi∈Vλi(s) and assume that v0=v1+⋯+vn. It suffices to prove that v0 = 0. For i=0,...,n, there exists an integer qi> 0 such that (r(s)−λi)qivi = 0. Consider the polynomials P(X)=∏i≥1(X−λi)qi and Q(X)=(X−λ0)q0. We have Q(r(s))v0=0, and
P(r(s))v0=∑i=1nP(r(s))vi=0. Since P and Q are relatively prime, the Bezout identity proves that v0=0.
c)S is finite and non-empty. We argue by induction on the cardinal of S. Let s∈S and S′=S--{s}. Let (vλ)λ∈P be a finitely-supported family of elements of V such that ∑λ∈Pvλ=0 and vλ∈Vλ(S). Let λ0∈P. Let P′ be the set of λ∈P such that λ∣S′=λ0∣S′. By the induction hypothesis applied to S′, we have ∑λ∈P′vλ=0. If λ,μ are distinct elements of P′,λ(s)=μ(s).
Since the sum ∑α∈kVα(s) is direct by b), and since vλ∈Vλ(s)(s),vλ=0 for all λ∈P′, and in particular vλ0=0, which we had to prove.
d) General case. Let (vλ)λ∈P be a finitely-supported family of elements of V such that ∑λ∈Pvλ=0 and vλ∈Vλ(S). Let P′ be the finite set of λ∈P such that vλ=0, and let S′ be a finite subset of S such that the conditions λ∈P′, μ∈P′,λ∣S′=μ∣S′ imply that λ=μ. We have vλ∈Vλ∣S′(S′); applying c), we see that vλ=0 for λ∈P′, which completes the proof.
Recall that, if x∈ End(V), we denote by ad x the map y→xy−yx=[x,y] from End(V) to itself.
(i) Assume that V is finite dimensional. Then x is triangularizable if and only if V=∑a∈kVa(x).
(ii) If there exists an integer n such that (ad x)ny=0, each Va(x) is stable under y.
(iii) Assume that V is finite dimensional. If V=∑a∈kVa(x) and if each
Va(x) is stable under y, there exists an integer n such that (ad x)ny=0.
Part (i) follows from Algebra, Chap. VII, §5, no. 2, Prop. 3.
Let E = End(V). Let B be the bilinear map (u,v)→u(v) from E×V to V. By the definition of ad x,
x(B(u,v))=B(u,x(v))+ B((ad x)(u),v)
for x∈E,u∈E,v∈V. Let x operate on E via ad x. By Prop. 2 (ii), B(E0(x),Va(x))⊂Va(x) for all a∈k. If (ad x)ny=0, then y∈E0(x), so y(Va(x))⊂Va(x), which proves (ii).
To prove (iii), we can replace V by Va(x) and x (resp. y) by its restriction to Va(x). Replacing x by x−a, we can assume that x is nilpotent. Then, (ad x)2dimV−1=0 (Chap. I, §4, no. 2), which proves (iii).
Assume that V is finite dimensional. The following conditions are equivalent:
(i) Condition (AC) is satisfied and, for all s∈S,r(s) is triangularizable.
(ii) For all λ∈P, Vλ(S) is stable under r(S), and V=∑λ∈PVλ(S).
If V=∑λ∈PVλ(S), then V=∑a∈kVa(s) for all s∈S, and it follows from
Lemma 1 that (ii) implies (i). Assume that condition (i) is satisfied. Lemma 1 and formula (1) imply that each Vλ(S) is stable under r(S). It remains to prove that V=∑λ∈PVλ(S). We argue by induction on dim V. We distinguish two cases.
a) For all s∈S,r(s) has a single eigenvalue λ(s). Then V=Vλ(S).
b) There exists s∈S such that r(s) has at least two distinct eigenvalues. Then V is the direct sum of the Va(s) for a∈k, and dim Va(s)< dim V for all a. Each Va(s) is stable under r(S), and it suffices to apply the induction hypothesis.
Assume that V is finite dimensional and that condition (AC) is satisfied. Let k′ be an extension of k. Assume that, for all s∈S, the endomorphism r(s)⊗1 of V⊗kk′ is triangularizable. Let P′ be the set of maps from S to k′. Then V⊗kk′=∑λ′∈P′(V⊗kk′)λ′(S).
Let r′:S→ End(V⊗kk′) be the map defined by r. If s1,s2∈S, there exists an integer n such that (ad r(s1))nr(s2)=0, hence (ad r′(s1))nr′(s2)=0. It now suffices to apply Th. 1.
Assume that V is finite dimensional and that condition (AC) is satisfied. Denote by V+(S) the vector subspace ∑s∈S((⋂i≥1r(s)iV). Then:
(i) V0(S) and V+(S) are stable under r(S);
(ii) V=V0(S)⊕V+(S);
(iii) every vector subspace W of V, stable under r(S) and such that W0(S)=0, is contained in V+(S);
(iv) ∑s∈Sr(s)V+(S)=V+(S).
Moreover, V+(S) is the only vector subspace of V with properties (i) and (ii). For any extension k′ of k,(V⊗kk′)+(S)=V+(S)⊗kk′.
The last assertion is immediate. Thus, taking Prop. 1 into account, in proving the others we can assume that k is algebraically closed. By Th. 1, V = ∑λ∈PVλ(S), and the Vλ(S) are stable under r(S). If s∈S, the characteristic
polynomial of r(s)∣Vλ(S) is (X−λ(s))dimVλ(S); it follows that ⋂i≥1r(s)iVλ(s)
is zero if λ(s)=0 and is equal to Vλ(S) if λ(s)=0; hence,
V+(S)=λ∈P,λ=0∑Vλ(S)(3)
which proves (i), (ii) and (iv). If W is a vector subspace of V stable under r(S), then W=∑λ∈PWλ(S) and Wλ(S)=W∩Vλ(S). If W0(S)=0, we see that W⊂V+(S), which proves (iii).
Let V′ be a vector subspace of V stable under r(S) and such that V′∩V0(S)=0. Then V′0(S)=0, so V′⊂V+(S) by (iii). If, in addition, V=V0(S)+V′, we see that V′=V+(S). Q.E.D.
We sometimes call (V0(S),V+(S)) the Fitting decomposition of V, or of the map r:S→ End(V). If S reduces to a single element s, we write V+(s) or V+(r(s)) instead of V+({s}). We have that V=V0(s)⊕V+(s),V0(s) and V+(s) are stable under r(s),r(s)∣V0(s) is nilpotent and r(s)∣V+(s) is bijective.
Let V and V′ be finite dimensional vector spaces, and let r:S→ End(V) and r′:S→ End(V′) be maps satisfying condition (AC). Let f:V→V′ be a surjective linear map such that f(r(s)v)=r′(s)f(v) for s∈S and v∈V. Then f(Vλ(S))=V′λ(S) for all λ∈P.
In view of Prop. 1, we are reduced to the case in which k is algebraically closed. We have V=⨁λ∈PVλ(S), V′=⨁λ∈PV′λ(S) by Th. 1, and V′=f(V)=∑λ∈Pf(Vλ(S)). Finally, f(Vλ(S))⊂V′λ(S) by Prop. 2 (i), hence the corollary.
Assume that k is perfect. Let V be a finite dimensional vector space, u an element of End(V),us,un the semi-simple and nilpotent components of u(Algebra, Chap. VII, §5, no. 8).
(i) For all λ∈k,Vλ(u)=Vλ(us)=Vλ(us).
(ii) If V has an algebra structure and if u is a derivation of V,us and un are derivations of V.
(iii) If V has an algebra structure and if u is an automorphism of V, then us and 1+us−1un are automorphisms of V.
In view of Prop. 1, we can assume that k is algebraically closed, so
V=λ∈k∑Vλ(u)
The semi-simple component of u∣Vλ(u) is the homothety with ratio λ in Vλ(u). This proves (i).
Assume from now on that V has an algebra structure. Let x∈Vλ(u), y∈Vμ(u).
If u is a derivation of V, then xy∈Vλ+μ(u) (Prop. 2 (ii)), so
us(xy)=(λ+μ)(xy)=(λx)y+x(μy)=(usx)y+x(usy)
This proves that us is a derivation of V. Then un=u−us is a derivation of V.
If u is an automorphism of V, Ker(us) =V0(u)=0, so us is bijective. On the other hand, xy∈Vλμ(u) (Prop. 2 (iii)), so
us(xy)=(λμ)(xy)=(λx)(μy)=us(x).us(y)
This proves that us is an automorphism of V; but then so is
1+us−1un=us−1u
2. The case of a linear family of endomorphisms
Assume now that S has a vector space structure, that the map r:S→ End(V) is linear, and that V and S are finite dimensional.
Assume that condition (AC) is satisfied, and let λ:S→k be such that Vλ(S)=0. If k is of characteristic 0, the map λ is linear. If k is of characteristic p=0, there exists a power q of p dividing dim Vλ(S), and a homogeneous polynomial function P:S→k of degree q, such that λ(s)q=P(s) for all s∈S.
Since Vλ(S) is stable under r(S) (Lemma 1 and formula (1) of no. 1), we can assume that V=Vλ(S). Let n= dim V. Thus, for s∈S,
det(X −r(s)) =(X−λ(s))n.
On the other hand, the expansion of the determinant shows that
det(X −r(s)) =Xn+a1(s)Xn−1+⋯+ai(s)Xn−i+⋯
where ai:S→k is a homogeneous polynomial function of degree i. Write n=qm where q is a power of the characteristic exponent of k and (q,m)=1. Then (X−λ(s))n=(Xq−λ(s)q)m; hence −mλ(s)q=aq(s), which implies the result.
Assume that k is infinite and that condition (AC) is satisfied. Let k′ be an extension of k. Put V′=V⊗kk′,S′=S⊗kk′. Let r′:S′→ End(V′) be the map obtained from r by extension of scalars. Then
V0(S)⊗kk′=V′0(S)=V′0(S′)
The first equality follows from Prop. 1. To prove the second, we can assume that V=V0(S) and so V′=V′0(S). Let (s1,...,sm) be a basis of S and (e1,...,en) a basis of V. There exist polynomials Pij(X1,...,Xm) such that
for 1≤j≤n and a1,...,am∈k′. By hypothesis, r′(s)n=0 for all s∈S, in other words Pij(a1,...,am)=0 for 1≤i,j≤n and a1,...,am∈k. Since k is infinite, Pij=0. Consequently, every element of r′(S′) is nilpotent and V′=V′0(S′).
Assume that k is infinite and that condition (AC) is satisfied. Let S be the set of s∈S such that V0(s)=V0(S). If s∈S, let P(s) be the determinant of the endomorphism of V/V0(S) defined by r(s) (no. 1, Cor. 2 (i) of Th. 1).
(i) The function s→P(s) is polynomial on S. We have S={s∈S∣P(s)=0}; this is an open subset of S in the Zariski topology (App. 1).
(ii) S is non-empty, and V+(s)=V+(S) for all s∈S.
The fact that s→P(s) is polynomial follows from the linearity of r. If s∈S,V0(s)⊃V0(S), with equality if and only if r(s) defines an automorphism of V/V0(S), hence (i).
Now let k′ be an algebraic closure of k, and introduce V′,S′,r′ as in Prop. 6. We remark that S′ satisfies condition (AC) by continuation of the polynomial identity (ad r(s1))2dimV−1r(s2)=0 valid for s1,s2∈S (no. 1, Remark ). Applying Th.1, we deduce a decomposition
V′=V′0(S′)⊕i=1∑mV′λi(S′)
with λi=0 for 1≤i≤m. For 1≤i≤m, there exists a polynomial function Pi non-zero on S′ and an integer qi such that λiqi=Pi (Prop. 5). Since k is infinite, there exists s∈S such that (P1...Pm)(s)=0, cf. Algebra, Chap. IV, §2, no. 3, Cor. 2 of Prop. 9. Then λi(s)=0 for all i, so V′0(S′)=V′0(s) and consequently V0(S)=V0(s) (Prop. 6), which shows that S=∅. If s∈S, the fact that V+(S) is stable under r(s) and is a complement of V0(s) in V implies that V+(S)=V+(s) (Cor. 2 of Th. 1).
3. Decomposition of representations of a nilpotent Lie algebra
Let h be a Lie algebra and M an h-module. For any map λ from h to k, we have defined in no. 1 vector subspaces Mλ(h) and Mλ(h) of M. In particular, if g is a Lie algebra containing h as a subalgebra, and if x∈g, we shall often employ the notations gλ(h) and gλ(h); it will then be understood that h operates on g by the adjoint representation adg.
Let h be a nilpotent Lie algebra and M a finite dimensional h-module. Denote by P the set of maps from h to k.
(i) Each Mλ(h) is an h-submodule of M. If xM is triangularizable for all x∈h, then M=∑λ∈PMλ(h).
(ii) If k is infinite, there exists x∈h such that M0(x)=M0(h).
(iii) If k is of characteristic 0, and if λ∈P is such that Mλ(h)=0, then λ is a linear form on h vanishing on [h,h], and Mλ(h)=0.
(iv) If f:M→N is a surjective homomorphism of finite dimensional h-modules, then f(Mλ(h))=Nλ(h) for all λ∈P.
(v) If N is a finite dimensional h-module, and B a bilinear form on M×N invariant under h, then Mλ(h) and Nμ(h) are orthogonal relative to B if λ+μ=0. Moreover, if B is non-degenerate then so is its restriction to Mλ(h)×N−λ(h) for all λ∈P.
Part (i) follows from no. 1, Lemma 1 and Th. 1. Part (ii) follows from no. 2, Prop. 7. Part (iv) follows from no. 1, Cor. 3 of Th. 1. We prove (iii). We can assume that M=Mλ(h). Then, for all x∈h,λ(x)= (dim M)−1Tr(xM); this proves that λ is linear (which also follows from Prop. 5) and that λ vanishes on [h,h]. Consider the map ρ:h→ Endk(M) defined by
ρ(x)=xM−λ(x)1M
from the above, this is a representation of h on M, and ρ(x) is nilpotent for all x∈h. By Engel’s theorem (Chap. I, §4, no. 2, Th. 1), there exists m=0 in M such that ρ(x)m=0 for all x∈h, so m∈Mλ(h).
The first assertion of (v) follows from no. 1, Prop. 2 (ii). To prove the second, we can assume that k is algebraically closed in view of Prop. 1 of no. 1; it then follows from the first and the fact that M=∑λMλ(h),N=∑λNλ(h), cf. (i).
Assume that k is perfect and of characteristic 2. Let h=sl(2,k), and
2(ab)
let M be the h-module k (for the identity map of h). If x= is an
ca
arbitrary element of h, denote by λ(x) the unique λ∈k such that λ2=a2+bc. A calculation shows immediately that M=Mλ(h); on the other hand, Mλ(h)=0 and λ is neither linear nor zero on [h,h], even though h is nilpotent.
Let h be a nilpotent Lie algebra, and M a finite dimensional h-module such that M0(h)=0. Let f:h→M be a linear map such that
f([x,y])=x.f(y)−y.f(x) for x,y∈h.
There exists a∈M such that f(x)=x.a for all x∈h.
Let N=M×k. Make h operate on N by the formula
x.(m,λ)=(x.m−λf(x),0)
The identity satisfied by f implies that N is an h-module (Chap. I, §1, no. 8, Example 2). The map (m,λ)→λ from N to k is a homomorphism from N to the trivial h-module k. By Prop. 9 (iv), it follows that N0(h) contains an element of the form (a,1) with a∈M. In view of the hypothesis on M,
(M×0)∩N0(h)=0
so N0(h) is of dimension 1 and hence is annihilated by h. Thus, x.a−f(x)=0 for all x∈h, which proves the corollary.
Let g be a Lie algebra, h a nilpotent subalgebra of g. Denote by P the set of maps from h to k.
(i) For λ,μ∈P, [gλ(h),gμ(h)]⊂gλ+μ(h); in particular, g0(h) is a Lie subalgebra of g containing h, and the gλ(h) are stable under ad g0(h). Moreover, g0(h) is its own normalizer in g.
(ii) If M is a g-module, gλ(h)Mμ(h)⊂Mλ+μ(h) for λ,μ∈P; in particular, each Mλ(h) is a g0(h)-module.
(iii) If B is a bilinear form on g invariant under h,gλ(h) and gμ(h) are orthogonal relative to B for λ+μ=0. Assume that B is non-degenerate. Then, for all λ∈P, the restriction of B to gλ(h)×g−λ(h) is non-degenerate; in particular, the restriction of B to g0(h)×g0(h) is non-degenerate.
(iv) Assume that k is of characteristic 0. Then, if x∈gλ(h) with λ=0, ad x is nilpotent.
The map (x,y)→[x,y] from g×g to g is g-invariant by the Jacobi identity, hence h-invariant. The first part of (i) thus follows from Prop. 2 (ii). Part (ii) is proved similarly.
If x belongs to the normalizer of g0(h) in g, (ad y).x=−[x,y]∈g0(h) for all y∈h, so (ad y)n.x=0 for n sufficiently large. This proves that x∈g0(h). Assertion (i) is now completely proved.
Assertion (iii) follows from Prop. 9 (v).
To prove (iv), we can assume that k is algebraically closed. Let x∈gλ(h), with λ=0. For all μ∈P and any integer n≥0, (ad x)ngμ(h)⊂gμ+nλ(h); let P1 be the finite set of μ∈P such that gμ(h)=0; if k is of characteristic 0 and λ=0,(P1+nλ)∩P1=∅ for n sufficiently large, so (ad x)n=0.
Assume that k is of characteristic 0. Let g be a semi-simple Lie algebra over k, B the Killing form of g,m a subalgebra of g. Assume that the following conditions are satisfied:
1) the restriction of B to m is non-degenerate;
2) if x∈m, the semi-simple and nilpotent components1 of x in g belong to m.
Then m is reductive in g (Chap. I, § 6, no. 6).
By Chap. I, §6, no. 4, Prop. 5d),m is reductive. Let c be the centre of m. If x∈c is nilpotent, then x=0; indeed, for all y∈m, ad x and ad y commute, their composition ad x∘ ad y is nilpotent, and B(x,y)=0, so x=0. Now let x be an arbitrary element of c; let s and n be its semi-simple and nilpotent components. We have n∈m. Since ad n is of the form P(ad x), where P is a polynomial with no constant term, (ad n).m=0 and so n∈c, and then n=0 by the above. Thus ad x is semi-simple. Consequently, the restriction to m of the adjoint representation of g is semi-simple (Chap. I, §6, no. 5, Th. 4b)).
Assume that k is of characteristic 0. Let g be a semi-simple Lie algebra, h a nilpotent subalgebra of g. The algebra g0(h) satisfies conditions (1) and (2) of Lemma 2 ; it is reductive in g.
1 By Chap. I, §6, no. 3, Th. 3, every x∈g can be written uniquely as the sum of a semi-simple element s and a nilpotent element n that commute with each other; the element s (resp. n) is called the semi-simple (resp. nilpotent) component of x.
Let x,x′∈g,s and s′ their semi-simple components, n and n′ their nilpotent components. We have
x′∈g0(x)⇐⇒ (ad s)(x′)=0 (Prop.4)
⇐⇒ (ad x′)(s)=0
=⇒ (ad s′)(s)=0
⇐⇒ (ad s)(s′)=0
⇐⇒s′∈g0(x) (Prop.4)
so x′∈g0(x)⇒n′∈g0(x) and (2) is proved. The Killing form of g is non-degenerate, so its restriction to g0(h) is non-degenerate (Prop. 10 (iii)). The fact that g0(h) is reductive in g thus follows from Lemma 2.
4. Decomposition of a Lie algebra relative to an automorphism
(i) For λ,μ∈k,[gλ(a),gμ(a)]⊂gλμ(a); in particular, g1(a) is a subalgebra of g.
(ii) If B is a symmetric bilinear form on g invariant under a,gλ(a) and gμ(a) are orthogonal relative to B for λμ=1. Assume that B is non-degenerate. Then, if λ=0, the restriction of B to gλ(a)×g1/λ(a) is non-degenerate.
Assertion (i) and the first half of (ii) follow from Prop. 2 (iii) applied to the composition law g×g→g and the bilinear form B. To prove the second half of (ii), we can assume that k is algebraically closed. Then g=⨁ν∈kgν(a). In view of the above, gλ(a) is orthogonal to gν(a) if λν=1; since B is non-degenerate, it follows that its restriction to gλ(a)×g1/λ(a) is also.
Assume that k is of characteristic zero and that g is semi-simple. Then the subalgebra g1(a) satisfies conditions (1) and (2) of Lemma 2 ; it is reductive in g.
Condition (1) follows from part (ii) of Prop. 12; condition (2) follows from Prop. 4 of no. 1.
5. Invariants of a semi-simple Lie algebra relative to a semi-simple action
In this no., k is assumed to be of characteristic zero.
Let g be a semi-simple Lie algebra, a a subalgebra of g reductive in g, and m the commutant of a in g. The subalgebra m satisfies conditions (1) and (2) of Lemma 2 of no. 3; it is reductive in g.
By Prop. 6 of Chap. I, §3, no. 5, applied to the a-module g, we have g=m⊕[a,g]. Let B be the Killing form of g, and let x∈a,y∈m,z∈g. Then,
B([z,x],y)=B(z,[x,y])=0 since [x,y]=0,
which shows that m is orthogonal to [a,g] relative to B. Since B is non-degenerate, and since g=m⊕[a,g], this implies that the restriction of B to m is non-degenerate; condition (1) of Lemma 2 is thus satisfied.
Now let x∈m and let s and n be its semi-simple and nilpotent components. The semi-simple component of ad x is ad s, cf. Chap. I, §6, no. 3. Since ad x is zero on a, so is ad s, by Prop. 4 (i). Thus s∈m, so n=x−s∈m, and condition (2) of Lemma 2 is satisfied.
Let g be a semi-simple Lie algebra, A a group and r a homomorphism from A to Aut(g). Let m be the subalgebra of g consisting of the elements invariant under r(A). Assume that the linear representation r is semi-simple. Then m satisfies conditions (1) and (2) of Lemma 2 of no. 3; it is reductive in g.
The proof is analogous to that of the preceding proposition:
Let g+ be the vector subspace of g generated by the r(a)x−x,a∈A,x∈g. The vector space g′=m+g+ is stable under r(A). Let n be a complement of g′ in g stable under r(A). If x∈n,a∈A,r(a)x−x∈n∩g+=0, so x∈m and then x=0 since m∩n=0. Thus, g=g′=m+g+. Let B be the Killing form of g and let y∈m,a∈A,x∈g. Then
B(y,r(a)x−x)=B(y,r(a)x)−B(y,x)
=B(r(a−1)y,x)−B(y,x)
=B(y,x)−B(y,x)=0
Thus m and g+ are orthogonal relative to B. It follows that the restriction of B to m is non-degenerate; hence condition (1) of Lemma 2. Condition (2) is immediate by transport of structure.
Exercises
All Lie algebras and modules over them are assumed to be finite dimensional over k; from §3 onwards, k is assumed to be of characteristic zero.