1.5 Rational Representations
1.5.2 Differential of a Rational Representation
LetG⊂GL(n,C)be a linear algebraic group with Lie algebrag⊂gl(n,C). Let (π,V)be a rational representation ofG. ViewingGandGL(V)as Lie groups, we can apply Proposition 1.3.14 to obtain a homomorphism (of real Lie algebras)
dπ:g //gl(V).
We call dπ thedifferentialof the representationπ. Sincegis a Lie algebra overC in this case, we haveπ(exp(tA)) =exp(dπ(tA))forA∈gandt∈C. The entries in the matrixπ(g)(relative to any basis forV) are regular functions onG, so it follows thatt7→π(exp(tA))is an analytic (matrix-valued) function of the complex variable t. Thus
dπ(A) = d
dtπ(exp(tA)) t=0
,
and the mapA7→dπ(A)is complex linear. Thus dπis a homomorphism of complex Lie algebras whenGis a linear algebraic group.
This definition of dπ has made use of the exponential map and the Lie group structure on G. We can also define dπ in a purely algebraic fashion, as follows:
View the elements ofgas left-invariant vector fields onGby the mapA7→XAand differentiate the entries in the matrix forπusingXA. To express this in a basis-free way, recall that every linear transformationB∈End(V)defines a linear function fC
on End(V)by
fC(B) =trV(BC) forB∈End(V).
The representative function fCπ=fC◦π onGis then a regular function.
Theorem 1.5.2.The differential of a rational representation (π,V)is the unique linear mapdπ:g //End(V)such that
XA(fC◦π)(I) =fdπ(A)C(I) for all A∈gand C∈End(V). (1.45) Furthermore, for A∈Lie(G), one has
XA(f◦π) = (Xdπ(A)f)◦π for all f ∈O[GL(V)]. (1.46) Proof. For fixedA∈g, the mapC7→XA(fC◦π)(I)is a linear functional on End(V).
Hence there exists a uniqueD∈End(V)such that
XA(fC◦π)(I) =trV(DC) =fDC(I).
But fDC=XDfC by equation (1.36). Hence to show that dπ(A) =D, it suffices to prove that equation (1.46) holds. Let f ∈O[GL(V)]andg∈G. Then
XA(f◦π)(g) = d
dtf π(gexp(tA)) t=0
= d
dtf π(g)exp(tdπ(A)) t=0
= (Xdπ(A)f)(π(g))
by definition of the vector fieldsXAonGandXdπ(A)onGL(V). ut Remark 1.5.3.An algebraic-group proof of Theorem 1.5.2 and the property that dπ is a Lie algebra homomorphism (taking (1.45) as the definition of dπ(A)) is outlined in Exercises 1.5.4.
LetGandH be linear algebraic groups with Lie algebrasgandh, respectively, and letπ:G //Hbe a regular homomorphism. IfH⊂GL(V), then we may view (π,V)as a regular representation ofGwith differential dπ:g //End(V).
Proposition 1.5.4.The range ofdπ is contained inh. Hence dπ is a Lie algebra homomorphism fromgtoh. Furthermore, if K⊂GL(W)is a linear algebraic group andρ:H //K is a regular homomorphism, thend(ρ◦π) =dρ◦dπ. In particular, if G=K andρ◦π is the identity map, then dρ◦dπ is the identity map. Hence isomorphic linear algebraic groups have isomorphic Lie algebras.
Proof. We first verify that dπ(A)∈hfor allA∈g. Let f ∈IHandh∈H. Then
1.5 Rational Representations 51 (Xdπ(A)f)(h) =L(h−1)(Xdπ(A)f)(I) =Xdπ(A)(L(h−1)f)(I)
=XA((L(h−1)f)◦π)(I)
by (1.46). ButL(h−1)f∈IH, so(L(h−1)f)◦π=0, sinceπ(G)⊂H. Hence we have Xdπ(A)f(h) =0 for allh∈H. This shows that dπ(A)∈h.
Given regular homomorphisms
G−→π H−→ρ K,
we setσ=ρ◦πand takeA∈gandf ∈O[K]. Then by (1.46) we have (Xdσ(A)f)◦σ=XA((f◦ρ)◦π) = (Xdπ(A)(f◦ρ))◦π= (Xdρ(dπ(A))f)◦σ. Taking f=fCforC∈End(W)and evaluating the functions in this equation atI, we
conclude from (1.45) that dσ(A) =dρ(dπ(A)). ut
Corollary 1.5.5.Suppose G⊂H are algebraic subgroups ofGL(n,C). If(π,V)is a regular representation of H, then the differential of π|G is dπ|g.
Examples
1.LetG⊂GL(V)be a linear algebraic group. By definition ofO[G], the represen- tationπ(g) =gonV is regular. We call(π,V)thedefiningrepresentation ofG. It follows directly from the definition that dπ(A) =AforA∈g.
2.Let(π,V)be a regular representation. Define thecontragredient (ordual) repre- sentation(π∗,V∗)byπ∗(g)v∗=v∗◦π(g−1). Thenπ∗is obviously regular, since
hv∗,π(g)vi=hπ∗(g−1)v∗,vi forv∈V andv∗∈V∗.
If dimV =d (thedegreeofπ) andV is identified withd×1 column vectors by a choice of basis, thenV∗is identified with 1×drow vectors. Viewingπ(g)as ad×d matrix using the basis, we have
hv∗,π(g)vi=v∗π(g)v (matrix multiplication).
Thusπ∗(g)acts by right multiplication on row vectors by the matrixπ(g−1).
The space of representative functions forπ∗consists of the functionsg7→f(g−1), where f is a representative function forπ. IfW ⊂V is aG-invariant subspace, then
W⊥={v∗∈V∗:hv∗,wi=0 for allw∈W}
is a G-invariant subspace of V∗. In particular, if (π,V) is irreducible then so is (π∗,V∗). The natural vector-space isomorphism (V∗)∗∼=V gives an equivalence (π∗)∗∼=π.
To calculate the differential ofπ∗, letA∈g,v∈V, andv∗∈V∗. Then
hdπ∗(A)v∗,vi= d
dthπ∗(exptA)v∗,vi t=0= d
dthv∗,π(exp(−tA))vi t=0
=−hv∗,dπ(A)vi.
Since this holds for allvandv∗, we conclude that
dπ∗(A) =−dπ(A)t forA∈g. (1.47) Caution: The notationπ∗(g)for the contragredient representation should not be confused with the notationB∗for the conjugate transpose of a matrixB. The pairing hv∗,vibetweenV∗andV iscomplex linearin each argument.
3.Let(π1,V1)and(π2,V2)be regular representations ofG. Define thedirect sum representationπ1⊕π2onV1⊕V2by
(π1⊕π2)(g)(v1⊕v2) =π1(g)v1⊕π2(g)v2 forg∈Gandvi∈Vi. Thenπ1⊕π2is obviously a representation ofG. It is regular, since
hv∗1⊕v∗2,(π1⊕π2)(g)(v1⊕v2)i=hv∗1,π1(g)v1i+hv∗2,π2(g)v2i
for vi∈Vi andv∗i ∈Vi∗. This shows that the space of representative functions for π1⊕π2 is Eπ1⊕π2=Eπ1+Eπ2. If π=π1⊕π2, then in matrix form we have
π(g) =
π1(g) 0 0 π2(g)
.
Differentiating the matrix entries, we find that dπ(A) =
dπ1(A) 0 0 dπ2(A)
forA∈g. Thus dπ(A) =dπ1(A)⊕dπ2(A).
4.Let(π1,V1)and(π2,V2)be regular representations ofG. Define thetensor product representationπ1⊗π2onV1⊗V2by
(π1⊗π2)(g)(v1⊗v2) =π1(g)v1⊗π2(g)v2
forg∈Gandvi∈V. It is clear thatπ1⊗π2is a representation. It is regular, since hv∗1⊗v∗2,(π1⊗π2)(g)(v1⊗v2)i=hv∗1,π1(g)vihv∗2,π2(g)v2i
forvi∈Vandv∗i ∈Vi∗. In terms of representative functions, we have Eπ1⊗π2=Span(Eπ1·Eπ2)
(the sums of products of representative functions ofπ1 andπ2). Setπ =π1⊗π2. Then
1.5 Rational Representations 53
dπ(A) = d dt
n
exp tdπ1(A)
⊗exp tdπ2(A)o t=0
=dπ1(A)⊗I+I⊗dπ2(A). (1.48)
5.Let(π,V)be a regular representation ofGand setρ=π⊗π∗onV⊗V∗. Then by Examples 2 and 4 we see that
dρ(A) =dπ(A)⊗I−I⊗dπ(A)t. (1.49) However, there is the canonical isomorphismT:V⊗V∗∼=End(V), with
T(v⊗v∗)(u) =hv∗,uiv.
Setσ(g) =Tρ(g)T−1. IfY ∈End(V)thenT(Y⊗I) =Y T andT(I⊗Yt) =TY. Henceσ(g)(Y) =π(g)Yπ(g)−1and
dσ(A)(Y) =dπ(A)Y−Ydπ(A) forA∈g. (1.50) 6.Let(π,V)be a regular representation ofGand setρ=π∗⊗π∗onV∗⊗V∗. Then by Examples 2 and 4 we see that
dρ(A) =−dπ(A)t⊗I−I⊗dπ(A)t.
However, there is a canonical isomorphism betweenV∗⊗V∗and the space of bilin- ear forms onV, whereg∈Gacts on a bilinear formBby
g·B(x,y) =B(π(g−1)x,π(g−1)y).
IfVis identified with column vectors by a choice of a basis andB(x,y) =ytΓx, then g·Γ=π(g−1)tΓ π(g−1) (matrix multiplication) . The action ofA∈gonBis
A·B(x,y) =−B(dπ(A)x,y)−B(x,dπ(A)y).
We say that a bilinear formBisinvariantunderGifg·B=Bfor allg∈G. Likewise, we say thatBisinvariantundergifA·B=0 for allA∈g. This invariance property can be expressed as
B(dπ(A)x,y) +B(x,dπ(A)y) =0 for allx,y∈V andA∈g.