Bourbaki

Proposition 3

Tag 09GZ, permanent. Cite it as [Bourbaki, Tag 09GZ].

Integration, VII, p. 56, § 3. Applications and examples

Let GG be a locally compact group. Let HH be a closed normal subgroup of GG, isomorphic to Rn\mathbf{R}^n and such that G/HG/H is compact.

a) There exists a closed subgroup LL of GG such that GG is the topological semi-direct product of LL and HH.

b) If MM is a compact subgroup of GG, there exists an element x∈Hx \in H such that x−1Mx⊂Lx^{-1}Mx \subset L.

c) Every compact subgroup of GG is contained in a maximal compact subgroup.

d) The maximal compact subgroups of GG are the subgroups that are the transforms of LL by the inner automorphisms of GG.

Let π\pi be the canonical homomorphism of GG onto K=G/HK = G/H. By passage to the quotient, the mapping (s,h)↦shs−1(s, h) \mapsto shs^{-1} of G×HG \times H into HH defines a continuous mapping (σ,h)↦σ⋅h(\sigma, h) \mapsto \sigma \cdot h of K×HK \times H into HH such that shs−1=π(s)⋅hshs^{-1} = \pi(s) \cdot h. We shall identify HH with Rn\mathbf{R}^n (and will therefore employ, as the case may be, either the multiplicative or the additive notation for the group law in HH). By the Cor. of Prop. 2, there exists a continuous mapping ff of GG into HH such that f(xh)=f(x)+hf(xh) = f(x) + h for x∈G,h∈Hx \in G, h \in H. For every x∈Gx \in G, let p(x)=x⋅(−f(x))p(x) = x \cdot (-f(x)), which depends only on the coset of xx with respect to HH. Set

(1)F(x,y)=p(xy)−1p(x)p(y)=f(xy)y−1x−1x(−f(x))y(−f(y))=f(xy)[y−1(−f(x))y](−f(y))=f(xy)−π(y)−1f(x)−f(y). \begin{align*} (1) \quad F(x, y) &= p(xy)^{-1} p(x)p(y) = f(xy)y^{-1}x^{-1}x(-f(x))y(-f(y)) \\ &= f(xy)[y^{-1}(-f(x))y](-f(y)) \\ &= f(xy) - \pi(y)^{-1}f(x) - f(y). \end{align*}

We see that if F(x,y)=0F(x, y) = 0 for all x,yx, y in GG, then p(G)=Lp(G) = L is a subgroup of GG that intersects each coset of HH in one and only one point. Since pp is continuous, GG is then the topological semi-direct product of LL and HH (GT, III, §2, No. 10).

Now, for any h,h′∈Hh, h' \in H,

F′(xh,yh′)=f(xhyh′)−π(y)−1f(xh)−f(yh′)=f(xhy)+h′−π(y)−1f(x)−π(y)−1h−f(y)−h′=f(xy(π(y)−1h))−π(y)−1f(x)−f(y)−π(y)−1h=f(xy)−π(y)−1f(x)−f(y)=F(x,y). \begin{align*} F'(xh, yh') &= f(xhyh') - \pi(y)^{-1}f(xh) - f(yh') \\ &= f(xhy) + h' - \pi(y)^{-1}f(x) - \pi(y)^{-1}h - f(y) - h' \\ &= f(xy(\pi(y)^{-1}h)) - \pi(y)^{-1}f(x) - f(y) - \pi(y)^{-1}h \\ &= f(xy) - \pi(y)^{-1}f(x) - f(y) = F(x, y). \end{align*}

Therefore FF defines, by passage to quotients, a continuous mapping φ\varphi of K×KK \times K into HH.

On the other hand, for all x,y,zx, y, z in GG, we have

F(z,xy)+F(x,y)=f(zxy)−π(xy)−1f(z)−f(xy)+f(xy)=−π(y)−1f(x)−f(y)=π(y)−1f(zx)−π(xy)−1f(z)−π(y)−1f(x)+f(zxy)=−π(y)−1f(zx)−f(y)=π(y)−1F(z,x)+F(zx,y), \begin{align*} F(z, xy) + F(x, y) &= f(zxy) - \pi(xy)^{-1}f(z) - f(xy) + f(xy) \\ &\phantom{=} - \pi(y)^{-1}f(x) - f(y) \\ &= \pi(y)^{-1}f(zx) - \pi(xy)^{-1}f(z) - \pi(y)^{-1}f(x) + f(zxy) \\ &\phantom{=} - \pi(y)^{-1}f(zx) - f(y) \\ &= \pi(y)^{-1}F(z, x) + F(zx, y), \end{align*}

therefore, for all x′,y′,z′x', y', z' in KK,

−φ(x′,y′)=φ(z′,x′y′)−y′−1φ(z′,x′)−φ(z′x′,y′). -\varphi(x', y') = \varphi(z', x'y') - {y'}^{-1}\varphi(z', x') - \varphi(z'x', y').

Let us integrate with respect to z′z' by means of the normalized Haar measure α\alpha of KK. Setting ψ(x′)=∫φ(z′,x′) dα(z′)\psi(x') = \int \varphi(z', x')\, d\alpha(z'), ψ\psi is a continuous function on K, and (on observing that the operations of K in Rn\mathbf{R}^n respect the vector space structure of Rn\mathbf{R}^n by GT, VII, §2, No. 1, Prop. 1), one obtains

−φ(x′,y′)=ψ(x′y′)−y′−1ψ(x′)−ψ(y′). -\varphi(x', y') = \psi(x'y') - {y'}^{-1}\psi(x') - \psi(y').

In other words, setting k−ψ∘πk - \psi \circ \pi, which is a continuous function on G,

(2)

−F(x,y)=k(xy)−π(y)−1k(x)−k(y). -F(x, y) = k(xy) - \pi(y)^{-1}k(x) - k(y).

Comparing (1) and (2), one sees that if ff is replaced by the continuous function f+kf + k (which leaves verified the property f(xh)=f(x)+hf(xh) = f(x) + h), F is replaced by 0 and, as we saw earlier, this completes the proof of a).

For every g∈Gg \in G, let lgl_g (resp. hgh_g) be the unique element of L (resp. H) such that g=hglgg = h_g l_g. If h1∈Hh_1 \in H and g∈Gg \in G, then

gh1=hglgh1=hg(lgh1lg−1)lg, gh_1 = h_g l_g h_1 = h_g(l_g h_1 l_g^{-1})l_g,

thus hgh1=hg+lgh1lg−1h_g h_1 = h_g + l_g h_1 l_g^{-1}. For every g∈Gg \in G, let ψg\psi_g be the mapping of H into itself defined by

ψg(h1)=hg+lgh1lg−1. \psi_g(h_1) = h_g + l_g h_1 l_g^{-1}.

One sees that the mapping (g,h1)↦ψg(h1)(g, h_1) \mapsto \psi_g(h_1) of G×HG \times H into H is continuous and makes H a homogeneous space for G, in which the stabilizer of the origin is L. We observe moreover that when H is identified with Rn\mathbf{R}^n, ψg\psi_g is an affine mapping of H into itself. This said, let M be a compact subgroup of G; by Lemma 2, there exists an x∈Hx \in H such that ψm(x)=x\psi_m(x) = x for all m∈Mm \in M. For y∈Hy \in H, ψy\psi_y is the translation with vector yy; it follows that for every m∈Mm \in M, ψx−1∘ψm∘ψx\psi_{x^{-1}} \circ \psi_m \circ \psi_x transforms the origin of H into itself, therefore x−1mx∈Lx^{-1}mx \in L. This proves that x−1Mx⊂Lx^{-1}Mx \subset L, whence b).

Let L′L' be a closed subgroup of G containing L. Then L′L' is the topological semi-direct product of L and L′∩HL' \cap H. If L′L' is compact, then L′∩HL' \cap H is compact hence reduces to a point (GT, IV, §2, No. 2, Cor. 1 of Th. 2), therefore L′=LL' = L. This proves that L is a maximal compact subgroup of G; the same is therefore true of the subgroups that are the transforms of L by the inner automorphisms of G. The assertions c) and d) of Prop. 3 are then immediate consequences of b).

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Cites

Transcribed from the printed volume. Not translated.