Bourbaki

Proposition 9

Tag 0ARW, permanent. Cite it as [Bourbaki, Tag 0ARW].

Lie Groups and Lie Algebras, chapter V, § 4. The geometric representation of a Coxeter group

With the preceding notations, WW is the subgroup of O(E)\mathbf{O}(E) generated by the reflections with respect to the hyperplanes of H\mathfrak{H}. It is an essential group (§3§ 3, no. 7) and C0C_0 is a chamber of EE relative to H\mathfrak{H}.

The first assertion is trivial. On the other hand, if x∈Ex \in E is invariant under WW, it is orthogonal to all the ese_s, and hence is zero; this shows that WW is essential. Finally, the isomorphism E→E∗E \to E^* defined by BMB_M transforms C0C_0 to the set CC of no. 4; the property (Pn)(P_n) proved there shows that, for all

2^2 The roots of the equation z3−1=0z^3 - 1 = 0 are 1 and −1±i32\frac{-1 \pm i \sqrt{3}}{2}. Thus cos⁡2π3=−12\cos \frac{2\pi}{3} = -\frac{1}{2} and consequently cos⁡π3=12\cos \frac{\pi}{3} = \frac{1}{2}. Note also that sin⁡2π3=32\sin \frac{2\pi}{3} = \frac{\sqrt{3}}{2}, hence

sin⁡π3=32,cos⁡π6=−cos⁡5π6=32,sin⁡π6=sin⁡5π6=12. \sin \frac{\pi}{3} = \frac{\sqrt{3}}{2}, \quad \cos \frac{\pi}{6} = -\cos \frac{5\pi}{6} = \frac{\sqrt{3}}{2}, \quad \sin \frac{\pi}{6} = \sin \frac{5\pi}{6} = \frac{1}{2}.
Similarly, the roots of the equation z2−i=0z^2 - i = 0 are ±1+i2\pm \frac{1+i}{\sqrt{2}}, so
cos⁡π4=sin⁡π4=22and consequentlysin⁡3π4=−cos⁡3π4=22. \cos \frac{\pi}{4} = \sin \frac{\pi}{4} = \frac{\sqrt{2}}{2} \quad \text{and consequently} \quad \sin \frac{3\pi}{4} = -\cos \frac{3\pi}{4} = \frac{\sqrt{2}}{2}.

w∈Ww \in W and all s∈Ss \in S, w(C0)w(C_0) does not meet HsH_s. It follows that C0C_0 is contained in the complement UU of the union of the hyperplanes of H\mathfrak{H}, and since C0C_0 is connected, open and closed in UU, it is a chamber of EE relative to H\mathfrak{H}. Q.E.D.

We can apply to WW and C0C_0 all the properties proved in § 3. In particular, C0‾\overline{C_0} is a fundamental domain for the action of WW on EE (in other words, the cone UU defined in no. 6 is equal to the whole of EE).

Conversely, let EE be a finite dimensional real vector space, provided with a scalar product (x∣y)(x|y) and let WW be an essential finite group of displacements of EE leaving 0 fixed; assume that WW is generated by reflections. Let C0C_0 be a chamber of EE with respect to WW (cf. § 3), and let SS be the set of orthogonal reflections relative to the walls of C0C_0. Then (W,S)(W, S) is a finite Coxeter system (§3§ 3, no. 2, Th. 1). Moreover, if s∈Ss \in S, denote by HsH_s the wall of C0C_0 corresponding to ss, and denote by ese_s the unit vector orthogonal to HsH_s and on the same side of HsH_s as C0C_0. If (m(s,s′))(m(s, s')) denotes the Coxeter matrix of (W,S)(W, S), Props. 3 and 7 of §3§ 3 show that

(es∣es′)=−cos⁡(π/m(s,s′)) (e_s|e_{s'}) = -\cos(\pi/m(s, s'))

and that the ese_s form a basis of EE. The natural representation of WW on EE can thus be identified with the representation σ\sigma of no. 3.

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

Cited by

Nothing.

Cites

Transcribed from the printed volume. Not translated.