Bourbaki

Proposition 8

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

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

If WW is finite, the graph of (W,S)(W, S) is a forest (Chap. IV, Appendix).

If not, this graph would contain a circuit (s1,…,sn)(s_1, \ldots, s_n), n≥3n \geq 3. Putting mi=m(s1,si+1)m_i = m(s_1, s_{i+1}), 1≤i<n1 \leq i < n, and mn=m(sn,s1)m_n = m(s_n, s_1), this means that mi≥3m_i \geq 3 for all ii. Let

x=es1+⋯+esn. x = e_{s_1} + \cdots + e_{s_n}.
Then BM(x,x)=n+2∑i<jBM(esi,esj)B_M(x, x) = n + 2 \sum_{i<j} B_M(e_{s_i}, e_{s_j}). Now2^2
BM(esi,esi+1)=−cos⁡πmi≤−cos⁡π3≤−12, B_M(e_{s_i}, e_{s_{i+1}}) = -\cos \frac{\pi}{m_i} \leq -\cos \frac{\pi}{3} \leq -\frac{1}{2},
and similarly for BM(esn,es1)B_M(e_{s_n}, e_{s_1}). Since the other terms in the sum are ≤0\leq 0, we obtain
BM(x,x)≤n−n=0, B_M(x, x) \leq n - n = 0,
contrary to the fact that BMB_M is positive non-degenerate.

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.