In this section, W denotes a group written multiplicatively, with identity element 1, and S denotes a set of generators of W such that S = S^{-1} and 1 \notin S. Every element of W is the product of a finite sequence of elements of S. From no. 3 onwards we assume that every element of S is of order 2.
1. Length and reduced decompositions
Definition 1
Let w \in W. The length of w (with respect to S), denoted by l_S(w) or simply by l(w), is the smallest integer q \geq 0 such that w is the product of a sequence of q elements of S. A reduced decomposition of w (with respect to S) is any sequence s = (s_1, \ldots, s_q) of elements of S such that w = s_1 \cdots s_q and q = l(w).
Thus 1 is the unique element of length 0 and S consists of the elements of length 1.
Proposition 1
Let w and w' be in W. We have the formulas:
l(ww′)≤l(w)+l(w′),(1)
l(w−1)=l(w),(2)
∣l(w)−l(w′)∣≤l(ww′−1).(3)
Let (s_1, \ldots, s_p) and (s'_1, \ldots, s'_q) be reduced decompositions of w and w' respectively. Thus l(w) = p and l(w') = q. Since ww' = s_1 \ldots s_p s'_1 \ldots s'_q, we have l(ww') \leq p + q, proving (1). Since S = S^{-1} and w^{-1} = s_p^{-1} \ldots s_1^{-1}, we have l(w^{-1}) \leq p = l(w). Replacing w by w^{-1} gives the opposite inequality l(w) \leq l(w^{-1}), proving (2). Replacing w by ww'^{-1} in (1) and (2) gives the relations
l(w)−l(w′)≤l(ww′−1),(4)
l(ww′−1)=l(w′w−1).(5)
Exchanging w and w′ in (4) gives l(w′)−l(w)≤l(ww′−1) by (5), proving (3).
Corollary
Let s=(s1,…,sp) and s′=(s1′,…,sq′) be two sequences of elements of S such that w=s1…sp and w′=s1′…sq′. If the sequence (s1,…,sp,s1′,…,sq′) is a reduced decomposition of ww′, then s is a reduced decomposition of w and s′ is one of w′.
By hypothesis, l(w)≤p, l(w′)≤q and l(ww′)=p+q. By (1), we must have l(w)=p and l(w′)=q, hence the corollary.
Remark In view of formulas (1) and (2), the formula d(w,w′)=l(ww′−1) defines a distance on W, invariant under right translations.
2. Dihedral groups
Definition 2
A dihedral group is a group generated by two distinct elements of order two.
Example
Let M be the multiplicative group {1,−1}, and let m be an integer ≥2 (resp. m=∞). Then M acts on the group Z/mZ (resp. on Z) by (−1).x=−x, and the corresponding semi-direct product of M by Z/mZ (resp. of M by Z) is denoted by Dm. The elements of Dm are thus the pairs (ε,x) with ε=±1 and x∈Z/mZ (resp. x∈Z); the group law on Dm is given by the formula
(ε,x).(ε′,x′)=(εε′,ε′x+x′).
We denote by ι the class of 1 modulo m (resp. ι=1) and set
ρ=(−1,0),ρ′=(−1,ι),π=(1,ι).
Then ρ2=ρ′2=1 and π=ρρ′. The formulas
πn=(1,nι),ρπn=(−1,nι)
show that Dm is a dihedral group generated by {ρ,ρ′}.
Proposition 2
Assume that S consists of two distinct elements s and s′ of order 2.
(i) The subgroup P of W generated by p=ss′ is normal, and W is the semi-direct product of the subgroup T={1,s} and P. Moreover, (W:P)=2.
(ii) Let m be the order (finite or infinite) of p. Then m≥2 and W is of order 2m. There is a unique isomorphism φ from Dm to W such that φ(ρ)=s and φ(ρ′)=s′.
(i) We have sps−1=sss′s=s′s=p−1, and hence
spns−1=p−n
for every integer n. Since W is generated by {s,s′}, and hence by {s,p}, the subgroup P is normal. It follows that TP is a subgroup of W, and since TP contains s and s′=sp, we have W=TP=P∪sP. To prove (i), it is therefore enough to show that W=P. If W=P, the group W would be commutative, so p2=s2s′2=1. But then the only elements of W=P would be 1 and p, contradicting the hypothesis that W contains at least three elements, namely 1, s and s′.
(ii) Since s=s′, we have p=1 and so m≥2. Since P is of order m and (W:P)=2, the order of W is 2m. If m is finite (resp. infinite), there exists an isomorphism φ′ from Z/mZ (resp. Z) to P taking π to p. Moreover, there exists an isomorphism φ′′ from M={1,−1} to T taking −1 to s. The group W is the semi-direct product of T and P. In view of the formulas (9) and ρπnρ−1=π−n, φ′ and φ′′ induce an isomorphism φ from Dm to W such that φ(ρ)=s and φ(π)=p, and hence φ(ρ′)=s′. The uniqueness of φ follows from the fact that Dm is generated by {ρ,ρ′}.
Remark
Consider a dihedral group W of order 2m generated by two distinct elements s and s′ of order 2. Denote by sq (resp. sq′) the sequence of length q whose odd (resp. even) numbered terms are equal to s and whose even (resp. odd) numbered terms are equal to s′, and let wq (resp. wq′) be the product of the sequence sq (resp. sq′). We have
w2k=(ss′)k,w2k+1=(ss′)ks,
w2k′=(s′s)k=(ss′)−k,w2k+1′=(s′s)ks′=(ss′)−k−1s.
If s=(s1,…,sq) is a reduced decomposition (with respect to {s,s′}) of an element w of W, then clearly si=si+1 for 1≤i≤q−1. Hence, s=sq or s=sq′.
If m=∞, the elements (ss′)n and (ss′)ns for n∈Z are distinct. Hence, the elements wq (q≥0) and wq′ (q>0) are distinct, and if s is a reduced decomposition of wq (resp. wq′) we necessarily have s=sq (resp. s=sq′). It follows from this that l(wq)=l(wq′)=q and that the set of reduced decompositions of the elements of W consists of the sq and the sq′. Moreover, every element of W has a unique reduced decomposition.
Suppose now that m is finite. If q≥2m, we have wq=wq−2m and wq′=wq−2m′; if m≤q≤2m, we have wq=w2m−q′, wq′=w2m−q. Hence, neither sq nor sq′ are reduced decompositions if q>m. It follows that each of the 2m elements of W is one of the 2m elements w0=w0′, wq and wq′ for 1≤q≤m−1, and wm=wm′. These 2m elements are thus distinct and it follows as above that l(wq)=l(wq′)=q for q≤m and that the set of reduced decompositions of elements of W consists of the sq and the sq′ for 0≤q≤m. Every element of W except wm has a unique reduced decomposition; wm has two.
3. First properties of Coxeter groups
Recall that from now on we assume that the elements of S are of order 2.
Definition 3
(W, S) is said to be a Coxeter system if it satisfies the following condition:
(C) For s, s' in S, let m(s, s') be the order of ss' and let I be the set of pairs (s, s') such that m(s, s') is finite. The generating set S and the relations (ss')^{m(s,s')} = 1 for (s, s') in I form a presentation¹ of the group W.
When (W, S) is a Coxeter system, one also says, by abuse of language, that W is a Coxeter group.
Example 1
Let m be an integer ≥2 or ∞ and let W be a group defined by a set of generators S = {s, s'} and the relations s2=s′2=1 when m=∞, s2=s′2=(ss′)m=1 when m is finite. Consider on the other hand the dihedral group D_m (no. 2, Example) and the elements ρ and ρ′ of D_m defined by (7). Since ρ2=(ρ′)2=1 and (ρρ′)m=1 when m is finite, there exists a unique homomorphism f from W to D_m such that f(s)=ρ and f(s′)=ρ′. Since ρρ′ is of order m, it follows that ss' is itself of order m. Hence, (W, S) is a Coxeter system, W is a dihedral group of order 2m and f is an isomorphism (Prop. 2).
By transport of structure, it follows that every dihedral group is a Coxeter group.
Example 2
Let Sn be the symmetric group of degree n, with n≥2. Let si be the transposition of i and i+1 for 1≤i<n, and let S = {s_1, ..., s_{n-1}}. One can show (§ 2, no.4, Example and § 1, Exerc. 4) that (Sn,S) is a Coxeter system.
Example 3
For the classification of finite Coxeter groups, cf. Chap. 4, § 4.
Remark
Suppose that (W, S) is a Coxeter system. There exists a homomorphism ε from W to the group {1, -1} characterized by ε(s)=−1 for all s∈S. We call ε(w) the signature of w; it is equal to (−1)l(w). The formula ε(ww′)=ε(w)ε(w′) thus translates into l(ww′)≡l(w)+l(w′)mod2.
¹ This means that (W, S) has the following universal property: given a group G and a map f from S to G such that (f(s)f(s′))m(s,s′)=1 for (s, s') in I, there exists a homomorphism g from W to G extending f. This homomorphism is unique because S generates W. An equivalent form of this definition is the following. Let W be a group, f a homomorphism from W to W and h a map from S to W such that f(h(s))=s and (h(s)h(s′))m(s,s′)=1 for (s, s') in S and such that the h(s) (for s∈S) generate W. Then f is injective (and hence is an isomorphism from W to W).
Proposition 3
Assume that (W, S) is a Coxeter system. Then, two elements s and s' of S are conjugate2 in W if and only if the following condition is satisfied:
(I) There exists a finite sequence (s1,…,sq) of elements of S such that s1=s,sq=s′ and sjsj+1 is of finite odd order for 1≤j<q.
Let s and s' in S be such that p=ss′ is of finite order 2n+1. By (9), sp−n=pns hence
pnsp−n=pnpns−p−1s=s′ss=s′,
and s′ is conjugate to s.
For all s in S, let As be the set of s′∈S satisfying (I). With the hypotheses in (I), the elements sj and sj+1 are conjugate for 1≤j<q by the above, hence every element s′ of As is conjugate to s.
Let f be the map from S to M={1,−1} equal to 1 on As and to -1 on S−As. Let s′ and s′′ in S be such that s′s′′ is of finite order m. Then f(s′)f(s′′)=1 if s′ and s′′ are both in As or both in S−As. In the other case, f(s′)f(s′′)=−1, but m is even. Thus (f(s′)f(s′′))m=1 in all cases. Since (W, S) is a Coxeter system, there exists a homomorphism g from W to M inducing f on S. Let s′ be a conjugate of s. Since s belongs to the kernel of g, so does s′, hence f(s′)=g(s′)=1 and finally s′∈As. Q.E.D.
4. Reduced decompositions in a Coxeter group
Suppose that (W, S) is a Coxeter system. Let T be the set of conjugates in W of elements of S. For any finite sequence s=(s1,…,sq) of elements of S, denote by Φ(s) the sequence (t1,…,tq) of elements of T defined by
tj=(s1…sj−1)sj(s1…sj−1)−1for 1≤j≤q.
Then t1=s1 and s1…sq=tqtq−1…t1. For any element t∈T, denote by n(s,t) the number of integers j such that 1≤j≤q and tj=t. Finally, put
R={1,−1}×T.
Lemma 1
(i) Let w∈W and t∈T. The number (−1)n(s,t) has the same value η(w,t) for all sequences s=(s1,…,sq) of elements of S such that w=s1…sq.
(ii) For w∈W, let Uw be the map from R to itself defined by
Uw(ε,t)=(ε⋅η(w−1,t),wtw−1)(ε=±1,t∈T).
The map w↦Uw is a homomorphism from W to the group of permutations of R.
(2) Recall that two elements (resp. two subsets) of a group W are said to be conjugate if there exists an inner automorphism of W that transforms one into the other.
For s∈S, define a map Us from R to itself by
Us(ε,t)=(ε.(−1)δs,t,sts−1)(ε=±1,t∈T),
where δs,t is the Kronecker symbol. It is immediate that Us2=IdR, which shows that Us is a permutation of R.
Let s=(s1,…,sq) be a sequence of elements of S. Put w=sq…s1 and Us=Usq…Us1. We shall show, by induction on q, that
Us(ε,t)=(ε.(−1)n(s,t),wtw−1).
This is clear for q=0,1. If q>1, put s′=(s1,…,sq−1) and
But Φ(s)=(Φ(s′),w′−1sqw′) and n(s,t)=n(s′,t)+δw′,−1sqw′,t, proving formula (14).
Now let s,s′∈S be such that p=ss′ has finite order m. Let s=(s1,…,s2m) be the sequence of elements of S defined by sj=s for j odd and sj=s′ for j even. Then s2m…s1=p−m=1 and formula (11) implies that
tj=pj−1sfor 1≤j≤2m.
Since p is of order m, the elements t1,…,tm are distinct and tj+m=tj for 1≤j≤m. For all t∈T, the integer n(s,t) is thus equal to 0 or 2 and (14) shows that Us=IdR. In other words, (UsUs′)m=IdR. Thus, by the definition of Coxeter systems, there exists a homomorphism w↦Uw from W to the group of permutations of R such that Us is given by the right-hand side of (13). Then Uw=Us for every sequence s=(s1,…,sq) such that w=sq…s1 and Lemma 1 follows immediately from (14).
Lemma 2
Let s=(s1,…,sq), Φ(s)=(t1,…,tq) and w=s1…sq. Let Tw be the set of elements of T such that η(w,t)=−1. Then s is a reduced decomposition of w if and only if the ti are distinct, and in that case Tw={t1,…,tq} and Card(Tw)=l(w).
Clearly Tw⊂{t1,…,tq}. Taking s to be reduced, it follows that Card(Tw)≤l(w). Moreover, if the ti are distinct, then n(s,t) is equal to 1 or 0 according to whether t does or does not belong to {t1,…,tq}. It follows that Tw={t1,…,tq} and that q=Card(Tw)≤l(w), which implies that s is reduced. Suppose finally that ti=tj with i<j. This gives si=usju−1, with u=si+1…sj−1, hence
w=s1…si−1si+1…sj−1sj+1…sq.
This shows that s is not a reduced decomposition of w.
Lemma 3
Let w∈W and s∈S be such that l(sw)≤l(w). For any sequence s=(s1,…,sq) of elements of S with w=s1…sq, there exists an integer j such that 1≤j≤q and
ss1…sj−1=s1…sj−1sj.
Let p be the length of w and w′=sw. By the Remark of no. 3, l(w′)≡l(w)+1 mod. 2. The hypothesis l(w′)≤l(w) and the relation
∣l(w)−l(w′)∣≤l(ww′−1)=l(s)=1
thus leads to l(w′)=p−1. Choose a reduced decomposition
(s1′,…,sp−1′)
of w′ and put s=(s,s1′,…,sp−1′) and Φ(s′)=(t1′,…,tp′). It is clear that s′ is a reduced decomposition of w and that t1′=s. The elements t1′,…,tp′ being distinct by Lemma 2, we have n(s′,s)=1. Since w is the product of the sequence s, we have n(s,s)≡n(s′,s) mod. 2 by Lemma 1, hence n(s,s)=0. Consequently, s is equal to one of the elements tj of the sequence Φ(s), hence the lemma.
Remark
The set Tw defined in Lemma 2 consists of the elements of the form w′′sw′′−1 corresponding to the triples (w′,w′′,s)∈W×W×S such that w=w′′sw′ and l(w′)+l(w′′)+1=l(w).
5. The exchange condition
The "exchange condition" is the following assertion about (W,S):
(E)Let w∈W and s∈S be such that l(sw)≤l(w). For any reduced decomposition (s1,…,sq) of w, there exists an integer j such that 1≤j≤q and
ss1…sj−1=s1…sj−1sj.
We assume in this number that (W,S) satisfies (E). By Lemma 3, this is so if (W,S) is a Coxeter system. The results of this number thus apply to Coxeter systems.
Proposition 4
Let s∈S,w∈W and s=(s1,…,sq) be a reduced decomposition of w. Then one of the following must hold:
a) l(sw)=l(w)+1 and (s,s1,…,sq) is a reduced decomposition of sw.
b) l(sw)=l(w)−1 and there exists an integer j such that 1≤j≤q, (s1,…,sj−1,sj+1,…,sq) is a reduced decomposition of sw and (s,s1,…,sj−1,sj+1,…,sq) is a reduced decomposition of w.
Put w′=sw. By formula (3) of no. 1, we have
∣l(w)−l(w′)∣≤l(s)=1.
We distinguish two cases:
a) l(w′)>l(w). Then l(w′)=q+1 and w′=ss1…sq, so
(s,s1,…,sq)
is a reduced decomposition of w′.
b) l(w′)≤l(w). By (E), there exists an integer j such that 1≤j≤q and (16) is satisfied. Then w=ss1…sj−1sj+1…sq and hence
w′=s1…sj−1sj+1…sq.
Since q−1≤l(w′)≤q, it follows that l(w′)=q−1 and that (s1,…,sj−1,sj+1,…,sq) is a reduced decomposition of w′.
Lemma 4
Let w∈W have length q≥1, let D be the set of reduced decompositions of w, and let F be a map from D to a set E. Assume that F(s)=F(s′) if the elements s=(s1,…,sq);s′=(s1′,…,sq′) of D satisfy one of the following hypotheses:
a) s1=s1′ or sq=sq′. b) There exist s and s′ in S such that sj=sk′=s and sk=sj′=s′ for j odd and k even.
Then F is constant.
A) Let s,s′∈D and put t=(s1′,s1,…,sq−1). We are going to show that if F(s)=F(s′) then t∈D and F(t)=F(s). Indeed, w=s1′…sq′ and hence s1′w=s2′…sq′ is of length <q. By Prop. 4, there exists an integer j such that 1≤j≤q and the sequence u=(s1′,s1,…,sj−1,sj+1,…,sq) belongs to D. We have F(u)=F(s′) by condition a); if j=q we would have F(s)=F(u) for the same reason, and hence F(s)=F(s′) contrary to our hypothesis. Thus j=q and hence t=u∈D and F(t)=F(s′)=F(s).
B) Let s and s′ belong to D. For any integer j with 0≤j≤q+1 define a sequence sj of q elements of S as follows:
s0=(s1′,…,sq′)
s1=(s1,…,sq)
sq+1−k=(s1,s1′,…,s1,s1′,s1,s2,…,sk)
for q−k even and 0≤k≤q
sq+1−k=(s1′,s1,…,s1,s1′,s1,s2,…,sk)
for q−k odd and 0≤k≤q.
Denote by (H_j) the assertion "s_j \in D, s_{j+1} \in D \text{ and } F(s_j) \neq F(s_{j+1})". By A), (H_j) \implies (H_{j+1}) for 0≤j<q, and (H_q) is not satisfied by condition b). Hence, (H_0) is not satisfied. Since s0=s′ and s1=s, it follows that F(s)=F(s′).
Proposition 5
Let M be a monoid (with unit element 1) and f a map from S to M. For s, s' \in S, let m(s, s') be the order of ss' and put
If a(s,s′)=a(s′,s) whenever s=s′ are in S, there exists a map g from W to M such that
g(w)=f(s1)…f(sq)
for all w∈W and any reduced decomposition (s1,…,sq) of w.
For any w∈W, let Dw be the set of reduced decompositions of w and Fw the map from Dw to M defined by
Fw(s1,…,sq)=f(s1)…f(sq).
We are going to prove by induction on the length of w that Fw is constant, which will establish Prop. 5. The cases l(w)=0,1 being trivial, we assume that q≥2 and that our assertion is proved for the elements w with l(w)<q. Let w be of length q and s,s′∈Dw; by Lemma 4 it suffices to prove that Fw(s)=Fw(s′) in cases a) and b) of that lemma.
for w′=s1…sq−1 and w′′=s2…sq and the induction hypothesis show that Fw(s)=Fw(s′) if s1=s1′ or sq=sq′.
b) Suppose that there exist two elements s and s' of S such that sj=sk′=s and sk=sj′=s′ for j odd and k even. It suffices to treat the case s=s′. The sequences s and s' are then two distinct reduced decompositions of w in the dihedral group generated by s and s'. By the Remark in no. 2, the order m of ss' is necessarily finite and, in the notation of that remark, s=sm and s′=sm′. Consequently, Fw(s)=a(s,s′) and Fw(s′)=a(s′,s) and hence
Fw(s)=Fw(s′).
6. Characterisation of Coxeter groups
Theorem 1
(W, S) is a Coxeter system if and only if it satisfies the exchange condition (E) of no. 5.
Lemma 3 of no. 4 shows that any Coxeter system satisfies (E).
Conversely, suppose that (E) is satisfied. Let G be a group and f a map from S to G such that (f(s)f(s′))m=1 whenever s and s′ belong to S and ss′ is of finite order m. By Prop. 5, there exists a map g from W to G such that
g(w)=f(s1)…f(sq)
whenever w=s1…sq is of length q. To prove that (W, S) is a Coxeter system, it suffices to prove that g is a homomorphism, which is a consequence of the formula
g(sw)=f(s)g(w)for s∈S,w∈W
since S generates W. By Prop. 4 of no. 5, only two cases are possible:
a) l(sw)=l(w)+1: if (s1,…,sq) is a reduced decomposition of w, then (s,s1,…,sq) is a reduced decomposition of sw, hence (21).
b) l(sw)=l(w)−1: put w′=sw; then w=sw′ and l(sw′)=l(w′)+1. By a), g(w)=f(s)g(sw) and hence f(s)g(w)=g(sw) since (f(s))2=1.
7. Families of partitions
Suppose that (W, S) is a Coxeter system. For any s∈S, let Ps be the set of elements w in W such that l(sw)>l(w). We have the following properties:
(A) ⋂s∈SPs={1}.
Indeed, let w=1 be in W and let (s1,…,sq) be a reduced decomposition of w. Then q≥1 and (s2,…,sq) is a reduced decomposition of s2w, so l(w)=q and l(s1w)=q−1. Hence, w∈/Ps1.
(B) For any s in S, the sets Ps and sPs form a partition of W.
Let w∈W and s∈S. By Prop. 4 of no. 5, we must distinguish two cases:
a) l(sw)=l(w)+1: then w∈Ps. b) l(sw)=l(w)−1: put w′=sw so w=sw′; then
l(w′)<l(sw′)
hence w′∈Ps, that is w∈sPs.
(C) Let s, s' be in S and let w be in W. If w∈Ps and ws′∈Ps then sw=ws′.
Let q be the length of w. From w∈Ps it follows that l(sw)=q+1; and from ws′∈Ps it follows that l(sws′)=l(ws′)−1≤q. Since l(sws′)=l(sw)±1, we have finally that l(ws′)=q+1 and l(sws′)=q.
Let (s1,…,sq) be a reduced decomposition of w and sq+1=s′; then (s1,…,sq,sq+1) is a reduced decomposition of the element ws′ of length q+1. By the exchange condition, there exists an integer j with 1≤j≤q+1 such that
ss1…sj−1=s1…sj.
(22) If 1≤j≤q, we would have sw=s1…sj−1sj+1…sq contradicting the formula l(sw)=q+1. Thus j=q+1 and formula (22) can be written sw=ws′.
Conversely, we have the following result:
Proposition 6
Let* (Ps)s∈Sbe a family of subsets ofWsatisfying (C) and the following conditions:
(A') 1∈Psfor alls∈S. (B') The setsPsandsPsare disjoint for alls∈S.
Then, (W,S)is a Coxeter system andPsconsists of the elementswofWsuch thatl(sw)>l(w).
Let s∈S and w∈W. There are two possibilities:
a) w∈/Ps. Let (s1,…,sq) be a reduced decomposition of w and
wj=s1…sj
for 1≤j≤q; also put w0=1. Since w0∈Ps by (A') and since w=wq is not in Ps, there exists an integer j with 1≤j≤q such that wj−1∈Ps and wj=wj−1sj does not belong to Ps. By (C),
swj−1=wj−1sj.
We have thus proved the formula
ss1…sj−1=s1…sj−1sj
which implies that sw=s1…sj−1sj+1…sq and l(sw)<l(w).
b) w∈Ps: put w′=sw, so that w′∈/Ps by (B'). By a), we then have l(sw′)<l(w′), that is l(w)<l(sw).
Comparison of a) and b) proves that Ps consists of those w∈W such that l(sw)>l(w). The exchange condition follows from this as we have seen in a), hence (W,S) is a Coxeter system (Th. 1 of no. 6).
8. Subgroups of Coxeter groups
In this number, we assume that (W,S) is a Coxeter system. For any subset X of S, we denote by WX the subgroup of W generated by X.
Proposition 7
Let w be in W. There exists a subset Sw of S such that {s1,…,sq}=Sw for any reduced decomposition (s1,…,sq) of w.
Denote by M the monoid consisting of the subsets of S with the composition law (A,B)↦A∪B; the identity element of M is ∅. Put f(s)={s} for s∈S. We are going to apply Prop. 5 of no. 5 to M and f. We have a(s,s′)={s,s′} for s,s′ in S if m(s,s′) is finite, hence there exists a map g:w↦Sw from W to M such that g(w)=f(s1)∪⋯∪f(sq), that is Sw={s1,…,sq} for any w∈W and any reduced decomposition (s1,…,sq) of w.
Corollary 1
For any subset X of S, the subgroup WX of W consists of the elements w of W such that Sw⊂X.
If w=s1…sq with s1,…,sq in S, then w−1=sq…s1; hence
Sw−1=Sw.
Prop. 4 of no. 5 shows that Ssw′⊂{s}∪Sw′ for s∈S and w′∈W, which implies the formula
Sww′⊂Sw∪Sw′
by induction on the length of w. By (23) and (24), the set U of w∈W such that Sw⊂X is a subgroup of W; we have X⊂U⊂WX, hence U=WX.
Corollary 2
For any subset X of S, we have WX∩S=X.
This follows from Cor. 1 and the formula Ss={s} for s in S.
Corollary 3
The set S is a minimal generating set of W.
If X⊂S generates W, then W=WX and hence X=S∩WX=S by Cor. 2.
Corollary 4
For any subset X of S and any w in WX, the length of w with respect to the generating set X of WX is equal to lS(w).
Let (s1,…,sq) be a reduced decomposition of w considered as an element of W. We have w=s1…sq and sj∈X for 1≤j≤q (Cor. 1); moreover, w cannot be a product of q′<q elements of X⊂S by definition of q=lS(w).
Theorem 2
(i) *For any subset X of S, the pair (WX,X) is a Coxeter system.
(ii) Let (Xi)i∈I be a family of subsets of S. If X=⋂i∈IXi, then WX=⋂i∈IWXi.
(iii) Let X and X′ be two subsets of S. Then WX⊂WX′ (resp. WX=WX′) if and only if X⊂X′ (resp. X=X′).
Every element of X is of order 2 and X generates WX. Let x∈X and w∈WX with lX(xw)≤lX(w)=q. By Cor. 4 of Prop. 7, we have
lS(xw)≤lS(w)=q.
Let x1,…,xq be elements of X such that w=x1…xq. Since (W,S) satisfies the exchange condition (Th. 1 of no. 6), there exists an integer j such that 1≤j≤q and xx1…xj−1=x1…xj−1xj. Thus, (WX,X) satisfies the exchange condition and is therefore a Coxeter system (Th. 1 of no. 6). This proves (i).
Assertions (ii) and (iii) follow immediately from Cor. 1 of Prop. 7.
9. Coxeter matrices and Coxeter graphs
Definition 4
Let I be a set. A Coxeter matrix of type I is a symmetric square matrix M=(mij)i,j∈I whose entries are integers or +∞ satisfying the relations
mii=1for all i∈I;(25)
mij≥2for i,j∈I with i=j.(26)
A Coxeter graph of type I is (by abuse of language) a pair consisting of a graph Γ3 having I as its set of vertices and a map f from the set of edges of this graph to the set consisting of +∞ and the set of integers ≥3. Γ is called the underlying graph of the Coxeter graph (Γ,f).
A Coxeter graph (Γ,f) is associated to any Coxeter matrix M of type I as follows:
the graph Γ has set of vertices I and set of edges the set pairs {i,j} of elements of I such that mij≥3, and the map f associates to the edge {i,j} the corresponding element mij of M.
It is clear that this gives rise to a bijection between the set of Coxeter matrices of type I and the set of Coxeter graphs of type I.
To assist the reader in following our arguments, we often represent a Coxeter graph of type I by the diagram used to represent its underlying graph, and write either next to or above each edge {i,j} the number f({i,j}). We generally omit these numbers if they are equal to 3.
3 See the appendix for the definition and properties of graphs used here.
If (W, S) is a Coxeter system, the matrix M=(m(s,s′))s,s′∈S, where m(s,s′) is the order of ss′, is a Coxeter matrix of type S which is called the Coxeter matrix of (W, S): indeed, m(s,s)=1 since s2=1 for all s∈S, and m(s,s′)=m(s′,s)⩾2 if s=s′ since ss′=(s′s)−1 is then =1. The Coxeter graph (Γ,f) associated to M is called the Coxeter graph of (W, S). We remark that two vertices s and s′ of Γ are joined if and only if s and s′ do not commute. For example, the Coxeter matrix of a dihedral group of order 2m is (1mm1) and its Coxeter graph is represented by
∘—m∘
when m⩾3 (or
∘—∘
if m=3) and by
∘∘
when m=2. \* The Coxeter graph of the symmetric group Sn is represented by
∘—∘—∘⋯∘—∘
(n−1 vertices.) \*
We show later (Chap. V, § 4) that, conversely, any Coxeter matrix is the matrix of a Coxeter system.
A Coxeter system (W, S) is said to be irreducible if the underlying graph of its Coxeter graph is connected (Appendix, no. 2) and non-empty. Equivalently, S is non-empty and there exists no partition of S into two distinct subsets S' and S'' of S such that every element of S' commutes with every element of S''. More generally, let (Γi)i∈I be the family of connected components of Γ (Appendix, no. 2) and let Si be the set of vertices of Γi. Let Wi=WSi be the subgroup of W generated by Si (cf. no. 8). Then the (Wi,Si) are irreducible Coxeter systems (no. 8, Th. 2) called the irreducible components of (W, S). Moreover, the group W is the restricted direct product[^1] of the subgroups Wi for i∈I. Indeed, this follows from the following more general proposition:
Proposition 8
Let (Si)i∈I be a partition of S such that every element of Si commutes with every element of Sj if i=j. For all i∈I, let Wi be the subgroup generated by Si. Then W is the restricted direct product of the family (Wi)i∈I.
It is clear that for all i∈I the subgroup Wi′ generated by the union of the Wj for j=i is also generated by Si′=⋃j=iSj. Thus
Wi∩Wi′=W∅={1}
by Th. 2 of no. 8. Since W is generated by the union of the Wi, this proves the proposition.
[^1]: A group G is the restricted direct product of a family (Gi)i∈I of subgroups if, for any finite subset J of I, the subgroup GJ of G generated by the Gi for i∈J is the direct product of the Gi for i∈J and if G is the union of the GJ. Equivalently, every element of Gi commutes with every element of Gj for i=j and every element of G can be written uniquely as a product ∏i∈Igi with gi∈Gi and gi=1 for all but finitely many indices i. This last condition is equivalent to saying that G is generated by the union of the Gi and that Gi∩GJ={1} for all i∈I and all finite subsets J of I such that i∈/J.