15.01.2013 Views

Simplicial Structures in Topology

Simplicial Structures in Topology

Simplicial Structures in Topology

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

36 I Fundamental Concepts<br />

where 0 ≤ k ≤ n and a stands for any element of S. LetF(S)=WS/E be the set of<br />

equivalence classes def<strong>in</strong>ed by E <strong>in</strong> WS; we denote the equivalence class of a word<br />

w with [w]. We give here the def<strong>in</strong>ition of a product <strong>in</strong> F(S) by juxtaposition:<br />

[s ε1 1 ...sεn n ][s εn+1<br />

n+1 ...sεn+m n+m ]=[sε 1<br />

1 ...sεn+m n+m ].<br />

(I.2.3) Lemma. The set F(S) with the operation juxtaposition is a group.<br />

Proof. Clearly, the juxtaposition is associative; the identity is the class of the empty<br />

word and is denoted with 1; f<strong>in</strong>ally, the <strong>in</strong>verse of [ω] =[s ε1 1 sε 2<br />

2 ...sεn n ] is [ω] −1 =<br />

[s −ε1 n s −εn−1 n−1 ...s −ε1 1 ]. �<br />

The group F(S) is the free group generated by S; the elements of S are the<br />

generators of F(S). When S has a f<strong>in</strong>ite number of elements, we may also write<br />

F({s1,s2,...,sl})=〈s1,s2,...,sl〉.<br />

It is useful to write the elements of F(S) without the square brackets. In practice,<br />

it is usual to simplify the notation by means of <strong>in</strong>tegral exponents, not necessarily<br />

±1, such as <strong>in</strong> s1 1s11 s11 s11 = s41 or s−1<br />

2 s−1<br />

2 = s−2<br />

2 . It is also customary to write s <strong>in</strong>stead<br />

of the word s1 def<strong>in</strong>ed by s ∈ S; all this allows us to write equalities such as s2s−1 = s,<br />

ss−1 = 1, and so on.<br />

(I.2.4) Remark. Awordω = s ε1 1 sε 2<br />

2 ...sεn n is reduced if, for every element a of S,the<br />

“subword” aεa−ε does not appear <strong>in</strong> ω (that is to say, if it is not possible to do any<br />

“cancellation”). In each class [ω]=[s ε1 1 sε 2<br />

2 ...sεn n ], there is one and only one reduced<br />

word. We may therefore def<strong>in</strong>e F(S) by consider<strong>in</strong>g only the reduced words def<strong>in</strong>ed<br />

by S.<br />

Given a nonempty subset R ⊂ F(S), letR be the <strong>in</strong>tersection of all normal subgroups<br />

of F(S) that conta<strong>in</strong> R; the quotient group<br />

F(S)/R = F(S;R)<br />

is the group generated by the set S with the relations R <strong>in</strong> F(S). The elements of<br />

F(S;R) are the lateral classes (modulo R)ofelementsofF(S); for each element ω ∈<br />

F(S), ωR denotes its class modulo R. In practice, to construct F(S;R) we only take<br />

reduced words of S and free them from all the subwords of R. IfS = {s1,s2,...,sl}<br />

and R = {w1,w2,...,wr}, we normally write<br />

〈s1,s2,...,sl | w1 = w2 = ···= wr = 1〉 = F(S;R).<br />

Examples: 1. S = {s}, F(S) =〈s〉 �Z; this group could also be described as<br />

S = {s,t}, R = {t}, 〈s,t |t = 1〉 = F(S;R) � Z.<br />

2. S = {s}, R = {s 2 }, 〈s | s 2 = 1〉�Z2.<br />

3. S = {s,t}, R = {sts −1 t −1 }, 〈s,t | sts −1 t −1 = 1〉�Z × Z, thefree Abelian group<br />

generated by the elements s and t.<br />

4. Any group G may be viewed as a group generated by elements and relations: take<br />

S = G and RG = {(st) 1 t −1 s −1 | s,t ∈ G}; then, G � F(G;RG).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!