Group Theory, Topology, and Physics

William Gordon Ritter

Group Theory

1. A Group is a set G with a binary operation (denoted just as ordinary multiplication) satisfying three axioms: (I) The operation is associative: g (hk ) = (gh)k for all g, h, k G. (II) There exists an identity 1G G such that ( g ) 1G g = g 1G = g . (III) ( g G) g 1 G such that gg 1 = g 1 g = 1G . 2. Example: let Zn = {0, 1, . . . , n 1} with addition modulo n. This is clearly a group, but what about multiplication? One can easily see that only elements of Zn that are relatively prime to n have multiplicative inverses. Therefore, Zn will be a group under multiplication i n is prime. Even if n is not prime, the set of invertible elements in Zn is a group. 3. If A, B are two sets, the direct product A B is dened to be A B = {(a, b) : a A, b B } . If A and B also happen to be groups, then the set A B has a natural group structure dened by performing all operations componentwise. The identity of A B is, of course (1A , 1B ). 4. Let G, H be groups. A map : G H is called a homomorphism if (g1 g2 ) = (g1 )(g2 ). Exercise: It follows from this that (1G ) = 1H and (g 1 ) = (g )1 , so it is not necessary to assume these as additional properties. A homomorphism is said to be an isomorphism if its 1-1 and onto. 5. Example: let Rn denote the set of nth roots of unity, i.e. the complex roots of the polynomial xn 1. They form a regular n-gon on the unit circle including the point 1, and they are a group under multiplication. Let z = exp(2i/n) be the fundamental root. Then : Zn Rn dened by (m) = z m turns out to be an isomorphism.

2 6. A subgroup is a subset H G which is also closed under multiplication and inverses, and contains the identity. In other words, H is also a group with the same operation. A left coset of H is a set of the form gH = {gh | h H } .
1 Theorem 1. Two left cosets g1 H and g2 H are the same if and only if g1 g2 H . 1 Proof. Suppose g1 H = g2 H . Then h H, h1 H such that g1 h = g2 h1 . Then g2 g1 h = h1 1 1 which proves that g2 g1 H H . By the cancellation law, this can only happen if g1 g2 H . 1 Now for the converse, suppose g1 g2 = h H . Then g2 = g1 h hence g2 H = g1 hH = g1 H .

7. Let G/H denote the set of left cosets of H in G, and investigate when the operation (g1 H ) (g2 H ) = (g1 g2 )H (1)

is well-dened. This is easy given Theorem 1. Suppose we chose dierent representatives 1 g1 H and for the two cosets above: g1 H and g2 H instead of g1 H and g2 H . Then g1 1 g2 g2 H . In order that the product (g1 H ) (g2 H ) yield the same answer, we must have (g1 g2 )1 (g1 g2 ) H . But
1 1 1 (g1 g2 )1 (g1 g2 ) = g2 g1 g1 g2 = g2 hg2

(2)

1 where we dene h = g1 g1 H . We also know that g2 = g2 h1 for some h1 H . So we can 1 1 hg2 h1 , and the latter is in H if and only if rewrite (2) as (g1 g2 ) (g1 g2 ) = g2 1 g2 hg2 H .

We are led to the inevitable conclusion that if H is a special kind of subgroup, for which 1 g2 hg2 H whenever h H (such a subgroup is said to be normal) then (1) is well-dened.

Limits and Continuity

8. Let X and Y be spaces with distance functions dX and dY respectively. (For example, the torus inherits a distance function from an embedding in R3 ). A map f : X Y is said to be continuous if, given > 0, there exists > 0 such that dX (x, y ) < dY (f (x), f (y )) < . 9. We give a second, equivalent denition of continuity. Let (an ), n N be a sequence of points in a space X . We write a = limn an in the case that dX (an , a) 0 as n . A map f : X Y is continuous if f (a) = lim f (an ) whenever a = lim an . One might say that f commutes with the limit operation.

3
Loops and Homotopy

10. A based space is an ordered pair (X, x0 ) where X is a topological space, and x0 X is a point. 11. A space X is said to be (path) connected if, given any two points p, q X , there exists a continuous map : [0, 1] X with (0) = p, and (1) = q . 12. A loop in a based space (X, x0 ) is a continuous map : [a, b] X , such that (a) = (b) = x0 . Here, [a, b] is any closed interval. Note the following important re-parameterization invariance: if : R R is a continuous real function, which maps some other interval [c, d] onto [a, b], then : [c, d] X is also a loop which has the same image in X as the original loop . We will not distinguish between loops which are related to one another by re-parameterization. 13. Let and both be dened on the interval [a, b]. A homotopy of two loops , is a continuous map F : [a, b] [0, 1] X such that F (s, 0) = (s) and F (s, 1) = (s). One also sometimes writes Ft (s) = F (s, t) so that each Ft is a loop. Thus a homotopy is a continuous path in the space of loops. 14. An equivalence relation, denoted , is a relation that is reexive (f f ), symmetric (f g g f ), and transitive (f g & g h f h). Homotopy of loops is an equivalence relation. Let [] denote the class of all loops homotopic to .

The Fundamental Group

15. Given a based space (X, x0 ), the fundamental group is denoted 1 (X, x0 ). It is the set of homotopy classes of loops based at x0 , with a group operation that we now describe. Let and be loops from [0, 1] X . Dene a new loop by ( )(s) = (2t) t [0, 1 ] 2 1 (2t 1) t [ 2 , 1]

We then dene the product of two loop classes [] [ ] = [ ]. One must check that this is operation is well-dened on equivalence classes, i.e. that [ ] only depends upon and through their homotopy class. This is the promised group operation for 1 (X, x0 ). 16. Let f : X Y be continuous. Note that if is a loop in X then f : [0, 1] Y is a loop in Y . Dene f : 1 (X, x0 ) 1 (Y, f (x0 ))

4 by the formula f ([]) = [f ]. One can check that f is well-dened on equivalence classes of loops, and is (in fact) a homomorphism, although it need be neither surjective nor injective in general. Let us check that f is a homomorphism. f ([ ]) = [f ( )] while f ([]) f ([ ]) = [(f ) (f )] (4) However f ( ) = (f ) (f ) up to a possible re-parameterization of the intervals used to dene the loops. Therefore, the two expressions (3) and (4) are equal. 17. A connected space X is said to be simply connected if 1 (X, x0 ) = {id} for some (hence all) x0 X . 18. A special case of van Kampens theorem. Let X = U V , where U and V are open sets with U V connected and simply connected. Then 1 (X ) = 1 (U ) 1 (V ), where denotes the free product. (3)