To show the reverse identity (x-1)-1 = x.
For any object w, wow-1 = e. This is also true for w = x-1.
Then wow-1 = e ==>
xowow-1 = xoe ==>
xowow-1 = x ==>
xox-1o(x-1)-1 = x ==>
eo(x-1)-1 = x ==>
(x-1)-1 = x.
aba = aba = aa2b = aa2b = eb = b.
aba2 = aba2 = aab = aab = a2b = a2b.
abab = abab = aa2bb =
aa2bb = a3b2 = ee = e.
aba2b = aba2b = aabb = aabb =
a2b2 = a2e = a2.
a2ba = a2ba = a2a2b =
a2a2b = a4b = a3ab
= eab = ab = ab.
a2bab = a2bab = a2a2bb =
a2a2bb = a4e = a
a2ba2 = ... = ... = b
a2ba2b = ... = ... = e
Last two equalities left as exercises.
Since y can be anything in H, y=x is possible. So e = xx-1 = xy-1 is in H. Therefore H has the identity.
x can be anything in H, so x=e is possible. Then for any y in H, y-1 = ey-1 = xy-1 is in H. therefore every object in H has its inverse back in H.
Let z = y-1. Then z-1 = (y-1)-1 = y. For every x,z in H, xy = xz-1 is in H. (Closure).
if e is the identity, then eS = S, because when the product of e and each object in S is that same object.
Let H be a subgroup. And let h be any object in H. Then hH = H because multiplying all the objects in H by h merely rearranges the objects in H. Also, HH = H.
If w is any object in G, then wH is called a left coset of H and Hw is called a right coset of H. As pointed out above, in a non-commutative group a left coset and a right coset of a subgroup may or may not be the same.
A subgroup of a group is a normal subgroup if every left coset of the subgroup is equal to the corresponding right coset.
Notation: Let H be a subgroup of a group G. H is normal if and only if for every object x in G, xH = Hx.
Example1. Let G be the non-commutative group of six objects. Let H = {e, a, a2}. Then bH = {b, a2b, ab} = Hb. The reader can verify that abH = Hab and a2b = H a2b. (Why is it not necessary to test cosets with multiplication with e,a, a2? Therefore xH = Hx and H is normal.
Exmple2. This time let H = {e, b}. The reader can verify that aH ≠ Ha. So H is not a normal subset.
==================
If there is a binary operation, then it is possible to combine a object with a subset of objects to get a subset.. Let subset S = {a,b,c,d} and let w be any object in the set. Then
In most discussions the set will be a group and the subset will be a subgroup. In [2.3] an additive group structure was imposed on the coordinate plane. Any line W through the origin is an additive subgroup of the plane. If W is the y-axis, then a point is in W if and only if its x-coordinate is zero. Let (0,u) and (0,v) be points in W. then
Let c be any real number. Add c to every point (x,y) on W to get a new point (x + c, y + c). (If W is the y-axis then add c to every point (0,y) on W.) If the W has slope m then the equation of W is y = mx. After adding c to every point of W, the new line has equation y = mx – c. (Click here to see why this is so.) (If W is the y-axis, then x = c is the new line.) The new line, denoted by c + W is parallel to W because both lines have the same slope. (If W = y-axis then W and c + W are vertical lines.) As c changes values, a family of parallel lines is produced. (In fact as c assumes all real numbers, a partition of the coordinate plane is formed from all these parallel lines.) Each of these lines parallel to W is called a coset of W. The coset c + W is not a subgroup.
It is possible to impose a "natural" addition operation on all these cosets of parallel lines:
Define a relation ~ on G as follows: x ~ y if and only if x-1y is in H, where x and y are any objects in G.
Part I: First show that ~ is an equivalence relation.
(a) ~ is reflexive x ~ x because x-1x = e and identity e is in subgroup H.
(b) ~ is symmetric x ~ y ==> x-1y is in H ==> y-1x is in H ==> y ~ x.
Note about implication ==>: the inverse of x-1y is y-1x because their product = e. H contains inverses of all its objects.
(c) ~ is transitive x ~ y and y ~ z ==> x-1y is in H and y-1z is in H ==> (x-1y)(y-1z) ==> x-1z is in H ==> x ~ z
Note about implication ==>: H is closed under multiplication.
Part II: Next show that the equivalence classes of ~ coincide with the left cosets of H
x is in [y] <==> x ~ y [definition of equivalence class [y]
<==> x-1y is in H [definition of ~]
<==> y-1x is in H [H contains all inverses of its objects]
<==> y-1x = h for some h in H
<==> x = yh
<==> x is in yH
Therefore, [y] = yH. This is true for all y in G. Therefore, the equivalence classes of ~ coincide with the left cosets of H. Hence the left cosets form a partition of G because equivalence classes do that.