02 - Properties of Groups

Last time we covered Groups.

Theorem

In a Group G the identity element is unique.

Proof

Suppose e1,e2G are both identity elements. Then:
e1e2=e2
because e1 is an identity element. Similarly e1e2=e2 since e2 is an identity. Thus e1=e2.

Theorem

Each element in a Group G has a unique inverse.

Proof

Take any aG and suppose α1,α2G are both inverses of a. Then:
α1a=α2a
Due to Cancellation Theorem (right cancellation) then α1=α2.

As a result, we'll instead of using

Note that now we won't use a symbol at all. Namley:

abab

to save on space. Also we'll use 1 as the identity element, and for inverses we'll use a1 to denote it. Namely:

aa1=a1a=1

Note that a1 doesn't literally mean 1a, but is just a denotation to get that same feeling.

Thought

Given an element aG we can compute a1=a,a2=aa,a3=aaa, and for the inverses a1,a2=a1a1, and denote a0=1. Then:

am+n=aman

This leads to the following theorem:

Theorem

For any Group G then:
(ab)1=b1a1

*Proof

Check the inverse property:
(ab)(b1a1)=a(bb1)a1=a(1)a1=(a1)a1=aa1=1

As a corollary:

The product of invertible elements is invertible

(a1a2an)1=an1a11 by induction on our previous theorem.

Sometimes it happens that:

an=1

for some nN. When it does we make the following definition:

order (algebraic)

The order of aG for some Group G is the smallest positive integer nN such that:
an=1

Note that some notations for order are:
ordG(a)
You may think having the G subscript isn't necessary, but later we'll be dealing with families of groups so this is helpful. Another option is:
ord(a)
or even:
|a|
where is the absolute value coming from? The idea is that absolute value usually just means distance, and for groups it dictates how many elements in the group. Namely:
$
|G| = \text{ # elements in G}
$

Extended Example: Dihedral Groups

See Dihedral Groups

Look at a plane square:

Let D8 be the (soon to be) Group of symmetries of this square. What we mean is all the operations that bring the square back to it's original "look" (even thought the points will move around). For this, this only consists of rotations and reflections.

Let's list these operations:

  1. Do nothing: identity transformation I
  2. Rotate 90 degrees CW:
to denote this we use $r$ 3. Rotate 180 degrees CW: $r^{2} = rr$ 4. Rotate 270 degrees CW: $r^{3}$ (equivalent to 90 degrees CCW) $\equiv r^{-1}$ 5. Rotate 360 degrees CW: $r^{4} = I$. 6. Flip on the vertical midline:
notice that this can't be created from just the $rs so we need to have a new denotion of the operation, which we call $s$ 7. Flip on the horizontal midline:
8. Flip on the diagonal (descending midline)
9. Flip on the ascending diagonal midline

Okay we want to know, what are the needed transformations to get all of these operations? For example:
7. Flip on the horizontal midline is just: r2s
8. Flip on the descending diagonal midline is just: r3s
9. To get the flip on the ascending diagonal midline we can do: rs

Notice that:

  • Every symmetry just re-orders the vertices. Namely, each symmetry causes a permutation
  • Not every permutation is physically possible. For example since 1,3 are points that are never next to each other, you can't have a square where 1,3 share some edge.
  • D8={1,r,r2,r3,s,rs,r2s,r3s} is the dihedral group, and lists all unique permutations.
  • Notice that means that so |D8|=8 (since there's 8 unique elements). Some books call this D4 since the Order (Groups) of this group is 4 (looking at r4=1)
    • r4=1 so |r|=4
    • s2=1 so |s|=2
    • r3s=sr (see the figure below)
## Definition
dihedral group

This Group is the symmetries of the n-gon. It's called the dihedral group of order 2n "generators". Namely as we did above we have:

  • r is rotation by 2πn about the center.
  • s is a reflection across some midline (just one is needed)
    D2n={1,r,r2,,rn1,s,rs,r2s,,rn1s}

The "key relations" here are that:

  • rn=1 (this implies that r1=rn1)
  • s2=1 (this implies that s=s1)
  • sr=rn1s (this implies that sr=r1ssrs1=r1)

Specifically the srs1 part is called the conjugate of r by s.

The book will decide to put the $rs on the right of the $ss rather than the left here. You can use the sr=rn1s relation.
Note

Another property is that:
ris=sri
for all 0in by doing induction on the previous relation.

Mini Example

If we are considering the dihedral group when n=4 then we have:
sr3=srrr=r1srr=r1r1r1s=r3s=(r1)3s=(r3)3s=r9s=rs

There's an idea from Linear Algebra that is taken where we have a "presentation" for D2n. Namely:

D2n=r,s|rn=1,s2=1,sr=rn1s
here the r,s part are called the generators.

Danger

Suppose:
G=x,y|x4=1,y3=1,xy=y2x2
We could ask what the orders of x,y are, and other things, but actually x,y=1 here (and thus G={1} as the trivial group). Hence we try to avoid this notation since it may obscure less obvious facts.

When n3 then the group D2n is not commutative (since srrn1s), so it's not an Abelian Group.

Let n be a fixed positive integer. Define a relation on Z by:
abn|(ba)
clearly aa,abba so the relation is trivially reflexive and symmetric. Further work shows it's transitive too, so then is an equivalence relation, so we can write:
abab(modn)
For any kZ we'll denote the equivalence class of a as a, named the congruence/residue class of a. It consists of the integers from a that by an integral multiple of n (using the relation above), ie:
a={a+kn|kZ}={a,a±n,a±2n,}
There are precisely n distinct equivalence classes modn, namely:
0,1,,n1
determined by the possible remainders after division by n.

ZnZ

The groups ZnZ are the sets of equivalence classes 0,1,,n1 such that each one describes the integers with remainder of the equivalence class.

Z6Z

There are 6 possible remainders here:
0,1,2,3,4,5
Where, for example, 2={,10,4,2,8,}.

A cool thing here is that any element of the equivalence class can represent it. Namely:
=4=0=6=12=

Some Properties

  • For (ZnZ,+) then a+b=a+b (so it's an Abelian Group of order n)
  • For (ZnZ,) then ab=ab (almost an Abelian Group, but it lacks the inverse for some elements like 0)
    • Here aZnZ has in inverse under iff gcd(a,n)=1.
Z6Z

Here:
2=2=4
11=1
55=25=1
Notice that 1,5 are the only elements with other.

  • The group (ZnZ)×={aZnZ|gcd(a,n)=1} is an Abelian Group under .
    • Back in the day, they used a=bab(modn). In ZnZ this is "a congruent to b modulo n".

Permutation Group

Consider the Permuation Group. Any finite set X will have perumations:

permuation

A permutation of X is a bijection σ:XX. It's essentially a shuffling of the elements in X.

X={1,2,3,4,5}

Count the number of bijections σ:XX. Using principles of proofs, this is the number of injections σ:XX (since X is finite). Once we know where 1 goes we have 5 options, then for 2 we have 4 options, ... so then we have 54321=5!=120 possibilities.

There are some notational options here (which is where the old-timey modulo notation comes into play):

  1. (Functional Notation:) one permuation is: σ(1)=2,σ(2)=3,σ(3)=5,σ(4)=1,σ(5)=4. This is obviously super slow so we don't want to do this one.
  2. (Modified Functional Notation:) write it instead with : 12,23,.
  3. (Use A Table:) write where the top row maps to the bottom row:
    (1234523514)
  4. ("Cycle Notation":) write out a visual graph of what number maps to where:

We write this as:
σ=(12354)
Another example would be:
τ=(135)(24)
Creates the following:

Here τ has a 3-cycle and a 2-cycle (or a transposition). This τ is a product of disjoint cycles. Notice that since we have a cycle where n3 then the group isn't abelian with operation τ.