0 - Ordered Sets, Suprema, Infima

The field of rational numbers are problematic.
As for construction, one can do it by defining equivalence classes on Z2, where (a,b)(p,q) if aq=bp. If we let [a,b] denote the equivalence class of (a,b) under the above relation, addition is defined as [a,b]+[p,q]=[aq+bp,bq], multiplication as [a,b][p,q]=[ap,bq] (both stand on addition and multiplication of integers), then It's not difficult to check that this is the familiar field of the rational numbers.

But we won't do this rational numbers as equivalence classes of Z2 nonsense here.

2 is irrational

There is no pQ such that p2=2.

proof:
For the sake of contradiction, suppose there is a rational number p for which p2=2. Now, p=ab where GCD(a,b)=1. Then, we have a2=2b2, which means 2|a2 hence 2|a. writing a as 2m, we have 2m2=b2 hence 2|b as well. Therefore, GCD(a,b)2. Arriving at the desired contradiction.


The gap at 2 in the rationals

Let A={pQ:p2<2}, B={pQ:p2>2}. Then A has no largest element, and B has no smallest element.

proof: Allow q=p(p22)p+2=2p+2p+2. Then, q22=2(p22)(p+2)2.

Now Suppose p is the maximal element of A. Using the rightmost equation, given that p2<2, we see that q22<0 Hence qA. However, using, the left most equation, we have that q>p, hence we arrive at a contradiction.

Now suppose p is the minimal element of B. Again, using the rightmost equation, given that p2>2, indeed, q22>0 Hence qB.
However, using the leftmost equation, we have q<p, arriving at a contradiction.


Holes in the rational numbers

Even though the rational numbers are "dense" in the sense that between any two a,bQ, we have a+b2Q, it seems that there are holes, certain equations have no solutions in Q, moreover, it seems like we can get arbitrarily close to these holes from the left and from the right, as illustrated for the hole at "2".

Ordered sets

A set S with a relation < on S is called an orderd set if:

  • For each x,yS exactly one of x<y, x=y, y<x is true. This is often called the Trichotomy law.
  • If x<y and y<z for some x,y,zS, then x<z, meaning that order is transitive.
  • Sometimes, we write x>y instead of y<x, and xy is used to denote x<y or x=y.
Upper and Lower bounds

For the following discussion, let S denote an ordered set.

A subset E of S is said to be bounded above if there exists bS such that for all eE, e<b. In such a case b is called an upper bound for E.

If γ is an upper bound for E such that for any upper bound b of E, γ<b, then γ is known as the least upper bound or supremum of E. Often denoted γ=supE.

A subset F of S is said to be bounded below, if there exits cS such that for all fF, f>c. In such a case, c is called a lower bound of F.

if β is a lower bound of F such that for any lower bound c of F, β>c, then β is known as the greatest lower bound or infimum of F. Often denoted β=infF.

A subset P of S is said to be bounded, if it is both: bounded above and below.

More on the gap at 2 in the rationals

Let A={pQ:p2<2}, B={pQ:p2>2}. Then A has no supremum in Q and B has no infimum in Q.

proof: Firstly, its clear that all upper bounds of A are in B. and that all lower bounds of B are in A. But form this proposition, it is clear that B has to smallest element, hence A has no supremum, and that A has to largest element, hence B has on infimum.


The least upper bound property

An ordered set S has the least upper bound property, or the suprema property, if for any ES, if E is bounded above, then E has a supremum in S.
Analogously, S has the infima property if E bounded bounded below implies E has an infimum in S.

Suprema property implies infima property

Let S be an ordered set with the suprema property. Then S has the infima property.

proof: Let ES be bounded below. Collect all the lower bounds of E into a set L. then, for each eE, e is an upper bound for L. meaning that L is bounded above, now let α=supL. Then, any lower bound β of E is at most α. Hence α=infE.


The real numbers

There exists an ordered field R with the least upper bound property, moreover Q is a subfield of R.

Informal

Of course, we don't exactly know what an ordered field is yet, and we will not prove this theorem, although one can construct the real numbers in many ways, we will not dive into the details of this here.