Lecture 5 - Some Consequences of Completeness

Recall from last time that:

R

an ordered field R which contains Q which satisfies the the axiom of completeness.

Here the axiom of completeness says that it satisfies the Least-Upper-Bound Property.

Last time we proved:

Archimedean Property

i. Given any xR nN where n>x.
ii. Given any real number y>0, nN where 1n<y.

As a corrallary we get:

Corollary

For xR, x=0 iff ε>0(|x|<ε).

Proof
Clearly so suppose ε>0(|x|<ε). Assume x0 for contradiction:

  1. If x>0 then |x|=x. Then x>0 so by our given then |x|=x<x is false.
  2. If x<0 then |x|=x. Then x>0so by our given then |x|=x<x is false.
Corollary

Let xR. Then nZ such that n1x<n.

Proof
If x is an integer then we are done. We only have to show n1<x<n. By the archimedian property, then n where n>x. Clearly any n+1,n+2,... could also work, so say that n is the smallest integer to choose. Thus that requires that n1<x as required.

Density of Q in R

We say that Q is dense in R. For more info see The Density of Q in R. A review is as follows, but here's the theorem:

Density of Q in R

a,bR where a<b, then rQ where a<r<b.

Let's do some scratch work:

We know nothing about x,y, but we do know that since x<y then 0<yx. Using the archimedean principle, then nN where 1n<yx. At some point we'll hit fractions that hit into the interval:

By the corollary then aZ where a1nx<a. As a a result:

  1. an<x
  2. We want a<ny. Notice that above from 1n<yx1<nynx1+nx<ny. So then:
a1+nx<ny

So here's the proof:
Proof
Let x<y. By the archimedean property, then nN s.t. 1n<yx. But our corollary says that aZ s.t. a1nx<a. Now notice that:

  1. nx<ax<an
  2. a1nxa1+nx<ny since 1n<yx1+nx<ny. Then clearly a<ny so y>an.
    Therefore:
x<an<y

Thus choosing q=an finishes the proof.

We then talked about the Nested Interval Property. It's an alternative to the least upper bound property which was a fundamental axiom for R.

Before we ended, let's look at trying to prove the following

αR s.t. α2=2

We can't go through the whole proof here (for time) but here's the idea:

  1. Let A={xR|x>0x2<2}. Clearly A isn't empty (1A) and bounded (2A). By the LUB property then α=sup(A)R.
  2. The goal is to show that α2=2. The idea is to show α22 and α22. If α2<2 then we should show that α is not an UB by finding a small positive ε>0 where α+εA. For this we'd need:
α+εA(α+ε)2<2α2+2αε+ε2<2

For the other part showing α22, we want to contradict the least part of being a supremum.