The reals (constructed as classes of Cauchy sequences of rationals) form a field

https://arbital.com/p/reals_as_classes_of_cauchy_sequences_form_a_field

by Patrick Stevens Jul 4 2016 updated Jul 5 2016

The reals are an archetypal example of a field, but if we are to construct them from simpler objects, we need to show that our construction does indeed have the right properties.


The real numbers, when constructed as equivalence classes of Cauchy sequences of rationals, form a totally ordered field, with the inherited field structure given by

Proof

Firstly, we need to show that those operations are even well-defined: that is, if we pick two different representatives (xn)n=1 and (yn)n=1 of the same equivalence class [xn]=[yn], we don't somehow get different answers.

Well-definedness of +

We wish to show that [xn]+[an]=[yn]+[bn] whenever [xn]=[yn] and [an]=[bn]; this is an exercise. %%hidden(Show solution): Since [xn]=[yn], it must be the case that both (xn) and (yn) are Cauchy sequences such that xnyn0 as n. Similarly, anbn0 as n.

We require [xn+an]=[yn+bn]; that is, we require xn+anynbn0 as n.

But this is true: if we fix rational ϵ>0, we can find N1 such that for all n>N1, we have |xnyn|<ϵ2; and we can find N2 such that for all n>N2, we have |anbn|<ϵ2. Letting N be the maximum of the two N1,N2, we have that for all n>N, |xn+anynbn||xnyn|+|anbn| by the [-triangle_inequality], and hence ϵ. %%

Well-definedness of ×

We wish to show that [xn]×[an]=[yn]×[bn] whenever [xn]=[yn] and [an]=[bn]; this is also an exercise. %%hidden(Show solution): We require [xnan]=[ynbn]; that is, xnanynbn0 as n.

Let ϵ>0 be rational. Then |xnanynbn|=|xn(anbn)+bn(xnyn)| using the very handy trick of adding the expression xnbnxnbn.

By the triangle inequality, this is |xn||anbn|+|bn||xnyn|.

We now use the fact that [cauchy_sequences_are_bounded], to extract some B such that |xn|<B and |bn|<B for all n; then our expression is less than B(|anbn|+|xnyn|).

Finally, for n sufficiently large we have |anbn|<ϵ2B, and similarly for xn and yn, so the result follows that |xnanynbn|<ϵ. %%

Well-definedness of

We wish to show that if [an]=[cn] and [bn]=[dn], then [an][bn] implies [cn][dn].

Suppose [an][bn], but suppose for contradiction that [cn] is not [dn]: that is, [cn][dn] and there are arbitrarily large n such that cn>dn. Then there are two cases.

Additive commutative group structure on

The additive identity is (formally, the equivalence class of the sequence ). Indeed, .

The additive inverse of the element is , because .

The operation is commutative: .

The operation is closed, because the sum of two Cauchy sequences is a Cauchy sequence (exercise). %%hidden(Show solution): If and are Cauchy sequences, then let . We wish to show that there is such that for all , we have .

But by the triangle inequality; so picking so that and for all , the result follows. %%

The operation is associative:

Ring structure

The multiplicative identity is (formally, the equivalence class of the sequence ). Indeed, .

is closed, because the product of two Cauchy sequences is a Cauchy sequence (exercise). %%hidden(Show solution): If and are Cauchy sequences, then let . We wish to show that there is such that for all , we have .

But by the triangle inequality.

Cauchy sequences are bounded, so there is such that and are both less than for all and .

So picking so that and for all , the result follows. %%

is clearly commutative: .

is associative:

distributes over : we need to show that . But this is true:

Field structure

To get from a ring to a field, it is necessary and sufficient to find a multiplicative inverse for any not equal to .

Since , there is some such that for all , . Then defining the sequence for , and for , we obtain a sequence which induces an element of ; and it is easy to check that . %%hidden(Show solution): ; but the sequence is for all , and so it lies in the same equivalence class as the sequence . %%

Ordering on the field

We need to show that:

We may assume that the inequalities are strict, because if equality holds in the assumption then everything is obvious. %%hidden(Show obvious bits): If , then for every we have by well-definedness of addition. Therefore .

If and , then , so it is certainly true that . %%

For the former: suppose , and let be an arbitrary equivalence class. Then ; ; but we have for all sufficiently large , because for sufficiently large . Therefore , as required.

For the latter: suppose and . Then for sufficiently large , we have both and are positive; so for sufficiently large , we have . But that is just saying that , as required.