3
$\begingroup$

I know "What are real numbers" has probably been asked before, and the answer would be "the unique complete ordered field" BUT, isn't there some subtlety going on here? In the sense that (at least the way I interpret it) when saying complete ordered field, we really mean complete as a metric space. But to define a metric space you need the notion of real numbers (in order to define a metric). So it seems like saying complete ordered field already assumes the real numbers to be constructed. The question I suppose is: is there some sort of concise way of defining real numbers, without appealing to real numbers? :) (of course, you can describe their construction vie Dedekind cuts for example, but that isn't exactly as succint - guess I'm looking for something closer to a definition rather than a construction)

$\endgroup$
3
  • 4
    $\begingroup$ There are other definitions of “complete”, and there is one in particular that is used for “complete ordered field”. $\endgroup$
    – Zhen Lin
    Commented Jul 24, 2022 at 11:08
  • 9
    $\begingroup$ "Complete" in "complete ordered field" means least-upper-bound property, that requires only order to define, not metric. $\endgroup$
    – mihaild
    Commented Jul 24, 2022 at 11:11
  • $\begingroup$ Ahh, right. That makes sense now! Thanks for the answer $\endgroup$
    – ham_ham01
    Commented Jul 24, 2022 at 11:26

1 Answer 1

4
$\begingroup$

Expanding the above comments into an answer:

There are multiple notions of completeness. When we say that $\mathbb{R}$ is the unique (up to isomorphism) complete ordered field, we are using Dedekind completeness: that is, $\mathbb{R}$ is the unique (up to isomorphism) ordered field in which every nonempty bounded-above subset has a least upper bound. This makes no reference to the real numbers, so the circularity you are concerned about does not arise. (It's separately worth noting that we can talk about metric spaces even before introducing the reals, by using "rationals-only" relations like "$d(a,b)<{p\over q}$" instead of an exact distance function, but that circumlocution is unnecessary here.)

In fact, metric completeness is a very weak notion in this context and does not suffice to characterize $\mathbb{R}$: see the discussion here.

$\endgroup$
0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .