If I remember right, the fact he was referring to was proved by Tarsky.
Doesn't that mean that you can't even define the reals using the language he started with? If your language doesn't even let you express the difference between the reals and the rationals, it seems to me that the thing to do is to extend your language until it can.
This depends on what you mean by define. If you mean a unique model than this is impossible. The reason is compactness theorem (every theory in which every finite set of formulas has a model has a model). The basic idea is to add constants and introduce axioms stating they are different. This will allow models of, say reals where there are plenty of reals that are not real reals (sorry for the pun, I could not resist). Nonstandard analysis takes it a bit further and makes it a bit more precise and useable.
If you mean you want to deal with (naively) definable reals only then intuitively there are only countably many of those that you can define (by formulas, etc) and you are missing a huge chunk of the real line again.
As I understand it, there is no compactness theorem in second-order logic, only in first-order logic. So your objection would not apply to extending one's language by using second-order logic.
Having said that, you were originally right that no theory of the reals can be satisfied by the rationals, but that's for a fairly unrelated reason.
Yes, agreed.
> you were originally right that no theory of the reals can be satisfied by the rationals, but that's for a fairly unrelated reason.
Can you elaborate?
You missed the crucial property that rules out the rationals (more precisely, the rationals with their standard ordering): one way of stating it is that every sequence that has an upper bound in the set, must have a least upper bound in the set. The rationals do not satisfy this property (for example, consider the sequence of successive decimal expansions, each one to one more decimal place, of sqrt(2)), but the reals do.
The challenge for me is to understand how there can still be countable sets that also satisfy that property of the reals. (Obviously any countable set can be put into one-to-one correspondence with the rationals, but for a countable set that satisfies the least upper bound property of the reals, such a correspondence with the rationals would put an ordering on the rationals that was not the standard one.)